The Developing Individual in a Changing World, Vol. 2: Social and environmental issues 9783111532417, 9783111164403

Citation preview

The developing individual in a changing world

New Babylon

Studies in the Social Sciences

24

Mouton • The Hague • Paris

The Developing Individual in a Changing World Volume II: Social and Environmental Issues Edited by

KLAUS F. RIEGEL and JOHN A. MEACHAM

Aldine • Chicago, Illinois / Mouton • The Hague • Paris

Copyright © 1976 by Mouton & Co. All rights reserved. N o part of this publication may be reproduced, stored in a retrieval system, or transmitted, in any form or by any means, electronic, mechanical, photocopying, recording or otherwise without the written permission of Mouton Publishers, The Hague Published in the United States of America and Canada by Aldine Publishing Company, Chicago, Illinois ISBN 90-279-7502-7 (Mouton, 2 Vols.) 0-202-26083-6 (Aldine, Vol. 1) 0-202-26084-4 (Aldine, Vol. 2) Printed in the Netherlands

Contents

Section IV: Environmental conditions and the development of the individual 1. The responsive environment: interdisciplinary developmental issues by E. Kuno BELLER A. Environmental ontogeny: A cognitive view, by Willis OVERTON

B. Environmental reciprocity: A socio-emotional view, by E. Kuno BELLER C. Histories and futures: Aspects of a study of the child's relation to the land and the city, by David K . R U N Y O N 2.

Effects of nutrition on development by Elaine N . MCINTOSH A. Role of nutrition in human development, by Elaine N. MCINTOSH and Alexander R. DOBERENZ B. Malnutrition in infancy and intellectual development, by Marjorie F . ELIAS C . Protein malnutrition in monkeys, by David A . STROBEL and Robert R. ZIMMERMANN

3. Effects of communication media on child development by David R. OLSON A. The means of instruction in the attainment of educational goals, by David R. OLSON B . Cognitive effects of visual media, by Gavriel SALOMON C. Television and the development of social behavior, by Aimee Dorr LEIFER

413 421 431

445 457 465

475 487 495

VI

Contents

Section V: Social organizations and the development of the individual 1. Variations in home-based early education: Language, play, and social development by Greta G. FEIN A. The construction and selection of environments: Design of the study, by William KESSEN B. Play: The elaboration of possibilities, by Greta G. FEIN C . Language: The formation of discourse, by Susan STARR D. Social development: Enriching connections, by Alison CLARKE-STEWART

E. The Emperor's new clothes, by Joseph

507

515 523 531

541

GLICK

2. Comparison of model preschool programs by Robert P. KLEIN A. Remarks about curriculum implementation, by Jenny W. KLEIN

547

B. Short-term cognitive effects of eleven preschool models, by Marshall S. SMITH C. Dimensional analysis of preschool programs, by Louise B. MILLER D.

Comparing model preschool programs, by John WORTH, John E. DOPYERA and Margaret Z. LAY

556 566

T . NEIS-

579

Section VI: Interaction in social groups and the development of the individual 1. Social interaction and personality development by Ursula LEHR, Bernice L . NEUGARTEN and Jacqueline M . FALK A. Social perspective-taking training: empathy and role-taking ability of preschool children, by Cornelis F . M . VAN LIESHOUT, Gerard LECKIE and Betty SMITS-VAN SONSBEEK B. Levels and patterns of social engagement and disengagement from adolescence to middle adulthood, by Joel SHANAN C. Education as an aid to adaptation in the adult years, by Rolf H . MONGE and Eric F . GARDNER D. Interaction of personality, SES, and social participation in old age, by Reinhard SCHMITZ-SCHERZER and Ursula LEHR E. The generation gap: imagination or reality? by Bengt-Erik ANDERSSON

591 601 611 621 628

Contents F. Affluence, reciprocity and solidary bonds, by Irving Rosow

VII 640

2. Variations in infant-caretaker interactions by Arnold J. SAMEROFF

A. Family interaction in the newborn period: Some findings, some observations, and some unresolved issues, by Ross. D. PARKE and Sandra E. O'LEARY B. The relation of infant's temperament and mother's psychopathology to interactions in early infancy, by Penelope KELLY

C. Mother-infant interaction, attachment, and mother's psychopathology, by Rena ANAGNOSTOPOULOU D. Cohort effects and apparent secular trends in infant research, by Stephen W. PORGES 3. Sibling influences on the development of the individual by Victor G . CICIRELLI A. Sibling position, sex of child, and maternal involvement, by Mary K. ROTHBART B. On the extent of sibling influence, by Walter TOMAN C. Sibling interaction and cognitive development, by Victor G . CICIRELLI

D. Social class, family size, and cognitive performance, by Kevin MARJORIBANKS and Herbert J . WALBERG E. The assessment of stability and change in peer interaction of normal hearing and deaf preschool children, by Cornelis F . M . v a n LIESHOUT

Bibliography Part II Index to Part II

653

664 676

687

697 707 715 723

728

i xxix

SECTION IV

Environmental conditions and the development of the individual

E. KUNO BELLER

1

The responsive environment: Interdisciplinary developmental issues

A. ENVIRONMENTAL ONTOGENY: A COGNITIVE VIEW by Willis Overton,* Temple University, Philadelphia

As each human organism develops there is a concomitant development in the nature of the organism's functional environment. In terms of the focus of this symposium, this is to say that a 'responsive environment' is not a given entity, but rather, a system or set of systems which changes as the requirements of the organism change. Although this conception is not totally new, I think that fruitful consequences issue from a more systematic elaboration of some of its implications than it has often been given. In a recent article, Wohlwill (1973) noted that even the traditional environmentalists have paid relatively little attention to the specific nature of the environment as it influences behavior and development. In a somewhat paradoxical fashion traditional environmentalism, with its emphasis on quantitative influences, tended to homogenize the environment by conceptualizing it as a series of particular events (stimuli), each such event having a highly restricted range of effects. The environmentalist's perspective will not be maintained in the present discussion, but rather the nature of the environment will be examined through the lens of a developmental structuralism, and the particular structuralist position to be employed as such a vehicle will be Piaget's cognitive theory. The major theme * The author wishes to express his appreciation to Dr. Irving E. Sigel for his helpful comments made on this paper.

414 Environmental conditions I hope to develop is that qualitative changes in the nature of the organism, of necessity, involve qualitative changes in the nature of the organism's effective environment. Or stated more succinctly, qualitative organismic ontogeny entails a qualitative environmental ontogeny. In order to elaborate upon this theme it may be helpful to first briefly sketch key facets of a developmental structuralist position, for it is exactly such a position that ultimately results in the issue of qualitative changes in the organism's effective environment. The first and perhaps the most important point to be made - for the rest of the argument is based in its recognition - is the fact that structuralism as an approach to the understanding of man and his products entails radically different epistemologica! commitments than the traditional American empiricist is willing to accept. Thse differences have been detailed in other places (Overton and Reese 1973; Reese and Overton 1970), and here it is sufficient to point out that taking a structuralist perspective entails both the abandonment of the traditional analytic ideal of knowledge and the relinquishing of the attitude that all of man's behavior, culture, and development will ultimately be explained totally in terms of efficient causality, i.e., antecedent stimulus conditions. As a method, structuralism involves the construction of models to best represent the underlying organization of the objects or phenomena under study (Lévi-Strauss 1967; Piaget 1970a; Wozniak 1973). In the specific context of understanding human behavior, a structuralism attempts to cut through the flux of behavioral activity and describe the relatively stable patterns, or forms, or structures of this activity (Overton 1973; Wozniak 1973). Moving beyond structuralism as a general method to the specific structuralism of various psychological theories, it should be noted that the type of formalization that is constructed depends on the goals of the general area of investigation. Thus, for example, Piaget's theory of cognition describes psychological structures which are structures of knowing activity. Piaget defines structures generally as self-regulating systems of transformation (Piaget 1970a) and proceeds from there to a formalization in terms of logical-mathematical models which he maintains best represent the means through which the human organism knows his physical and social world. This brief outline of structuralism is not, however, sufficient for developmental considerations. For this purpose it is necessary to recognize that across the course of the life-span a single structural model may not

The responsive en vironmen t 415 be adequate to its task. Rather, it may be more appropriate to construct a sequence of structural models in order to represent major discontinuities in the individual's patterns of activity across the life-span. This, of course, is a major point upon which theorists such as Chomsky and Piaget disagree. Chomsky takes the position that a single structural model is adequate across the life-span while Piaget constructs several structural models which he presents as an ordered series or 'stages'. Thus, the concept of'stages' in Piaget's theory primarily makes reference to an ordered series of systems of transformation. Each such system has its own unique properties and these are imposed on information arriving from the environment. In summary then, a developmental structuralism leads to a stage theoretical position and includes ideas such as discontinuity and emergent properties of structures. It is at exactly this point that the relationship between these stage ordered structures and the environment assumes major importance. For in examining both the development and utilization of these different structures one must wonder whether it is not also essential to consider the role of environmental structures as well as the traditional environmental stimuli. Consider an analogy. It is well known that a computer may be programed to perform various sets of operations. If several computers, each programed differently, were joined in a series it would not be unlike the various stages of psychological structures. Now the data or stimuli fed into any or all of these computers could vary widely provided only that the language employed, i.e., the structure of the data, were compatible with each computer. If the situation existed such that each computer was not only programed differently but was] also sensitive to a different language system it would be imperative to know what languages (environmental structures) were compatible with what computers (organismic structures). Given that this analogy is a bit of an oversimplification and also that it does not deal with the specific problems of the development of organismic structures, it does set a general problem that is often ignored, i.e., the question of the importance of a mapping of environmental structures onto organismic structures. This does not mean that there has been no sensitivity to the issue of the match between a growth fostering environment and the organism's psychological structures. Certainly in the field of education, the whole era of progressive education might be thought of as a general attempt to find the environmental experiences that best match the developmental status of the child (Overton 1972). Also, present-day Montessori-and Piaget-based

416 Environmental conditions educational programs (Beller 1973) represent a similar orientation. However, in these and other attempts, including Hunt's (1961) explicit statement of the match between the organism's structures and the environment, the focus too often has remained on either specific environmental determinants or on one specific type of environmental structure. An example of this latter case is seen in the Montessori approaches to education. In this situation, although there is a very explicit series of ordered experiences prescribed for the child, all such experiences are generally limited to a sensory-perceptual type of environmental structure. In the context of this problem of the tendency to emphasize a single type of environmental structure, consider also three proposals that have been made for optimizing the cognitive development of the young child. Montessori opts for a sensory-perceptual structure-, Moore and Anderson (1968) through their concept of the 'responsive environment' call for experiences in which the child's behavior is closely tied to stimulus feedback or what might be called an action environmental structure; and Irving Sigel's 'distancing hypothesis' proposes an environment which creates, 'temporal, spatial and psychological distances between self and object' (Sigel 1971, p. 1). Certainly, as general environmental determinants of development there is some significant conflict here. The three proposals suggest what, on the face of it at least, appear to be quite contradictory hypotheses concerning the role of the environment. One alternative for dealing with a conflict of this sort is, of course, to make an appeal to further research in the expectation that research findings will ultimately produce sufficient data to resolve the question in favor of one or another of the hypotheses. Another, and I believe more productive, alternative is to view the three hypotheses as a set of orderable environmental structures each having differential effects depending upon the level and developmental status of the structure of the organism. The general proposal being put forth then is that an understanding of the environment as a set of orderable systems or structures along with an exploration of how these structures may be mapped onto developing organismic structures would lead to a more realistic and meaningful view of how the environment relates to the development and functioning of the organism. This proposal would ultimately yield what Wohlwill has recently described as a 'functional ecology of development' (1973), and its general outline would most probably not be unlike the outline of the organism-environmental relationships discussed by Werner and Kaplan

The responsive environment

417

(1963). In their view, developmental processes are conceptualized as moving from organismic reflex reactions to physio-chemical stimuli, through organismic sensorimotor action upon signaled things and finally to organismic contemplative knowledge about objects (Werner and Kaplan 1963, p. 9). In summary, the present proposal would lead to a series of discontinuous 'responsive environments' rather than a single homogeneous environment. In the remaining time, I would like to move from this paradigmatic orientation to a narrower focus in an effort to detail and provide some empirical support for at least a part of this somewhat speculative general proposal. In the voluminous and controversial literature which examines the effect of environmental variables, i.e., training, task, and situational variables, on the development of that structural period labeled 'concrete operational thought' there is one major fact which has impressed several authors (see, e.g., Beilin 1971). This is the strong suggestion that the traditional experimental environmental variables lead to successful performance only to the degree that the particular skill involved is, to some extent, already present. While this situation may be characterized in different ways, depending upon theoretical perspective, I believe it is fruitful to consider it from the framework of the earlier described organismic-environmental structuralism. More particularly, I think it is worthwhile to consider it as analogous to the development and functioning of anatomical structures. In the development of the eye, for example, there are a series of interactional effects between the differentiating embryo and its uterine environment which culminates in the fully formed structure. However, the eye does not function at this time, but rather, it is ready and awaits a very different environmental context in order to be activated and make adaptive responses. Thus, structures are subject both to what Carmichael has termed the law of anticipatory function (1970, p. 448), i.e., structures are completed and ready to function prior to the time of their first adaptive responses, and to heterogeneous environments which differentially affect their development, their activation, and their functioning. It is assumed that cognitive structures follow the same general course. A basic competence is developed through the interactions of the organism and its appropriate environment (organismic structure matching environmental structure). Following this development, there ensues a phase, generally referred to as the transitional period, during which the specific skills are not always demonstrated in overt performance (Flavell

418 Environmental conditions and Wohlwill 1969). During this phase the introduction of certain environmental variables results in the activation of newly formed structures and thus leads to an improvement in task performance (Overton, Wagner, and Dolinsky 1971). We have explored the activation component of this model in several studies. In an early investigation (Overton and Brodzinsky 1972), we reasoned that figural or perceptual environmental variables, while possibly having a positive effect during the acquisition of competence, might interfere with task solutions after logical structures had been acquired. This is not unlike Piaget's observation that an inappropriate use of the Cuisenaire rods in teaching number skills leads to a focus on perceptual rather than operational features and, thus, subverts the significance of this instructional aid (Piaget 1970b, p. 19). As a consequence we presented three groups of children ages 4-5, 6-7, and 8-9 with two forms of a multiplicative classificatory or matrix completion task. One form was the standard matrix completion task. The second form of the task reduced the perceptual features of the problem while maintaining the logical requirements. In support of our expectations, only the transitional or 6-7 year olds showed any effect of task differences. This group performed significantly better under the logical task structure. These findings have also been replicated in a recently completed study in our laboratory. A similar rationale formed the basis for an investigation in which we explored the development of spatial perspectives rather than multiplicative classification (Brodzinsky, Jackson, and Overton 1972). In this study, children at three age levels were presented a version of Piaget's mountain scene task. Half the children at each age level received the traditional form of the task, and half had the scene perceptually shielded from their view while they chose among alternative representations of the scene. As with multiplicative classification, it was found that reduction of the perceptual features led to enhanced performance primarily in the transitional group. In the later investigation, Brodzinsky and Jackson (1973) replicated this finding and went on to explore the relevant task features related to perceptual shielding. Further support for the hypothesis that the reduction of perceptual features of tasks has an activating effect on newly formed structures has appeared in several investigations beyond our own. For example, following from the Frank (1966) studies which explored the role of perceptual screening in conservation problems, Strauss and Langer (1970) demonstrated such a screening effect to be greatest for the transitional

The responsive environment

419

child. Also, Wohlwill (1968) found that for class inclusion, the presentation of each problem verbally rather than pictorially enhanced the performance of the transitional child. Going beyond the specific hypothesis of perceptual shielding there is also a good deal of evidence which suggests various specific training techniques play their primary role in the activation rather than the acquisition of cognitive structures. The majority of this evidence has been summarized by Beilin (1971) in an excellent general review of training studies. It is sufficient to mention here a few studies that have been published since this review. In one investigation, Youniss (1971) expanded an earlier class inclusion study conducted by Ahr and Youniss (1970). In the later study, children were pre-tested on a series of tasks considered to be precursors to class inclusion and then received a correction training procedure. The results supported a conclusion that the training was effective for those children who already had acquired the prerequisite structural competence. A study of multiplicative classification by Parker, Sperr and Rieff (1972) is particularly significant in that it supports the model being described, despite the fact that neither the earlier research nor structural' considerations generally played any role in the conceptualization of the investigation. In this study, children at ages 5^, and 1 \ were presented with either a thirteen-step sequenced feedback training program, a single-step feedback training program, or a control condition. Performance on a transfer task demonstrated significant training effects only at the 6J and 1 \ year levels. Finally, Kuhn (1972) explored the effects of modeling on a free classification task. Children were categorized according to Piaget's six levels of classificatory skills and then exposed to a model who classified objects at either the same, lower, or increasingly higher levels of skill. A post-test indicated that the child's performance improved most when the model classified the objects at one level above the child's initial sorting. This, then, is some of the evidence in support of the specific activation model which is in turn a part of the earlier described, more inclusive, model of organismic-environmental structural matching. Certainly, both models have been presented in a rather schematic fashion and require a good deal of elaboration and testing. The activation model has focused on environmental performance features. However, there is also evidence available which suggests that various organismic variables such as attention, motivation, and various cognitive styles might also profitably be considered as having primary impact on performance rather than com-

420

Environmental conditions

petence. Furthermore, although the acquisition-activation dichotomy may be adequate for concrete operational thought, where eventually the structural competence is demonstrated in virtually all tasks requiring logical solutions, this may not be the case at other periods of development. For example, at the period of formal operational thought, it is not clear that the formal structures ever reach a point of always being reflected in performance. This suggests that there may be a class of variables which are related to the utilization of such structures, and thus it may be necessary to consider the relationships among features of acquisition, activation, and utilization. In conclusion, I would suggest that beyond opening these new avenues of inquiry, the models described here have the advantage of providing alternatives to traditional 'nothing but' types of explanation which embellish a particular environmental variable with total causal efficacy. They should also provide educators involved in the designing of learning environments with a healthy skepticism towards simplistic unitary solutions.

The responsive environment

421

B. ENVIRONMENTAL RECIPROCITY: A SOCIO-EMOTIONAL VIEW by E. Kuno Beller, Temple University, Philadelphia

Socialization techniques and their emotional context. When one reviews the research literature on child-care practices with regard to early socialization of feeding and toilet training, it becomes clear that the study of techniques rarely yields consistent and meaningful findings unless the investigation includes the emotional or social context in which the technique is used. Findings from studies which investigate both suggest that the emotional context of the experience often determines the psychological consequences of child-care practices. For example, in one study carried out by Heinstein (1963), it was found that the mere distinction between breast feeding and bottle feeding had no importance for the child's subsequent adjustment. A significant and meaningful relationship between feeding practice and later adjustment of infants emerged only when feeding was studied in conjunction with the emotional component of mothering. Prolonged nursing by a cold mother was most detrimental to the later psychological adjustment of infants, and the same consequence applied to the nursing of girls by cold mothers regardless of the length. Similar evidence comes from a study by Sears et al. (1957). These investigators noted that feedingproblems in children were related to a combination of coldness of the mother, her use of punishment, and her resentment of clinging. The need for including the emotional context in studies of child-rearing practices applies also to studies of techniques of toilet training. Sears et al. (1957) report that variations in severity of toilet training alone did not have differential effects on personality development. However, severity of training in conjunction with the emotional relationship between mother and child had an appreciable effect on the child's later adjustment. Thus, when severity of training was accompanied by maternal coldness it had a more disturbing effect than when it was accompanied by maternal warmth or responsiveness. In another study, Hetherington and Brackbill (1963) found that severity of training alone was unrelated to so-called anal character traits in the child; however, more comprehensive patterns of parental function which included both the technique of socialization and the effect associated with such techniques had an appreciable effect. For

422

Environmental conditions

example, orderliness and obstinacy of parents were correlated with the same characteristics in their children. Findings such as these have several important implications for research on child care. First, as indicated earlier, these findings suggest that research in this area must take into account the practices and techniques of care-givers as well as the attitudes and emotional reactions associated with them. It also suggests that identification and modeling play a very important role in the effect of child care on the developing child. Identification and modeling. The literature suggests that techniques of socialization and their emotional context have a combined effect on both the direction and the extent of identification as a central process in personality development. For example, Levin (1958) found that the children of warm and responsive mothers who employed flexible discipline engaged significantly more often in adult role taking in fantasy play than children of cold or non-responsive mothers who exercised strict discipline. While Levin's investigation gives us a glimpse of the importance of paying attention to the emotional context in studying patterns of child rearing and their effect on identification in fantasy, an earlier classic study gives us a very similar insight into the relationship between patterns of child rearing and the overt manifestation of identification. I make reference to the study by Lewin, Lippit and White (1939) who reported that only in groups with democratic leaders who were responsive to the needs of the children and flexible in their controls did children continue to hold together and function in an organized way when the leader was absent. In groups with autocratic leaders who were non-responsive and inflexible in their discipline the boys became disorganized when the leader left the room. The implications of these findings for modeling and identification are clarified by two later studies, carried out by Winterbottom (1953) and Watson (1957) in which similar dynamics seem to have been at work. Winterbottom found that children who were encouraged towards independence by mothers who had a warm and accepting relationship with them became more independent. Watson reported that children reared in warm, permissive homes assumed more responsibility for their own behavior. To be responsive to and to be accepting of one's child's choices is indicative of material independence because of relative non-imposition of her own values on her child. Such a parent provides a model of independence for the child and supports identification on the child's part by encouraging him to be independent. A different but related meaning may apply to the

The responsive environment

423

parent-child relationship reported by Watson. Parental warmth and permissiveness in the form of acceptance of the child's own choices make it easier for children to assume responsibility for their own behavior because the parents avoid placing undue importance on their own reactions to the child's choices as criteria of acceptibility or non-acceptibility of the child's behavior. The implication for the child of such interaction is that the child is encouraged to assume responsibility for its own behavior which is not overly burdened with parental approval or disapproval. Conversely, parents who make their approval of their child's behavior the major criterion of its acceptibility make it more difficult and, finally, less necessary for the child to assume responsibility for his own behavior. Parent discipline and modeling. Although the studies discussed so far provide considerable evidence for the importance of modeling (and the emotional context in which it occurs) for socialization, two critical factors in this process did not emerge clearly until more recent investigations of socialization. I make reference to the firmness and consistency with which parents set limits or control antisocial behaviors of their children and to inner control of parents over their own impulses and behavior. Both Baumrind (1967, 1971) and Coopersmith (1967) provided relevant data in this area. In accord with Watson's and particularly Winterbottom's findings, discussed above, Baumrind (1967) also found that independent children who were characterized by high initiative and well-developed inner controls came from parents who were nurturant, communicative towards their children and made demands for the child's maturity in the form of respecting the child's decisions and encouraging his independence. However, Baumrind's study went a step further by extending her investigation to the area of parents' setting limits and standing their ground instead of withdrawing from power conflicts with their child. Baumrind found that the presence of these last parental characteristics was essential of facilitating the development of self-reliance, self-control, and initiative in children. Parents who were nurturant without setting firm limits and without making demands for independence and maturity tended to have immature children, while parents who made demands for maturity and independence without being nurturant or communicating with their children tended to have children who were both immature and anxious. Coopersmith (1967) found that parents of children with high self-esteem and high independence manifested themselves high esteem and high independence. Similarly, in a more recent study, Baumrind (1971) found that socially responsible

424

Environmental conditions

behavior was facilitated in children of parents who themselves behaved in socially responsible ways and demonstrated self-respect, as well as respect for the child as an individual. Evidence of the importance of impulse control for modeling comes from a study carried out by Sears et al. (1957). These investigators found that parental disapproval of aggression implemented by physical punishment (and presumably with negative affect) produced the most aggressive children; whereas parental disapproval of aggression implemented by techniques other than physical punishment produced the least aggressive children. Inconsistency and arbitrariness seem to be important factors in these relationships. They function as sources of frustration and therefore heighten instigation of aggression in the child, as well as parental modeling of the aggressive role. Parents who reject aggression in their children but utilize the very behavior which they reject in their children to control those same children achieve the opposite effect of what they profess to aspire to. One might say that the inherent inconsistency of the parents combined with their modelling of an aggressive role are more powerful than the intended effects of punishing the aggressive behavior in the child. A finding from another study tends to support this interpretation. Zunich (1966) reports that parents who overemphasized their discipline and were intent on breaking the child's will had children who were aggressive towards other children in the form of criticizing and dominating them. To sum up, it would appear that inconsistency, arbitrariness, and irrationality of environmental response to the child's antisocial and impulsive behavior interfere with the development of inner controls and results in reliance on networks of hurting and being hurt by others, and on controlling and being controlled by others. Further insight into the dynamics of the modeling process and its consequences for internalization comes from a study by Hoffman and Saltzstein (1967). These investigators studied how children internalize their parents' reaction to aggression of the child. Children reacted to their own aggression (or transgression) in the same way as they reported their parents to have reacted to their aggression, that is, either punitively by expecting retribution, or non-punitively by expecting tolerance and understanding. The mothers of children who responded with anticipated retribution to their own aggression were colder or less responsive than the mothers of children who reacted non-punitively to their own aggression. Thus we see that parental warmth, responsiveness, and modeling play an important role both in the modification of aggression and in the child's

The responsive environment

425

perceived or expected consequences of aggressive behavior - his own or that of someone else. Parent-child reciprocity. Careful scrutiny of the studies that we have discussed briefly and the type of findings which were highlighted suggest a critical further step in research on child rearing and socio-emotional development. I make reference to a reciprocal process in which there is mutual feedback and mutual determination between the adult and the child, rather than a simple one-way traffic. The parent, the educator, or the adult in some other role as care-giver and the child enter into a pattern of mutual interaction in which each elicits and supports patterns of attitudes and behavior in the other which often are similar to one's own. Although direct evidence has not been presented for such a process in the studies which have been discussed so far, the data clearly point to it. In the studies of Winterbottom (1953), Sears et al. (1957), Baumrind (1967, 1971), and Coopersmith (1967), the parent-child relationships reflect stabilized dyadic cycles in which each partner perpetuates the reactions of the other around a common theme. The parent who is pleased, rather than indifferent or worried, over the child's independent exploratory excursions is likely to have a similar attitude to such behavior on his or her own behalf. Therefore, the experience of seeing one's child engage in independent activity which the child enjoys not only pleases the parent but probably also strengthens the parent's own motivated behavior in this direction. It is not surprising that investigators who found a relationship of mutual independence between parent and child also found a heightened degree of warmth and nurturance to be characteristic of these parents. Warmth refers to a heightened positive responsiveness since it involves not only noticing and accepting but also appreciating and caring about the needs, values, and behaviors of the other person. Nurturance also heightens parental responsiveness partly because its defining characteristics overlap with warmth and partly because 'to be nurturant' represents adequacy and competence, which have similar connotations and consequences as independence, thus heightening the parental modeling effect of independence. Encouragement of independence by cold or indifferent parents would contribute either to a lower level of parental responsiveness 1 or heightened rejection, or both, 1. Lowered responsiveness has to be distinguished from inconsistent or from ambivalent responsiveness (e.g., to both appreciate and deprecate). Such a distinction has not yet been implemented in research studies, at least not in the studies discussed here.

426

Environmental conditions

since parental pressure for independence would be associated with nonresponsiveness to dependency demands or dependency anxiety. Under such conditions pseudo-independence is more likely to develop than genuine independence. Pseudo-independence manifests itself in avoidance of seeking help, attention, or contact, in compulsive independence or in using independence as a means toward dependency goals, i.e., getting attention, praise, and acceptance. Since such pseudo-manifestations of independence have rarely been assessed in past research, no evidence of its relationship with child-rearing practices is yet available in the literature. We do have evidence of the consequences of parental encouragement of independence without or with reduced nurturance and warmth (communication). Baumrind (1967) finds that such a pattern of parental functioning results in anxiousness, restlessness, depression, and disaffiliation rather than independence. How the child's reactions to parental practices and responsiveness affect subsequent parental practices is not immediately evident from existing research material, although some conclusions are strongly suggested. For example, in Winterbottom's study the failure of cold mothers to heighten independence in their children may reflect the result of a child's failure to manifest sufficient independent action and motivation. Such functioning (or malfunctioning, from the parents' perspective) might have reduced the mother's positive affect without changing her demands for independence. Similarly, the relationship of the same parental pattern to the child's dependency anxiety and avoidance of independent physical and social exploration in Baumrind's study may reflect the consequences of the child's reactions on the parent who continues to press for independence (maturity) but ceases to be nurturant and warm. One important extension of this line of reasoning is that such negative consequences of the child's actions and reactions for parental behavior probably involves a change of parental modeling of independence competence. Reduced nurturance and probably other immature behaviors which accompany parental negative reactions to their child's immaturity change the parent's own role of being a model of independence and strength. What emerges is a new reciprocity between parent and child in which the emotional component changes although some of the parental behaviors and goals persist. That is, the parent continues to demand and expect independence from the child without providing the necessary affect or model to make his demand effective.

The responsive environment

427

One of the studies which I have discussed suggests a mutual feedback and directional reversibility. In the Hoffman and Saltzstein study (1967), it is possible to interpret the relationships between children's reactions to their own transgressions and perceived parental attitudes towards such transgressions as internalization of parental standards by the children, as well as a projection of these internalized reactions to transgressions onto the parents. Teacher-child reciprocity. In order to present more evidence of adultchild reciprocity in education settings I shall draw from a study that I carried out with Head Start children (Beller 1969). The central question in the study was whether intellectual gain from that program related to dependence-independence interaction between teachers and children. The investigator first determined which children gained from the program and which children either failed to gain or even lost after eights months experience in an all year round Head Start program. After having classified the children into these different groups they were observed intensively in their interaction with their teachers. It was found that the children designated as gainers made more realistic as distinct from 'emotional' dependency requests of their teachers and, in turn, generally received more positive reactions to their requests. These gainers also made more constructive use of the help they received in response to their requests and reacted more constructively than other children when the teacher failed to respond to the child's request. When the observations focused on the instances in which a child was occupied by himself, it was found that the children who had gained from the program received more unsolicited gestures of interest and attention from the teachers than children who had not gained from the program. Finally, children who gained from the program were more often found in classrooms of teachers who stressed the idiosyncracies of individual children and manifested interest in as well as respect for the child's family. Two conclusions are indicated from this study: First, emotional components and interpersonal attitudes play an important part in the impact of the teacher-child relationship on the intellectual development of the child. Secondly, there appears to be much similarity between certain behaviors of teachers and of those children who benefit most from the educational efforts of the same teachers. Both are responsive to each other's requests and expectations, tolerant of each other's needs; in short, both are respectful and appreciative of each other. This suggests a positive cycle in which adult and child support each other's

428

Environmental conditions

needs, through mutual responsiveness to the autonomous needs of the other person. The reverse is equally true. Both the child and the adult may function in a negative cycle in which each supports insensitivity and non-responsiveness to each other's needs, a state of affairs which might be expected to result in a lack of gain for the child, as well as for the teacher in terms of her explicit functions and objective. There is further evidence for such cycles from research of other investigators of teacherchild relationships. For example, De Groat and Thompson (1949) found that more intelligent children with higher academic achievement more frequently received high approval from their teachers. A second investigator, Hoehn (1954), reported that teachers have more favorable contacts with children of higher academic achievement. A third investigator, Wilensky (1968), found that teachers spent more time with brighter children; and finally, in a fourth study, Battle (1957) found that high achievers and their teachers held more similar values than low achievers and their teachers. Implications and conclusions. The implications of the cycling or reciprocity model for research on socialization and education are that we must concentrate equally on both partners as interactive participants. The classical model for controlled laboratory research in which the independent variable is manipulated by the experimenter in order to study its effect on the dependent variable is much too limited for research on reciprocity. Parent and child, teacher and pupil, therapist and patient are both actors and reactors. To better understand, predict, and modify either partner in these pairs, research must concentrate on reciprocal processes which govern stability and change in both partners. With regard to theory, we must look for concepts that transcend the illusion of a one-way traffic under which we have labored for so long in theoretical formulations of socialization, education, and therapy. The mother or parent is the most powerful socializer of the young child precisely because she is more deeply affected by her child than almost any other adult is. Similarly, a son or daughter is more effective in changing or maintaining parental practices because the parent means more to them and vice versa. Thus both affective and structual components are closely intertwined in the reciprocal process. To reverse the unidirectional approach in socialization by making the child the independent and the parent the dependent variable (Osofsky 1973) has helped to correct some longstanding misconceptions paving the for realizing a design for studying reciprocal processes.

The responsive environment

429

The idea of reciprocity helps us understand a wide range of phenomena and findings in the research literature which have remained unclear without it, as, for example, the different effects of maternal responses to crying or of maternal warmth on the development of dependency and independence. The infant who cries responds egocentrically to his own discomfort. The mother who picks him up promptly and comforts him displays an other-directed response. She responds to the infant's stress differently than he does. The infant probably experiences the mother's response to his own stress a different from his own reactions to stress. He is tense and self-directed while she is calm and other-directed. More importantly, she responds effectively to the source of discomfort which allays the pain that produced stress reactions like crying or rage, which by themselves are less effective in removing the pain. Over time the infant learns that mother's response is more effective than his own. This process is particularly applicable when the infant's crying occurs in response to inner tension (rather than only to separation from the mother), and when the mother's response goes beyond holding the infant, that is, to looking for the source of tension and removing it. Some such interactional process may be an important reason why several recent studies have found that picking a crying infant up promptly instead of reinforcing the crying or spoiling the infant results in lowering the frequency of crying and in more active and effective coping of infants (Bell and Ainsworth 1972; Yarrow, Pedersen, and Rubinstein 1971). The role of parental warmth and nurturance in this situation is clearly similar to its role in reciprocal parentchild interactions around independence discussed earlier. For example, if a parent would pick up the crying infant promptly without nurturance, that is, without attempting to remove the source of pain, the infant's crying is likely to continue after a temporary delay. Similarly, if the parent, while picking up the crying infant, would respond with negative affect, she would reinforce the infant's own negative affective state. This would counteract a reduced stress response to pain and more effective coping behavior on the infant's part on subsequent occasions. Conversely, if the infant fails to calm down in response to being picked up and to efforts by the mother to remove the source of stress, the mother's affective reaction is very likely to turn negative, i.e., annoyance, impatience, despair, and rage. In such a case, depending on when a researcher starts his study, the finding may easily be a misconception of a directional cause-effect sequence due to an artifact of the moment in time in the reciprocal sequential process at which the researcher happened to commence his study.

430

Environmental conditions

In conclusion, I have tried to outline an interactional approach to the study of psychosocial development in which the concepts of responsive environment and environmental reciprocity play a central role, and in which motivational-emotional and perceptual-cognitive factors interact and converge.

The responsive environment 431 C. HISTORIES AND FUTURES: ASPECTS OF A STUDY OF THE CHILD'S RELATION TO THE LAND AND THE CITY by David K. Runyon, University of Wisconsin, Whitewater

This is a brief statement of the history of playgrounds in the 19th and 20th centuries. It is divided into four chronological units: A continuation of the English garden/park tradition; the active recreation movement; the European adventure/junk playground movement; current trends. The garden/park tradition democratized. The private forest preserve for hunting and the castle-palace-mansion garden of the ruling classes have been a paradisian adult park or playground since antiquity. Garden/parks were a luxury enjoyed exclusively by the rich with the exception of an occasional opening of these private gardens to the public, e.g., ancient Roman emperors, Renaissance princes, and the English crown in the 17th, 18th and 19th centuries. In 1847 the first private parks were opened permanently and presented to the public. They were Sheffield Park given by the Duke of Norfolk and London's Victoria Park (crown property), but the first park designed specifically for public use was Birkenhead Park, planned in 1843 by Joseph Paxton, architect of the famous Crystal Palace. Birkenhead Park also opened in 1847. This park was purposely located near the crowded and polluted slum dwellings of the workers community for sanitary and hygienic reasons. However, the financing of Birkenhead was accomplished by developing the perimeter of this land into elegant housing for the wealthy middle class who faced the park. In terms of playground development, Birkenhead was the first public park that made provision for active recreation, i.e., sports and games like archery and cricket. By contrast, the parks of the 1830s, crown property open to the public, were passive, walking parks. Frederick Law Olmstead, probably America's most noted 19th century landscape architect, saw Birkenhead on his tour of England in 1850. He and his partner, Calvert Vaux, used it as their inspiration for the design of Central Park in New York City in 1858. Three concepts that held true in the traditional garden/park remained constant in the Central Park plan: (a) escape the city; (b) salvation through nature (beauty); and (c) passive recreation. The same general visual ugliness, unsanitary con-

432

Environmental conditions

ditions, and social-spiritual depravity of the 19th century city in England were in existence in America as well. Olmstead had the opportunity and the vision to preserve a large piece of wilderness in what he foresaw would become the heart of a great city. Carriage traffic passed relatively concealed either over bridges or recessed below ground level to facilitate easy crosstown travel; thus the park maintained its mood of an undisturbed rustic open space of fresh air and sunlight. Man-made topography and planting design screened oif sight and sound of the city; however, it should be remembered that most buildings were not more than six stories high, and transportation was not motorized. Much of the space in the park was broken down into intimate areas for meditating as one strolled through changing settings, and for picnicking activities, which are of a decidedly passive orientation; however, as in Birkenhead, there were active areas. Central Park had paths for horseback riding, a reservoir for boating and ice skating, and meadows for cricket, informal play, etc., thus expanding Paxton's introduction of active play areas. Chalk lines were laid down for tennis on grass courts, and the area was hidden by a green belt of trees. Unity of design demanded a rustic theme be maintained uninterrupted throughout. To have destroyed this wild, rustic setting, would have been a desecration of a surrogate Garden of Eden. Public buildings, national monuments, and park reserves are repositories of sacred mass values. A walk through Central Park was a pantheistic variation of a religious liturgy, a sacred dance, where every man, as his own priest, received revelation through the beauty of untrampled, wild nature. A distinction between the word park and other green areas, like city squares, commons, and greens may be helpful. The park is an outgrowth of the traditional hunting preserve and/or garden area in or around a castle, palace or mansion. City squares, commons, and greens come from different traditions based on the respective functions of open markets, pasture lands, parade and military drill grounds, settings for important buildings, assembly spaces, and, incidentally, recreation in the park tradition. Boston Common, laid out in 1634, and probably the first city 'park' in the United States, is a very large, irregularly shaped green space; its function today, like other New England greens and commons, has changed from its original practical function to become primarily recreational in the escape-beauty-passive park tradition. Ironically, this change in green space use was brought about by industrial revolution which switched the commons orientation from the practical needs of its puritan non-leisure focus to a recreation area justified on the basis of

The responsive environment

433

health. In the 1830s and 1840s, American industrial cities were expanding, and farmlands were being relocated in exurbia. The convenience of nearby public open spaces for park use was almost nonexistent. Because of the basically dignified mood of the escape-beauty-passive character of the park tradition, we can more easily understand how cemeteries on the outskirts of the cities and towns were readily used as parks without any sense of impropriety. The park tradition from the last quarter of the 19th and all through the 20th century was expanded into the national and state park movements for both conservation and preservation of national beauty as well as recreation. Urban 'breathing-spaces' like green belts, parkways, and municipal parks have developed, sometimes by combining them with specific, active recreation areas. The active recreation movement; the moralists triumphant. Life-styles, the taste of the times, etc., are the larger contexts by which we can more easily perceive the meanings of smaller issues. Because we are born into one era, we often falsely assume that things have always been that way. Our current understanding of childhood is quite different from other ages and cultures and is, in fact, a product of a new style that appeared in the 19th century. Phillipe Aries' book, Centuries of Childhood, documents how, up until the last few hundred years, there was no concept of childhood as we understand it today. There were infants (from birth to six years) and adults (six or seven years old and upwards). The separation point was not puberty but work ability. There was no nuclear family but an extended family of sorts. Middle-class male children were apprenticed to another family at six or seven, whereupon they shared in an adult life-style; they lived, worked, and played alongside adults in their homes, shops, taverns, fields, and beds. Infants were allowed a separate life-style and had their own games (adult hand-me-downs) and toys, hobby-horses, windmills, birds on a leash, dolls, etc. For both boys and girls of six or seven and older, there was no separate existence from adults. Boys of seven and older joined in adult recreation of horseback riding and archery, as well as adult debauchery. It was not until the 1500s that a moralist movement began among a few lawyers, priests and other educators to shield the young from what these moralists considered corruption by the adult world. A greater consciousness of childhood developed as children's innocence was recognized. This innocence was increasingly protected by isolating the children gradually from a tainted

434 Environmental conditions adult world into a separate "antiseptic" world dominated by adults of the moralist persuasion. This new outlook was begun among the middle class and continued to grow in dominance as the middle class grew, until the 19th and 20th centuries when it reached fulfillment in a period of prolonged childhood. Careful attention was paid to the child's discreet environment, upbringing, and education. The economic and social position of the middle class has always been more precarious than the more stable peasant or noble classes who had less need to change their traditional child-rearing practices. Since their economic and social position was relatively stable, they had no reason to be defensive, but the middle class was quick to protect their achievements and anxious to stabilize and increase their new prosperity. The new separate life-style of the middle class centered on the private life of the home and the family rather than the communal orientation of the poor and the nobility. Out of this middle-class life style came a consciousness and a development of a period of childhood as a separate period between infancy and adulthood. In time, this new age designation (childhood) was further divided into childhood and adolescence. A separate formal education from college down to elementary school further isolated children from the wider adult community, first in private schools for the wealthy few and then, later on, as free, compulsory, and tax-supported schools for the masses. Thus developed a consciousness of the young as a separate group. American colleges were set up on European models. The Germans, in neo-classical (supposedly in ancient Greek and Roman) style, preferred physical training in the form of gymnastics. This spirit filtered down to lower grades through the city day school (gymnasium) which provided physical training but, usually, no playground. The German ideal was academic learning, but the English emphasized gentlemanly middle-class values. Games taught courtesy, fairness, subordination, and cooperation. Boarding schools for the sons of wealthy middle classes were in the country where meadows were available for informal play such as rugby, riding to the hounds, etc. The Americans began by copying the German model in colleges, but soon the German and English traditions blended, with an emphasis toward the English system. Several Eastern colleges and schools constructed outdoor gymnasiums between 1820 and 1840. The English playground, quite naturally, filtered into the English landscape park tradition, which was transplanted to America. William Cullen Bryant, editor of the Evening Post, and Andrew Jackson Downing, the American landscape gardener and propagandist for suburban living, played key

The responsive environment 435 roles in promoting the first major statement of this English ideal in America, i.e., a central park for New York City. It was an incredibly slow and difficult task to get playgrounds into primary and secondary American schools. Both parents and teachers saw playground recreation as a waste of time. In 1907, nine-tenths of American schools had little or no space around them. The little space available was to assemble and march the children up to their classrooms (Curtis 1907). The active playground movement as an organized lobby came into being around the turn of this century and was part of a general reform movement by the wealthy classes, who sought to escape from the inner city to the suburbs, and to refurbish their inner city working environment on a grand and noble scale. The best example of this was The City Beautiful Movement in Chicago, specifically, The Columbian Exposition of 1893 and Daniel H. Burnham's plan for Chicago published in 1909. Like the public park movement, the Active Playground Movement had its beginnings in the same unhealthy living and social conditions of the industrial city. The moralist tradition was carried on by a few concerned people amidst a larger middle class who thought playgrounds were a waste of time. Playgrounds were begun by progressive public officials, social workers, and educators who had an experimental and creative approach in meeting their civic responsibilities. The first official playgrounds were built in the Boston area. By action of the town meeting in 1862, Brookline, Massachusetts, purchased two tracts of land for playgrounds. Late in the 1860s a playground was sponsored by the Old First Church of Boston in the yard of a public school. It was approximately fifteen years later when a genuine movement developed on a larger scale to provide recreation for little children. In 1885 the Massachusetts Emergency and Hygiene Association financed the construction of sand gardens in the yard of the Children's Mission in Boston. This experiment was so successful that two years later ten such centers were opened, and paid matrons were employed for the first time. By 1889 the city of Boston contributed funds to the support of these centers (Doell and Fitzgerald 1954). The settlement or neighborhood house concept transported from England to America in 1887 supported the play movement. In 1892, Hull House in Chicago established a model playground three years after the settlement house was established. In New York City, Jacob Riis, a newspaper man, promoted play space and recreational activities, since slum housing had no lawns or vacant lots, and schools had no land around

436

Environmental conditions

them for play. By 1899, Boston had twenty-one lots for small children. By 1900 ten cities had established playgrounds. The moralist reformers took on two battles. The first was to prove that there was an educational type of play (play as virtue), which distracted youngsters from vice, sin, crime, or worse. The puritan recreation reformer saw play rigidly linked to competitive and organized games - sports and athletics where a trained adult was generally in charge. The goals were health, social training, and character building in the tradition of English middle-class values: good sportsmanship, fairness, courtesy, obedience to the laws of the game, subordination to the captain, and cooperation with other members, i.e., playing for the team, not for the grandstand. This was bourgeois survival training through competition, albeit restrictive competition. The second battle was more presumptuous. The reform movement was not only for middle-class children but was an invasion of middle-class values into lower classes. In Henry Curtis' U.S. Bureau of Education Report of 1907, his rhetoric for play and playgrounds exhibits a viewpoint which would be considered anachronistic. He writes that play is to be understood as physical training and as a social arena in which friendships are formed and cemented, where the child is fitted to become a member of society. Play is a school for character building where habits of childhood and youth are chiefly formed and their tendencies fixed. 'Play kept a boy from wrong acts and thoughts' (Curtis 1907). From Gregg (The World Book) comes the observation that boys fix things and play marbles and generally prefer more active strenuous activities and girls skip rope and cut out dolls. These two battles viewed together reveal the general 19th century spiritual milieu of business efficiency, paternalism, intrusiveness, and male chauvinism. It is easy to see how the Y.M.C.A., Boy Scouts, Girl Scouts, etc., grew out of this atmosphere. Another stimulus to the organized game emphasis of the recreation movement was the general sports orientation of recreation among adults after the Civil War. Certain adult games among the wealthy become democratized and were played by the middle classes in America, e.g., cricket, golf, tennis, etc. On the rural scene, people lived too far apart to get together conveniently; therefore, their recreation was more individual and informal, but in cities the organized game was the style, probably because people were naturally clustered together, and because it was the cheapest form of recreation. Thousands of foreign immigrants and indigenous American farm people migrated to cities to fill the needs of a

The responsive environment

437

growing industrial revolution. This exacerbated an already crowded, unhealthy slum condition. In the wider economic and social situation of the times, there were revolts and uprisings which threatened the security of the barons of industry. Like ancient Rome, the modern industrial city needed mass distractions, a release to ease tensions. Bicycles became popular in 1870, and bike races were the rage in the 1890s until the introduction of cars for the masses destroyed this enthusiasm. (The tricycle came in around the 1890s.) Baseball was introduced before football (1869) and became the most popular sport. When it was shown that a paid professional baseball team could win all its games against amateurs, professional sports became popular. The first world series began in 1900. From that moment, organized sports became a big business which helped legitimize this sort of play activity. The organized sport approach also exploited the need for heroes and winning teams which eventually helped consolidate the recreation movement into schools and recreation-park commissions throughout cities and suburbs. A system, which ranged from little league to university gate receipts to glory funneled talent into professional sports. This professional turn of events made athletics into a spectator sport for the masses, thus deflecting the original goals of active physical and social development of the recreation movement. The middle-class values of gentlemanly conduct which were something of a balance between process and end product were now changed into three business values: win, win, win! The same held true for professional sports of an individual character like boxing, auto racing, etc. With increased industrialization and urbanization, the home as a central focus for family life was destroyed; the father spent long hours working outside the home, and town and city culture took over traditional domestic middle-class responsibilities. The child spent much of his time away in school, church, play, part-time jobs, etc.; the home was no longer the dominant focus of his interests or activities. To fully appreciate the insensitivity to children's needs and interests inherent in the adult orientation of 19th century thinking which was the source of most of our current 20th century thinking about play and playgrounds, we have to focus on the new direction in the study of human behavior which has gone on since the turn into the 20th century. Today's focus is on the child, and on his normal growth as developmental stages which tend to emerge in fixed sequences. In the affective realm Freud postulated oral, anal, and genital sequences. Pioneer work in motor development began in the 1920s by Nancy Bayley and Myrtle McGraw,

438 Environmental conditions was followed up by Gesell in the 1940s, and is being revised by Anna Doudlah (Central Wisconsin Colony, Madison). The impact of hierarchical strategies of cognitive development began in the 1930s by Jean Piaget and Heinz Werner and their followers, as well as David Rapaport and other followers of the neo-Freudian persuasion. Having the difference between the 19th Century's adult focus and the 20th century focus clearly in our mind as a point of contrast, we can understand why most people in the 19th century saw children as small sized adults and why an adult seldom felt the need to pay attention to children's preferences. To the 19th century mind, what one really needed was a strong adult model to orgainize and command children at play. The focus was on motor experience, not motor development, and on inculcating social-emotional values not affective development. Vacant lots with interesting topography were leveled for organized court games. When dirt streets were paved with asphalt for reason of sanitation, so were many playgrounds. This loss of play potential was traded off for greater, yet unnecessary, sanitation and an understandable bureaucratic sensitivity to respectable looking public spaces under their jurisdiction. (The visual openness and public nature of these playgrounds is a continuation of the traditional public park tradition). Most of the equipment in the play yard, swings, see-saws, etc., were historically adult hand-me-down amusements. By the end of the first quarter of the century, a bureaucratic tradition had been established of investing big money into play equipment because adult supervision had to be curtailed due to wartime needs. The ultimate loss of contact with children occurred when playgrounds without adult play leaders became a general policy. In summer months when school was out of session, highly dense urban municipalities, probably out of fear, provided adult leadership as far as finances allowed. These programs were often bedeviled by public relations in situations where the recreation bureaucracy was more interested in preserving their sterile middle-class image than in catering to the interests of the children. Thus the crisis in American playgrounds is exemplified by the familiar sight of an abandoned play lot surrounded by a chain link fence, covered with asphalt and broken glass, with a few isolated pieces of iron mongery partially ripped out. The new reformers: The European adventure¡junk playground movement. No other century has devoted so much scientific research into the development of children as our own. This new openness, in attempting to

The responsive environment

439

understand who children are, has given us opportunities to allow children to show us who they are. We have become more careful observers. We are beginning to discover what has been going on under our noses for centuries. For city children, the streets have been a playground since cities began. The old fashioned open market with live animals, craft shops, horsedrawn transportation, and pedestrian dominance made urban streets exciting areas. The diversity of shops, things to see and do, and the general activity was a type of combined school and playground. Today our carefully zoned, residential suburban tracts and high-rise housing urban developments are bland by comparison. Rural children did not have the ideal life we often think they did. They worked long hard hours, and, like lower-class city children, they lacked physical and social development. It is a great irony that the moral reformers of the recreation movement extolled the rural life with one breath, and then in the next, complained that the deficiencies and debaucheries were just as bad, if not worse, than the urban scene; what vices rural children lacked in variety, they made up for in intensity (Curtis 1907). An historical study of informal play has yet to be written. Much of the information is to be found in novels like Tom Sawyer, Huckleberry Finn, the writings of A. A. Milne, etc. Bits and pieces, like the old swimming hole, the hay loft, the abandoned lot, the woods at the end of the housing tract, etc., should be assembled to see what has attracted children in the past. In addition to these activities, the ever-changing environment produces new interests. In lieu of that information, it is helpful to survey adult contributions to child directed play in the 20th century. The Japanese-American artist and environmental designer, Isamur Noguchi, noticed that children prefer land that has various elevations. Playgrounds built by the recreation movement, usually flat areas for court games, had failed to make use of the play possibilities provided by hills. Noguchi proposed this varied topographic idea as early as 1931, but it was not realized until twenty or thirty years later. In the meantime, he gave up hope of changing the fixed concepts advocated by the recreation movement establishment. In the 1920s housing projects in Copenhagen began to include garden space to give children the opportunity for the kind of creative play that was available in the country. Professor C.T. Serensen, a landscape architect, was hired to lay out a traditional play garden in one of these housing projects. The playground was an instant failure. Upset, yet

440 Environmental conditions concerned, Sorensen looked around for the neighborhood children and found them a few blocks away enjoying themselves in a construction site. In 1931 he wrote (Allen 1968): 'Perhaps we should try to set up waste material playgrounds in suitable large areas where children would be able to play with old cars, boxes and timber. It is possible there would have to be some supervision to prevent children fighting too wildly and to lessen the chances of injury, but it is likely that such supervision will not be necessary.' A new attitude was born towards both play materials and play leadership; this was child initiated and directed play with adult approval. Sorensen's suggestion was ignored until the crisis of Nazi occupation when there was no money for traditional playgrounds. Twelve years after Scrensen's initial suggestion, Dan Fink, an architect, 'designed' the first adventure playground in Emdrup, a suburb of Copenhagen. The playground was staffed by an adult leader whose role was more of a referee and monitor than a leader. Two years later (1945) Lady Allen of Hurtwood visited Emdrup and was convinced of the value of the junk playground. Upon her return to London, she became the major international advocate for the junk playground. There are thirty or forty of them in England, including three for the developmentally disabled. There are many of these playgrounds in Europe today, for both normal and developmentally disabled children, but in the United States, they have yet to win general acceptance. The adventure playground is not always a junk or waste material playground. Its focus is on the type of play, not the equipment as such. The view is that a child's natural curiosity should be encouraged by providing him with opportunities to explore, discover, and experiment for himself. The role of the designer and adult leader is to set up an environment for creativity and imagination in which the child has an opportunity to manipulate a great variety of moveable materials, to see things from various points of view (his own), to do things in a new or old way following his own direction, and to put things together in satisfying combinations. In a very real sense, the child is his own play designer and leader; he has a sense of control over his own world, to change it, to remake it, to destroy it and to rebuild it again and again. These challenging, creative, diverse, new, fun, often dangerous experiences are what make it educational. The child learns through his action, i.e., through his manipulation of himself and protean materials. The emphasis is on individual fulfillment at one's own stage of development rather than

The responsive environment

441

having to fit into a procrustean bed of doing the current (peer) fashionable thing or by winning games or (adult) prizes for personal achievement. The exploring process is its own reward and pleasure; there are few products in a materialist sense. There is no competitive system of rewards and punishments. The major task of the play leader is to encourage a child to do his own thing. The word freedom sums it up. However, when one person intrudes on another, that is license, and it is not permitted. The adult's job is to make three major provisions: first, to provide a constant and sufficient supply of fresh, stimulating materials to generate diverse explorations. Second, to help the child by providing him tools to explore the materials. Third, to provide reasonable safetly in activities from social relationships to construction pieces forclimbing,swinging,burning, and shelter. A successful adventure playground is in a state of constant construction and destruction, building and rebuilding; the play leader is in perpetual motion and is as unintrusive as possible. Some adventure playgrounds are not entirely built by the children themselves. These are exceptions but follow the same basic spirit. When one wishes children to have important experiences that they would otherwise not have, it becomes advisable to build the appropriate equipment. Lady Allen's adventure playground in Chelsea for handicapped children is an example. The same justification holds true for small children. An excellent example was the playground for small children at Expo '67 in Montreal, Canada, designed by the landscape architect, Cornelia Oberlander, which included a tree house, varying topography, tunnel, row boat moving along a narrow stream or floating in a small pond, gaint Lincoln logs, etc. Lady Allen's junk playgrounds have proved successful not only for children, but for the surrounding neighborhood, when the playground is carefully screened. Vandalism has dropped remarkably in areas where children had very little space to play. One gets a very strong feeling that Lady Allen's junk playgrounds are a new type of school. Her ideas seem to be in harmony with A. S. Neill, the British Infant School and Open Classroom movements, and the Swiss child developmental psychologist, Jean Piaget. The adventure play proponents strike a balance of psychomotor, affective, and cognitive opportunities for development. The traditional American physical education leader is an organizer; the junk playground leader is a facilitator; this is a major educational difference. Current trends. Puritan-influenced laws prohibiting all forms of amusemen

442 Environmental conditions and misuse of time appeared in the 1600s and represent America's earliest forms of recreation legislation. Contemporary American playground design for elementary and secondary school children is philosophically still back at the turn of the century although it is dressed-up in new clothes. At its best, it is correcting design mistakes made in earlier equipment. Since much of this equipment is somewhat appropriate for the young and/or handicapped child, it is a contribution. One could say that the American contribution has not been marred by genius. The most interesting designer on the American scene is M. Paul Friedberg. He has become the major spokesman for the concept of linked play. By contrast to the earlier practice of placing equipment in isolated areas of the playground to break up the neigborhood gang so that the play director could control the group more effectively, Friedberg's playgrounds are designed in an opposite way. There are no adult play leaders in his playgrounds, only adults passing by, i.e., public policing. Friedberg pulls the equipment together and organizes it in concentrated traffic patterns so that children are given more choices within a small area. These various pieces of equipment are like links in a chain that provide up and down or across choices of large muscle activity. As soon as a designer makes the decision to build a public playground, one that is visually exposed to pedestrian traffic, he has taken the path of no return. He has automatically cut out the use of most materials needed to create an environment which will be an ever-changing and constant source of exploration and discovery. The park-type designer is saddled with a demand for adult aesthetics by virtue of the playground's public nature. This is probably the most important initial design decision. It is no more appropriate to impose a junk aesthetic on adults than it is to impose a 'neat and clean' aesthetic on children. After this decision, other decisions logically follow which inhibit or encourage certain types of play. The junk playground is fenced in and locked-up when no adult is available and is characterized by an abundance of materials small enough for children to move; a public playground will have everything stolen if its materials are not permanently fixed. Thus, in a Friedberg playground, the child is limited to the manipulation of three things: his body, sand, and water (if the water has been turned on). These limitations have forced Friedberg to go back into the very limited, bureaucratic park tradition to correct the errors in its equipment. Friedberg's greatest contribution is the one least discussed. His parks are visually attractive, but they are urban in spirit, not rustic; he breaks

The responsive environment

443

company with the escapist English park tradition. Friedberg likes the city and wants to relate the playground to the surrounding urban fabric. He is in the pro-urban camp eloquently expounded through the writings of Jane Jacobs. His designs include areas for all ages. At Jacob Riis the park is divided into four sections, the playground, a j/iz-promenade, an amphitheatre, and a space that was to be an old people's gathering area away from the hub-bub and the noise of the playground. But old people would not go near the place he designed for them. Instead, they congregated near the playground to watch the children. Thanks to Friedberg's mistake, we have learned to locate geriatric areas near 'the action'. The original senior citizens' area is used by teens, who need a more private, intimate area to work out their social roles. By contrast, the amphitheater is for all ages; there are music fests, film festivals, drama productions, and political-community forums, somewhat like early-American town meetings. For the children, there is a waterfall and gutter-channels. Sycamore trees and pergolas enclose the amphitheater giving the area human scale and partial visual screening from the harsh architecture of the surrounding housing project. This canopy also serves as a climber for older children. Public School 166 is Friedberg's best example of a public park as a springboard into the city. Although the playground is a limited success, it is a meeting place for the community in general, and, more importantly, it is a point of departure for the transportation of children to various parts of the city for enrichment programs. Summary and conclusion. There have been two break-throughs in our thinking and practice with children and play in the 20th century. The first was a change in attitude and orientation from an adult to a child focus, with the accompanying understanding of child growth as occurring in developmental, sequential stages by the interaction between the child and the environment of the real world in every-day living. Secondly, the adventure playground concept aggressively encouraged children to manipulate an enriched, protean, moveable, changeable setting into new combinations of the child's choice and at his/her own level of development. It was a situation where the child was also exposed to and could come to terms with chance and the unexpected (reality) via junk and danger. Taken together we know the playground is a learning environment, a type of 'classroom', and its constituent parts comprise a curriculum: adults as facilitators and children as designers. It was a 'new day' in playground concepts when the user became the designer rather than

444

Environmental

conditions

either the patron or the landscape architect being the designer. The purpose of a properly designed playground, like the purpose of life itself, is not its deadly end result, but the quality of its ever-changing process. 'Why provide an enriched environment? Why provide help to overcome particular deficiencies? After all, don't all children end up with the same basic skills around eleven to fourteen years of age?' This view is spiritually bankrupt, for any sensitive person who has reflected upon his own childhood knows there were times when specific helps were needed and not available. As a result we felt differently about ourselves; we became different people.

Figure 1. Aerial equipment, with various ways to go up and down. Pacific Oaks, Pasadena, California Figure 2. Interaction with protean materials. Pacific Oaks, Pasadena, California Figure 3. The user designs an ever-changing environment. Notting Hill, London Figure 4. Small multi-level play space. Children's House (Wilmar), Madison, Wisconsin

ELAINE N. MCINTOSH

2

Effects of nutrition on development

A. ROLE OF NUTRITION IN HUMAN DEVELOPMENT by Elaine N. Mcintosh and Alexander R. Doberenz, University of Wisconsin, Green Bay

Nutritional status during early life can exert a lasting influence on subsequent physical development. Nutrition plays an important role not only with respect to physical growth and maturation but to the mental and behavioral facets of development as well. There are some data suggesting that malnutrition is associated with a decrease in both size and number of brain cells, depending on the timing and severity of the nutritional insult. That intellectual peformance is lower in severely malnourished children has been documented. However, since malnutrition occurs within a complex ecological setting, it is difficult to assess the extent to which malnourishment per se affects performance, compared with other environmental factors (Levitsky and Barnes 1973). Nevertheless, without optimal nutrition, the full expression of an individual's genetically determined potential cannot be realized. The effect of malnutrition on mental development takes on added significance as society becomes more complex. For example, the ability of a developing country to progress is directly related, in large part, to the intellectual capacity of its population. The ability to resist infections is influenced by nutrition, particularly among young children. Indeed, ability of the organism to withstand infectious diseases (stresses) of any kind is intimately interrelated with nutritional status throughout the life cycle. In considering the effects of

446 Environmental conditions nutrition on development, the entire life-span should be considered. In this paper, an attempt will be made to present a brief over-view of the effects of nutrition on physical development and disease during the various life phases. What is malnutrition ? Human beings require six basic nutrient groups for maintenance and growth: protein, fat, carbohydrate, vitamins, minerals, and water. In addition to other functions, protein, fat, and carbohydrate are the sources of energy for the body. Good nutritional status depends on adequate intake and utilization of the essential nutrients by the body. Malnutrition ('bad nutrition') results when nutrient intake is inadequate with respect to amount (under-nutrition) or kind, or if the nutrients are not properly metabolized. Malnutrition also results from the ingestion of too much food (over-nutrition) or from a nutrient imbalance. Individual countries have established standards for the recommended daily intake levels of various nutrients. These are based on research data obtained from large groups of individuals and therefore statistically relate back to populations as a whole rather than to individuals. Because of variations in metabolism, nutrient needs will vary among individuals and even in an individual from time to time. Thus, national dietary standards serve at best as a general guide with respect to nutrient intakes. Undoubtedly most people who experience nutrient deficiencies are subject to those which are 'sub-clinical' in nature, i.e., they are not severe enough to cause overt symptoms. Moreover, they usually occur in clusters, which render deficiency diagnoses even more difficult. In view of these considerations and others, it is easy to understand why the busy medical practitioner, when confronted with a patient who complains of various mild, general type symptoms of discomfort and malaise may simply prescribe a multiple vitamin-mineral supplement rather than having the patient undergo potentially costly and time-consuming tests. It is pertinent to note also that typically physicians are neither trained nor experienced in diagnosis of nutritional diseases. Fetal development. Nutrition plays an important role throughout the development of the embryo, including its maintenance, differentiation, and growth (Giroud 1970). During these stages, nutritional requirements vary. Since maintenance or survival requires low levels of nutrients, only very severe dificiencies will cause death of the embryo. Differentiation is less sensitive than growth to environmental influences. Therefore, if

Effects of nutrition on development

447

nutritional deficiencies occur differentiation may continue but growth is reduced. Differentiation occurs during the first 8 weeks and the fetal stage begins with the onset of the 3rd through 9th months. During the fetal stage, there is no production of new types of protein, but there is continued production of the same type of proteins produced during the embryonic period. During the embryonic stage, deficiencies can lead to malformations; during the fetal stage, deficiencies can lead to impaired growth. The malnourished fetus will exhibit a lower degree of organ development, including the brain. In addition, this fetus is more likely to have congenital abnormalities and problems of perinatal adjustment. The effects of malnutrition on cellular growth patterns are different among organs and appear to be more extensive when dietary restriction coincides with the period of most rapid cell division. The greatest rates of cell division and of growth occur in utero in some species but in the post-natal period in others. In man, birth occurs during the steepest part of the growth curve of the fetus. A number of studies have indicated a definite relationship among dietary deficiencies during pregnancy - especially protein and vitamins and the incidence of abortion, eclampsia, premature births, and neonatal deaths (Committee on Maternal Nutrition 1970, 1973). For example, either a deficiency or an excess of vitamin A can cause abnormalities (malformations). A lack of vitamin D can result in rickets in the human embryo. A lack of iodine leads to goiter, enlarged thyroid, and stunted growth of the fetus, with later development of cretinism (Giroud 1970). Contrary to popular belief, pregnancy does not constitute a true parasitic relationship, except possibly in instances of mild dietary deficiency of the pregnant woman: i.e., protection of the fetus does not appear to be unlimited. In a survey conducted in Boston in 1943 (Burke et al.) it was found that most of the infants with major congenital defects were born to mothers receiving nutritionally inadequate diets. If malnutrition occurs during the critical period of organogenesis, before the 10th week of gestation, a deficiency of moderate degree may result in a malformed baby (Millen 1962). The malnourished baby will probably be undersized. Winick (1967) has shown that intrauterine growth retardation is associated with poor placental growth which in turn is influenced greatly by poor nutrition. Numerous studies have indicated a positive relationship between maternal weight gain and new-born birth weight. A maternal average weight gain of 24 lb. is recommended as favoring an optimum course and outcome

448

Environmental conditions

of pregnancy, provided of course that the pregnant woman's weight is close to ideal at the onset of pregnancy. (Committee on Maternal Nutrition 1970, 1973). Weight gain which results from good nutrition rather than weight gain per se (Ademowore et al. 1972) is a most important factor. In certain cases, weight gain may be the result of edema or fat accumulation rather than to maternal and fetal tissue and fluid formation (Jacobson 1972). There is general agreement that under optimal conditions, the average birth weight for babies in this country is 3.3 kg., with a baby under 2.5 kg. being considered below the normal range and/or premature. It should be pointed out, however, that the baby 2.5 kg. and under may not be necessarily premature. This infant may be under-nourished, or born of an exceptionally small mother (e.g., under 46 kg.) (Ademowore et al. 1972). Early growth patterns. Despite individual variations as to timing and degree, the growth of a well-nourished baby follows a fairly predictable pattern from infancy through childhood and adolescence. Growth, though rapid during the first year, does not equal the fetal rate. The newborn may lose up to 10 % of his birth weight as a result of early postnatal adjustments, then gain rapidly at the rate of about 8 oz. per week during the early months. The average baby in the United States doubles its birth weight by 5 months or earlier, and triples it by the end of the first year. Growth then tapers off slightly during the last quarter of the first year, although occasionally one will observe a 'growth spurt' after about 9 months in a baby who has been breast fed and then introduced to solid foods. Starting with an initial length of approximately 48 cm., the baby grows 20-23cm. during the first year. The head increases rapidly in circumference during this period, and at six months, the head and chest in a normal baby are about equal. Thereafter, the chest begins to exceed the head in size, unless there are nutritional deficiencies such as proteincalorie malnutrition (marasmus). These measurements, along with the height and weight serve as good indicators of nutritional status in children. In addition, skin-fold thickness measurements can serve as indicators of fat-muscle ratio. Growth patterns of childhood. The growth rate tapers off during the second year, and the young child gains 3.5-4.5 kg. During the pre-school years, he gains only 1.8-2.2 kg. per year in weight. Thereafter, during the schoolage years (6-12) growth is slow but steady. A few children experience

Effects of nutrition on development

449

what is called the juvenile midgrowth spurt between the approximate ages of 6-8 (Tanner 1962). Some children during this period will experience a degree of 'catch-up growth', usually a child who previously was malnourished or experienced a chronic and debilitating illness. Failure to gain in height during infancy and early childhood may denote a serious lack of nutrients - particularly calcium, phosphorus, and vitamin D. Although not usually a problem, sufficient calories are also a necessity for normal growth to occur. Studies have shown that growth rate is not uniform throughout the year. Weight gains tend to be greater during the fall and early winter and less in the late winter and spring, when height gains tend to be greater. This is true at all ages, including adolescence. Diet, sunshine, exercise, nervous tension, and susceptibility to infections may all contribute to this irregular pattern. As adolescence approaches, the basal metabolism increases, and the child enters a second cycle of rapid growth. Nutrient needs increase markedly during this period. Unfortunately, despite its importance, this phase has not received the attention it deserves from parents and professionals involved in the care of the adolescent. In addition, the adolescent's eating patterns are often poor. The pre-adolescent growth spurt. All children in late childhood experience a year or so where, excluding infancy, growth exceeds that of any other year. This is the 'period of maximum growth'. The age of the individual growth spurt varies between sexes and individuals of the same sex. The female maximum growth tends to be approximately two years ahead of that of males. In females, the growth spurt may begin as early as 9 or as late as 12 years or more. In males, it may range from as young as 11 to 14 years or more. There can be up to 6 years difference between early and late maturation without any abnormality. Dr. Valadian and co-workers (1971) have reported that in both sexes, children experience greater height gain during the growth spurt when it is early than when it is late. In females, the peak of maximum growth is usually followed within a year by menarche. The male undergoes comparable but less dramatic sexual changes (Valadian 1971). Growth essentially ceases after females are about 16 years of age and after males are in their early 20s. However, the process of mineralization continues. This is very important from the standpoint of bone strength, and in females is important in the building up of calcium stores to support mineral needs during possible future pregnancies. Maximum calcification is reached by approximately 30 years

450 Environmental conditions of age. From then on, there is usually a progressive decrease in bone density. Segmental growth. Different parts of the body grow at different rates. The head, which earlier grew more rapidly than any other part of the body, changes little during the adolescent growth spurt and the growth that does occur takes place more in the face than in the skull. The feet achieve their full length three months after the growth spurt starts (Tanner and Taylor 1965). The growth of other parts of the body follow in the general order of chest, arms, and legs. Trunk length is the last to increase, after all other parts have ceased growing. This pattern of segmental growth leads to periods of disproportion between length of limbs and total height, resulting in clumsiness on the part of the rapidly growing adolescent. Effects of malnutrition on growth. Growth is a complicated and extensive process that requires good nutrition and is affected adversely by the lack of virtually any nutrient. Lack of protein, vitamins, or minerals will inevitably result in growth retardation. Adequate caloric intake is also a necessity for optimal growth. Macy and Hunscher (1951) have noted that a deficit of as little as 19 calories per kg. body weight was sufficient to cause failure in growth. Malnutrition is often associated with delay in both sexual and skeletal maturation. Generally speaking, malnutrition limits growth, delays adolescence, and decreases final size (Garn 1972). Malnutrition during infancy and childhood reduces the size of the individual prior to adolescence. Malnutrition, according to Garn (1972) delays adolescence so dramatically that it can be demonstrated in the thinner children of wellnourished populations. Garn and Frisancho (1970) maintain that true 'catch-up growth' is never really achieved, although the malnourished child does grow for a longer period of time than the one who has experienced continuous normal nutrition. According to their calculations, true 'catch-up' growth could occur if growth were extended for four additional years, rather than the usual additional year or so. They explain that Central American children, due to decreased growth-rate, experience a deficit of 20 centimeters, which at a growth rate of 5 cm. per year would require four additional years of growth. Instead, the added growth period is only a year or so, with the result that Central American adults are about 15 cm. shorter than normal.

Effects of nutrition on development 451 Achieving maximum growth potential. Animal experiments of Chow and Stephan (1971) indicate that dietary restriction of the mother during the gestation period alone yields growth-stunted pups. In malnutrition, the result is an individual who is sexually mature at a far smaller size. Garn (1972) has postulated that this results in an individual who is capable of supporting the conceptus possibly less long, with a resultant smaller birth weight. Failure of females to achieve maximum growth potential may also result in other problems. Mothers of short stature are more likely to suffer from pelvic disproportion than taller women, with resultant delivery problems, and increased perinatal deaths due to birth trauma (Cravioto and DeLicardie 1971). With respect to males, the importance of achieving full growth potential is more subtle. Studies have shown that taller men have a much greater likelihood of achieving high positions of leadership than men who are of less than medium height. Mitchell has reported in 1964 that improved nutrition can affect stature in a national group. Improved nutrition practices in Japan since 1948 have resulted in increased stature of school children. In 1962, males of age 14 were 3 inches taller and females of 11 were 2-1/2 inches taller than were children of those ages in 1948. There is a definite trend in the United States as well as in certain other parts of the world toward children growing taller and heavier, and maturing earlier. This trend is the result of several probable factors, including: (a) better nutrition; (b) better environment, e.g., better housing; (c) better medical care, including immunizations and sanitation. Other effects of malnutrition. Because of the rapid rate of growth with its accompanying high nutrient demands, the child is more susceptible to effects of nutrient deficiencies than is the adult. Moreover, the child lacks the body reserves which the well-nourished adult has built over time. Consequences of even severe malnutrition in adults have been remarkably transient (Cravioto 1963) in contrast to the infant who may incur permanent damage. Severe malnutrition results in a change in body composition which follows a general pattern: loss of fat, loss of nitrogen due to muscle atrophy, followed eventually by water retention. The retention of body fluids is due in large part to lessened plasma protein and resultant change in osmotic pressure. This tendency toward fluid retention may be aggravated by cardiac insufficiency, as in an advanced stage of beriberi (thiamine deficiency).

452

Environmental conditions

The five most common nutritional deficiencies on a global basis are (Lowenberg et al. 1968): (a) general under-nutrition; (b) protein-calorie malnutrition; (c) anemia - from several possible deficiencies; (d) blindness and eye conditions; and (e) dermatological disorders. In the United States there are four conditions which are sufficiently prevalent to call for special preventive efforts, particularly among the young (Foman and Egan 1972): iron deficiency anemia, dental caries, atherosclerosis, and obesity. Iron deficiency. Many studies indicate that iron deficiency is a common occurrence in children (ages 1-6), regardless of socioeconomic level, but the incidence tends to be greater among the poor. In addition to a correlation between anemia and lowered intellectual performance, the recent results of a study by Owen and associates (1970) indicate that iron deficiency is associated with growth retardation. They reported that children with growth below the 25th percentile had lower hemoglobin and transferrin saturation levels. Whether anemia is a cause of the growth retardation or is indicative of a chronic state of general malnutrition has not been established. Dental caries. During the growth period, control of dental caries is very important, since tooth formation is maximal during early and middle childhood. Tooth decay is a complicated and subtle problem involving many factors. Dental caries is a common occurrence in cases of malnutrition. The child may lack one or more of the following essential nutrients, each involved in the development and maintenance of teeth: calcium, phosphorus, vitamin A, vitamin D, and ascorbic acid. The incidence of cavities may also be reduced through the ingestion of fluoridated drinking water, and through a decreased consumption of sucrose (particularly in 'sticky foods'). Efforts should be directed toward diets which provide adequate and balanced amounts of the necessary nutrients (especially minerals and vitamins) and are low in carbohydrate, particularly in sucrose. The consumption of sticky types of candy between meals should be discouraged by substituting snacks of fruit, vegetables, and whole or skim milk. Obesity. It is well known that over-nutrition (over-eating) imposes needless strain on the digestive and excretory systems, in addition to leading to overweight and obesity. This causes also a resultant strain on the vital

Effects of nutrition on development

453

organs. Often this problem has its roots in habits which are established in babyhood and early childhood by parental influence based on ignorance. Many babies are over-fed by over-zealous mothers whose goal is maximum weight gain. Such a child learns early to develop an appetite that is not satisfied by the simple allaying of hunger; he learns to over-eat. Sweets are used as rewards and pacifiers in times of emotional stress, and these habits carry over into later life. Another factor which leads to obesity, even among our youngsters today, is our increasingly sedentary existence. Inactivity may actually be contributing more to over-weight than is overeating in many cases; this is well documented, particularly in the case of adolescents. Mothers need to be re-educated to appreciate that 'bigger is not necessarily better' with reference to youngsters, and activities which involve physical exercise and discourage eating should be followed. Atherosclerosis. Atherosclerosis is a condition wherein deposits of fatty material clog the arteries. These fatty deposits include cholesterol and related compounds and mineral salts. Cholesterol has its origin to some extent from the diet (exogenous sources) but to a greater degree from synthesis by the body (endogenous sources). The amount produced by the body will vary widely from person to person but is influenced greatly by the amount and type of fats deposited in the body. The more fat, the more raw material is available to the body for cholesterol synthesis. For this reason, atherosclerosis and obesity are associated. Dietary intake of cholesterol and long-chain, saturated fatty acids, as well as total fat in the diet are also factors in the development of atherosclerosis. Studies have shown that atherosclerosis builds up over a period of many years, even though it usually does not produce clinical symptoms until the middle years. Malnutrition and stresses. In addition to producing deficiency conditions, there is a synergistic effect of malnutrition and disease, including such conditions as measles, tuberculosis, intestinal parasitism, and gastrointestinal diseases. Often the incidence of an infectious disease such as measles will precipitate the onset of overt clinical symptoms of a chronic deficiency disease, such as kwashiorkor (protein deficiency) in a young child. The relationship between tuberculosis and malnutrition has long been known. Until the comparatively recent use of drug therapy, the usual treatment of this disease included a good diet and exposure to outdoor air. The relationship of parasitism to malnutrition is obvious, since the presence

454

Environmental conditions

of intestinal parasites further depletes a body which may already be low in nutrients. As a result of malnutrition, there are alterations in the morphology and function of the gastrointestinal tract (Hansen et al. 1971) which predisposes the individual to attacks of diarrhea. Although malnutrition may not directly cause a specific disease, it often influences its incidence, severity, and ultimate outcome. Such stresses include in addition to infectious disease, environmental factors, such as poor sanitation, personal hygiene, or lack of proper medical care. Unfortunately, during this period when the infant is suffering from an inadequate intake of nutrients, he or she is also subject to an unsanitary environment. This is at a time when the immunity obtained from the mother has been largely lost, and his or her own body has not yet developed its potential. Nutrition and aging. Aging is a continuous process of change throughout the life-span. As one ages, certain changes tend to occur, although at widely different speeds and to different degrees, among individuals. Many of these changes are related to the life-long nutritional status (particularly after 30 years of age) and not specifically to old age. Generally, there is an increase in, and redistribution of, body fat. There tend to be concomitant decreases in water content and in body cell mass. All of the metabolic reactions of the body tend to slow down. With this continuing reduction in 'basal metabolism' caloric needs are reduced by about 5 % per decade, amounting to about 20 % between the 20th and 60th years. Along with the general slowing down of body processes, there is decreased muscle tonus including intestinal tonus and motility. There may be loss of teeth (periodontal disease) or ill-fitting dentures which make chewing difficult. All of these factors and others make for lessened ability to digest and utilize food. With the reduction in acidity, there is lessened ability to absorb minerals. There may also be lessened ability to absorb or utilize folic acid and vitamin B 12 with resultant anemia. There may be some deterioration of vital organs as well, with the development of late onset diabetes and/or atherosclerosis, arthritis, and osteoporosis. All of these conditions are nutritionally related. Common nutritional deficiencies of the aging include calcium, thiamin riboflavin and ascorbic acid. It is a comnon belief among older adults that they don't need milk since they are 'full grown'. Failure to drink milk, along with reduced utilization of calcium, may be major causes of calcium deficiency and idiopathic osteoporosis among the elderly. There may also be an upset of the calcium metabolism among aging females as the result

Effects of nutrition on development

455

of hormonal changes following the menopause. Riboflavin deficiency, which can cause sore and watery eyes, and in extreme cases, blindness, may be a reflection of lack of milk and green leafy vegetables. Chewing problems may discourage the eating of meat, resulting in a decrease of both riboflavin and thiamin intake. Ascorbic acid deficiency is usually the result of too little intake of citrus fruits, although there are also substantial amounts of this vitamin in tomatoes and in such vegetables as potatoes and cabbage. Ascorbic acid deficiency also contributes to the gum problems (periodontal disease) of the aging. Vitamin and/or mineral deficiencies may also cause various skin conditions or nervous reactions. Since nutrient deficiencies tend to occur in clusters, the above outlined symptoms rarely occur singly. Nutritional deficiencies tend to aggravate any existing degenerative conditions associated with the aging process, such as cardiac conditions or cataracts. Nutrition and longevity. Tappell (1969) has proposed that certain antioxidants in the diet such as vitamins E and C, selenium, and various sulfhydryl compounds exert a protective effect against the process of aging. These compounds are claimed to retard harmful oxidative reactions in the cell which have a net aging effect. However, beneficial effects in man are not yet proven. A number of studies with rats, beginning with the classical work of McCay and associates (1935), have indicated that caloric restriction, particularly during the early periods of life, contributes to a lengthened life span. Knittle (1973) has noted that if undernourished animals are able to survive the first 3 or 4 months, they have a lengthened life span. Evidence from the work of Ross (1969) indicates that rapid growth rate shortens life expectancy in rats. The work of other investigators suggests that there may be more resistance to disease and lowered incidence of degenerative conditions among calorically restricted animals (Mickelsen 1971). The classical work of Sherman (1952) indicated that when supposedly adequate diets of white rats were improved by adding more calcium, ribflavin, and vitamin A, there was increased growth, a prolonged reproductive period, and a longer life. It would appear that a modest curtailment of total caloric intake coupled with a diet adequate in minerals and vitamins (including those having antioxidant qualities) is commensurate with prolonging the 'prime of life' and lengthening the total life span. It is generally agreed that the desirable goal is not to prolong the individual's life span per se but to prolong his or her period of productivity. In other words, rather

456

Environmental conditions

than putting 'more years into life', good nutritional practices should serve to put 'more life into years'. It is clear that adequate nutrition is essential for optimal physical growth and development. Good nutritional status enhances the human being's capacity to function throughout life. The ability of the human organism to withstand stress is markedly related to nutritional status. Of all the important environmental factors impinging upon growth and development, nutrition is the one most readily controlled and manipulated by man.

Effects of nutrition on development

457

B. MALNUTRITION IN INFANCY A N D INTELLECTUAL DEVELOPMENT by Marjorie F. Elias,* Harvard University, Cambridge

Children who are born into conditions of extreme poverty experience from their first hours a world that is different from the world of children born into affluence. Later in infancy and childhood those who have survived the hazards of infection and starvation are often impaired in their intellectual functioning. The deprivation suffered by such children could include every kind of social and psychological hazard as well as nutritional hazard. Yet it is frequently asserted that the intellectual impairment is caused by malnutrition. The assertion has significant consequences for policy in many parts of the world. It implies that of all the adverse influences to which the baby is exposed, it is the nutritional insult which is critically damaging. It also implies that feeding programs for infants will significantly improve long-term functioning as children grow up even if all other elements of their life experience are left unimproved. Extensive work has been carried on in the past decade on the problem of the relation of malnutrition to mental development in children. The problem has not been solved but some of its elements have become clearer. The purpose of this review is to trace the emerging clarification of questions. Studies published before 1970 were broadly exploratory and were addressed primarily to the question: Is there a relation between undernutrition in infancy and later intellectual functioning? By 1970, the answer to that question was becoming clear. There appeared to be such a relationship. It was demonstrated in many parts of the world and under many different conditions. However another element had also emerged whose importance had not been anticipated. That element was the complexity of factors occurring together in conditions of extreme poverty. Malnutrition was only one factor among many. In the early studies investigators had looked only for malnutrition as a factor associated with low test scores and had found a relationship. When they looked at nonnutritional factors, they found that children who were deprived nutritionally * The author wishes to thank Dr. Johanna T. Dwyer, Dr. M. Guillermo Herrera and Dr. Kenneth W. Samonds for their helpful criticism of this paper. This review was prepared as part of on-going studies supported in part by a grant from the Fund for Research and Teaching, Department of Nutrition, Harvard School of Public Health.

458

Environmental conditions

were also deprived in many other ways. Therefore it was not justifiable to attribute the intellectual deficiencies to malnutrition alone. By 1970, then, the question to be addressed in studying the relation of malnutrition to intellectual development had changed. It had become: What part of the variance in intellectual development is attributable to nutrition? The major studies undertaken since that time have attempted to deal with this more complex question. This question implies recognition of the presence of confounding factors along with the nutritional factor. The studies before 1970 have been well reviewed (Pollitt 1969; Ricciuti 1970; Frisch 1971; Klein, Habicht, and Yarbrough 1971; Kaplan 1972; and Birch 1972) so the focus of the present review is on reports published since that time. The primary issue in the human studies is that of confounding of nutritional with non-nutritional factors. Confounding does not arise as a severe problem in animal studies where laboratory control is possible, but it is a central problem in human studies in which malnutrition is studied as it occurs spontaneously in the community. For this reason, discussion in this paper is limited to the confounding issue and does not cover the problem area of possible critical periods of nutritional insult in infancy. Let us consider some recent studies in the light of the question with which they are concerned: What part of the variance in intellectual functioning can be attributed to nutrition? The problem is discussed from many points of view in a book edited by Kallen (1973). Four ways of dealing with confounding in an experimental design will be discussed: matching of subjects; identifying groups in terms of both nutritional and non-nutritional criteria; statistically separating the parts of the variance of test scores associated with nutritional and nonnutritional factors; and direct intervention with food supplementation and pre- and post-tests on the study population. Each of these ways of dealing with the confounding problem has been used in recent research projects. Matching. Matching of pairs of subjects is a technique used frequently to make experimental and control groups as similar as possible in all variables except the variable being experimentally manipulated. Chase and Martin (1970), in a study carried out in Denver, Colorado, matched each subject identified as malnourished with a control subject who was similar in birth date, birth weight, sex, and race. They found delayed development in the malnourished group, especially in those in whom the

Effects of nutrition on development

459

episode of severe malnutrition occurred at an older age. However, as described above, the authors found that the groups varied systematically on many social variables as well as in nutritional status. Matching, in that case, proved to be ineffective in controlling for non-nutritional differences. Sibling matching was used by Evans, Moodie, and Hansen (1971) in a study of survivors of kwashiorkor in South Africa. It was expected that use of well-nourished siblings as controls would control for many of the non-nutritional factors that influence intelligence since siblings share both genetic endowment and rearing experience. The authors found that children whose episode of severe malnutrition occurred before 15 months of age did not differ from their sibs on a South African adaptation of the WISC (Wechsler Intelligence Scale for Children) or on the HarrisGoodenough Drawing Test. Those whose illness had occurred at 16 to 48 months of age tested lower than their sibs on solving of arithmetic problems in the WISC and on the Harris test. Their achievement in school was also inferior. Unfortunately the nutritional status of the siblings in early childhood was not known except for the fact that they had not been hospitalized, nor were any other differences known which might have occurred between siblings. Sibling matching was used in a pilot study for a longitudinal malnutrition study in Colombia but was abandoned (Christiansen and Herrera, in preparation). Sibship was found to account for only about 25 % of the variance of Griffiths IQ scores, so other methods of statistical control were necessary. But these could not be easily applied in a matched sibling design because of different ages of sibs and the inaccessibility of the past history of the older sib. It was also found that families which contained both well- and malnourished children were hard to find and seemed to be atypical in various ways. The authors concluded that sibling matching was not an effective control mechanism in their study and that it imposed limitations on other means of control. If the matching strategem did not control for non-nutritional factors, the studies by Chase and Martin and by Evans et al. can be considered as comparable to the older studies in which two naturally occurring groups were compared. As such, they both offer evidence in contradiction with the commonly held hypothesis that early infancy is a particularly vulnerable period for lasting damage from a syndrome of deprivation. Use of both nutritional and non-nutritional criteria. Another way of coping with the problem of confounding variables is identification of groups in

460

Environmental conditions

terms of both nutritional and non-nutritional criteria. This method was utilized by Pollitt (1972) in a study of mild and moderate malnutrition carried out in Peru. Height was used as an indicator of nutritional deprivation and three bio-social factors (mother's education, mother's parity, and family stability) as indicators of environmental deprivation. By grouping children by both height and bio-social criteria, Pollitt was able to contrast four groups in a 2 x 2 factorial design. The cognitive level of the subjects was assessed by a categorization task. Thirty-nine children of 6 months to 4 1/2 years of age made up the sample. The doubly deprived groups, those both short in stature and coming from families who were low on the bio-social index, were found to be significantly less competent at the task than were the other three groups. This result suggests that deprivation in nutritional and other factors might interact in such a way that a combination of factors is required to impair functioning. Such a hypothesis would explain the ubiquitous appearance of both nonnutritional deprivation and malnutrition in children of low IQ. If this were true, improvement of conditions in either nutrition or the home environment might serve to ameliorate the behavioral damage caused by multiple deprivation. A difficulty with this design was mentioned by the author (personal communication). Stature appeared to be linked to bio-social conditions. He found it difficult to find subjects to fill the two groups requiring tall children with low bio-social ratings and short children with high biosocial ratings. The possibility that both are results of a common cause weakens the interaction hypothesis. Statistical techniques. One can use statistical techniques to attack the problem of confounding after having gathered data in the usual fashion on groups of well- and malnourished children as they occur in the population. Such an approach is being used in three longitudinal studies underway in Mexico, Guatemala, and Colombia. All three studies focus on mild and moderate malnutrition as it occurs in populations in which it is prevalent. The Mexican and Guatemalan studies are studying rural populations; the Colombian study is of an urban population in low-income areas of Bogota, the largest city in Colombia. DeLicardie and Cravioto (1972) have used statistical techniques in reporting results from a longitudinal study of the entire birth cohort of one year in one village in Mexico. The study design is described elsewhere (Cravioto 1972). Nutritional status was assessed by anthropometric methods. Social status was assessed by a home stimulation inventory

Effects of nutrition on development

461

developed by Caldwell. At the age of six months, before the onset of malnutrition, 19 children who subsequently suffered from severe malnutrition lived in homes which were rated significantly lower on the inventory than were the homes of children who remained well nourished. At 48 months, after rehabilitation, the previously severely malnourished children still lived in homes that were significantly inferior. The whole cohort was assessed for nutritional status by height, for social conditions by the home stimulation inventory, for intellectual development by a test of comprehension of concepts of opposite developed by Palmer. Associations between the variables were significant, but not strong. The correlation between social conditions in the home and intellectual development was .20 by a simple correlation. Nutritional status was correlated with the intellectual measure .26. When nutritional status was partialled out, the partial correlation between social status and the intellectual measure fell to .15. The partial correlation of nutritional status and intellectual development with social status partialled out was .23. So both nutritional status and social conditions as measured here account for only about 4 % to 5 % of the variance of intellectual scores. Klein, Freeman, Kagan, Yarbrough, and Habicht (1972), working in Guatemala, approached the problem in a comparable way but used different measurement techniques. The project is a longitudinal study of different villages in which some villages are given a nutritional supplement and others are not. The results published so far are cross-sectional in design and do not include the effects of the supplement, so they investigate a population in which mild and moderate under-nutrition are prevalent under naturally occurring conditions. The study design is described elsewhere (Canosa, Salomon, and Klein 1972). Nutritional assessment was by height and head circumference. Social measures included quality of house, father's occupation, mother's dress, mother's hygienic practices, instruction of the child in household tasks, and number of social contacts of the family. Intellectual measures used were developed for this study. They included language (naming pictures and stating relations between concepts), short-term memory for digits, and perception of embedded figures. Correlations between social measures and intellectual measures ranged from .17 to .39. Those between nutritional and intellectual measures ranged from .18 to .36, virtually the same range. No significance figures were reported for the correlations. Some of the correlations appear to have been slightly higher than are those of deLicardie and Cravioto, but

462 Environmental conditions they were still modest in terms of the proportion of variance of test scores explained. The parts of the variance attributable to nutritional and social factors were again about equal. A third longitudinal study being carried out in Colombia by Mora et al. (1973) in collaboration with Harvard University has provided comparable findings in a cross-sectional analysis of data taken before an intervention program was undertaken. A sample of children was drawn from an urban population in which mild and moderate malnutrition was prevalent. Nutritional status was assessed by height and weight; social status by a variety of social variables including mother's teaching of the child, socioeconomic status of the family, mother's age, and mother's attendance at Mass; and intellectual development by the Griffiths test of IQ. In this study the correlation of social with intellectual measures was .41. Nutritional status was correlated with intellectual measures .46. The correlations were significant at the .05 level. The relationship between factors was slightly stronger than in the other studies, but again the two factors of social and nutritional status were close to equal in strength. The proportion of the variance of intellectual scores explained by both factors together was 31 %. The proportion explained by nutritional measures alone was 14 %. Results in these three studies showed that neither the nutritional nor social measure predominated in the strength of its relation to intellectual functioning. The proportion of the variance of test scores which was explained by nutritional factors was not large, ranging from 5 % to 14 %. Nutritional and social factors in combination accounted for no more than 13 % of the variance. One must conclude that a great proportion of the variance of scores, and consequently of the explanation of impaired intellectual development in mildly malnourished children, must be due to factors which were not measured in these studies. Direct intervention. Direct intervention by means of food supplementation is yet another technique for separating environmental and nutritional factors. If both well- and malnourished children are randomly assigned to supplementation and control groups, and tested before and after a period of supplementation, changes in scores will indicate whether the supplementation program had any influence on performance. Mora et al. (1973) report the results of such an intervention in the Bogota sample. The food supplement was given for one year, and children were tested twice with the Griffiths test at an interval of one year. No change was

Effects of nutrition on development

463

found over the year for children who were originally well-nourished whether or not they received the supplement. For the malnourished children, improvements were greater for those who were supplemented. It was concluded that the improvements appeared to be associated with the supplementation program, but the conclusion was tentative because of a difference in initial scores between the supplemented and control groups. Another intervention study is being carried out in Cali, Colombia, on a sample of urban preschool children (McKay, McKay, and Sinisterra 1973). The focus of that study is on a combination of nutritional supplementation and environmental stimulation by means of a day-care program. Results are preliminary as yet since they do not have significance levels attached to them. Changes from pre-test to post-test seem to indicate that children given nutritional supplementation and a day-care program improved sharply in measures of school-related achievement such as use of letters. They attained a level of achievement comparable to that of children from some of the most advantaged families of Cali. This evidence, while not conclusive, suggests that intellectual rehabilitation is possible during the pre-school years when intervention in both nutritional and non-nutritional factors is given. It is to be hoped that more data will come in from all the longitudinal studies to add to these first findings. Unfortunately intervention studies have problems of their own. It is not known whether the effect found is due to the intervention itself or to other aspects of the program such as giving the families psychological support or making money available for them to use in other ways when some of their weekly food is provided free. Conclusions. In conclusion, studies reported since 1970 have shown an increase in sophistication of research design. The problem of covariance between nutritional and non-nutritional factors in the life experience of children reared under conditions of extreme poverty and deprivation has been recognized and dealt with in various ways. Matching of siblings or age-mates was not found to be an effective method. Identification of groups into high and low ratings on each dimension was effective, but the non-consistent groups were hard to fill. Pollitt's study using that technique revealed that only the doubly deprived were significantly impaired by their deprivation, but that deprivation in nutrition alone or bio-social environment alone was not damaging. Statistical separation

464

Environmental

conditions

of variables produced consistent results in three studies, but they suggested that the nutritional and social factors amenable to measurement could account for no more than one-third of the impairment in test scores so commonly found among malnourished children. Nutritional intervention was found tentatively in one study, and combined nutritional and environment stimulation in another study, to have ameliorating effects on such impairment. The picture that emerges is of a complex interaction of factors. Both nutritional and environmental deprivation have slight effects. When examined in a two-way design, double deprivation has a stronger effect. Prevention of one type of deprivation helps the most impaired children. So once again the question to be posed changes. The question dealt with by the studies described above was: What part of the variance in intellectual functioning is attributable to nutrition? The evidence described here suggests that in mild or moderate malnutrition the answer may be: very little. Yet the fact is well documented that children reared under conditions of deprivation are intellectually impaired. If nutritional deficiencies alone account for no more than 14 % of the intellectual impairment, and nutritional plus measurable social deficiencies account for no more than onethird of that impairment, what might account for the remainder? As we learn more, the question changes again. It is now becoming: What interactions among variables in a syndrome of deprivation are critical in order to produce impairment? This question is followed closely by its corollary: What types of intervention are effective in preventing or reversing such impairment? These are the questions which must be faced now if we are to use wisely the limited resources available for improving the lives of millions of under-nourished children.

Effects of nutrition on development

465

C. PROTEIN MALNUTRITION IN MONKEYS* by David A. Strobel, University of Montana, Missoula, and Robert R. Zimmermann, Central Michigan University, Mount Pleasant

The primate model. Attempts to characterize the effect of nutritional deficiencies on the behavior of developing organisms have centered on the possibility that inadequate nutrition leads to intellectual impairment or mental retardation. There is a considerable accumulation of evidence from human studies which supports this supposition. Children with a history of malnutrition appear to show defects in visual, motor, and pattern perception (Stoch and Smythe 1963, 1968); perception and abstraction (Champakan et al. 1968); sensory integrative ability (Cravioto et al. 1966); categorization sorting (Brockman and Ricciuti 1971); and general IQ scores (Cravioto and Robles 1965; Mönkeberg 1968). Furthermore, many investigators believe that the relative permanence of these mental deficiencies, even in the face of dietary rehabilitation, is dependent upon the timing and severity of the nutritional insult (Cravioto and Robles 1965). Human studies, however, have been plagued with problems of methodological complexity. The malnourished children, in addition to inadequate nutrition, tend to live under conditions of social degration, including poor sanitation, parental neglect, illegitimacy, and alcoholism. Even the most elaborately controlled experiments have not been free from the confounding of social and environmental variables that may have substantially contributed to observed differences in the performance of the subjects under study. Because of the difficulty in controlling these extraneous variables, many reviewers are in agreement that the present scientific evidence is inconclusive in defining the relationship between malnutrition and intellectual development (Coursin 1967; Frisch 1970). The use of experimental animals provides a number of distinct advantages in comparison to humans as subjects in nutritional research. The animal models afford control and specificity with respect to the genetic pool, experimental diets, onset and duration of the nutritional manipulation, and some definition of the environmental and experimental history of the * Preparation and original research presented in this manuscript was supported in part by grant No. 401 from the Nutrition Foundation, Inc. and grant No. HD-04863 from the National Institute of Child Health and Human Development.

466 Environmental conditions subject. Long-term effects of early protein malnutrition on maze learning performance and classical conditioning, for example, have been reported for rats (Griffiths and Senter 1954; Barnes et al. 1966; Zimmermann and Wells 1971; Levitsky and Barnes 1970, 1972) and swine (Barnes et al. 1967; Barnes, Moore, Reid, and Pond 1968). While the animal model provides a reasonable amount of control over the confounding variables of cultural and socioeconomic conditions and environmental deprivation, there are interspecies differences in response to malnutrition. Furthermore, the value of animal research in understanding the human nutritional effects is only as valuable as the validity of its analogue to the human situation. Non-human primates are logically the best compromise between man and animal for evaluating the effects of malnutrition on behavior. While the brain of the newborn infant rhesus monkey is close to two-thirds its adult weight, compared with one-third for the human infant at birth, the periods of rapid brain growth that appear sensitive to nutritional insult are more comparable between monkeys and humans than between monkeys and sub-primates (Dobbing 1968). Monkeys, similar to humans, have relatively protracted developmental histories and do not show maximal intellectual maturity before sexual maturity (Zimmermann and Torry 1965). Monkeys have a behavioral repertoire which is sufficiently advanced to provide a meaningful comparison with the complexities of the human condition. Rhesus monkeys may demonstrate simple instrumental learning and classical conditioning but are also capable of higherorder conditioning, performing complex problem solving tasks, perceptual, and concept learning problems. Finally, although pitted edema is a rarity in protein malnourished monkeys, except under special conditions, all other essential clinical signs associated with human kwashiorkor or mid-mild protein malnutrition have been reproduced in these animals (Ramalingaswami and Deo 1968). Acute versus chronic malnutrition. A large proportion of human studies tested their subjects during or following hospitalization for malnutrition or diseases associated with malnutrition. Children requiring hospitalization likely represented a population of severely malnourished subjects that presumably would have perished without medical treatment. The vast majority of under-nourished or malnourished children, however, are estimated by the World Health Organization (1965) to be suffering milder,

Effects of nutrition on development 467 but chronic forms of the disease. Lactational efficiency appears to protect the infant from early malnutrition, but after six months, or so, the child requires a supplementary diet. It is during this period and following that the growth rate of the child begins to fall in most economically impoverished societies and the consequences of chronic malnutrition may be expected. The present experimental program with developing monkeys was designed to evaluate changes in chronically malnourished animals during the interval between weaning and sexual maturity. These behaviors were then re-examined after dietary rehabilitation to determine if adult or maturing behavior was altered. Experimental groups and treatments. All of the subjects were laboratoryborn rhesus monkeys that were reared under a number of different environmental and dietary conditions. For the present series of experiments, only four groups of monkeys will be described. All of these animals were weaned to a solid food diet containing 25% protein (casein) by weight following maternal separation at 90 days of age. Two experimental groups were formed by placing five and six animals on a diet containing 3.5 % casein, at 120 and 210 days of age, respectively. Two control groups of four animals each were continued on the 25 % casein diet. Details of the diet composition and preparation are described by Geist, Zimmermann, and Strobel (1972). The monkeys on the low-protein diet showed a significant reduction in weight gain, total blood serum proteins, and serum albumin levels compared to controls within the first 38 days following onset of protein deprivation. These values did not return to normal over the four-year period the low-protein monkeys were on the diet (Geist et al. 1972). Anthropometric measures also showed the malnourished animals to be retarded in body length and head circumference. The malnourished animals adapted readily to the diets, remaining relatively healthy throughout the period of deprivation, and consuming food proportional to their body weights. Complete test batteries were given to the animals during the four years of protein deficiency and were designed either with nutritive or nonnutritive reward contigencies. In general, tests of curiosity and manipulation indicated that the malnourished animals appeared less interested in exploring and manipulating their environment than controls. The malnourished monkeys showed poorer performance, for example, on tests

468

Environmental conditions

of chain manipulation (Strobel and Zimmermann 1971) and puzzle solving. This reduction in manipulation was not thought to be merely the result of a decrease in general activity, as the groups did not differ significantly in movement around a cage (Geist et al. 1972). However, when food rewards were provided for puzzle solving, the malnourished monkeys showed almost perfect performance (Aakre, Strobel, Zimmermann, and Geist 1973). High rates of puzzle solving behavior for all groups could also be maintained by partial food reinforcement. The malnourished animals showed an immediate drop in performance when the food was removed completely, while high-protein animals continued to respond at high rates. This suggests that the malnourished monkeys were highly sensitive to food reinforcement contingencies compared to controls who appeared also to respond for the intrinsic value of the manipulation itself. While the malnourished monkeys were fed more of their diet than they would consume in a day, they showed a heightened motivation for food rewards. They demonstrated a consistent preference for high-protein foods in a self-selection experiment (Peregoy, Zimmermann, and Strobel 1972). Furthermore, in tests of dominance, the low-protein monkeys scored higher on tests involving competition for food but were inferior to control animals when dominance was measured under conditions of free social behavior or shock avoidance (Wise 1972; Wise, Zimmermann, and Strobel 1973). The malnourished monkeys also showed abnormal social behaviors when placed in an open room with animals of their own age and dietary condition (Zimmermann, Steere, Strobel, and Horn 1972). They showed fewer social contacts, less play and sexual behavior, and less grooming per interaction than high-protein animals. When the malnourished animals did make social contact, their behavior was characterized by fear and aggression. In terms of reaction to stress, the low-protein animals had lower shock thresholds than the high-protein monkeys (Wise and Zimmermann 1973). In addition, the malnourished animals showed fear and avoidance reactions to novel objects or novel situations (Zimmermann, Strobel and Maguire 1970; Strobel and Zimmermann 1972; Peregoy et al. 1972). In standard tests for learning set and long- and short-term memory using a Wisconsin General Test Apparatus (WGTA), the performance of the malnourished animals on the restricted diets was either indistinguishable or superior to controls. Zimmermann (1969) found that malnourished monkeys with a history of under-nutrition were superior to controls on a

Effects of nutrition on development

469

learning set task requiring the long-term memory of 100 pairs of objects. This relationship occurred prior to dietary rehabilitaion, but not after. Similar results were found using 50 pairs of stimuli (Zimmermann 1970). The 210 high- and low-protein groups were tested on a short-term memory task in which the response of the animal was delayed over intervals up to 60 seconds. The delayed-response task has been used as a test that was sensitive to certain types of brain damage in primates (Harlow et al. 1952). The low-protein monkeys were consistently superior to the controls under a number of delay conditions. The tests were repeated 28 months later, but this time the high-protein animals were mildly deprived of some of their daily rations. Under these conditions, the performance of the two groups became indistinguishable. Finally, Zimmermann (1973) found that monkeys tested before and within a month after going on the experimental diets did not show any learning deficit in reversal learning problems. Experiments on attention. The evidence from traditional WGTA tests would indicate that the protein-malnourished monkeys did not have impaired memory or deficits in learning ability. Only a single experiment showed a break from this pattern. A group of 150-day-old monkeys, on the protein-deficient diet for 60 days, was found inferior to controls on reversal problems, but not in acquisition. It was hypothesized that the primary difference between this experiment and other studies of reversal learning in the WGTA was the manner in which the stimulus objects were constructed. Deficits in reversal learning were found when the objects were mounted on identical plaques of hard board. Poorer performance on reversal problems did not occur between malnourished animals and well-fed controls when the objects were not mounted on plaques. In all tests of learning with the WGTA, the animal was presented with a pair of objects on a delivery tray and the aninal made a choice by pushing aside the object for a food reward located under one of the objects. Schuck, Polidora, McConnell, and Meyer (1961) demonstrated that if the cue indicating the correct response was spatially disparate from the locus of the animal's response, some impairment in learning resulted. Virtually all of the monkeys touched the edges of the hard board when they responded to the objects mounted on plaques. Presumably, the monkeys look at where they put their fingers. Schuck (1960) proposed a decreasing gradient of visual sampling around the region of the monkey's hand during the response.

470 Environmental conditions To test the hypothesis that the observed decrements in reversal learning were the result of the failure of the malnourished monkeys to sample the object cues at the center of the hard-board plaques, a series of stimuli was constructed with the discriminative cues located either centrally or peripherally. The monkeys were trained to respond to either the black or white cues on a series of stimuli that varied the area of the discriminative cue from 100 % to 5 % of the stimulus card. It was hypothesized that decreasing the area of peripherally located cues would have less effect on reversal learning than decreasing the area of centrally located cues if the monkeys had a narrow gradient of visual sampling centered at the edge of the stimulus card. As expected from earlier studies, the groups did not differ in the rate at which they learned the discrimination. However, on reversal trials the protein-malnourished group showed a significant decline in performance on centrally located cues as the area of the cue decreased. These results suggested that the difficulties incurred by the proteinmalnourished monkeys on reversal problems were the consequence of narrowed attention. To further test the generality of this hypothesis, the separation between the location of the discriminative cue and the locus of the response was radically extended. A sign conditional discrimination problem was also established for the WGTA. In this problem, two grey hard-board plaques covered food wells on either side of a stimulus card holder. On any one trial, the presence of the triangle was the cue for the animal to displace a specific plaque, for example on the right. On a successive trial, the square signaled the location of the reward under the opposite plaque, for example, the left-hand position. None of the younger low-protein monkeys reached the criterion for learning established at 23 out of 30 correct responses for two consecutive days. Only two malnourished monkeys altogether reached criterion, and it required over 1,700 trials. All of the high-protein monkeys learned this problem, averaging less than 800 trials to criterion. A third experiment was conducted to determine if malnourished monkeys could discriminate between forms that were masked or hidden within a more complex perceptual field. The monkeys were tested using the method of equivalence procedure Butter (1966) used successfully in training primates to disembed. The experimental and control animals were trained to discriminate between a triangle and square during concurrent presentations of these stimuli. After learning this problem to a criterion of 90 % correct responses, the animals were tested on stimuli

Effects of nutrition on development

471

that contained the original shapes, but on a background of irrelevant line patterns. The training stimuli were rehearsed between randomized presentations of the test stimuli until all pairs had been presented 20 times. The 20 test trials were divided into blocks for analysis of immediate transfer and subsequent learning to disembed. The high-protein animals transferred their responses above chance in the first 10 trials on each stimulus and then showed improvement with practice over the next 10 trials. The malnourished animals, on the other hand, did not score significantly above chance on the first 10 trials and showed a decline, or no change, in performance with further practice. Performance on rehearsal training stimuli during the test sessions averaged between 75 % and 80 % correct responses for all groups. It was hypothesized at this time that the apparent narrowing of attention in protein-malnourished monkeys was the result of their heightened interest in food rewards. That is, the food-oriented behavior was channelized or focused on locations that consistently yielded reinforcements in the past. According to this hypothesis, the protein-deprived animals should have particular difficulty on tasks which maximized response-reinforcement disparity. A task called the pattern strings problem was used to test this hypothesis. The monkeys were trained to pull on one of two chains to obtain food reward at the other end of the chain. All groups were tested on parallel, crossed, and pseudo-crossed patterns, in that order. If the malnourished animals had difficulty in understanding the perceptual relationships, then poor performance on the pseudo-crossed patterns might be predicted. However, if the proximity between reinforcement and response was the critical determinant of the choice response, then the parallel and pseudo-crossed problems should not be difficult for either high- or low-protein animals. In comparison, the crossed-pattern problem would be difficult for the malnourished monkeys, because the correct response occurred at the point farthest from the location of the reward. The low-protein monkeys were found to be significantly inferior to the controls on the crossed patterns across the 400 test trials. On the other hand, no significant differences were found in the parallel and pseudo-crossed patterns. Nutritional rehabilitation. The group of malnourished monkeys that were placed on the protein-deficient diets at 210 days of age was gradually shifted to a diet containing 25 % protein by weight at 1,500 days of age. This dietary transition took approximately eight weeks, and the percentage

472

Environmental conditions

of protein was increased by approximately 4% increments per week. Weight gain total blood serum protein levels returned to near control levels following 180 days of dietary rehabilitation. Preliminary results indicated that preference for high-protein food in the self-selection apparatus disappeared during the rehabilitation interval (Pettus 1973). Neophobic reactions to unfamiliar objects in the self-selection situation persisted for 60 days of rehabilitation but disappeared within six months. Finally, the conditional learning experiment was repeated using the identical procedure as before approximately 240 days after onset of rehabilitation (Lindvig and Zimmermann 1973). In contrast to previous behavior exhibited under the conditions of dietary deficiency, the performance of the rehabilitated low-protein monkeys was indistinguishable from controls in learning the conditional response. Discussion. A complete picture of the dynamics of adaptation and response to the conditions of long-term protein deficiency in non-human primates is speculative and premature. Nevertheless, examination of the present results is suggestive and provides a relatively consistent composite for interpretation. During the period of nutritional deprivation, the experimental animals adjusted to the loss of protein by a reduction in weight gain and by limiting their intake of the diet. Concomitantly, they showed increased interest in other sources of food. In particular they demonstrated a pronounced preference for foods containing higher concentrations of protein. The protein-malnourished animals showed a reduction in environmental manipulation or exploration when these operations did not have appetitive consequences. On the other hand, they readily performed these same manipulations if they resulted in food reward. Sensitivity to appetitive reinforcement contingencies may be hypothesized to be fundamental to the malnourished state. The channelizing of energy expenditures toward food-directed behaviors may additionally have influenced the responsiveness of the malnourished animals to social initiations. The consequences of failure to respond appropriately in social situations is often a vicious circle, resulting in a variety of abnormalities in development and maturation (Mitchell 1970). Zimmermann et al. (1972) suggested the possibility that protein malnutrition produced functional isolates. Aggression with brutality and excessive fear responses to novel stimuli are possibly the symptoms of the syndrome.

Effects of nutrition on development

473

The malnourished monkeys were not found inferior to high-protein monkeys on traditional learning problems in the WGTA and often showed superior levels of performance. Low-protein monkeys were thought to have higher levels of food motivation than well-fed controls. Barnes, Reid, Pond, and Moore (1968) also described an increased interest in food by malnourished, as well as previously malnourished, pigs. The results of tests which minimized stimulus-response or stimulusreinforcement spatial contiguity strongly suggest that protein-malnourished monkeys had difficulty on tasks where increasing demands were placed on attentional processes. It may be tentatively hypothesized that these animals were highly food oriented and developed a narrowed span of attention, focused on the shortest route of the goal and on locations with a history of 'pay off'. Klein et al. (1969) reported that previously malnourished children sometimes did poorer than controls on perceptual tests and short-term memory tasks, including embedded-figures problems. They hypothesized that these differences may not have resulted from altered cognitive capacity or intellectual ability but rather from lack of attention or willingness to concentrate on the task. These findings are consistent with the present interpretation, except that the malnourished monkeys did not maintain poorer levels of performance than controls on retests following rehabilitation. The transitory nature of the dietary effect in the monkeys suggests the influence of a more labile motivational response than for the human condition reported by Klein and co-workers (1969). It is not known if the timing or severity of the dietary deficiency may be responsible for these differences. Whether or not dietary rehabilitation is sufficient to reverse the social patterns acquired during the period of nutritional deprivation in the monkeys is of current experimental interest. In sum, recent experiments have barely scratched the surface of the myriad or subtle conditions and interactions that influence the course of development and maturation of the protein-malnourished organism. Hopefully, continued research with non-human primates will provide sufficient data to assist in the planning and execution of efficient and viable programs for the human condition.

DAVID R. OLSON

3

Effects of communication media on child development

A. THE MEANS OF INSTRUCTION IN THE ATTAINMENT OF EDUCATIONAL GOALS by David R. Olson, Ontario Institute for Studies in Education, Toronto, Canada This paper is concerned with the ways in which the mass media of communication, particularly radio and television, now function or may function as a complement to or as an alternative to formal schooling. This issue is not only of relevance to our understanding of child development; it has profound political implications as well. I shall attempt to sketch out the ways in which these developmental and political issues are related. The allocation of responsibility for education. First, the political issue. Formal education is one of the largest and most expensive enterprises in a modern society, outpaced only by defense and health services. Responsibility for education has by and large been exercised through the provision of school and the training and supporting of schoolmen. But this educational establishment is coming under the increasing attack of a former ally, the communication media, particularly radio and television producers and broadcasters. Illich (1971) has summed up this looming conflict accurately: 'Proponents of recorded, filmed and computerized instruction used to court schoolmen as business prospects; now they are itching to do the job on their own' (pp. 4-5).

476

Environmental conditions

Schoolmen see the threat to their traditional monopoly over formal education both in the increasing government support for educational broadcasting and, in Canada, in the increasing jurisdictional disputes between provincial and federal governments over the right to control educational broadcasting, education being a provincial matter and broadcasting being a federal concern. This is clearly manifest in the recent actions of the Council of Ministers of Education who have been engaged in just such a jurisdictional dispute with the federal government over educational broadcasting, while simultaneously imposing spending ceilings on the school boards under their control. Understandably, schoolmen are alarmed. The Secretary-General of the Canadian Teachers Federation, the national body representing professional schoolmen in Canada, is reported to have accused all provincial governments of downgrading and starving schools financially. According to the news report, 'He contrasted what he described as the provincial governments' "financial starvation" of schools with the diversion of millions of dollars in taxes "to create or support communication media systems that will serve only a tiny fraction of the school population" (The Globe and Mail, July 5,1973). Broadcasters, publishers, and producers are no less adamant. They cannot understand the resistance and delay in applying to educational problems the recent advances in communications technology - see, for example, the comments of the president of Ross Telecommunications Engineering Corporation at the recent conference on communications reported in Greenberger (1971). Informed exponents of the new educational technology, such as Henri Dieuzeide of UNESCO (1974), accuse the educational bureaucracies of being effective blocks to progressive educational changes that would permit the maximum utilization of both technological and human resources in the achievement of educational goals. This conflict documents the impending struggle for public funds for the support of formal education. What is overtly a concern for the improvement of education is covertly a power struggle between professional groups, the entrenched schoolmen versus the encroaching producers and broadcasters and publishers. Moreover, it is a struggle that proceeds without reasoned theory or empirical evidence as to the possible division of educational functions between the media and the schools. Can our psychological and educational enquiries contribute to an enlightened resolution of this issue? Henri Dieuzeide (1974) has stated the problem as follows:

Effects of communication media on child development

477

'Instead of attempting merely to recruit and train an ever increasing number of teachers, the time may have come to try to analyze the various educational functions with a view to redistributing the human and material resources available to wherever in the educational system their potential can be most fully realized. This, of course, implies the acceptance on our part that instead of continuing to let the machine do only what the teacher cannot do, we should ask ourselves what it is the teacher should do that the machine cannot do.'' But what are the educational functions that can be reasonalby turned over to broadcasters and what are the functions that must be retained by schoolmen? This is the psychological question the answer to which has important political implications. Without that evidence, the conflict between broadcasters and schoolmen will remain a simple power struggle between professional groups, with the fundamental concerns of the child and the culture as a whole lost in the shuffle. To state the question in more manageable terms, what is worth knowing and what can be taught via the mass media such as print or television? Conversely, what must be learned through one's own personal activity, through performance and reflection or praxis to use Freire's (1972) compelling term? The boundary between what is public and communicable and what is private and not communicable by the mass media is neither obvious nor fixed. An important form of investigation is the attempt to make communicable through the mass media types of information that have previously been assumed to be private, personal, and non-stateable. Indeed Sesame Street and The Electric Company should be viewed as such experiments, experiments devoted to determining the extent to which a powerful electric medium coupled with the research technique of formative evaluation can contribute to the achievement of a planned educational effect (Palmer 1974). Hence, technological developments and developments in the uses of that technology may serve to shift the boundary between what may be communicated and what may not. Nor is that boundary obvious. To illustrate, there are books and films on every conceivable subject and activity, even on subjects that one would assume cannot be told. (One of my unfulfilled aspirations is to make a film or videotape on How to Take a Walk. Perhaps it could be a whole series!) Yet some reflection will show that any skilled performance, whether it is doing, saying, or making something, resists complete description. Every sentence assumes more than it asserts. Polanyi (1958,

478

Environmental conditions

p. 4) was the first to insist that 'we know more than we can tell'. How then can we specify these boundaries? On the acquisition of knowledge and skill. Olson and Bruner (1974) have outlined a conception of the nature of experience that would at least permit us to begin an analysis of the question of what can be taught (via any mass medium, whether print, film, television, lecture, and so on) and what must be acquired through one's own performatory activity or praxis. Beginning with Piaget's (1970) conclusion that one's picture of the world is biased by or coded in terms of the actions one performs in the world, we have argued that one's activity yields two types of information: information about the world and information about the activity itself. In an aphorism: From the action of sitting on chairs, one learns both about chairs and about sitting. The first type of information, which reflects the features of objects and events that are more or less invariant across different activities, may be called knowledge. The second type of information, which reflects the features of the activity itself as it is performed across different objects and events, may be called skill or ability. Knowledge then reflects the invariants in the natural and social environment, while skills or abilities reflect the structure of the activity relevant to the medium or performatory domain in which that activity is carried out. It may be noted that Gilbert Ryle's (1949) famous distinction between propositional knowledge (knowing that) and procedural knowledge (knowing how) corresponds to the distinction drawn here between knowledge and skill. Knowledge and skill relate to experience in distinctive ways in that skills are tied to particular types of activity while knowledge is independent of those particular activities. To illustrate: One may acquire the knowledge that The stove is hot by (a) touching it, that is, through direct contingent experience (reinforcement, if you are a behaviorist); or (b) by seeing someone recoil from touching it, that is, through modeling or observational learning; or (c) by hearing, understanding, and believing the sentence 'The stove is hot'. As to the knowledge they convey, these forms of experience converge; they are to some extent equivalent and interchangeable. Hence, a wide variety of mediated, symbolically coded messages may substitute for direct, contingent experience in the acquisition of knowledge. On the other hand, the skills developed from these different forms of experience diverge; the mental processes whereby that information is extracted and the mental skills developed as a result are quite

Effects of communication media on child development

479

different in the three cases. One form requires discovery skills, another observational skills, and the third, linguistic skills. In terms of skill development, then, these forms of experience are not interchangeable. The differing mental processes called upon by these three modes of experience - direct contingent experience, modeling, and symbolically coded experience - are shown by the fact that they may be ordered in terms of evolutionary development. Bruner (1972) has pointed out that while all animals learn from contingent experience, primates are distinctive in their capacity to learn by observation - the young observe the old and incorporate what they have seen into a pattern of play. The human species is, of course, marked by its reliance on symbolically coded experience - so much so that the development of language is taken as a distinctive characteristic of the human species, and the development of literacy in various symbolic codes is the primary concern of formalized schooling. It follows that these three forms of experience differ greatly in the assumptions they make about the learner; that is, they differ primarily in terms of the skills they assume in the learner. Reciprocally, they differ in the skills they develop in their users - skills which our culture defines as intelligence. To summarize, any experience involves both knowledge and skill. But while quite different forms of experience can generate the same knowledge, every different form of experience generates or calls upon quite different mental skills. Educational means. Now let us apply this conception of experience to the effects of various educational means, particularly active performances as in discovery learning, creative writing, the visual arts, film-making, and so on and in such forms of receptive learning including auditing lectures and viewing films or television. My hypothesis is that acquisition of knowledge is related primarily to the 'content' or the message, while the development of skill is related primarily to the means or medium through which that content is coded and delivered. The claim is that both educational and developmental theory have overlooked the effects of the means of instruction on the acquisition of skills - so much so that general skills or abilities have often been treated as fixed properties of an individual attributable to heredity. Any means of mass or group instruction employs some communication medium which in turn relies upon the learner's competence in one or more symbolic systems - ordinary language, either oral or written, some

480

Environmental conditions

formal language such as mathematics or logic, spatial representations such as representative drawing, maps, or geometry, and so on (cf. Goodman 1968; Gardner, Howard and Perkins 1974). The use of each of these not only conveys a certain content, messages, or 'knowledge', but also calls upon a distinctive set of mental skills - various forms of literacy. Learning by means of the symbolic systems of the culture, like learning from one's own practical activities, involves distinctive patterns of skill and, as McLuhan (1964) has shown, results in a biased view of reality. Sentences, to the extent that they are about something, that is, have a message, carry information common to the other forms of experiencing; sentences draw upon and contribute to the general picture of reality that we have called knowledge. But the skills involved in using sentences are unique to that particular mode of expression and communication. The skillful use of a symbolic system involves the mastery of both its structure and its rules for transformation. From a linguistic point of view, once mastered this linguistic structure may be considered as generative - the old structure may be used to generate novel utterances. From a cognitive point of view, these linguistic skills, once mastered, may be considered as a form of intelligence primarily because the range of applicability is virtually open. To generalize, verbal, numerical, and spatial abilities reflect skills in such cultural symbolic activities as speaking and writing, counting, and manipulating Euclidean space, an observation which has led me to define intelligence as 'skill in a cultural medium' (Olson 1970). The implicitness of skill. Furthermore, the knowledge conveyed through the sentence (or through any other symbolic means) is explicit, employing a set of agreed upon conventions, but the skills underlying the comprehension and production of that sentence are implicit, unknown, as such, to the speaker and listener. We don't know how we produce a novel sentence; how we do it is the problem, as yet largely unsolved,of linguists and psychologists. Skills are primarily tacit, as Polanyi (1958, 1966) has argued. He states: 'The aim of a skillful performance is achieved by the observance of a set of rules which are not known as such to the person following them' (1958, p. 49). Generalizing this to all symbolic processes, Gross (1974) has stated: 'Symbolic competence, whether at the level of basic decoding or at more complex levels of appreciation and creativity ... is dependent upon extensive and continual action. Skillful action in a symbolic mode is intelligence and knowledge itself and at the same time it is the only way

Effects of communication media on child development

481

in which such knowledge can be acquired, maintained, extended, and utilized; in creative and productive activity. All such skills are largely tacit, transparent, serve to involuntarily structure perception, memory and cognition, and are generative.' An interesting description of the achievement and exercise of this symbolic competence has been given by Gardner: '[The artist attempts to] capture within a symbolized medium those properties - modes, vectors, feelings, insights, rhythms, and other forces which permeate (his) subjective experience and which ... [he] wants to convey to others. The artist does not have to think explicitly about these qualities, nor to engage in conceptualizing thought about factors; rather, through a series of experiments and executions within a symbolic medium he attempts to embody a desired mode or quality in an effective way' (1973, p. 274). But notice that while skilled action is private and inexplicit, communication systems are restricted to phenomena that are public. Not knowing explicitly how we do it, it is impossible to convey or communicate that skill explicitly by symbolic means. Communication systems, as they rely upon shared symbols, are by their nature restricted to phenomena that are public. Skills are, therefore, non-communicable via the mass media. A further indication of the tacit nature of skills and their resulting non-communicability is provided by a consideration of the ways in which skills are taught. Skill is acquired, not through being told, but rather through practice, that is, through performatory activity coupled with feedback from an expert. I cannot tell you how to tie a knot or to distribute your weight while skiing or how to make up an appropriate summarizing aphorism but I can illustrate it, provide occasions for practicing it, and provide feedback as to how your attempt may be improved. To again cite Gross (1974): 'All competence in a symbolic mode is acquired on the basis of constant practice and repetition (as well as observation and appreciation). While at the higher levels of competence practice may get you to Carnegie Hall, at the initial stages of learning it is the only way to get anywhere. One achieves competence in a medium by slowly building on routines which have been performed over and over until they have become tacit and habitual.' The limitations of the mass media. Here then is the paradox. Skills, because of their generalizability, must stand at the heart of the educational

482 Environmental conditions enterprise. Yet they are largely inappropriate for expression via the conventional symbol systems conveyed by the mass media. The mass media, radio, television, the press, are by their nature restricted to the communication of knowledge and not the development of mental skills. Schools, too, perhaps for reasons of convenience rather than of necessity, have tended to rely upon things that can be told, communicated via the group or mass media, regardless of the utility of that knowledge. The dominance, in schools, of things that can be told, reflects a simple fact - teachers are confronted with large numbers of children; in order to maintain order, they have to be taught by some variation of the recitation method generally in large groups. What can be taught to a large group of children? Primarily information which can be coded in one of the mass media. Knowledge is defined in terms of statements and propositions and is therefore communicable. Skills are not; hence they are poorly represented in conventional schools. Musical performances, whether playing an instrument or composing a score, creative writing, painting, the arts in general, as well as scientific thinking and problem-solving are not codifiable in the mass media (except in the most superficial ways, as in the statement of the steps of the scientific method). Hence, if taught at all, they are not taught well via the mass media or in the ordinary classroom. That is not to say that no skills are developed from watching the mass media or in conventional schools. Rather, it is to claim that the skills which are developed are primarily those called upon in the process of extracting the message conveyed by that medium. As indicated earlier, while the explicit content of any instructional program is related to the knowledge acquired, the means or medium of instruction determines the skills that are implicitly developed by that program. As most of the subject matter taught in conventional schools is taught via the oral and printed word, students are likely to develop considerable linguistic and literacy skill. To the extent that they watch film and television media, they will also develop some visual literacy skills. Even this aspect of skill acquisition (through passive or receptive methods) may have to be qualified in that skill is developed primarily not through listening, reading, and viewing but rather through actual performatory attempts in these media of expression and communication as Piaget (1970), Gross (1974), and others have argued. Freire (1972) makes a similar point in his Pedagogy of the oppressed by contrasting the conventional, narrative or 'banking' approach to education with that of praxis, action coupled with reflection and discussion, the latter for a liberating education.

Effects of communication media on child development

483

How and to what extent the observation of a skilled performance or the hearing or reading of a description can influence the acquisition of skill is an important subject for research. Studies by Salomon (1974) and by a student of mine, Joanne Rovet (Rovet, in preparation), have shown that filmed animations of spatial transformations, in Rovet's case of the spatial rotation of three dimensional objects (cf. Shepard and Metzler 1971), can facilitate the acquisition or at least the refinement of a mental skill of some generality. Yet, such descriptions and demonstrations do not translate directly into skilled performance - otherwise the simple act of watching the winter Olympics would change us all into championship skiers. Descriptions and models facilitate the acquisition of skills in an, as yet, unspecified way. Piaget (1970) has categorically differentiated 'figurative' from 'operative' knowledge and argues that the former, which may be acquired through listening or watching, does not result in intellectual skills while the latter, which is acquired through activity, does. The distinction drawn in this paper between knowledge and skills preserves that difference, but it fails to answer the question as to how models and descriptions facilitate skilled performance. This question warrants much further research of the sort begun by Salomon, Rovet, Feldman, and others in order to determine just what is learned from remote media and how that learning occurs. (Modeling theory seems to be concerned primarily with showing that models have some influence on behavior, but not with how the effect occurs). My conjecture, which I shall not pursue here, is that descriptions and demonstrations provide a representation or mental model of an act against which the representation of our own performances may be compared. But how we generate those performances remains implicit - we know only and compare only the outcomes of those integrated performances. To state this in another way, models and descritions indicate goals and sub-goals, not the means for their achievement: 'If you did that right, your weight will now be on your downhill ski.' Such an interpretation would keep activity, personal attempts at performance, at the center of the acquisition of skill. To summarize the argument to this point, formal education is concerned with the acquisition of both knowledge and skill. Only the former can be symbolically coded and conveyed through the mass media. The mass media, books, films, television programs, and, to a lesser extent, the schools, are dependent upon public, shared symbolic systems including ordinary language for the communication of information. Hence, while they adequately serve the educational goals pertaining to the acquisition

484

Environmental conditions

of knowledge, they serve poorly those pertaining to the development of skills. Yet it is the development of these mental skills that is responsible for the generalizable effects of experience and that we designate as intelligence. Skills may be acquired through individual practice coupled with suitably tailored feedback, namely, the kinds of experience that the mass media are incapable of providing and that the schools provide only with great difficulty, if at all. The educational consequences of means. To this point, the claim that the content of any means of instruction, including the mass media, is the primary determinant of the knowledge acquired while the means itself, including the medium of communication, is the primary determinant of the skills that are acquired remains a theoretical argument. We have recently begun an attempt to provide some empirical evidence for these arguments. While the more important claim advanced above is that the medium of instruction will specify the skills developed by an instructional program, the more general claim is that the means of instruction even within a single medium of communication will result in a distinctive pattern of skills. The extreme difficulty of equating content across different media led us to experimentally examine the differences between means within the same linguistic medium as a preliminary test of the hypothesis. In one experiment, Ruth Pike and I taught the same content, in this case a list of 20 word pairs, by three distinctively different means to groups of subjects aged approximately 10 to 12 years. Most such studies in the past have then looked to see which means resulted in the best acquisition of the content; we looked to see which means resulted in the acquisition of the most powerful skill. The means were selected to call upon quite different mental activities on the part of the children in the acquisition of that common content. The 'Rehearsal' group learned the word pairs by simply rehearsing them, 'apple-car, apple-car, apple-car'; the 'SentenceGiven' group of subjects learned the same word pairs by means of hearing a sentence relating the words, 'the apple fell on the car'. The 'SentenceGenerated' group learned the pairs by means of constructing for themselves a sentence relating the words of each pair. The hypothesis was that while all three means result in the mastery of the same content, some more effectively than others (Rohwer 1971; Bower 1969), they would result in quite different patterns of skill that would show up on subsequent trials. Specifically, subjects who had mastered the skill of generating sentences linking the word pairs would be able to use this skill on subsequent recall

Effects of communication media on child development

485

tasks. The evidence we have collected to this date support that hypothesis; different means of instruction may result in the acquisition of similar content knowledge while contributing to the development of quite different skills. The role of the means of instruction on the development of skills must be more completely examined. Especially interesting will be studies that employ different media of communication to see if the same results may be obtained. In both of these cases what is required is the study of the effects of various means and media on the development of skills and competencies rather than the traditional focus on the acquisition of content knowledge. In a similar way, the usual tests of transfer which examine the generalizability of the content knowledge communicated by the instructional programs must be complemented by tests of transfer which examine the generalizability or applicability of the skills to new contents. As these skills are largely implicit, considerable ingenuity will be required to detect and specify the role of means in the acquisition of skills. Allocating educational responsibilities. Assuming for the moment that the case can be established, namely, that the content of a medium determines the knowledge acquired while the medium itself determines the mental skills that are developed, what implications follow for the political issue with which we began? The mass media, print, radio, film, and television, by their very nature rely upon what can be explicitly coded in their respective symbolic systems. It is these explicit, public and shared symbolic statements that constitute knowledge. That is, knowledge is that information that can be coded in a form appropriate for transmission by the mass media. Responsibility for the acquisition of knowledge, therefore, should be turned over to the mass media and their agents, whether publishers or broadcasters, who would contract directly with the learners thereby circumventing the schoolmen's traditional monopoly on knowledge (Cooper 1971). However, according to the argument outlined here, the acquisition of knowledge per se is of definite second-order significance. It is the acquisition of skills in the symbolic systems of the culture that is primarily responsible for the achievement of a generalized intelligence, the primary goal of a general education. Those mental skills are acquired primarily through active attempts at performance coupled with feedback tailored to that performance. Children require an abundance of occasions for attempting various kinds of performance whether of a manual or a

486 Environmental conditions symbolic sort, attempts at formulating logical arguments, creatively expressing ideas in the various forms of language, in carrying out controlled experiments, in applying number and spatial systems of representation to their casual and formal observations, in performing, appreciating, and criticizing one or more of the arts. It is in the provision of occasions for performance and in the provision of suitable feedback that schoolmen have such an advantage over the mass media. Not that schoolmen usually seize that advantage (see Olson 1973). Because it has not been clear that knowledge and skills are acquired in quite different ways and serve quite different purposes, schoolmen have been willing to sacrifice competence in the arts and sciences for the simpler acquisition of factual knowledge. Furthermore, teachers are confronted with increasingly large classes which seriously jeopardize their attempts at individualization. Someone has pointed out that teachers can afford to let their students write for 15 minutes per day; more writing would make the task of 'correcting' prohibitive. To carry out their distinctive functions, the role of schoolmen would have to be altered to resemble that of a tutor rather than that of a lecturer with class size reduced accordingly. Teachers' roles would resemble those of art masters or piano teachers - a brief but intensive individual tutorial, much practice, and feedback tailored to individual performance. The mass media, then, have an important part to play in the education of children. It is already undoubtedly the case that children acquire the bulk of their knowledge about the world in a symbolically coded form delivered largely through the mass media. But the mastery of the skills fundamental to a wide-ranging intellectual competence will remain in the hands of suitably reoriented teachers.

Effects of communication media on child development 487 B. COGNITIVE EFFECTS OF VISUAL MEDIA by Gavriel Salomon, Hebrew University, Jerusalem, Israel

The possibility that media cultivate mental skills. When the effects of the mass media, particularly TV, are discussed, the readily available associations are 'aggression', 'family relations', 'knowledge', 'attitudes', 'consumption', and the like. It is interesting to note that most of the research up to the present concerned with the effects of TV dealt with only three classes of factors: The typical contents of commercial TV, particularly as they affect attitudes, aggressive and anti-social behavior (e.g., Eisenhower 1969; Feshbach and Singer 1971), knowledge, and the like (Schramm et al. 1961); the televiewing situation, particularly as it affects homework, family habits, and the like (e.g., Furn 1971); and the technology of televising - outstanding in its interest among educators (e.g., Allen 1971), as it affects achievement. However, there is one factor which, for some reason, has received hardly any empirical treatment, namely, the coding system (i.e., the symbolic-communicational formats) of the medium. It was Edmund Carpenter (1960), followed later by Marshal McLuhan (1964), who first drew our attention to the possibility that the communicational 'language' of media may have unique effects, above and beyond the effects of content or situations of exposure. Yet in spite of the widespread publicity of McLuhan's writings, and possibly also due to his style, the question as to how (if at all) the 'language' formats of media can affect one's psyche was never brought to an empirical test. Still, the possibility has not escaped the attention of researchers. Munsterberg (1972) has observed as far backas 1916thatfilm'externalizes' our modes of recall, association, and thinking. Bruner (1961,1964) much later speculated that man has to develop internal systems to serve as counterparts of external media and technologies. Thus, media are not just the extensions of man, but man adapts himself, psychologically, to their novel demands. Olson (1970) speaks of intelligence as being skill in a medium. Expanding this argument (1972a, 1972b) he asserts that media 'cultivate' mental skills. Feldman (1972), taking a Piagetian point of view, theorizes that media, as well as other types of experiences, accomplish a 'crystallizing' function in conceptual development, thus enhancing movement from stage to stage.

488

Environmental conditions

In short, the possibility that media affect cognitive skills has received some theoretical consideration, but little empirical evidence to back it up. In the present paper I wish to further examine this cognition-cultivating function of media and provide initial findings from laboratory and field studies done in Israel. Some assumptions. That different media have different types of coding systems, or 'language' formats, seems to be a common observation. It is quite obvious that digital codes, which characterize language and mathematics, are very different from analogical codes of pictures, drawings, films, and TV. Similarly, the spatial representation which constitutes maps is quite different from that of aerial photos or of photographed landscapes. Clearly, some media have more similar coding elements while others are more dissimilar. It could be possible to suggest even a hierarchy of media and classes of media according to the extent of their shared and dissimilar coding elements. 1 However, for our present purpose it suffices to agree on the assumption that media differ not only in terms of their technologies of transmission, but also (and perhaps more importantly) in terms of their coding systems, or 'language' formats. Another assumption which we could safely make is that as coding systems differ, so also do the information extracting and processing activities which are applied to them. In other words, the way one extracts and processes information from one coding system is quite different from the way he does it from another. The question of whether the resultant knowledge or meaning is or is not the same is somewhat irrelevant to this point. Indeed there is evidence to show that storage, recall, solution of problems, recognition, even kinds of mediators which are activated, differ as a function of the coding system, spatial arrangement, or format structure of the input information (e.g., Paivio and Yarmey 1966; Samuels 1970; Fleming and Sheikhian 1972; Brooks 1968; Huttenlocher 1968). Given these two assumptions, it follows that different media (not technologies) have to call upon different modes of mental activity, and that these have to undergo constant changes as one is continuously exposed to new and changing coding systems. Could we undergo such 1. Particular chapters in the 1974 NSSE Yearbook (D. Olson, Ed.) deal more directly with this issue.

Effects of communication media on child development

489

changes as the result of exposure to media which challenge our mental capacities with new demands? Cross-cultural studies repeatedly demonstrate the high media-illiteracy of non-Western people. Yet, as exposure to Western technologies and media increases also media-literacy develops. Does exposure to the media cause the shift from illiteracy to literacy? And if so, what constitutes that shift? As Lloyd (1972) observed: 'Socialization in the Western pictorial mode is important although we are as yet unable to specify the precise dimensions of this experience.' In another place (Salomon 1972), I have argued that if learning to cope with new coding systems means the cultivation of appropriate mental 'counterparts' (Bruner 1961, 1964), then becoming media-literate entails more than just learning some associations between formats and meanings. It could entail internalization of the code, its schematization, and its use as a 'mental tool', mediator, or covert representation. Thus, exposure to maps ought to lead to changes in spatial representation and exposure to film to better ability to think in terms of the filmic codes or formats. Olson (1972a) reaches a similar conclusion, although by means of a different rationale, suggesting that 'experience has its effect on behavior through the information it provides'. Any activity requires information for the guidance of every component in it. One acquires both the information provided by the experience as well as the skills involved in its extraction. Thus, sitting on chairs teaches us both about chairs as well as about the skill of sitting. Two major questions can now be raised. First, can media be made to cultivate mental skills, and if so, how is this function accomplished for different people? This is an educational question. Second, do media, and particularly the mass media, cultivate mental skills? And if so, how is this accomplished? This is a developmental question. Some preliminary experiments. To answer the first question, a series of experiments was carried out. In these experiments a particular coding element of film was chosen and presented to children in varying degrees of explicitness. For example, in one case we chose the element of the zoomin-and-out on details. This element is quite unique to film. However, the psychological function that it accomplishes is apparently isomorphous to the skill of selecting perceptual details out of a complex and noisy context. Thus, such a filmic element could be internalized and could be used as a covert mental schematized operation. In another study we have used the element of laying-out solid objects. This is less unique to film,

490 Environmental conditions but the filmic range of codes, or formats, allows us to lay out even the most unusual objects, in quite simple ways. We have reasoned that such highly explicit presentations of filmmediated operations can be imitated by observers, and that once imitated, they are internalized and can serve as modified (or newly acquired) mental skills. We have contrasted such explicit presentations with others (usually slides) which call upon or activate rather than model the skill to be learned. Our experiments indicated two things: (a) Exposure to films which explicitly model a particular operation, that is, part of the filmic coding system, leads indeed to improved mastery of that operation. Learners showed improved ability to carry out tasks which required the mental skill modeled by the film. Thus, it appeared that a medium such as film can modify mental skills. (b) Systematic interactions between learners' initial ability and explicitness of presenting the operation were found. The poorer the mastery of the specific operation the more is learned from the highly explicit modeling presentation. The better the initial mastery, the more is learned from a presentation that calls upon one's already existing mental skills. We were in the position to suggest that media can affect mental skills in two ways: They can model an operation, which may be internalized2 by those who start out with poor mastery of that operation, and they can modify the operation by calling upon it, or activating it, in those of initial better mastery of it. These findings suggest that media can modify mental skills in differential ways. But do the mass media, in fact, lead to such cognitive changes? That is what the study of Sesame Street in Israel came to answer. A field study: The effects of'Sesame Street' in Israel. Sesame Street, the educational but commercial-TV-like program, was originally designed for American 3-6 year old children. What characterizes these, in comparison with, say, Israeli children is the large quantity of commercial TV broadcasts which they consume daily. Acknowledging this fact, the producers of Sesame Street have tried very hard to present their educational messages via the most sophisticated, commercial-like programs, and make them resemble, in format, the most advanced commercial TV programs. 2. J. S. Bruner, while commenting on these results, suggested that the modeling films enhanced the mental skills through 'scaffolding' them, rather than leading to the internalization of the model (private communication).

Effects of communication media on child development

491

The educational messages themselves were much more pedestrian and resembled the typical contents in every common kindergarten. When such a program was brought to Israel, whose children have had only a very brief exposure to TV with quite old-fashioned programs on it, the question of TV-literacy came immediately to mind. If the program's formats are geared to play upon the televiewing habits and to call upon the mental skills of the American TV generation, what kinds of mental requirements do they make for the relatively TV-illiterate Israeli children? More importantly, to what extent does exposure to such a program modify, or cultivate, specific mental skills in Israeli children? It is important to note that the encounter of children, like the Israeli children of 1971 with an extremely unique TV program like Sesame Street provides a rare opportunity for the study of the question of whether a medium's specific formats do affect mental skills under realistic and natural conditions. The study was rather complex, and only a few of its major findings will be reported here3; 317 children, aged five, seven, and eight, and representing low income and middle classes, took part in the study. A battery of 'TV-literacy' tests was developed following a careful analysis of the program's most outstanding formats. Thus, e.g., we found that the show is highly fragmented. We reasoned, therefore, that it calls upon one's skill of synthesizing the various elements and imposing some logic on it. Hence, a test for this skill was developed. Similarly, we found that the camera in the program often takes the viewer around objects to show him the object from different points of view. We hypothesized that the program models the skill of changing one's point of view. Hence, a test of this skill was developed. Other tests of additional skills, such as disembedding a complex figure, relating components to wholes, were developed. Some of those skills we assumed to be called upon and thus expected them to be cultivated in children who had already some mastery of those or equivalent skills. Other skills were assumed to be modeled, and we expected them to be cultivated in children with initial poor mastery of them. However, since the program also had specific instructional objectives, we tested their attainment (using the ETS battery, Ball and Bogatz 1970) as well. We wanted to examine the extent to which attainment of the 3. A more detailed description can be found in G. Salomon 'Sesame Street' in Israel. Its instructional and psychological effects on children. New York, The Children's Television Workshop, 1974.

492

Environmental conditions

contentual knowledge (the intended objective of the show) is related to cognitive changes resulting from the novel 'language' formats of the show. The execution of such a study poses many methodological problems, not the least of which is the absence of an adequate control group. We solved that problem by using the technique of multiple regression (Cohen 1968; Walberg 1971) to analyze our data. This technique enabled us to partial out variance which was due to background and initial ability differences, and to examine the extent to which different amounts of exposure, and different components thereof, accounted for variance in learning. In other words, how much difference in mental skills and in the intended knowledge-areas does exposure to the program make? Major findings. Briefly stated, exposure to the program has a significant effect on TV-literacy mental skills. This effect was far larger on highly specific mental skills, such as changing points of view or relating components to wholes, than on more general abilities, or styles, such as field dependency. But even in the latter case, exposure to the program accounted for up to about 10 % of the after-the-show ability variance. In the former cases exposure was found to account for up to 23 % of the variance. In sum, then, exposure to a psychologically demanding program such as Sesame Street has rather substantial cognitive consequences, thus indicating that TV does, apparently, cultivate certain mental skills. There were, of course, differences among the age and SES sub-groups. Older children were more affected by the program's 'language' formats than younger ones, and more well-to-do children benefited more than low-income ones. This comes as no surprise, as the older and the more well-to-do children were more exposed to the program. More important, they were mentally better equipped from the outset and thus understood the show much better from the early days of its presentation. Indeed, it was not sheer amount of exposure to the program, but rather the intelligent viewing which made it have so profound a cognitive effect. 4 Naturally, intelligent viewing cannot exist without exposure but the latter can, and often does, take place without the former. Another finding of the study further strengthens this point. Half the mothers of our five-year-old group were urged to view the show with their children, explain it, and elaborate on its messages. This did not 4. While, for instance, amount of exposure over the four-month period of the study accounted for hardly any significant portion of the learning variance, it was the knowledge of what has been seen to which much of the variance could be ascribed.

Effects of communication media on child development

493

make much of a difference for middle-class children. Encouraging their mothers to join them in viewing the show did not make them learn more or be more strongly affected. However, encouraging mothers made a large difference for the lower-class children. The result was that SES differences, a dominant factor in the non-encouraged group, ceased to play a role in the encouraged group. That is, lower and middle-class children benefited to a similar extent as a result of encouraging their mothers. The active participation of the lower-class mothers made their children view the show more intelligently. How did the program cause the cognitive effects which we have found? Claiming that it is 'novel' and 'demanding' (in formats, not contents) explains very little. Thus, we have tried to distinguish, as best as we could, between those formats that modeled skills and those which called upon skills. We hypothesized, on the basis of our former experiments, that while the less skillful children ought to benefit more from what is modeled, the latter ought to profit more from what calls upon their skills. Indeed, we found negative or zero correlations between pre- and post-viewing measures of skills which were modeled, meaning that the initially less-skillful children learned more from what was modeled. On the other hand, we found positive correlations between pre- and postviewing measures of those skills which were called upon, suggesting that the more one mastered a skill at the outset, the more it was developed. These findings suggest the conclusion that media cannot only be made to affect cognition, but that, under particular conditions, they do in fact have such effects. Furthermore, these are differential effects, interacting with children's capabilities. Less-able ones are more affected when the TV 'language' formats model a skill, while initially better-able ones are more affected when the formats call upon, or activate, existing skills. In these respects, our findings in the Sesame Street study are in full agreement with those obtained from the experiments. Finally, what about contentual learning and mental cultivation? Two findings suggest a close relationship between the two. Firstly, we found little relationship between the two groups of variables before the program was broadcast. The absence of relations continued after the broadcasting season was over, among the little-exposure children. However, among those who were intensively exposed to the program, mastery of TVliteracy skills and contentual learning became strongly interrelated. The second relevant finding was that, other things being equal, exposure to Sesame Street made the children better able to learn from another instruc-

494

Environmental conditions

tional film. That is, they were better able to extract and process scientific information presented to them via film, as a result of their exposure to Sesame Street. Taken together, these two findings support the hypothesis that Sesame Street did, indeed, affect TV-related mental skills, and that these - once developed - facilitated contentual learning from the medium. Concluding remarks. Media have their effects on human beings. But this assertion means very little. Specific components of media, such as typical contents, situations of exposure, or particular technologies, have some kinds of effect on some types of people in some social settings. To this list of factors and interactions, we have added - what we consider to be the very heart of media - the cognitive effects of particular 'language' formats which characterize one medium or another. As a result of our experiments and field study, we are now a bit better able to claim that media can be made to affect mental skills, and that given sufficient novelty and cognitive stress - they do cultivate specific abilities. It now becomes more evident that children's exposure to TV is a rather important force in their cognitive development, although we know as yet too little to state exactly how this takes place, with whom and under what conditions. The distinction we have made between formats that accomplish the function of modeling skills, and those that call upon skills, has received empirical support. The findings show also who learns more through modeling and who from skill activation. However, all this does not tell us yet what role media play in the overall process of cognitive development. The study of Sesame Street in Israel provided a rare opportunity to see how media, in a non-laboratory setting, affect cognitions. But it would take a longitudinal cross-cultural study to make us understand the role modern media play in developing abilities. The 'coca-colonization' of the world, through the wide dissemination of Western media, makes the world into a 'global village', not only in terms of shared knowledge, interests, habits, aspirations and attitudes, but also in terms of shared mental skills.

Effects of communication media on child development

495

C. TELEVISION AND THE DEVELOPMENT OF SOCIAL BEHAVIOR by Aimee Dorr Leifer, Harvard University, Cambridge

Most children in the United States have the opportunity to observe much social behavior on television. This opportunity is afforded children in only a few other countries in the world. Even in those countries where television is available to children, the amount of time they spend with it may be much less than it is here. Moreover, the content and its presentation format may differ dramatically (Ambrosino and Fleiss 1971). There are increasingly frequent suggestions now that television, while entertaining American children, also plays a role in their social development. American television provides our children with one view of our social system. Most of the social information on television is encountered in entertainment programs since they account for most of the programing time (Gerbner 1969; Steiner 1963) and most of the viewing time of children (Lyle 1972; Schramm, Lyle and Parker 1961) and adults (Steiner 1963). The characteristics of the social system presented on television have been partially defined by a number of different investigators. Violence is very frequent, often used to resolve interpersonal conflict, often unaccompanied by emotions other than humor, and often counter to the norms of most teenagers and adults in our society (Baker and Ball 1969; Barcus 1971, 1972; Gerbner 1972a). Women and members of minority groups are generally under-represented and often portrayed as either less competent or as more nearly one-dimensional than are white males (Gerbner 1969, 1972b; Mendelson and Young 1972; Sternglanz and Serbin 1973). The frequency and nature of socially positive behavior and of intellectual content have not been analyzed, although they are both apparently less frequent than aggressive content. These content characteristics are not necessarily the same as those of television in other countries. Most observational learning theorists postulate that children need simply to observe the social behavior of others - either live or mediated to acquire it for later reproduction. Most television programs provide at least two kinds of social information, each probably requiring somewhat different methods of encoding and utilization. There are specific social behaviors and there are general rules by which most people operate in our society. Specific norms or modes of operation are often stated. When they

496

Environmental conditions

are not stated, they are presented so consistently in program plots that children could eventually acquire them. This is the more socially significant aspect of American television. My own work has examined the acquisition of both specific behaviors and also models for behavior. I have been continually interested in developmental aspects of the acquisition of social behavior from television and will consider these in more detail later. First, I would like to relate work which indicates that children will acquire specific behaviors and general norms for behavior from media presentations. Acquisition of behavior from one exposure. In one study I examined the acquisition by three-, five-, and seven-year-old children of sex-typed toy play from a short film of ten-year-old boys and girls playing with certain toys (Leifer 1966). Subjects all performed specific behaviors presented in the films, with the number of imitations increasing with age. When we asked whether some toys and actions were more appropriate for boys than girls and vice versa, 90 % of the judgments corresponded to those in the films. There were, however, only weak indications in the performance data that children adopted this generalization. Boys displayed more sextyping of the toys and actions than girls did, but the differences were small and diminished among older children. Thus, children acquired some specific behaviors and some information about sex-typed toys and actions from a film, but exposure was not sufficient generally to sex-type their play behavior. Later work with aggression indicated again the acquisition of both specific behaviors and also models for behavior, this time from videotapes. One study showed preschoolers videotapes of twelve-year-old boys playing with toys (Leifer and Roberts 1973). In line with earlier work (Bandura, Ross and Ross 1963) three scenarios were constructed. In one Jamie did not share with Rocky when Rocky asked, Rocky aggressed against Jamie's toys, aggressed against Jamie, and finally took his toys. In a second, Jamie did not share, Rocky aggressed against the toys, aggressed against Jamie, Jamie retaliated, and kept his toys. In the third, Rocky and Jamie played actively together with the toys. We found specific imitations of the aggressive toy play. We also found that children had acquired the message from one tape to play aggressively with the toys and from another tape not to play aggressively. We also looked for the social message children had acquired about how to resolve conflict with another person and found that those who viewed either of

Effects of communication media on child development

497

the first two tapes generally adopted the norm that aggression was the way to resolve conflict. This was how conflict was resolved in these tapes. This was not the norm acquired by the children who saw active play without interpersonal conflict. In additional work with aggressive behavior in actual television programs we found acquisition of some norms about social behavior (Leifer and Roberts 1972). We presented children in kindergarten, third, sixth, ninth, and twelfth grades with one of six television programs and then tested for the frequency with which they would resolve interpersonal conflict by physical or verbal aggression, leaving the field, or positive nonaggressive interaction with the annoying person. The television programs varied in the sheer amount of violence shown, the social acceptability of the motives for the violence, and the desirability of the consequences to those who participated in violence. We were most interested in seeing if children acquired ideas about the social acceptability of violence and the consequences for it and if these would influence their choice of strategies for resolving interpersonal conflict. Children in this study apparently obtained only one generalization from the programs: The more aggression in the program the more desirable it must be. Children who were exposed to the more violent programs were more aggressive in resolving interpersonal conflict than were children who saw less violent programs. Messages such as 'crime doesn't pay' and 'only bad guys fight' were messages adults believed were in the programs, but they were not messages children received. Thus, there is clear evidence that at some ages children will adopt social behaviors and behavioral patterns presented once on film, television, or videotape. Other researchers have presented similar results from exposure to media displays of aggression (c/. Bandura 1969; Goranson 1970; Surgeon General's ... 1972). There are, of course, numerous studies using live performers which provide supporting data which we will not cite here. We have, therefore, ample evidence that television could have an impact on the development of social behaviors. The question then becomes whether it does, in fact, have an impact and if so, how this impact occurs. Acquisition of behavior from multiple exposures. There is now some indication that long-term exposure to television content in the United States can impact social development. Some of the work commissioned for the Surgeon General's study of television and social behavior demon-

498 Environmental conditions strated that long-term exposure to aggressive television programing increased the probability of aggressive behavior (Surgeon General's ... 1972). Steuer, Applefield and Smith (1971) also demonstrated such effects with preschool children exposed to a series of aggressive cartoons. A study commissioned by the Children's Television Workshop demonstrated that exposure over a three-week period to prosocial behavior on Sesame Street increased the cooperative play of children attending daycare programs (Paulson, McDonald and Whittemore 1972). The work I'm currently conducting also suggests that children can adopt more cooperative patterns of social interaction (Leifer 1973). Preschool children saw specially constructed half-hour videotapes of Sesame Street material on positive social behavior. They saw three tapes a week for two weeks. Information was presented on how to draw one house cooperatively, how to combine boxes of two different colors to produce one very tall tower, and how to interact cooperatively in a number of different situations using a number of different behaviors. We tested children in a nearly exact duplicate of the house-drawing situation and a toy-play situation similar to the box stacking vignette. Then we observed their social behavior during free play in their daycare centers. We hope to learn more about the characteristics of television material that is effective in encouraging positive social behavior and how much the initial social and cognitive skills of the child determine these characteristics. The data have not been altogether analyzed yet, but there were some indications of desirable behavior change. For instance, children were able to cooperate more in drawing a house and thus produce a better product. We also saw in some children clear imitation of strategies for resolving conflict which were presented in the programs. This was particularly gratifying in the behavior of one boy. This child was very big for his age and typically resolved any conflict with an aggressive outburst. The television programs he watched showed potential conflict situations resolved with a vigorous, prolonged handshake. The day we observed him, after two weeks of television, he began his typical pattern. Suddenly he stuck out his hand and shook hands with great animation and the episode ended. The pattern was new in his repertoire and clearly derived from the television he had seen. Thus, we have ample evidence that television does provide some social input as children develop in the United States. It can with one exposure provide examples children will use in subsequent test situations, and with

Effects of communication media on child development

499

repeated exposure it can provide examples children will use in their day to day lives. These influences may be substantial for certain children and not for others or at certain times in a child's development. It is important now for us to try to understand the factors that determine the magnitude of television's effects on the social development of American children. Developmental considerations. The years from two to sixteen, during which children watch a lot of television, also encompass major changes in their cognitive and social functioning. Many researchers have characterized the cognitive changes which occur during this period (Flavell 1963; Piaget 1970; Piaget and Inhelder 1969; White 1965) and have hypothesized, and provided support for, various stages of cognitive functioning. Less work has been done to characterize social development and its possible stages, although there are stages in Freudian and moral development theories. Still there is agreement that social behavior changes significantly as children mature. Three of these changes in cognitive and social functioning which may have profound implications for the possible effects of television and which will be discussed in more detail here are: (1) Information acquisition, (2) persuasibility, and (3) cumulative experience. (1) Over the past few years I have looked at the social information children of different ages acquire from media presentations, asking first simply whether there are age differences in the quantity acquired and second whether the types of information acquired change with age. In this we are, of course, ignoring the obvious fact that some presentations are more easily understood and remembered than others, regardless of the viewer's age. The variables that promote such understanding and retention will not be discussed here, although they are also important in considering the social effects of television. In one of our studies, four-, seven-, and ten-year-old children saw a twenty-minute professionally produced film of a fairy tale (Leifer et al. 1971). They were then questioned about the sequence of main events, the motives of the characters, and the feelings of the characters. In another study children between the ages of five and eighteen years were shown a half-hour television program with commercials and asked about the motives for aggressive behavior and the consequences of it (Leifer and Roberts 1972). In another, children were shown edited versions of commercially broadcast television programs and questioned about the

500 Environmental conditions overall motives for and consequences of aggression (Leifer and Roberts 1972). Over all three studies older children were more able to sequence the events of a film, understand the motivations for aggression, understand the feelings of various characters, understand the consequences of aggression, and remember the denouement for various characters. In all these studies we had expected to find developmental differences in the types of information acquired. In no case did we. We looked in one study for increased learning of the actions of same-sex characters (Leifer et al. 1971) and in the other two for increasing learning of the motivations for action (Leifer and Roberts 1972). Subsequent work by others has demonstrated that children will pay relatively more attention to the motivations for action, rather than the consequences, as they mature (Berndt, Collins and Hess 1973; Rule and Duker 1973). The differences be tweenour approach and those of these researchers are not clear. Other researchers have also demonstrated developmental differences in the acquisition of information either central or incidental to the plot of a program (Collins 1970; Hale, Miller and Stevenson 1968; Hawkins 1973). These studies all indicate that children retain more information about the main events of the plot as they get older, that until adolescence they retain increasingly more incidental information, and that at early adolescence they begin to forget the incidental information. What we have shown here is that older children almost always acquire more correct information from an audiovisual presentation than do younger children. Yet older children do not necessarily perform more than younger children of the specific behaviors or behavior types in a media presentation. Moreover, one cannot necessarily find a linear relationship between how much is learned and how much is performed. In the series of studies on television and aggression referred to previously we failed to find support for the hypothesis that children who learned more from television would be more influenced by it in their behavior (Leifer and Roberts 1972). This was true both across ages and within any age. Why should this be? The answer is not obvious. How much information a child acquires will partially determine how much he or she performs, but we must also look at how persuasible children are, including the credibility they grant to television, and to how many other social experiences the child has had in the particular behavioral domain under study. (2) There is evidence that children become less persuasible as they grow up. For instance, Abelson and Lesser (1959) reported that children

Effects of communication media on child development

501

became less persuasible during the elementary school years; Orth (1969) reported less attitude change as children matured; Roberts (1968) also reported less attitude change from an audiovisual presentation as children matured from fourth through tenth grades; and Dysinger and Ruckmick (1933) postulated the growth of adult discount offilmsto explain decreases in emotional responsiveness to film content. In my work with television and aggressive behavior there were two instances in which it seemed clear that television became a less potent source of influence as children matured. In one study we showed children between kindergarten and twelfth grade television programs that varied in the amount of violence shown (Leifer and Roberts 1972). We found clear effects over all subjects for the amount of violence in the programs. When we looked at the effects for each grade separately, however, we found that the programs were maximally differentiated with third graders and then steadily decreased in differentiation. By twelfth grade there seemed to be no differential impact of the various types of programs. Older students acquired at least as much information from the program as the younger students, so we postulate a decrease in credibility of the medium and in persuasibility of the teenagers as a partial explanation of the age-related differences. In the second study (Leifer and Roberts 1972) we found that preschoolers and fifth graders were discouraged from aggression by a program which showed only undesirable motives for and consequences of aggression but not by programs with other motives and/or consequences. Twelfth graders, on the other hand, were not influenced by any of the programs. This argues for a general decrease among older children in the ascribed credibility of television. This is not to say that television cannot have an impact on older children. It clearly can (e.g., Alper and Leidy 1970; Walters and Thomas 1963). However, the impact is generally less and is probably more specific to content and program formats than it is with younger children. We know that television is generally granted high credibility in the United States by both children and adults (Lyle 1972; Schramm, Lyle and Parker 1961; Steiner 1963). Most of this work has, however, focussed on television's credibility when it is a source of information and news, rather than a source of social behavior. We also know little about the relationship between ascribed credibility of a source and its influence on behavior. Research in attitude change suggests that a more credible source would have more influence than a less credible source, but that the information

502 Environmental conditions source could lose importance with time. How directly these findings translate to children and television is not clear now, although there is a suggestion that television loses credibility as a source of social information for older children. (3) A third determinant of developmental differences in the social effects of television is the amount of information or experience a child has accumulated. As children mature they are exposed to different people, actions, and norms. This provides a range of choices when they have an opportunity to perform or opine. For example, children who encounter Puerto Ricans only on Sesame Street may learn that interactions with them are always cordial and use that as a model for a first encounter with a live Puerto Rican. In contrast, a child who grows up close to a Puerto Rican enclave will understand much more about the variety of ways in which he or she could interact with Puerto Ricans. Hopefully, Sesame Street will provide information for this child too, but the information will take its place alongside other relevant information rather than providing the single model for interaction. To date there have been few attempts to examine the interaction between previous social experience and social behavior presented on television. Knowledge of such relationships would allow us to use television more effectively to encourage socially valued behavior and to protect some children from the undesirable effects of exposure to some television content. Turiel's work on changing stages of moral judgment is most interesting in this regard. He has reported that children may move from a lower stage of moral reasoning to a higher stage after exposure to reasoning at the higher stage (1966) with the most significant change when the modeled moral reasoning is one stage higher. If this principle were also to apply to social behavior, one could begin to specify more exactly the content which could affect social behavior and the level of social action to present to affect maximally the child viewer. My current work examines this possibility for socially valued behavior. Conclusion. As we have seen, American television probably acts as one of the socializing agents of our children. It presents a view of the social system to children. Our children will spend more time watching this system in operation than they will spend experiencing the social system of the schools. The impact of exposure to this social system apparently changes as children mature. As yet we have done little to elucidate the

Effects of communication media on child development

503

age-related variables that account for these changes. I have suggested three variables - information acquisition, persuasibility, and cumulative experience - which contribute to age-related differences in the effects of television on social development. It remains to provide the research data to explain the operation of any one or a constellation of these variables.

SECTION V

Social organizations and the development of the individual

GRETA G. FEIN

1

Variations in home-based early education: Language, play, and social development

A. THE CONSTRUCTION AND SELECTION OF MENTS: DESIGN OF THE STUDY

ENVIRON-

by William Kessen, Yale University, New Haven

These are strange times for people who study the young child. On one side, we are reassured, almost ritualistically, that the early years are critical to development and that it is in the interest both of science and of humanity for us to understand the ways in which the events of the first years make their mark on tender tissue. On the other side, and with increasing volume, we hear the voices of restraint - perhaps even of retrogression - that tell us that the early times of the child are well buffered from casual or consequential variation in the world around him and, further, that much of the injury that may be done early on can be repaired by nature or by schools or by some other later remediation. Another example, I fear, among the many in psychology in which we ask the wrong questions, then use the wrong methods to get the answers we intended to get when we asked the wrong questions in the first place. Developmental psychologists, particularly those of us who are concerned with change in the first years of life, are often polemicists and protectors of ideology; it is no wonder that the world is beginning to wish us bad cess. We join in the hope that the Ann Arbor meetings will move developmental studies a step or two closer to the researcher's proper tasks of restrained and sceptical analysis, even of his own ideas.

508

Social organizations

Background of the study. The study - or better, the enterprise - that we will talk about this morning has many tangled personal and conceptual beginnings; it is only the clarity of looking back that permits us to state some of the more important ones. What the enterprise became, in the end, was a six-group study of home-based education for about one hundred children between a year of age and two and a half. It started quietly. We, all of us, recognized that the second and third years of life were busy ones developmentally, when the child achieves that subtle transition from baby to person. We also, though with different force from one to another, felt that the centers of that developmental transition were language, play, and social development. Each of us believed, in addition, that there ought to be some way of bringing the orderliness and clarity of the laboratory into natural settings - a conviction that has been sorely tested and blunted in part but which remains our bulwark, however battered, against the forces pressing toward some variant of the uniqueness argument - the impenetrability, the ineffability of human personality. More on that issue later. What we were not prepared for was the unique demand in studies of early education for a shifting focus, for taking at least three different points of view toward what we are doing - often in rapid succession, and, occasionally, simultaneously. To study the child at home, you have to be a developmental psychologist, calling on your knowledge of data and ideas in the field. To influence that child at home, you have to be an agent of change, a persuader and communicator. Some of the issues here are, to be sure, technical, but, far more critically, the issues are normative and moral. Finally, to judge your impact, you have to turn back on your earlier activity and become assesser and evaluater, now coldly indifferent to the historical justification of the study. Clearly a tangled net, but in it are caught all serious students of early education. If I tried to find our specific place in the net, incidentally, it would be more in the developmental research corner than in either the teach and change corner or the test and measurement corner. Now in this knotted skein of issues, one that confronted us early and often was the painful question of the morality of what is called, in its ugly way, 'intervention'. We had seen enough of, and sad to tell, been enough of, the plantation liberal - let us go repair the disadvantage of the poor folks - that we were wary to the point of paralysis on the issue. Our approach to a solution had two main components - first, we sidled up to our educational curricula very slowly and second, we organized

Variations in home-based early education 509 the curricula around strategies for the mother rather than around specific behaviors, theories, or instructions. Let me say a word about each of these themes because they have relevance to our methods as well as to our sensitivity about the morality of intervention. Two relevant steps preceded our design of curricula for babies in homes. Stage one: Two independent studies were carried out in New Haven several years ago, one by Alison Clarke-Stewart on the interaction of young babies and their mothers, the other by Katherine Nelson on the early development of language in children. Their reports will appear in the current series of Monographs of SRCD. These studies provided us with some guidelines for our later observations and with some important messages about method. Stage two: Before we put together anything that could be called curricular, we set up three longitudinal studies - we have called them panel studies - of play, language, and social development. In other words, the panel studies were our surveys of the terrain, our exploratory probes into the phenomena we wanted to study through later systematic variation. You will not hear much about the panels today; their importance was in providing us with data - data collected both in homes and laboratory settings - that were the raw material from which we fashioned our curricula. Stage three: Only after we had digested the results of the ClarkeStewart and Nelson studies, only with the panels leading the way did we begin to put together our educational programs, the central part of our work that you will hear about shortly. Let me first say something about the notion of strategies of our early education. It is the most important message we have to bring to you today. Strategies of early education. Slowly, all too slowly, psychologists are discarding the image of the child as an input-output system, part of a linear arrangement of the time and events in which something happens in what we evasively call the environment, to be followed by something that happens in what we evasively call behavior. Typically the somethings have been very narrowly defined, with a hope that we have found the right pieces of environment and behavior to represent some grand process like memory or learning or language, and more, that the pieces we had chosen were context-independent, revealing the underlying process unambiguously in the rather restricted circumstances under which we observed them. But that, of course, is not the child at all.

510 Social organizations He is much more a field of events, complexly interconnected in ways that we can only presently guess at. What seems beyond guessing is that there are several separable theories of the field of the child operative when a psychologist looks at or attempts to influence the development of the child in his home. There are, at least, the psychologist's theory of the field (which has evaluative and normative components as well as the analytic ones we advertise), there is the child's theory of the field (almost the exclusive province of Piaget until the day before yesterday), and there is the mother's (or other caregiver's) theory of the field. Now, I recognize that I am elementalizing some profoundly important and difficult issues in developmental theory, but I want today only to provide a framework for you to hang our study on. My colleagues will probably forgive me the following hyper-reduction of our research goals - we wanted to use our emerging psychological theory of the child's theory to influence the mother's theory of the field of the child. More prosaically, we used the insights we gained in our earlier observations, particularly in the panel studies, to elaborate some general principles about how the child changed over the months between 12 and 30 in the special areas of language, play, and social development. We then, as our major educational and curricular theme, tried to inform the mothers of our sample about the discovered general developmental principles. There are, as you will have guessed by now, several reasons for such a curricular strategy. First, and often foremost for us, trying to tell mothers what we have found out or believe about babies seems less directive and manipulative than prescribing a set of materials or routines or behavioral objectives. Of course, it was impossible for the planners of the curriculum to suppress their undocumented prejudices (for example, that it is unwise for mothers to intrude always into the child's ongoing play), and it was impossible for the home visitors not to reveal their attitudes toward children (in particular example, most of our home visitors were shaken by the persistence among our families of what might be called classical attitudes toward gender differentiation). But our focal attempt was to inform the mother about child development, to draw out her goals and intentions for her child, to make her as aware as we could of the intricate relation between her life and the baby's and to (in imitation of our curriculum) intrude on her relation with her baby as lightly as we could. The added joy of such a research procedure is that it can be justified not only on ethical grounds but on theoretical grounds as well. If you believe, as we do, that the parents are the major agents of change in the life of the young

Variations in home-based early education

511

American child, and if you have as we do, an image of the child as a field of events interrelated to the phenomena around him, then the basic educational strategy must be to modify the mother's theory of the child in ways that are held to be developmentally benign. In a word, we must change the mother's theory of her child (at least, some mothers; others, far wiser than we, became informal consultants on the development of curriculum). At the outset, we recognized that human beings - especially, perhaps, mother human beings - are not influenced greatly by didactic manipulations. You don't send them a book, or read them a lecture, or show them a film, and then say, 'Now you go do it'. But, on the other side, we were not sure what would work. So we adopted a complex strategy with several basic components: Having the home visitor model interactions with the baby, involving the mother as research assistant (keeping records, making occasional observations), and centrally of keeping the notion of interactional strategies in the forefront of the exchange among mother, baby, and home visitor. The home visitor was, therefore, a diagnostician and decision-maker in her own right. You will hear of the specific forms of these procedures shortly. Another word, if you will, on our prejudices of method. The child is not only a complicated field, he is also a highly adaptable one. That is, the social sensitivity of the child is so highly developed during the second year that there are many ways in which the study itself would begin to change him - his relation to the home visitor, his perception of his mother's attitudes toward the study, the impact of unusual materials and procedures. And, more, let it be said outright, the field of the home visitor is changed too as the study moves along. For this multiplex of reasons, we wanted to see the babies in a number of settings and with a variety of observational procedures. Our encounters with the children ranged from home to laboratory, from parties to experiments, from checklists to developmental tests, from maternal reports to videotapes. We tried to see the child and his mother from as many angles as we could feasibly manage, both because we recognized his likely variety and because we were deeply concerned about the involvement of any particular observer in the process of observation. Our hope, certainly not unique in the field, was to arrange a 'wraparound' of observational procedures that would provide a good first statement, at least, of the stability and variety of each child. First, a brief description of our assessment procedures. Mothers and

512 Social organizations children are seen four times during the year-and-half of their participation in the study (a pretest followed by three other assessments at six-month intervals), and each of the four assessments involves three separate meetings (each about an hour-and-a-half) with every family. The first two sessions stress the assessment of language and social interchange (these take place in the home). The third session stresses play and takes place in the laboratory. A final word on the problem of what to do with several thousand columns of data and interrelated sets of variables. Because it is expected and the way it has always been done, we will begin by selecting a small number of summary variables for univariate analysis. The more interesting and relevant task comes next. If we talk about the child as a field of events complexly interconnected to the world around him, and if we have proposed to bias that field, we are compelled to analyze a field of data rather than a series of independent, isolated points. Even more, our interest in the susceptibility to change of particular aspects of the child leads us to anticipate conclusions expressed as an elaborate matrix of change possibilities. Fortunately, our earlier studies and the panel research have helped us to appreciate the power of multivariate procedures. The adaptation of these procedures to educational designs is, perhaps, our most challenging next task. Finally, and of central conceptual status, these several procedures gave us an unusual opportunity to address the issues underlying the polemics I spoke about in my first words. We did not set out to ask whether or not the child can be changed significantly in the first years of life; rather, we tried to make a first assessment of the susceptibility to change of particular aspects of the child. It is, you see, our conviction that the classical argument about early experience has been wrongly drawn. Our central task as students of the young child is to make a systematic analysis of the possibilities of change - which will surely vary widely from one behavior system to another - and to relate that structure of possibilities to characteristics of the child, his parents, the setting of his early life, and, at last, whatever educational innovations are made in his first years. The final answer on the effects of early education will not be 'yes'> o r 'maybe' but rather will be an elaborate matrix indicating the likelihood that particular aspects of the child can be influenced in particular dimensions by particular kinds of situational or educational change.

Variations in home-based early education

513

Design of the study. Now, let me sketch out the educational study that has grown from our interest in young children. There are three groups of primary interest - children who are seen with a curricular concentration on play, language, or social development. For each of these groups, there is a three-way exchange among mother, child, and home visitor. The home visitor goes to each home 32 times, first on a weekly basis then on a two-weekly basis, and then once a month. My colleagues will tell you something about what happens during these visits. The basic theme of the curricula, however, is easily stated - to inform, to involve, to commit mothers. We are seeing the families in three replications of 6 babies each. Thus, when we are done, there will be 18 children in each of the three curriculum groups. In addition to the three groups of primary interest, there are three comparison groups, two serious, one proforma. In the two serious comparison groups, elements of the curricula are brought into the home but not in a way that will commonly involve mother, home visitor, and child, but only two of the three. Thus, the comparison group called mother-only is one where the baby is treated politely but not as part of the team (the analogy is obviously to the social worker) and in the second, the home visitor concentrates on the baby with the appropriate materials but keeps the mother at a polite distance (the analogy is to the baby teacher). Both these groups, let me emphasize, are designed to assess the impact of our materials when the exchange is two-person mother-home visitor or two-person child-home visitor rather than the three-way arrangements of our primary groups. Finally, there is the required testonly group, which gets a periodic assessment but no home visits. We recognize that this is, like others have been elsewhere, a risky design any conversation with a lonely young mother may have the same commanding effect - but we think it the right way to test propositions about the differential effect of different patterns of exchange with mothers and babies. Thus, the full design involves some 108 children in six groups of 18. In addition to his thirty-odd home visits, each child is seen in our laboratory or in his home several times over the 18 months of his involvement with us for systematic assessment - probes of changes in his behavior that might reflect curriculum effects. As you might guess, quite a few people have been workers in the study. In order to reduce the impact of observer variation on the data, we arranged for every home visitor to see babies in all experimental

514

Social organizations

groups. In other words, a home visitor would see a language baby, a social baby, a play baby, a mother-only baby and baby-only baby. Further, we have tried to schedule assessment observations in such a way that babies are seen by members of the staff who have not seen them in the homes and who do not know their group assignment. Of course, to the happy degree that our curricula made any difference, to that degree the observers can detect which group the child comes from. Let me underline here a couple of points that I hinted at earlier. First, the variety of humanity in its second-year forms is so great that any educational intervention is, on the face of it, unlikely to make a dramatic differential impact. We had mothers and babies marvelously sensitive to playful exchange but they were in the language group; we had responsive speakers and listeners in the social curriculum and so on. We are persuaded that, if you did not hold us to ramdom assignment and let us instead make the assignment of babies on the basis of one or two shot observations, we could show spectacular group effects. But, seriously, facing the wealth of diversity and having available only our present knowledge, we recognize with so many other early educators that we are at best adding a flute obligato to a fully functioning natural symphony when we bring our curricula to the homes of one-year-olds. The second point is the commanding character of the baby as a unit of observation. Our observers and home visitors have become sophisticated students of young children, and they well understand the analytic task of trying to make sense of variables, of differences among groups, of variations in standard test and observational instruments. And yet, even now, when the staff comes together to talk about the study, the home visitors want to talk about particular mothers and particular children, dividing the world not across psychological dimensions but across specific individual (to give that abused word its exact due) families. There is, in brief, a unity and integrity and (more to the point) sensibleness about the particular child in his particular family that makes him the natural unit of consideration. We have not solved this eternal problem, either administratively or in terms of method.

Variations in home-based, early education

515

B. PLAY: THE ELABORATION OF POSSIBILITIES by Greta G. Fein, Yale University, New Haven

The play curriculum is linked to the oldest but perhaps most fragile tradition in early education. Since the 17th century, philosophers have defended educational schemes based on the importance of physical activity, sensory stimulation, and experiences with ordinary, everyday objects for children's development. The opportunity to observe and manipulate a stick, a leaf, a cup, a shoe - simple encounters with natural or people-fashioned things - have been held to be essential experiences for the development of mental life. At least two issues divided these early viewpoints. The first issue was how thing-ideas were represented in the child's thought. Pestalozzi's position was perhaps the most static and physicalistic, whereas Dewey's was the most dynamic and functional. The second issue was whether planned adult intervention was necessary or beneficial. Here Rousseau emerged as an extremist on behalf of the child discovering the world on his own, whereas Pestalozzi eventually became the creator of a graded, adult-managed curriculum. Clearly, these issues are with us still today. Our play curriculum shares the assumption of these earlier views that the thing-environment of nature and human manufacture provides the raw materials for intellectual growth. But we assume more complex structures and processes. On the one hand, we do not view cognitive development as the accumulation of thing-ideas or correct, functional responses. The curriculum does not focus on the transmission of specific responses to a prescribed set of object dimensions. Clearly, all normal children acquire the perceptual categories of color, form, density, and texture, and the conceptual categories of object, space, time, and number. It is unlikely, however, that this knowledge results from explicit training to choose red rather than green or to sort circles and squares. It is even less likely that the more sophisticated structures arise from playing hiding games for a few minutes a day, from putting together puzzles, or from counting to ten. On the other hand, we believe that an environment geared exclusively to the convenience and requirements of adults is not necessarily useful to the development of children. The goal of the play curriculum can be conceptualized as the elaboration of possibilities. For the child, this means

516 Social organizations the elaboration or varied activities - the probes, tests, and finding out procedures from which the child derives his knowledge of the complex interrelations among the properties of things, actions, and consequences. For the mother, it is the elaboration of the child's opportunities to encounter an environment suited to his intellectual and physical capacities. The play curriculum is based on three quite simple assumptions about the implications of play for cognitive development. Consider for a moment the question 'What can you do with a cup?' A child's answer (in actions or words) might be bang it, wave it, fit it into another cup, put something into it, take something out of it, drink from it, or use if for a hat. Or, consider the complementary question 'What things can you drink from?' The child's answer might include a cup, and, in addition, a thimble, a nutshell, or a paper bag. The first assumption of the play curriculum is that play can be usefully conceptualized as the child's encounters with problems of actions and problems of things, that possibilities can be elaborated on either the activity side or the thing side; that many activities can be performed on a cup and many things can be either banged, fitted, or used as containers. Within this framework, the child's functional knowledge develops at the intersection of elaborated categories of actions and things. The scheme we are proposing is similar conceptually to that used by Wallach and Kogan (1965) and others who have investigated creativity in older children and adults. Recent studies by Yarrow, Rubenstein, Pederson and Jankowski (1972) and Clarke-Stewart (1973) suggest that the object experiences of young children may have important consequences for intellectual development. A second assumption of the play curriculum is that qualitatively different kinds of thinking appear in the child's solutions to action/thing problems. For example, the general problem of topological relations is at issue when the child fits one cup into another. Simple ordering schemes are involved when children arrange objects in a row; ordering and classifying schemes are involved when children build one tower of red blocks and another of blue blocks. Symbolic schemes, disconnections, and transformations may be involved when the child treats the cup as if it were a hat, when he feeds himself, his mother or a doll from an empty cup, a bottomless tube, or a tongue depressor. In varying degrees, these spontaneous play behaviors have been translated by other investigators into standardized test items with specially designed materials, detailed administration procedures, and carefully established performance criteria. The assumption that these behaviors reflect or index intellectual activity

Variations in home-based early education

517

is therefore not novel or controversial. More unusual, perhaps, is the added assumption that there is a fundamental tension between two major categories of problems which children pose, between problems which concern things 'as they are' and problems which concern things 'as they might be'. In Piagetian terms we are proposing a tension between the child's accommodation to the physical properties of material things and his assimilation of material things to an intellectual organizational structure. Recent research has confirmed Piaget's observations regarding the developmental sequencing of these activities. A report by Sinclair (1970) notes the emergence of pretend play at about 18 months of age and spontaneous ordering activities a few months later. Our preliminary findings suggest that, with age held constant, pretending and relational activities go together. That is, children who use things out of context, who transform one thing into another, who deal with things 'as they might be but are not', are the same children who fit keys into keys holes, fit caps on bottles, stack cups into or on top of one another, in brief, who deal with things 'as they are'. The third assumption of the play curriculum is that play flourishes in an approving, familiar environment in which the material resources are interesting and diverse, in which the child commands the initiative, adult expectations are scaled to the child's capacities, and 'yes-no' rules are clear and skewed toward 'yes'. Some recent findings from experimental studies of our own suggest that pretend play is enhanced by materials which clearly represent familiar, functional objects (a cup with a handle, a detailed toy crib), and by the presence of a familiar adult. In general, the object manipulation and exploration of children during the second and third years of life is enhanced by a change of materials (Mendel 1958) and is inhibited by the presence of strangers or the departure of the mother (c/. Ainsworth and Bell 1970). With respect to play opportunities provided in the home, our own observations and those of other investigators (Watts 1973; Clarke-Stewart, 1973; Yarrow et al. 1972) suggest that it is not unusual for children to be barred from cupboards and kitchen drawers, while their own, store-bought toys are inaccessibly housed in a toy chest or playpen. Indeed, toys are in playpens as often as children, and the children and the toys are not always there together. In addition, criticism, correction, pressure to conform to an adult standard seems to dampen children's enthusiasm for materials and to interfere with their openness to the creative potentials of material things. The general tenor of these findings is rather surprising in one respect. The support and

518

Social organizations

elaboration of play does not seem to require initiating, directive, teaching adults. Even the ameliorative program developed by Smilansky (1968) was highly effective when composed of the teacher's extensions of ongoing sequences initiated by the children. The available evidence thus suggests that the play of young children could be promoted by perceptive and unobtrusive maternal involvement and by a thoughtfully planned material environment. In sum, the overarching goal of the play curriculum is to encourage mothers and children to appreciate the possibilities of things. The program evolved to implement this goal was based on three assumptions about the play of young children: (1) That it is a generative system of action-object relationships; (2) that the juxtaposition, in play, of things 'as they are' and 'as they might be' has implications for cognitive development; and (3) that the environmental factors-both immediate and long term-which influence the level and diversity of children's play behaviors can be modified by informed and unobtrusive interventions by the mother. Curriculum strategies. Five general strategies emerged from our goals, assumptions, and research findings. (1) The first strategy is to deepen the mother's respect and appreciation of what the child is doing. The mother is taught a simple observational language to describe the child's activity. The language is adopted from our own research and involves verbs (such as bang, wave, fit, feed, sleep) and object labels. Sets of toys are brought into the home, and mother and home visitor record activities and objects on a simple observation form. For example, the toy set for one visit includes coffee cans and cardboard tubes of different sizes, along with different size balls and spools and a cigar box. These materials are presented to the child. Then mother and home visitor discuss and record what happens. For example, during one home visit the child banged one coffee can on another, banged the tube on the coffee can, picked up and examined 5 balls one after another, put the last ball in the coffee can, and stirred the ball with the tube. When the mother expressed surprise that young children do so much, the home visitor took advantage of the mother's reaction to develop the idea that these simple activities can be seen as child-initiated experiments through which the child obtains information about the properties of objects. The written observations provide recorded, concrete evidence of what the child has done, opportunities for home visitor and mother to explore the significance of his activities, to share their appreciation and admiration,

Variations in home-based early education 519 and, since these observations are saved in a baby book, opportunities to note changes over time. (2) The second strategy is to demonstrate that materials have a biasing influence on children's activities. Toy sets are planned to enhance particular types of activities. The set described above was intended to provoke children's exploration of topological and size relationships (small cans fitting into large cans, small balls fitting into small tubes, and large balls into large tubes - the balls roll through the tube and out the other end, an intriguing phenomenon for 15 month olds). Another set, designed to stimulate pretend dressing games, included an array of hats (straw hat, baseball hat, cowboy hat, helmet), large work gloves, chiffon scarves, bangles, necklaces, a spaghetti mop, a doll, a stuffed animal, and to it was added from the home a pair of mother's shoes, a shopping bag or a purse, and a pair of snow boots. The biased toy sets have two purposes: The first is to introduce the idea that children's play is modified by the materials given to them. The second is to introduce the categories 'pretend', 'relational', and 'manipulative' play. A new toy set is brought every week, and left in the home. Discussions emphasize categories of play and pose questions such as 'how do we know the child is pretending?' or 'does he really understand the relation between the size of things and how they fit together?' These questions are posed to highlight the problem of interpreting the child's intentions. (3) The third group of strategies involves the provision, management, and arrangement of materials and space in the home. In the early phases of the curriculum the question 'what is a toy?' is raised. The proposed answer is that to young children 'toyness' is what can be done with a thing. Many of the materials brought into the home are either made out of everyday objects (e.g., used coffee cans, empty plastic containers, toilet paper tubes) or ordinary household things (spoons, cups, coffee pots, keys, scarves, mops, brushes). Later in the series, manufactured toys, specially designed to illustrate physical and functional concepts are introduced. The point is to sensitize mothers to the 'toy' possibilities of a wide range of materials. Another issue concerns the organization of the home as a stimulating, safe place for a child to play, with rules clearly specifying what is permitted and what is forbidden. The mother is encouraged to set up play areas in convenient places, with toys stored in drawers, cupboards, and baskets to which the child can have easy access. The mother and the home visitor construct 'yes-no' play maps of the home which compare where the child wants to play with where the mother

520 Social organizations wants him to play. Mothers are encouraged to maintain consistent and clear 'nos' while offering ample and interesting 'yeses'. (4) The fourth strategy is to specify the ways adults can participate in children's play so as to maintain or elaborate it. 'Holding' techniques such as imitating and helping are distinguished from elaborating techniques. Within our analytical framework, the definition of 'elaboration' is relatively straightforward. Suppose a child is banging on a coffee can the activity is banging and the object is coffee can. The adult can propose an elaboration to the child in one of two ways: She can vary either the activity or the object. In the above example, the mother might change the object by proposing that the child bang on a pail, or she might change the activity by proposing that the child roll the can. The value of elaborative techniques is that the adult remains within the framework of the child; the child sets the theme which the adult then modifies with carefully chosen suggestions. The notion of elaborations is initially introduced by the home visitor in her own play with the child. The home visitor might comment on the child's activity. 'He put the hat on his head' and then add 'Let's see if he'll put a hat on the doll's head'. Eventually, the home visitor introduces the term 'elaboration', defines it in terms of the relationship between the child's object-actions and the adult's proposal. The elaborative proposals of the home visitor are discussed and noted on an observational form designed for that purpose. For many home visitormother pairs, elaborations have become a shared challenge. The following scene was recorded by a home visitor: 'The child put the bangles on his arm and the mother put a bangle on his foot. The child studied his bangled toes and the home visitor put a bangle on his other toe. At this point, mother and home visitor sat back and watched. Should they put a bangle on the doll's head? On the bottom of the cup? Or should they switch objects? What about a necklace on the child's neck or gloves on his hand? In the meantime, the child transfered a bangle from his arm to his foot, whereupon the mother covered his foot with a scarf.' The point, of course, was not to convince the child to put bangles or scarves on his feet, but rather to propose a variation and a possibility which the child can choose to accept, reject, or elaborate upon. Throughout, adult proposals are punctuated by appreciative observations of the child's behavior. (5) The fifth strategy pertains to the needs of a changing, growing child for developmentally appropriate experiences with materials. The curricu-

Variations in home-based early education 521 lum is organized into three major phases. The first phase emphasizes what children do when they play. In addition, mothers are asked to record the materials children avoid and those they find highly attractive, or that some materials lend themselves to more varied activities than others do. In this phase, too, mothers are introduced to adult play techniques which support or elaborate children's activities. The first phase of the curriculum is thus concerned with the autotelic, materials-oriented aspect of play, and with unobtrusive contributions of the adult. In the second phase, the emphasis shifts to manufactured toys - toys deliberately created to illustrate certain concepts, or materials geared to constructive or 'making' activities. These materials offer both opportunities and dangers. On the one hand, they offer children the opportunity to explore colors, shapes, and patterns. On the other hand, they often tempt adults to impose inflexible standards - the puzzle must be put together correctly, a picture of a house looks like this. The notion of elaborations is stressed when these materials are introduced. Mothers are encouraged to consider how the child uses the materials. For example, a puzzle with four duck pieces of different sizes is presented as a toy which might stimulate pretend play, or simple sensorimotor manipulation, as well as the activity of fitting the pieces into the proper spaces. We do not teach the mother to teach the child to conform to the apparent educational purpose of the toy. Rather, we suggest that the apparent purpose of the toy is but one of the ways in which the toy can have educational value. The third phase of the curriculum retains preceding themes but adds a new emphasis on the child's participation in household activities. The last series of home visits contain units on 'helping'. In one, the home visitor brings refreshments, paper cups, napkins, plates, plastic spoons, a small pitcher, and a sponge. The child is encouraged to help set the table, serve refreshments, and clean-up afterwards. Cognitive components of this activity (grouping things, establishing one-to-one correspondences), sensorimotor components (pouring), socialization components, and sheer pleasure are stressed. On different visits, the children polish shoes, plant seeds, and receive an introduction to household carpentry. In sum, the play curriculum employs five general strategies to enhance children's play opportunities: (1) Deepening the mother's appreciation of the child's interests and intentions; (2) demonstrating the biasing influence of materials on children's activities; (3) enriching the available materials and play space in the home; (4) promoting unobtrusive, elaborative play techniques; and (5) extending the sense and spirit of play to areas of

522 Social organizations mother-child contact which often become demanding and directive. Assessment. The variables of interest in the play assessment adhere closely to the goals of the curriculum. In examining the impact of the curriculum, we are concerned with the diversity, depth, and mode (pretend or relational) of children's involvement with objects, and the elaborative, facilitative proposals of the mother. Children's performance in free play and structured task situations is examined in order to contrast functional and test behaviors. The following variables are assessed at six-month intervals: (1) The diversity of the child's object encounters is assessed in two freeplay situations. Spontaneous activities with standardized test objects are scored according to the same criteria. In addition, observers rate the child's play (at various times and under various circumstances) along this dimension. (2) The depth of object involvement is assessed in the free-play situation and during a preliminary period in which mother and child are first alone, and then with a stranger. Two measures, the number of activities recorded for the most frequently used object and the number of 10-second intervals that the child has maintained contact with an object are recorded. (3) Mode of object contact (prented or relational) is coded in the freeplay situation, during the standardized tasks, and (at two assessment periods) during the mother-child play situation. Structured probes explore additional dimensions of pretend and relational problem-solving activity. Children's preference for objects which vary in their problem-posing characteristics is also examined. (4) The Bayley, Binet, and Leiter tests are used to examine children's performance in a structured task situation. Items from the Griffith's test, a one-one correspondence task, and several probes of problem-solving strategies are also used. (5) Mother's play style and overall level of play involvement is assessed in a mother-child play period. Maternal behaviors such as elaborative, helping, imitative, reciprocal, and unrelated intrusions are coded.

Variations in home-based early education

523

C. LANGUAGE: THE FORMATION OF DISCOURSE by Susan Starr, Yale University, New Haven

Unlike social development and, to some extent,the development of play, language is a frequent subject of contemporary early education curricula. Programs like those of Levenstein, Painter, and Schaeffer, are devoted to improving the vocabulary and syntax of children from one to two and a half. The curriculum reported here shares one major goal of its predecessors: A focus on vocabulary skills. Children's accumulation of a vocabulary of 200 to 300 words is one of the more striking developments of the second and third years of life. Since some children accomplish that challenging but necessary task far more easily than others, improving vocabulary becomes an important aim of early language programs. In addition to promoting vocabulary, our goal is to help children appreciate the multifold uses of speech. Our research and that of Nelson (1973) indicate that in this period the child becomes aware of the possibilities of language. Some children develop a language orientation toward describing objects. They learn nouns and adjectives. Others focus on social and expressive terms; their speech is devoted to expressing needs and desires. In our research we have also followed the child's growing awareness of language as a tool for communication. Initially speech seems to occur independent of the speaker's need to communicate, then at around 19 months the child becomes aware that speech can be used to tell some one something. Speech now occurs most frequently in interactive situations. At 21 months some children make the further discovery that language can be used to communicate with self; language becomes an accompaniment to play. Finally at around two years speech use is activated by both the needs of the speaker and those of the listener. Some children direct speech toward another when the other is listening and toward themselves when no one is listening. The work of Loban (1966) and Bernstein (1970) suggests that the child's early hypotheses concerning the function of language may have important implications for his later ability to communicate effectively with others. A child who focuses exclusively on expressive, social language, for example, might later have difficulty communicating in problemsolving situations. Thus our program is devoted to familiarizing the child with all the roles which language may perform.

524 Social organizations Like the programs of our predecessors, we regard the mother as the major source of linguistic information for her child. Her speech and activity are the primary variables through which we can affect the child's language. Operating from a Piagetian framework, we believe that the essential function of maternal speech is to provide a source of information to which the child must accommodate his growing knowledge of language. The child's language changes radically in this age span. If the mother's speech is to provoke accommodation, it must also change. Yet little is known of the precise characteristics of maternal speech which serve this function. Some tentative propositions have been made in formulating the curriculum. In the early stages of language acquisition linguistic information is provided when the mother acquaints her child with both the referential and expressive functions of speech; language can be used to describe the world and to make social contacts. In this period we encourage the mother to describe the child's activities to him as they occur so that he may see this as a proper function of speech. She also uses language in social games in order to suggest its expressive function. Description seems to be particularly beneficial for linguistic development; according to our data children who emphasize referential speech have larger spoken vocabularies and comprehend a greater number of descriptive terms at age two than those who focus more exclusively on expressive speech. In some respects the syntax of expressive children also appears to be more limited and stereotyped; in the period of two-word utterances a majority of their speech focuses on their desires, whereas referential children can syntactically express not only what they want but also what they are doing, what others do, what others look like, etc. For these reasons the early curriculum has placed particular emphasis on descriptive referential speech. As the child starts to speak the emphasis of the curriculum shifts from the child's hypotheses about the function of language to the development of language itself. Our research and that of others (Nelson, Pfuderer, Snow) suggests that the crucial characteristic of optimal maternal speech now becomes its responsiveness. At this stage the mother bases the content of her speech on the focus of the child's attention. She bases the complexity of her speech on the child's level of understanding. She is sensitive to his incorrect and tentative classification system. When he misclassifies objects she responds by gently correcting him. When he calls a truck a car, she says, 'Yes, it's a kind of car, that kind is called a truck'. Rather than, for example, merely informing him, 'No, that's not a car'.

Variations in home-based early education

525

When the child starts to speak in sentences, the quality of maternal speech apparently changes again (Halliday 1969). Now the mother engages her child in dialogues, asking him questions, responding to his speech, and asking him new questions to elicit more speech. By her example she teaches the child how to engage in adult conversation. Our data suggests that responding to what the child says is particularly important. Presumably the mother's expression of interest encourages the child to initiate conversation himself. Curriculum strategies. The focus of the language curriculum has thus been to promote children's vocabulary development and an effective use of language by affecting maternal speech. It has tried to influence the mother to create a linguistic environment which may promote her child's knowledge of words and his ability to communicate. Four major strategies have been used to modify maternal speech. (1) A primary strategy was to model the speech of the 'ideal' mother. Models were provided both by the home visitor and by other mothers and children on tape. In a typical visit using this strategy, the home visitor first described to the mother the general outlines of a specific speech style. She next played a tape of another mother using that style. While the mother and home visitor listened to the tape and looked at a transcript, the home visitor pointed out examples. Finally the home visitor again modeled the style. This time the mother kept a record of the home visitor's utterances and her child's responses. These records served to focus the mother's attention on the relevant aspect of the home visitor's speech as she recorded those utterances of the home visitor to which her child responded. Finally, the mother was asked to try the style herself. In the early part of the curriculum the home visitor modeled descriptive referential speech. She described the objects that interested and absorbed the child. She talked about what these objects did and what the child was doingwiththem.The motherwas encouraged to do likewise. Modeling was also used extensively when the child began to speak in sentences. The home visitor then engaged the child in conversations. Records of children in our research project indicated that certain subjects were most likely to produce extended dialogues between mothers and their children. These topics included what the child saw, what he had done, and, at a later stage of development, what other people had done. Tapes such as the following were played for the mother:

526 Social organizations Mother Show Sue how you read the book, alright? Ok. That's right What is that? What does the wow-wow say? Say doggie, not wow-wow. Say doggie. What's that? Butterfly, yeah. Count 'em. One, two, three, four. Do it again. One, two, three, four. What is this Chris? No, kitty-cat. What does the kitty say? Meow. No, kitty-cat. That's not a doggie, that's a kitty-cat. Kitty-cat. Tell mommy what the kitty-cat says? Balloon, that's right. Balloon. What's this? What's this? No, those are duckies. What does the duckie say? Yeah, balloon, that's right.

Child Ok. Ok, wow-wow. Wow-wow. Wow-wow. wow-wow. Un-un. voc., fly. Da doo. Da doo. Voc. Wow-wow? Woof. Voc. Wow-wow. That a wow-wow. Oh. A wow-wow. Meow, meow, meow. Voc. That. Balloon? Balloon. That. Chair.

Tapes of such conversations were played for the mother. The mother then tried to engage her child in a similar conversation. (2) With this first strategy the curriculum introduced the mother to new speech styles. In a second strategy opportunities were created for the mother to actually use these styles. At the initial stages of language development, the mother and home visitor identified situations which would be conducive to descriptive referential speech. These situations included dressing, feeding, taking walks, etc. The mother recorded objects that she could describe during these activities and kept a record of the child's response. Did he learn to say the word or to understand its meaning? The home visitor also provided books and toys with names that were easy to say. The mother was encouraged to describe them to the child. Several children quickly learned the name of a friction lady 'bug' for example, thereby reinforcing their mother's use of descriptive language. Some mothers found descriptive referential speech extremely difficult. This was clearly not the type of speech they normally addressed to a child.

Variations in home-based early education 527 When social expressive speech was emphasized, they were far more at ease. The home visitor provided games, rhymes, andsongs forthemothertoplay with her child. Hickory Dickory Doc, Old MacDonald were a few of the games which were presented as examples of how speech could be an adjunct to sociable, enjoyable, activity. The mother was also encouraged to show her child how to linguistically express his desires by teaching him words like 'want' and 'more'. Towards the latter half of the curriculum the mother and home visitor created situations for conversations. They took walks and talked to the child about what they saw during the walk. They let the child play alone and then asked him about his activities. Books also provided occasions for conversations and several were given during this period. (3) In the first two strategies, mothers were exposed to descriptive, social, and conversational speech. A third strategy addressed itself to the problem of maternal responsiveness. If the 'ideal' mother is one who bases her speech on the linguistic abilities of the child, then she must be aware of those abilities. The third strategy tried to inform the mother about her child's language development. Mothers kept records of their child's language development. For example, they kept a list of the child's first 50 words. Mother and home visitor noted what type of words the child seemed to be learning - names for people, food, animals, etc. Mothers were encouraged to teach their child the types of words in which he was currently interested. Records were also kept of the way in which the child communicated. For example, mothers noted how a child expressed a desire. They recorded the way this expression changed from a non-verbal communication, to a one-word utterance, to a more complete sentence. Many mothers expressed surprise at the communication abilities of their child. Tapes also provided an opportunity to inform the mother of the child's linguistic abilities. Some mothers did not believe their children were speaking until they heard them on tape. Early in the curriculum the mothers taped their children and identified the sounds the child used most frequently. They then tried to teach their child words using those sounds. Later mothers taped the child in a variety of speech situations. They noted how the child's growing sensitivity to the use of speech was shown in the way he changed his speech as the situation changed. Mother and home visitor also explored the child's language ability by giving him tasks to perform. They gave him commands and evaluated his comprehension. They recorded his understanding of contrasts, like big

528

Social organizations

and small, and remarked on his vocabulary. The looked at the way in which he defined words and classified objects. These tasks were usually repeated with several variations. The mother could then see how certain circumstances made language easier for her child. For example, the comprehension task was given with and without context cues to illustrate how dependent early comprehension is on context. (4) In a fourth, and final, strategy, mothers were provided with specific activities which might promote vocabulary growth. Books, of colors, shapes, and textures, were given to the children. Mothers and home visitors made scrapbooks of pictures to read to the child. Puzzles with shapes and colors were made available. Mothers and children also played games matching pictures to objects. This strategy, of course, lacked the general quality of those mentioned earlier. An example of descriptive referential speech can be generalized to many other examples, while a scrapbook cannot. However, some mothers seemed to need and to want the specifics of books, of songs, books and animals, and special language tasks to perform with their children. With these four strategies the curriculum has endeavored to create an environment which will foster language proficiency. In comparing these strategies with those of the play curriculum, the central contrast becomes the role of the mother. In play the mother is encouraged to let the child take the lead. The child suggests and the mother imitates, aids, extends, and elaborates his suggestions. This philosophy reflects the theory that play is autotelic and personal. There is no correct way to play and no socially sanctioned goal of play. Maternal responsiveness is thus a key concept in the play curriculum and in social development as well. The mother is encouraged to base her actions on those of the child. As pointed out earlier, responsiveness has an important role in linguistic development. However, unlike play, language has a fixed goal, set by others. The child, after all, is learning to speak in the way agreed upon by his society. Thus the mother cannot only respond to the child's idiosyncratic language; she must also initiate language learning situations in which she can provide a model for her child. Through her speech she shows him how language can be used and what things are called. Through her conversational technique she gives him practice in how to communicate with others. Thus the language curriculum has, like play, placed emphasis on informing the mother of what her child can do. But it has also struggled to provide her with ideas about what she can do to promote language development.

Variations in home-based early education

529

Language assessment. Each language assessment includes a measure of the child's vocabulary, his use of speech, and his comprehension and production abilities. Maternal speech is also evaluated. (1) The assessment of the child's production ability takes place in a 20-minute period during which the child's speech is taped while the observer also keeps a written record of what he says. Using the tape, the written record is later corrected. The final transcript is scored on a number of dimensions; The number of words in the child's speech, and his mean proficiency in language production. (2) The tapes are also used to assess vocabulary development. The words produced are classified according to parts of speech and according to whether they refer to people, places, acts, objects, etc. These classifications provide information on the variability and the specificity of the child's emerging vocabulary. At 18 months few words are produced in 20 minutes, therefore, the mother is also given a list of children's early words and asked to indicate those her child says. This list is classified and evaluated for variability and specificity in the same manner used for the transcripts. At 24 and 30 months the Palmer Concept Inventory is used as a more standardized measure of vocabulary development. (3) Comprehension is assessed with tests developed in our longitudinal research. Since the general comprehension level of the children changes, these tests vary in content from assessment to assessment. They all consist of a series of commands read by the mother to her child while the child plays with a standard set of toys. The tests have good reliability and are highly correlated with other measures of language development. (4) The assessment of language use is also based on measures developed in our research study. In this study we have been able to trace the child's hypotheses concerning the proper function of speech from their early diffuse nature to their more sophisticated later forms. For example, the sophisticated child at 24 months uses speech both egocentrically, in play, and for interactive purposes. Thus at the 24-month assessment the observer keeps a record of the child's speech. During this period she notes every 10 seconds whether the child is playing or interacting. The proportion of utterances spoken during play and during interaction can then be calculated for each child and their use of both types of speech compared. Earlier probes looked at a more primitive development, the dependence of speech on whether anyone is listening. Later probes focused on the child's ability to carry on a conversation.

530 Social organizations (5) Maternal speech is measured during a 10-minute segment in which the mother reads a book to her child and then shows him a standardized set of toys. This segment is taped, and the transcript is later scored for the complexity of maternal speech, its style, and the variability and specificity of her vocabulary. In addition, during the period when the child's speech is taped, the observer notes how frequently the mother responds to her child's utterances.

Variations in home-based early education 531 D. SOCIAL DEVELOPMENT: ENRICHING CONNECTIONS by Alison Clarke-Stewart, Yale University, New Haven

The third curriculum in our study is different from all other programs of early education in that it focuses exclusively on the child's social development. When one thinks about the problems involved in designing such a curriculum, it is immediately obvious why this type of program has not been popular. For one thing, an emphasis on social development is clearly out of synch with the current mood of American education, which stresses cognitive development and the acquisition of academic skills. But even more, there are problems with such a program (1) because it is difficult to conceptualize - perhaps because we have no universally accepted standards for social development; (2) it is difficult to formulate specific program principles because our knowledge about social development, even more than about cognitive or language development is sketchy and has lacked a productive theoretical framework; (3) it is difficult, even once conceptualized, to communicate - unaccustomed as we are to talking about such abstract and diffuse concepts as love and responsiveness, and finally (4) it is difficult - even if these problems can be overcome - to implement such a program because it is by its very nature so deeply personal and individual. Consequently, the social curriculum described here consists of a collection of rather tentative and exploratory strategies which we expect will foster children's social development. The goal of the social curriculum is conceptualized as the enrichment of interpersonal connections - in particular, and primarily, the connection between the child and his mother. The program does not presume to create bonds - but to supplement, enrich, and make more enjoyable those which already exist. The mother-child relation is central in the social curriculum because it appears to be the key to the child's social development. Research suggests that it is the most important single bond for children under three, at least for those living at home with the mother as primary caretaker. It is also likely that this relationship critically affects the child's relations with people other than the mother and is therefore an important factor in the child's continued social development beyond the initial tie. This suggestion, as well as being a basic tenet of psychoanalysis, is suppported by evidence from the observational study of children's social development which was conducted in conjunction with this educational program.

532

Social organizations

From this study it appears that the mother-child connection can provide security for the child and joy for the mother; it can increase the effectiveness of mother and child in regulating each other's behaviors; it offers a model for subsequent relations between the child and other people. If we accept the mother-child relationship as fundamental and central in the child's social development at this early age (1 to 2\ years), the main questions then become: What does an ideal or optimal mother-child connection look like, and how can we deliberately foster that ideal? In answer to the first question, Mary Ainsworth has provided one vivid glimpse in her observations of children's attachment behaviors in unfamiliar settings. One-year old children range in their behavior toward their mothers in a strange room from totally ignoring her and spending all their time exploring the surroundings, to clinging to her continuously and refusing to be separated. But the optimal attachment behavior appears to be that of children who are able to use the mother as a secure base; they are happy to explore as long as they can return periodically to the mother, especially in stressful circumstances, for reassurance. Similar patterns of attachment behavior were observed in the study of mother-child interaction which was completed prior to this project, referred to by Mr. Kessen (Clarke-Stewart, 1973). In that study 36 mothers and children were repeatedly observed as they interacted more-orless naturally at home from the time the children were 9 months until they were 18 months old. Children who were most competent in all areas of development showed a pattern of development in which attachment to the mother was balanced with independence. Although they spent over 80 % of the time they were awake in the same room as the mother, they were physically close to her (within 4 feet) only about half of their time, and in actual physical contact only from 7 % to 20 % at 11 months and even less (4 % to 14 %) at 17 months. Children who were lowest on measures of overall competence, on the other hand, either never touched their mother or were in physical contact with her more than 30 % of the time. The competent children responded positively to their mother's approach and negatively to her leaving - but in moderation (they responded positively to 35-60 % of occurrences of 'mother comes' and negatively to 10-30 % of 'mother leaves'). The responses of poorly developing children were more extreme in either direction. Competent children demonstrated the highest proportions of smiling at mother, having eye-to-eye contact with her, and playing with her. In fact, these mother-child pairs interacted more frequently in every mode except physical contact and routine

Variations in home-based early education

533

caretaking. Moreover, during these interactions the behaviors of mother and child were more likely to be contingent upon and responsive to the actions of the other person. Neither mother nor child was operating in isolation; their contact was reciprocally interactive. The development of an optimal mother-child relationship appeared to depend upon the pairs working out mutually satisfying and balanced interaction patterns. The next question, for the curriculum then, was how to promote or enrich such mutually rewarding mother-child relationships. Again, the data from this observational study (which were congruent with observations by prior investigators such as Ainsworth, Yarrow, Lewis, Moss and Robson), were helpful.' In particular, the following results were incorporated into program strategies for the social curriculum: (1) The child's optimal attachment to his mother and his positive involvement with her were positively related to indices of his competence in other spheres of functioning, particularly language and cognitive development. (2) There were three kinds of maternal behavior which were most closely related to the child's competence. They were social stimulation (looking, playing, and talking to the child), expression of affection (smiling, caressing, and speaking warmly and positively to the child), and responsiveness to the child's social behavior (responding immediately, contingently, positively, and appropriately to the child's expressions). (3) On the average, in these families (which were comparable to those in our educational sample) affectionate and playful maternal behaviors were very infrequent (accounting for only 4 % of the observed time the child was awake). (4) When patterns of causal direction over this 9-month period were examined, it was discovered that the child's social behavior was a potent elicitor of positive maternal attention at a later time. If the infant looked, smiled, and vocalized to the mother frequently at 11 months, his mother stayed in the same room more, was more responsive to his distress and demands, and was more affectionate to him at 17 months. This clearly demonstrated the child's contribution to the mother-child relationship. (5) The examination of causal patterns also suggested the possibility of cyclic role reversal by mother and child. Over the course of the particular 9-month period of observation, it appeared that maternal attention at 11 months led to infant attachment at 14 months, which in turn increased maternal attention at 17 months.

534 Social organizations Curriculum strategies. From these empirical results, then, emerged the following set of program strategies for the social curriculum: (1) One strategy was to provide opportunities for enjoyable social interaction between mother and child which promoted mutual smiling, laughing, touching, talking, and eye-to-eye contact. Such opportunities were used to encourage the mother to initiate games and social activities with the child (since such stimulation enhances children's competence) and to maximize the likelihood the child would smile and 'turn on' the mother (since the child's social expressions increase the closeness of mothei and child). The home visitor offered a rich supply of nursery rhymes and children's songs, familiar finger games and less known physical exercises, pastries and birthday cakes, trips to the park, the zoo, the museum or a neighbor's house, picture books and simple stories, and play materials like balloons, playdough, finger paints, jack-in-the-boxes, puppets, mirrors, dress-up clothes, bathtub toys, and old magazines - all intended to make playing together exciting and fun for mother and child. Not only did the home visitor provide opportunities for new social activities, however, she also encouraged the mother to elaborate playfully on the child's spontaneous social activities, to elaborate on traditional games like peek-a-boo or pat-a-cake which mother and child had already discovered, and to incorporate social play into otherwise boring or disagreeable caretaking routines. Social activities were an essential part of each visit; they were also supported between visits. Occasionally, to make the importance of social play more salient, the mother was asked to keep a social diary for the child during the week. Or sometimes the home visitor left toys or suggested activities for mother and child to try out during the week. At first, social activities during home visits involved only mother and child. After the first few visits, when the child seemed ready to accept her participation, the home visitor joined in too - but the mother-child relationship remained central. Later, after 6 months or so, the mother was encouraged to invite father, grandmother, other family members, and, still later, other children to play too. Parallel to this trend of enlarging the mother-child relationship was a trend toward increasing the physical distance between mother and child. The suggested social activities for mother and child became decreasingly dependent on physical contact and increasingly verbal in nature. Mothers were informed about the rationale for such social activities and were encouraged by conversation and example to be uninhibited in expressing their enjoyment during such interactions.

Variations in home-based early education 535 (2) A second curriculum strategy was to provide opportunities for social interaction with other mothers and children. On special occasions like Christmas, Halloween, Easter, or Mothers' Day, social parties were held, and all 18 families in the social curriculum groups were invited. Similarly, friends and neighbors were often asked to participate in home visits in the last third of the program. The activities at these social gatherings included movies, games, an appearance by Santa Claus, and deliberate manipulation of children's prosocial and antisocial behaviors (for example, by modeling, sharing, demonstrating aggression with a Puncho doll, providing few or many toys, and so on). They were accompanied by observation and discussion between mother and home visitor, either concurrently or at the next home visit. As well as just being lots of fun, these social occasions offered social stimulation for the mothers, a chance for the children to experience some independence and contact with a variety of responsive adults, and provided a source of information for the mothers about other children and about the dynamics of children's interactions. (3) The third general strategy was to create situations which would foster the mother's awareness and appreciation of the child's unique individual qualities, his particular abilities, and his increasing maturity especially in the area of social relations. This strategy was implemented in several ways: (a) One tactic was for the home visitor to structure the situation during the visit in order to demonstrate specific abilities of the child, while the mother filled out an observation form. For example, the home visitor wound up and put near the child a scary toy robot, or the mother was asked to ignore the child briefly, to leave the room, or to wave good-bye and leave the house - in order to elicit attachment behavior. Or, at another visit, a stranger was invited to the house, and the mother was asked to record the child's reactions to a new person. (b) These observational records were preserved in a 'Baby Book' which the mother kept, and such demonstrations were repeated at different ages to provide 'then and now' comparisons. (c) The home visitor also provided opportunities for the mother to compare the child with other children in similar situations. During the visit, for example, the mother and home visitor sometimes took the child to a neighborhood park or playground, where they could observe and compare ongoing behaviors of different children. During the home visits following social 'parties', too, mothers and home visitors were able to

536

Social organizations

discuss contrasting behaviors of different children. Another source of comparisons was 'home movies' which illustrated the reactions of different children in identical situations. In one movie, for instance, three children, 8, 15, and 22 months old, were each shown in a sequence which consisted of interacting with the mother, being offered a cookie, having the cookie taken away, looking at a jack-in-the-box, and being approached by someone wearing a Hallowe'en mask and growling loudly. This same scenario was then carried out with the study mother and child at the home visit. In all these comparative situations the home visitor avoided evaluative judgments and focused on the individual strengths of each child particularly of the child in the study. (d) The home visitor also provided information about child development to prepare the mother for anticipated changes in her child's behaviors (for example, toward increasing independence) while helping the mother to keep records of the child's current status in the Baby Book. (e) Finally, the mother's appreciation of the child was fostered by the home visitor's judicious but enthusiastic praise of the child, throughout. The desired outcomes of this general strategy to increase the mother's awareness, respect, and appreciation of the child as an individual, were for the mother to give the child greater independence and initiative rather than imposing her own desires and standards on his behavior, and for her to express her appreciation through positive behaviors to the child. (4) A fourth strategy guiding the social curriculum was to demonstrate to the mother her own importance in the child's development, to convince her of her influence on his behaviors. (a) One way this strategy was implemented was by demonstrating and discussing how the child used his mother as a secure base, especially in unfamiliar or threatening situations. (b) Another tactic was to demonstrate the mother's impact as a model. This maternal role was introduced by the home-visitor's reading the poem 'Recognition'. '/ don't want to hear another word!' I hear my daughter scold. 'Dear me/' I think, 'She's awfully strict For a playful three-year-old!' She rolls her big eyes heavenward And sighs with great disdain. ' What am I going to do with you!'

Variations in home-based early education Her dolls hear her complain. 'Sit down! Be still! Hold out your hands! Do you have to walk so slow? Pick up your toys! Go brush your teeth! Eat all your carrots! Blow/' I start to tell her how gentle A mother ought to be When blushingly, I realize She's imitating ME! (Barbara

537

Burrow)

Situations were deliberately set up to elicit the child's imitation of the mother's words, sounds, and actions - for instance, in the context of action songs like 'Here We Go Round the Mulberry Bush', 'I'm a Little Teapot', 'Did You Ever See a Lassie?', and 'All Around the Kitchen'. The mother might then be asked to keep a record of the child's imitations during the next week. (c) Another tactic for demonstrating the mother's importance focused on her role as 'communications expert'. In this role it is necessary for adults to both give and receive effective communicative messages. To illustrate the impact of the messages she sent the mother was asked to observe and record the immediate effects of her expressive behaviors to the child, for instance, smiling, crying, waving, saying 'no, no', showing the child a toy, or giving mixed messages by smiling and saying 'no' at the same time. The meanings of nonverbal as well as verbal communications were examined and discussed frequently by mother and home-visitor. The social curriculum also stressed the importance of the mother's reception of messages from the child. A number of observation exercises were designed to increase the mother's sensitivity to the child's expressive behavior, particularly his nonverbal signals, in different situations, at different times, on different levels. Examples of these observation exercises can be seen in the following description of some typical scenes from a visit (this one at 15 months) which focused on the question of how the child expresses his needs, desires, and emotions. The home-visitor suggests, in simple language, that the child will develop social skills only if his attempts to communicate are recognized and responded to appropriately by the mother. She mentions changes which have already occurred since the program began three months ago. Then she and the mother look together at a form entitled 'What might these behaviors of a young child indicate?' The form includes behaviors

538

Social organizations

like lifting up arms to mother, stamping feet, lying on the floor, holding on to mother's leg, banging spoon on table, rubbing eyes, and so on. The mother tells the home-visitor how she would interpret such behaviors (in specific contexts) or how she would find out what the child intended. To illustrate further the interpretation of nonverbal expressions mother and home-visitor try matching photographs of children with descriptions of specific emotional reactions in the child (for example, by offering him a cookie, taking away a toy, or approaching him with a scary mask) which the home-visitor photographs with a polaroid camera. These photographs are examined, discussed, and put in the Baby Book. To follow up the theme of communication, the mother is asked to fill out a form during the following week which asks 'how does your child tell you - that he is angry, afraid, tired, that he wants to play, to be alone?' At other visits, exercises included observation of the child's behaviors related to independence, the frequency and circumstances of his expressions of intense enjoyment, of anger, of fearfulness, observation of his behaviors toward the mother and other people in a free-play situation with other children. At the same time as these attempts were made to sensitize the mother to the meanings of the child's behaviors, her awareness of her own responses to these behaviors was also promoted, and she was encouraged to respond immediately, positively, and contingently. Techniques to foster responsiveness included guided observation of behavioral chains of sequences in mother-child interaction, recording of the mother's use of 'no', 'don't', and 'stop it' directed to the child, and demonstration of the effect of positive reinforcement on a single, quantifiable behavior such as vocalizing. (5) The final strategy guiding the social curriculum was one of general education or consciousness raising. Mothers were encouraged to think about and helped to articulate their attitudes and goals for social development (for example, they were asked to describe their notion of an 'ideal' four-year-old), and then they were given information about alternative ways of looking at issues, alternative roles, styles, and so on. The issues which were thus discussed included sibling rivalry, violence, sex roles, social goals, social rules, and discipline. The issue of changing roles for men and women was particularly popular, and materials to spark such discussions were plentiful in the press. Mothers, and home-visitors, were also eager to share experiences, memories, and attitudes related to this issue. Mothers were exposed to other ways of thinking - for instance, the

Variations in home-based early education 539 ideas of child development 'experts', pediatricians, other mothers, and the home-visitor herself - but there was never any pressure to accept these other ideas. A word in conclusion - you may rightly infer that more than any other curriculum, the social curriculum reflects the individual characteristics of mother, of child, and of home-visitor. The home-visitor gives more of herself to the interaction - confidences, openness, frankness, warmth; she is guided more by the mother's goals, tempo, tolerance, receptivity, and ability to comprehend and utilize curriculum ideas; and she is more responsive to the child's spontaneous social activities. Sensitivity and flexibility are essential characteristics not only for good mothers but for effective home-visitors as well. Assessment of social development. The most important source of information about the effectiveness of the social curriculum is the observation of naturally occurring mother-child interaction in the home. A minimum of one hour of such observation is recorded at each test point. From these observation records are derived measures of the relative frequencies, duration, and absolute amounts of specific maternal and child behaviors, particularly looking, talking, touching, smiling, and playing. Since the method of recording observations allows for the analysis of sequences of behaviors, measures of the contingent responsiveness of each partner to the other's behaviors are also available. Three probe situations are also included in the social assessment: (1) A stranger - either an unfamiliar adult or an unfamiliar mother and her young child - comes to the home, and behaves toward the child in certain specified ways (looking, talking, approaching, playing with, ignoring, and so on). As the stranger goes through her paces, the reactions of the child to the stranger are recorded. (2) After the stranger leaves, the mother is asked to carry out certain activities vis-à-vis the child. These include leaving the room, leaving the house, returning, approaching the child, talking to him, playing with him, ignoring him, and so on. From these procedures are derived measures of the child's attachment to his mother and his relative independence, and indices of the mother's ability and eagerness to enter into playful social interactions with the child. (3) The third probe situation consists of a series of choices presented to the mother and to the child in order to ascertain their positions on various dimensions. Choices for the mother are designed to elicit indications of

540

Social organizations

the mother's stress on stereotypic sex-role definitions, her stress on independence for the child, her own attachment to the child, her preferred mode of social interaction with the child (caretaking, physical contact, teaching, playing with toys, talking, conventional games, and so on), and her play or teaching style (does she use reinforcement, demonstration, verbal direction, description, physical manipulation, or let the child learn or play on his own?) The specific choices include things like: Would you rather brush Susan's hair or read her a story? Would you prefer to have the observer play 'house' with Peter or play 'house' with him yourself? Would you rather have John play with a truck or a doll ? Would you prefer to show Lisa how these blocks fit in the holes or let her figure it out for herself? Would you rather teach Darren to sort these red and blue beads into two piles or play Ring-Around-a-Rosie with him? After the mother is asked for her choice she indicates how difficult it is for her to make a decision, and then her ease of carrying out the activity is observed and rated. The extent of the child's participation and cooperation in these mother-child activities is also rated. Choices for the child are presented nonverbally to discover his preference for his mother over the observer, the observer or his mother over a novel toy, and his choice among various toys.

Variations in home-based early education 541 E. THE EMPEROR'S NEW CLOTHES by Joseph Glick, City University of New York

The distinguishing feature of these papers is their collective bravery. From the outset the contributors recognize that they are treading on the edges of a conceptual abyss and have somehow found the strength to go on. Since there is little as yet to document the effectiveness of their endeavor, I should like to devote this discussion to the architecture of the abyss as a way of documenting how brave these hardy souls are. Although the local topic involves the discussion of a variety of curriculum plans oriented toward biasing the manner in which parents interact with their children, I believe that the conceptual problems of this sort of endeavor are ones that are general to psychology (or education or any discipline dealing with the unruly human animal). Accordingly, it is to this more general problem that I address myself today. One of the central features of the modern bind has been indicated in Kessen's and Fein's introductory remarks: The environment is not a static and completely specifiable entity but is rather constructed and selected by the action of the organism. The notion that psychologists must deal with the psychologically effective and not merely the physically present environment has been with us for a good length of time. I can remember cutting my graduate student's eye teeth on Uexkiill's penetrating 'stroll through the worlds of animals and men' and Kafka's chilling report of an adventure on Lake Constance. Yet it has fallen on the last fifteen years, with the recovery of interest in cognitive psychology and the American's re-discovery of Piaget, to fully impress upon us the gravity of the environment problem. The message has been loud and clear - the organism selectively attends to particular features of an externally describable environmental array. Not only that, since organisms are active, and are capable of action on the environment, they are also capable of creating different sorts of environmental events by the very character of their actions. A ball may be a ball to us - but to a kid it can be a throwable, a suckable, a bouncable and, in the brave new world of the new math, countable. The same environmental object, depending upon its knit into a skein of ongoing activity can take on a number of different, and for the moment, unrelated properties. This state of affairs makes a psychology that needs a starting point in

542 Social organizations some fixed property particularly difficult. Where is the stimulus? What is the stimulus? And in what sense is the for-now stimulus the same when our actions take on different directions and become located in different contexts? These assertions about the stimulus world add up to the notion that somehow our descriptions of a stimulus must take into account the active and selective properties of the stimulus recipient - and hence we can no longer take solace or even find hope in psychologies that depend upon a separation of stimulus and response. A stimulus must depend in part on the response propensities of the organism. With this we step a step closer to limbo. Now, these remarks have pointed meaning for those who would take the giant step of moving from the 'stimulus' (that seemingly isolatable entity) to the 'curriculum' - which can be likened to a stimulus on a grander scale. From what has been said above, it would seem that we need to be very cautious in believing for one minute that there is such a thing as a curriculum that can be presented to people. We can try, and try again, but we must be unsettled by the realization that our stimulus, or curriculum, is something that is understood 'for us' and is not necessarily that which is understood either by parents or children who are the object of the curriculum. The significances as presented may not be those as received. In my contacts with this research group I have often found them puzzling over the indignities that a random assignment design strategy imposes upon the clarity of their efforts. One often finds a 'play mother' in a 'social curriculum' and a 'social mother' 'languaging' to beat the band. Not only that, but there are a notable few who are ideal 'mother onlies' languishing in a curriculum plan that must include the child. The parents seem to select that aspect of the curriculum that most meets their interests, needs, and capabilities and then respond to that. When assessment time comes it might be an interesting endeavor to analyse results in several ways - organizing curriculum groups by the intended curriculum and then re-organizing the groups by the describable features of the received and selected curriculum. I wonder where the greater consistency lies? Some of these papers have made a notable effort to recognize problems of this sort and it is part of this recognition that leads them to define their curriculum in generative, rather than substantive, terms. I believe that they are correct in believing that if any chance exists in curriculum imposition

Variations in home-based early education

543

it is on the level of attempting to alert parents to the significance of certain developmental areas and hoping that this will form a basis for the parents' generating their own curriculum. But I am still not sure how much can really get through. I do have some hope for the generative notion. I, myself, am involved in a parent education project wherein we hang around a pediatric outpatient clinic and offer our Black and Puerto-Rican mothers a short course in child development and early curriculum. I recently completed an interview with one of our mothers who claimed that her attitude toward her child had been transformed. At first she viewed her two-year-old as a child with limited capabilities for learning and as the course progressed began to see the child as capable of more and more. She claimed that this in fact transformed the manner in which she approached the child - expecting learning and providing opportunities for the child to do what she expected. One might hope that this sort of change (mirrored in this case by a measured IQ jump of some 35 points) had something to do with what we had tried to get across as a generative notion of a curriculum. However, there are, as well, disquieting elements in this case. The mother's parents had run reading courses and Head Start programs. Could it be that all we had done is take a theory of child behavior that the mother had and merely applied it to a new case - own child who was below Head Start age? This last observation serves to introduce another major bugaboo in our psychology. Just as the stimulus or curriculum has a particularly indeterminate characteristic so, too, does the response-to-stimulus or learningfrom-curriculum have an indeterminate characteristic. We have learned with increasing force over the last few years that just as the stimulus has an interpreted, selected, and constructed characteristic, so too must action (or response) be described as action to some purpose, response for some end, or under some construal of what is expected. This sort of view has been forced upon me by some cross-cultural work that I had done with Michael Cole, John Gay, and Donald Sharp some years back (Cole, Gay, Glick and Sharp 1971). We discovered that for many different classes of activity (learning, memorization, classification, and logical problem solution) subjects were capable of giving responses that looked stupid under certain conditions and antithetical performances under slightly altered conditions. One anecdote that particularly sticks out in my mind (and which has not grown stale by its repetition) concerns

544

Social organizations

a classification task that we were doing. We had compiled a motley array of objects which were partially classifiable into foodstuffs, tools, clothes and utensil categories. Subjects were asked to approach the aggregated heap of these objects and to 'put the ones together that go together'. Our subjects produced functional categories - a knife was placed with an orange because 'the knife cuts the orange' etc. When asked for the rationale for these sortings, the subjects would claim 'this is the way a wise man would use these objects' etc. Finally, in desperation, we asked 'how would a fool do it?' The answer was 'a fool would put all the food together, all the tools together, etc'. The point of this anecdote is clear - we are often dealing not with cases of capacity or incapacity but with a realm of cognition that might be, for want of a better term, called the domain of 'rules of application'. What do people consider to be reasonable behavior in such and such a context under such and such a set of instructions. Vis-á-vis these curricula, there exists the parallel problem of applying these general rules to particular circumstances. One can say 'be contingently responsive' but that takes on many meanings when a mother lives out that proposal. We know practically nothing at this point about people's considerations of reasonable and wise behavior. We know less about how these construals relate or do not relate to our refined professional knowledge or experimental expectations. Without knowing these things, we are playing a game without knowing the rules. The relevance of these considerations to a curriculum project should be manifest. Does the curriculum, if it is successful, operate to alter the array of capabilities that people have - or does it rather operate on the domain of rules of application of things that are known but not applied? If it is the latter, then the demand upon our sensitivities is great - for we must serve as diagnosticians to find where the capability that we wish to apply to a certain occasion already exists - and where the barriers to its application exist, and work with these to produce criterion performances. These considerations serve to transform our understanding of psychology to one which necessarily involves a good deal of 'clinical case understanding' (another characteristic met with in the working of this project). For a recognition of the theoretical and methodological understandings of recent psychological work can no longer allow us to arrogantly divide our field into experimental searchers after basic truth and those

Variations in home-based early education

545

clinicians who are stuck in the vagaries of the particular case. Where I have gotten to is the feeling that the vagaries of the particular case are the essential tools with which we can build a more general theory of psychological functioning. We must know the stimulus as construed, the response as intended, and the rules governing the relationship between these two. And we will not get to this understanding by pretending that the stimulus exists and that the response is clear. My hope is that by deepening our understandings of particular cases we may emerge with some truly general and meaningful laws. And if we do not, I will still not despair. It is reassuring to think that somewhere beyond our intentions, our curricula, and our imperial dreams of the Complete Psychology, there are individual people shaping their own lives by rules of their own. It may be, that by a weird twist of the old story, the new clothes may show us that the emperor was never really with us.

2

ROBERT P. KLEIN

Comparison of model preschool programs

A. REMARKS ABOUT CURRICULUM

IMPLEMENTATION

by Jenny W. Klein, Department of Health, Education and Welfare, Washington, D.C.

This paper deals with implementation of various curriculum approaches into the education component of Project Head Start. As the person responsible for managing the Planned Variation project, I was able to visit many of the project sites, observe in classrooms, talk to classroom and administrative staff. In addition, I was privileged (and sometimes burdened) to read the very voluminous reports of Planned Variation consultants. I will divide my presentation into three parts: (a) A description of the project; (b) a presentation of some of the findings on curriculum implementation which were analyzed and reported by Carol Lukas and Cindy Wohleb, under the direction of Marshall Smith of the Huron Institute; (c) my own observations and conclusions about curriculum implementation into Head Start programs. Description of the project. The Planned Variation experiment was an effort to explore the effects of a variety of curriculum approaches to preschool education. A similar study was being conducted by Project Follow Through, a program designed to continue in elementary schools the comprehensive compensatory educational efforts begun in Head Start. The data obtained from the coordinated programs of Follow Through and Planned Variation was expected to provide information in three

548

Social organizations

general areas: The effects of various educational strategies on children in Head Start, the effects on the child of a continuous intervention program begun in Head Start and extended into the primary grades, and the process of curriculum implementation (including the variables affecting implementation). The curriculum models or approaches chosen for the study generally span the spectrum of opinion in early childhood education. Each model or curriculum is based on strong philosophical convictions and has origins in accepted learning theory. One can attempt to order the curriculum models in a variety of ways. The Office of Child Development has divided them into five basic groups: The pre-academic models, the cognitive discovery models, the discovery models, the parent educator models, and the enabler model. These basic groups vary in their theoretic perspective from behavioral psychology to psychosocial developmental thinking. They generally differ in basic philosophy, approach to learning and teaching, goals and steps to reach goals, classroom procedures, role of the teacher, teacher training, and curriculum materials. Not all well-established curriculum models were included in the study. Program sponsors selected for participation in the Planned Variation Program were required (a) to be operating ongoing Follow Through models in communities where Full Year Head Start was available; (b) to have a curriculum approach defined for preschool children (ages 3-5); and (c) to be willing to work within the framework of Head Start. The Planned Variation study was begun in the 1968/69 school year and officially ended in 1972. A total of about 2000 children participated in the first wave, and 5000 in each succeeding wave. One-third of these children comprised the comparison group and attended regular Head Start classes. The rest comprised classes implementing the 12 experimental curriculum models in thirty-seven communities. (See the Appendix for a description of the curriculum models.) The 1972/73 year was considered a phase-out year, and only implementation was observed. We are now planning to have a look at at least a random selection of classes in the 1973/74 school year to see what parts, if any, of the curriculum are still in place. Findings on curriculum implementation. The Planned Variation study was initiated by the Office of Child Development. Data is now available for the first two years. Stanford Research Institute was responsible for most of the data collection during the entire study and for the analysis of the

Comparison of model preschool programs

549

data for the first year. Since spring 1971, the Huron Institute has been handling the sampling design, data analysis, and reporting. Curriculum implementation can be defined as the extent to which classroom practice matches both model theory and prescription. Thus full implementation would mean that all aspects of the model are present in the classroom. During the first two years implementation data was collected in the following ways: Individual classrooms were rated on three instruments - sponsor ratings of teachers, teacher questionnaires, and teacher aide questionnaires. Classroom observations, using an observation instrument developed by Jane Stalling, were obtained in some classes. Consultant reports described individual classrooms and also contained information on model implementation in each site (rather than in each class). A great deal of clinical and descriptive information was obtained from reports of early childhood education consultants who visited each site on a regular basis. In addition, individual child case study reports by staff from the Department of Human Development at the University of Maryland have contributed a rich source of information on children in the various Planned Variation models. Thus, data about implementation is available in a variety of forms. However, the only direct measures of degree of implementation come from sponsors' ratings of teachers. Each sponsor rated every head teacher on her performance in the specified curriculum model on a 0-9 scale. The scales were administered both in February and May. The results show that generally all sponsors rated models as moderately well implemented in May with a mean ranging from 3.7 to 6.5. None of the models was rated as outstandingly well implemented. Second-year programs were not rated higher on implementation than first-year programs. Interestingly enough, there is little difference in degree of implementation among models. However, large standard deviations point to significant differences among sites within the same model. To me the most significant finding, however, is the large difference in degree of implementation within each site. Approximately 58 % of the variance in ratings lies within sites. Available data indicates that many models were not clearly specified and thus were not only difficult to implement but implementation was difficult to measure. Lucas concludes her very detailed analysis of the data by saying that implementation is an interactive process depending on sponsor input, staff input and reaction, and on the operational context of the staff.

550

Social organizations

Observations about curriculum implementation into Head Start programs. Implementation of a model into a preschool organization such as Head Start is a complicated process. In fact, it would probably be easier to start an altogether new program with a particular specified curriculum. Introducing change is always difficult. Gilbert and Mosteller suggest that 'most innovations don't work, even when they are introduced with the best will in the w o r l d . . . A few years ago, Gross, Giacquinta and Bernstein (1971) made clear that implementing organizational innovation is complicated and in fact frequently just doesn't occur. One factor which complicated implementation in this project was that although staff and parents of Head Start programs had a choice in accepting or rejecting the Planned Variation program, they had no choice in the selection of a particular model. Thus, in fact, each community in the study was given a specific model and expected to implement it in their Head Start classrooms. Many problems and unresolved issues became evident soon after the Planned Variation program was initiated. Some of the models were not clearly specified and thus could not be easily disseminated. The lack (or absence) of a simple set of operational specifications for most models meant it was difficult if not impossible to accurately determine the degree of model implementation. In order to achieve model implementation, not just a single variable was involved as many program people expected, but rather a comprehensive educational program had to be instituted. The Head Start programs into which these models were implemented were not a single entity consistently comparable in size, philosophy, available resources, staff training, political climate, etc. Further, the Head Start programs as well as the curriculum models continued to change throughout the study and thus were part of an ongoing developmental process. Sponsor support varied widely in type, frequency, and in the sponsor's flexibility. All of these problems and unresolved issues made it quite evident that implementing and testing a curriculum model in an existing Head Start program presented a very different situation from implementing and testing such a model in a laboratory situation. Consultant reports describe the reality with which sponsors, staff, children, and parents had to deal. For example, in one community the church which housed classes had burned down and classes could not be held for several weeks. Curriculum materials arrived several months late in a few programs. In many projects where resources were sparse, Planned Variation curriculum materials were shared with other classes

Comparison of model preschool programs

551

and were not always available to model classes. Many program budgets did not allow for teacher substitutes, and when a trained teacher was ill, classes were combined and model implementation made difficult, if not impossible. Trained staff members left the program for better jobs and more certain future employment, and implementation of the model ceased until others could be trained. Sponsor's field representatives sometimes had difficulties relating to local program staff, parents, or visitors and sometimes found that many people were in charge. In 1970, Planned Variation consultants who had been visiting programs on a regular basis exchanged ideas and reached agreement regarding some of the most important variables which facilitate or inhibit curriculum model implementation. These variables pertain to the model, the sponsor or modeler, the program site, and teacher/staff factors. Model variables include clarity of model requirements as well as the complexity and inclusiveness of model specifications. Models vary in their vulnerability to physical plant limitations. There are variables between models in the complexity and clarity of teacher role, teacher style, and teacher skill required by the model. Another model variable is the ease with which parents can learn model specifications and subsequently participate in classrooms. Sponsor or modeler variables which affect implementation include those which have to do with intensity and adequacy of pre-service, in-service and on-site training offered by the sponsors. There are sponsor variables related to the intensity of 'moral' support given on-site and related to the skill, knowledge, resourcefulness, and degree of commitment to local staff and Head Start Community of the sponsor's field representatives. The clarity of the field representative's role varies between sponsors. Other sponsor variables relate to the extent of sponsor's and field representative's experience with early childhood programs and young children. Among the on-site variables affecting implementation are measures of adequacy of the Head Start program before adoption of the model and of adequacy of physical facilities, materials and equipment as well as the accessibility to resources, materials, and equipment needed by the model itself. Program site variables include such areas as knowledge of and receptivity to the model in the local community, the local political climate (e.g., race relations, school politics), and the suitability and adaptability of the model to local culture and customs. There are program variables related to administrative smoothness vi. friction and variables describing regional Office of Child Development relations. Variables in receptivity

552

Social organizations

of local public school system to the model can be noted as well as in communication between Head Start and Follow Through personnel (i.e., preschool and public school personnel). Teacher and other staff variables reflect such areas as classroom staff's attitude and commitment to the model's central concept, goals, and techniques as well as staffs' ability, personality, prior training, and experience. Smoothness vs. friction among teaching staff is another teacher/ staff variable. It is likely that each model is especially vulnerable or resistant to some of these variables. However, in retrospect, all of the variables are important primarily because they have an impact on the teacher's behavior in the classroom. The teacher, i.e., the classroom staff, is the prime determinant of curriculum implementation. Although each model has its own curriculum and characteristic way of implementation, in each one the teacher plays the central role. In all models for successful curriculum implementation, the teacher must not only understand the model requirements but must also be fully aware of his role as prescribed by the model, and be willing to carry out that role at all times. I recall visiting programs and observing teachers whose incomplete or incorrect understanding of the model hindered successful curriculum implementation. Some teachers followed the model's basic philosophy during specified classroom time but slipped back into their own natural role and style on the playground and at other times. Thus, a teacher in a reinforcement theory model would forget about reinforcing correct responses except during 'lesson time' and a teacher in a model which required a nondirective teaching approach would become very directive during snack time, outdoors, or on a trip. I watched one teacher whose model prescribed programed materials 'teach' an entire lesson as directed by the instructor's manual, but from the wrong lesson sheet without realizing they did not correspond. Another teacher failed to understand her model's principle of providing children with concrete experiences wherever possible before introducing them to related abstract concepts and symbols. Consequently, there were no books available for children's use, since the teacher thought she should not expose them to symbols of experiences they had not had. A third teacher clearly misinterpreted her model's principle of encouraging freedom by providing opportunities for children to choose and make decisions. The classroom was chaotic: Guidance and order were completely absent. Curriculum implementation is difficult. While it has been successful

Comparison of model preschool programs

553

in laboratory situations, it presents a variety of problems in the real world. One of the problems we face in many of our educational programs is that we look for easy solutions to complicated problems. We turn to new curricula to help us improve our education, new systems, new books, and often new gimmicks. We try one curriculum, and if results can't be seen in a year or two we drop it and try a new one. The literature is full of new reading methods, new ways to learn this or that. N o doubt, some new methods have value, and new models, systems, and ideas should be developed and explored. However, I would hope that researchers would turn to study the implementation process and help us join information about the processes that occur in the classroom. If the treatment, i.e., the model, differs from classroom to classroom, and from site to site, we cannot easily compare results. One important fact we certainly learned from this study is that what goes on in a classroom is a complex process. Implementing a model curriculum is not a panacea. The Bank Street college model developed and sponsored by the Bank Street College of Education in New Y o r k City represents a developmental approach in which the ultimate objective is to enable each child to become deeply involved and self-directed in his learning. Activities are planned both for groups and for individual children. The classroom is the child's workroom where he is free to investigate objects and explore materials. Opportunities for concrete, sensory, and motor activities are interrelated with opportunities for functional and expressive use of language. The behavior analysis model was developed and is sponsored by Donald Bushell of the University of Kansas. The goal of the program is to teach children needed skills by means of systematic reinforcement procedures. The teacher's role is that of a behavior modifier. Individual instruction is emphasized through the use of programed materials such as the Sullivan reading program (McGraw-Hill) and the Suppes mathematics program (Random House). The academically oriented preschool model is sponsored by Wesley Becker and Siegfried Engelmann of the University of Oregon. Originally developed by Mr. Engelmann and Carl Bereiter, this model is premised on the belief that every child can achieve well academically if he receives adequate instruction and if there is a payoff for learning. The approach used is structured and academic. Programed materials are used to teach essential concepts and operations in reading, arithmetic, and language. The cognitively oriented model, developed and sponsored by David Weikart of the High Scope Educational Research Foundation, presents a

554 Social organizations cognitively oriented preschool program derived from the theories of Piaget. The program has three main foci: (a) The curriculum, which is cognitively oriented; that is, the teacher plans activities that help children develop the ability to think about physical and logical relationships; (b) the teacher, who participates actively in developing classroom programs; and (c) the home, where the teacher works with the mother to promote cognitive growth in the children. Learning objectives are stated as behavioral goals. A pragmatic action oriented model sponsored by the Education Development Center at Newton, Massachusetts, was inspired by the English Infant Schools. The objectives of this model are (a) to provide a flexible curriculum and schedule; and (b) to fashion classroom environments responsive to the individual needs of children as well as to the talents and styles of each teacher. The Institute for Developmental Studies model originated by Martin Deutch was one of the first models specifically designed for the education of disadvantaged preschool children. It emphasizes both the cognitive and affective development of young children. Specific areas of attention include concept formation, perception, language, self-image, and socialemotional growth. Instruction is individualized through classroom management techniques, continuous teacher assessment, and work through small groups. The Florida Parent-Educator Model, developed and sponsored by Ira Gordon of the University of Florida, uses both classroom and home instruction through parent-educators. A parent-educator is a mother from the local community who works with the parents in their homes, presenting each with weekly tasks. These tasks are individualized for each child to develop cognitive skills. The parent-educators also assist in the classroom. The Primary Education Project Model is developed and sponsored by the Learning Research and Development Center of the University of Pittsburgh. The most critical component of this model lies in an individual progress plan for each child, which allows the child to work through finely graded steps of curriculum at a rate and in a manner suited to his own needs. The curriculum emphasizes the basic skills and concepts the child needs to learn to master a variety of subjects. The Responsive Model, designed by Glen Nimnicht of the Far West Laboratory for Educational Research and Development, is focused on helping children develop both a positive self-image and intellectual ability.

Comparison of model preschool programs

555

Emphasis is placed on increasing the child's sensory and perceptual acuity, language development, concept formation, problem solving, and abstract thinking. The child is encouraged to explore and find answers from responsive people and responsive materials and equipment. The Responsive Environment Corporation Model uses specially designed learning materials and educational technology within a flexible classroom structure. The approach maintains a balance between structured and nonstructured activities and emphasizes individual, independent work. The talking page learning system (a teaching machine that 'reads' to children) is used for language development. Self-correcting sensorial learning materials are provided for learning basic concepts and developing abstract reasoning. The Tucson Early Education Model, designed by Marie Hughes, is sponsored by the University of Arizona. This model emphasizes the development of language competence, intellectual base, motivational base, and social arts and skills. The curriculum is carefully structured but is flexible enough to allow for choices in activity and behavior, thus providing the children with opportunities to develop individual skills at individual rates. The Enabler Model, designed by the National office of Child Development Office, emphasizes consistent technical assistance, rather than a specific curriculum. During monthly visits to the local programs, an early childhood educator first helps the program's staff and the children's parents formulate program goals and then assists them in developing means of attaining the goals. Emphasis is placed on using staff and other local resources. Since this model requires no specific curricular methods, the consultant's records serve as documentation of the techniques being used.

556 Social organizations B. SHORT-TERM COGNITIVE EFFECTS OF ELEVEN PRESCHOOL MODELS by Marshall S. Smith, Harvard University, Cambridge

During the early months of 1969 the Office of Child Development planned a three-wave longitudinal study designed to assess the relative impacts of a variety of preschool curricula. The study was called Head Start Planned Variations (HSPV) and began in the fall of 1969. Plans called for the systematic assignment of a number of specified curricula, each to two or more sites throughout the country. Selected sites were to meet three criteria. First, each site was to contain an on-going Head Start Program. No funds were allocated for serving children other than those already being served by Head Start. Second, each site was to draw participant children from a preschool population living largely within the attendance area of a school or schools where older children attended a Follow-Through program. By fall 1969, most Follow-Through schools had adopted one of a number of well-defined educational curriculum models. Follow-Through is being evaluated by the Office of Education. Third, a selected Head Start site had to agree to adopt the curriculum model being used in the Follow-Through schools in its site. Aid in implementing the models was provided by consultants responsible to the original architects of the models. In addition, extra funds for purchasing equipment and for hiring teacher aides were provided to the selected Head Start classes. Overall, the cost of implementing a Planned Variations model is estimated to be $350 per child above the cost of conventional Head Start (McMeekin 1973). The design of the Planned Variations study called for children in all three waves to be tested at the beginning and end of their Head Start experience. Following Head Start, the children would enter the FollowThrough program in their community and be evaluated at the beginning and throughout their Follow-Through experience. The records of the Head Start and Follow-Through testings could then be linked. The linkage would provide data for a longitudinal assessment of the combined pre-school and early elementary school experience of the Planned Variations children. Testing was also planned for other groups of Head Start children in every Planned Variation site. These children would attend Head Start classes without a designated curriculum component and

Comparison of model preschool programs

557

serve as a local comparison group for the study of the Planned Variations Head Start classes. Progress reports on Planned Variations were planned at three times during the course of the study: First, at the end of the Head Start experience for each of the three waves of children; second, at various times during the Follow-Through experiences of the three waves; and third, in 1976, after the final wave of Head Start children had completed FollowThrough. A preliminary report on implementation of the models during the first year of 1969-1970 was prepared in 1971 by the Stanford Research Institute (SRI) for the Office of Child Development. This paper focuses on data from the second wave of Head Start Planned Variations. 1 It summarizes results pertaining to the short-term (one-year) cognitive effects of Head Start models on children. Methods and procedures. Thirty-seven sites (communities) used Planned Variations curricula in 1970-71. Twelve curricula models were represented. 2 There were comparison classes at 14 of the 37 sites and at seven (off-site) locations not having Planned Variations classes. Fifteen of the 37 sites were also in the Planned Variations study in 1969. There are a number of things to note about the design. First, nine of the twelve models have at least three sites. Although it would have been preferable to have greater replication, the fact that some systematic replication did occur strengthens inferences about model effects. Second, the study is not confined to one region of the country. Although there is some regionality associated with some of the various models, e.g., the Florida model sites are all in the South, by and large, the models are being tested in a variety of locations. Third, a number of sites do not have comparison classes in their same location. This tends to complicate analysis of the data. Due primarily to economic constraints, not all sites were assessed on all measures. There were three levels of evaluation activities. Table 1 below 1. Four reports were prepared by the Huron Institute on these data. The first report considers the process and success of implementing the Planned Variations curricula in the various sites (Lucas and Wohlleb 1972). The second report presents a detailed summary of the different measuring instruments used in all three waves of the study (Walker et ed. 1973). The third report explores the possibility that different characteristics of children interact with particular curricula to produce different outcomes (Featherstone 1973). The final report focuses on the short-term main effects of Head Start in general, the Head Start Planned Variations models taken together, and on the specific main effects of the various models (Smith 1973). 2. For description of the models see the chapter by Klein in this volume.

558

Social organizations

shows the number of sites per sponsor included in each of the levels of evaluation. Level I was the most primitive. Nine Planned Variation sites fell into this category. Data gathering at this level did not directly involve children. Teachers completed demographic information forms and filled out the California Social Competency Scale for each child in their classrooms in both the Fall and in the Spring. Teachers and teacher aides responded to questionnaires requesting information about their backgrounds, teaching experience, training, and attitudes. Sponsors and Head Start Directors rated the level of implementation in the classrooms in every site. Since direct measures of child outcomes were not included in data collection activités, sites at this level were not included in analyses of effects of HSPV on children. This paper does not report on findings from the Level I sites. Thus we report here findings from only eleven of the twelve sponsors (NYU had only one site and it had Level I data collection activities). Tabel 1. Number of sites per sponsor at each level of evaluation Far Ari- Bank Ore- Kan- High Flor- EDC IPI West zona St. gon sas Scope ida Level I 2 Level II 1 Level III 2

1 2

1 1 2

1 2

1 2

1 1 2

1 1 2

REC NYUEnablers 1

1 2

1

2 3

1

Data collected at Level I were also collected at Level II. In addition, three other sets of data were gathered at Level II. Classroom observations were made in both the Fall and Spring by observers using the SRI Classroom Observation Instrument. All children in tested classrooms were administered the Basic Test Battery in both the Fall and Spring. Four tests were included in the Basic Battery - the Caldwell Preschool Inventory, NYU Booklet 3D (a 17 item test of achievement - pre-math, pre-science and linguistic concepts), NYU Booklet 4A (an 18 item test of achievement letters, numerals, prepositions), and the Motor Inhibition Test. Children of Black and Spanish origin, whose parents agreed, were administered a test assessing their knowledge of their ethnic heritage. Ten Planned Variation sites were classified as Level II. Of the ten, six had comparison classes also tested at Level II. Level III sites had all data collection activités carried out in Level I and Level II sites and, in addition, four other activities. A randomly

Comparison of model preschool programs

559

chosen one-half of the children in each tested Level III classroom were administered the Stanford-Binet in both the Fall and Spring. The same children received the test both times. The children in the other half of the class - along with one of their parents or guardians - were administered the 8-Block Sort Task in the Spring. Additionally, the parents or guardians of the children in this group completed a parent questionnaire which asked about attitudes toward Head Start, their child, and the Planned Variations model used in their child's classroom. Eighteen Planned Variation, ten on-site comparison and five off-site comparison, sites were assigned to Level III. The data were collected by the Stanford Research Institute (SRI). By and large, we have great confidence in the quality of the data. (See Walker et al. 1973, for details.) The strategy for analysis was dictated in large part by constraints on the study. First, we focused principally on the analysis of cognitive growth. To do otherwise would have been to seriously overplay the existing data. In doing this we recognized that we were not even attempting to capture the richness of a preschool experience or the largest part of the differences among preschools. Second, also as described earlier, Planned Variations sites were not randomly assigned to models. The sites were selected on two criteria unrelated to the requirements for an adequate experimental design, and then given the opportunity to accept or reject the assigned curriculum. Moreover, the local community had control over the specification of which classes within a site were to employ the Planned Variations curriculum. Since the selection of comparison classes within the PV site occurred after selection of the Planned Variations classes, the treatment and comparison classes cannot be assumed to be random samples drawn from one population. Thus, randomization did not occur at eitherof the two critical design points - at the level of assignment of curricula to sites or at the level of assignment of treatment and comparison groups within sites. Since we did not have random assignment, all estimates are potentially biased in some unknown fashion. There are numerous statistical techniques to help reduce bias (matching, covariance, crossed designs and standardization techniques). Each may be helpful depending on the adequacy of the structural model we are trying to fit. That is where the essential problem lies, for we have no a priori way of determining the appropriate model. Given this state of affairs, we followed Tukey's advice: 'As in the famous discussion between Student and Fisher and the interjections by Sir Harold Jeffreys, it may not be a bad thing to use all the allowed principles of witchcraft and not just one set' (Tukey 1972, p. 112).

560

Social organizations

We did not use all the principles of methodological witchcraft, but we did use a number. In particular, our strategies for removing bias in the data depended on (a) our choice of a statistical model; (b) our choice of variables; and (c) our assessment of the accuracy with which the data are measured. Different decisions in these areas of judgment led to a variety of estimates of 'effects'. The variability of estimates from different analyses gives us a sense of confidence limits for the reported effects. Such an approach will generally inspire caution in interpretation, for most of the effects found in this study are small. On the other hand, large effects which turn out to be insensitive to variations in analysis methods presumably should inspire confidence. Summary of findings in main effects study. Three main questions were addressed in this study: (a) What are the overall short-term effects of a Head Start experience on children? (b) Are there discernable differences between the effects on children of a Head Start Planned Variations experience and a conventional Head Start experience? (c) Do Planned Variations models differ in their effects on Head Start children? Five measured outcomes were used to assess each question. The PSI is a general standardized achievement test for pre-school children. The N Y U Booklet 3D and N Y U Booklet 4A are tests of specific achievement areas. The Stanford-Binet is a well known test of general intelligence. The Motor Inhibition test assesses a child's ability to control motor behavior. Overall short-term effects of Head Start. With regard to the question of overall short term effects of Head Start we reached four conclusions: 1. The Head Start experience substantially increased children's test scores on all five outcome measures. On four of the five outcome measures children's scores were estimated to increase 'naturally' over the seven or eight months of the Head Start program - even had the children not been exposed to Head Start, their scores would have risen. For two of these measures (PSI and Booklet 3D) the Head Start experience was estimated to double the 'natural' rate of growth. For two other measures (Booklet 4A and the Motor Inhibition tests) the Head Start experience was estimated to better than triple the 'natural' rate of growth. Increments attributable to Head Start ranged from 0.26 standard deviations (for the Motor Inhibition test) to 0.82 standard deviations (for the Booklet 4A test). On the fifth measure, the Stanford-Binet, our estimates indicate that the scores of children in this sample would have 'naturally' decreased by

Comparison of model preschool programs

561

about 0.30 standard deviations had they not attended Head Start. The Head Start experience arrested this apparent decrease and further increased Head Start participants' Stanford-Binet scores by roughly 0.35 standard deviations (see Table 2). Table2. Gains for the total analysis sample on five outcome measures*

Test

Observed gain (total)

PSI Book 3D Book 4A Motor inhibition Stanford-Binet

0.94 0.73 1.15 0.36 0.35

Attributable to maturation (estimated) 0.50 0.36 0.33 0.10 —0.30

Attributable to Head Start (estimated) 0.45 0.36 0.82 0.26 0.64

* Observed gains are partitioned into two components - gains attributable to maturation and gains attributable to Head Start experience. All gains are expressed in individual level pre-test standard deviations.

2. Children who had a prior preschool experience gained less overall ('natural' + Head Start related growth) than children for whom 1970-71 Head Start was their first year of preschool. This effect held for all outcome measures and for most of the subgroups. If, however, we allocate the total gains for the two groups of children between 'natural growth' and the Head Start experience, we find that the effects attributable to Head Start are roughly equal for children with and without prior preschool experience. This indicates that the expected 'natural growth' for children with prior preschool experience is less than for children without prior preschool. In other words, a common preschool experience may partially overcome the effect of age differences among children on the five outcome measures. Some support for this notion comes from the fact that variances on four of the five outcome measures are somewhat smaller at post-test time than at pre-test time. Preschools may have a 'fan-close' rather than a 'fan-spread' effect on children. 3. Children who would enter first grade (El) directly from Head Start tend to gain more than children who would enter kindergarten (Ek) directly from Head Start on the Booklet 4A, Booklet 3D, PSI, and Stanford-Binet tests. On the Motor Inhibition test the Ek children gained more. (The average age of El children when they entered Head Start was 65 months - Ek children were roughly one year younger.) The greater gain for El children was most pronounced for the Booklet 4A test and least for the Stanford-Binet (see Table 3). When the gains attributable

562 Social organizations to Head Start were examined, the effect appears to strengthen, though they are still small for the Stanford-Binet. These effects are probably due to a combination of two things. First, the larger gains attributable to Head Start for El children on the cognitive measures and particularly the Booklet 4A test (a measure of letters, numerals, and shape names) may be due to older children's advanced academic readiness. Second, there may be a greater interest by Head Start teachers in El sites in preparing children for reading and arithmetic. 4. There are no consistent differences among Mexican American, Black, and white children in their Head Start gains on the five outcomes. A brief mention of the methodological procedures used to arrive at these conclusions is in order. Since we did not have a group of 'control' children (children who did not have the benefit of an Head Start experience), our estimation procedures relied on natural variations in prescores for children of different ages. The reader, therefore, is warned to treat these data as rough estimates and to evaluate for himself the assumptions of the procedures. Analyses of the third wave of data (1971-72), when there was a real 'control' group, indicate that the estimates are generally accurate (Smith 1973). Effects of HSPV vs. comparison Head Start. The second major question regards overall differences in effects for Planned Variations and conventional Head Start programs. Although we consider the question, it should not be thought of as very important. For while we might expect there to be differences among PV programs in their effects on the five outcome measures, we have little reason to suspect that there should be systematic differences between an overall PV effect and an overall effect of conventional Head Start programs. This question totally obscures systematic differences among treatments. One rationale for studying the question is to determine whether the extra funds allocated to PV Head Start programs had a consistent effect on the measured outcomes. Our conclusion supports the findings of a large number of recent research efforts which have failed to detect any systematic relationship of gross expenditures to variations in outcomes. We conclude there are no differences in effects between the PV programs (taken together) and the comparison Head Start programs on any of the five outcome measures. Differential effects of HSPV models. The third question addresses dif-

Comparison of model preschool programs

563

ferences among HSPV programs in their effects on Head Start children. Table 3 crudely summarizes our findings regarding differential model effectiveness. The eleven HSPV models are the rows of the table while the five outcome measures are each represented by a column of the table. The cell entries indicate effectiveness relative to the other HSPV models and to appropriate comparison HeadS tart classrooms. Five general conclusions may be reached after inspection of this table. 1. We began the study with the expectation that there would be few strong differences among the models in effectiveness as assessed by our five outcome measures. This expectation was realized. Table 3 indicates that for each of the outcome measures we have classified the majority of the models as having average effectiveness. No model stands out as either more or less effective than the others on more than two of the five outcomes. There are no overall winners or losers. Table 3. Summary ofplanned variation model effectiveness on five outcome measures* Model Far West Laboratory Arizona Bank Street University of Oregon University of Kansas High Scope University of Florida EDC University of Pittsburgh REC Enablers

Book 3D

Book 4A

PSI

Stanford Binet

Motor in!

0 0 0

0 0 0

0 0 0

0

+

0

0

0

0

++

0 0

0

0

++

+

0 0

0 0

0 0

+

0 0 0

+ +

+ —

0 0 —

—

0

—

0 0 —

0 0

+

0

0 0 —

0

+

* Zero (0) indicates model is of average effectiveness on outcome measure. Minus (—) indicates model may be of below average effectiveness. Plus ( + ) indicates model may be of above average effectiveness. Double plus ( + + ) indicates model is probably highly effective.

2. A second expectation was that models which emphasized academic drill combined with systematic reinforcement would be more effective than other models on the four cognitive outcome measures. This expectation was realized only for one of the four cognitive measures. Only

564

Social organizations

for the Booklet 4A measure - a test assessing knowledge of letters, numerals, and shape names - is there evidence of greater effectiveness for the models emphasizing drill and reinforcement. The University of Kansas model is the clearest example of this finding. We found it to be clearly superior to all of the other models and to the comparison classes in its effectiveness in raising Booklet 4A test scores. The two other models which emphasize academic drill (University of Oregon and University of Pittsburgh) both appear to be above average in their impact on this test. On the other cognitive tests there is no indication of special effectiveness of these three models. These findings are at some variance with the finding of other researchers in the preschool area (see Bissell 1970 and White et al. 1972). These researchers indicated that there may be a general positive effect of structured academic emphasis and drill on cognitive tests. Our data, however, indicate that the effect is specific rather than general. In particular, it appears as if this approach may be more effective for imparting information that is easily taught through systematic drill while it is only of average effectiveness in other cognitive areas. 3. One model clearly stands out as more effective than the others in raising Stanford-Binet test scores. The High Scope PV model appears to increase Stanford-Binet scores by an estimated twelve to fifteen points, roughly 0.9 individual-level standard deviations. The average effect of other PV and comparison models is on the order of three to four points or roughly 0.3 standard deviations. The effect of the High Scope model is particularly strong in one Southern rural site where the measured average gain is slightly over thirty points. Although we can probably attribute some of the measured gain to testing and regression effects, the 'corrected' gain is still on the order of a very substantial twenty points. Analyses of the item profiles of children in the High Scope sites indicate that the gains apply over the entire range of appropriate Stanford-Binet items. The particular effectiveness of the High Scope model on the StanfordBinet does not appear to generalize to the other outcome measures used here. For three of the four remaining tests the model appears to be of only average effectiveness. On the fourth test, Booklet 3D, there is some indication that the High Scope model may be of above average effectiveness, but no firm conclusion may be reached from the data. 4. Two of the eleven models (University of Pittsburgh and REC) account for 40 % of the 15 cells in Table 3 where there is an indication that a model has other than average effectiveness on an outcome measure.

Comparison of model preschool programs

565

No other model is rated as other than average on more than two of the measures. Three things are common to REC and Pittsburgh. Each uses some form of programed instruction, each was a first-year model in 1970-71, and each has only one site in this study. Although the first two common elements may be important, our inclination is to view the fact that each model has only one site as the principal reason that these models have more than their share of 'other than average' effects. This may be due to differential effectiveness of models in different sites or to uncontrolled biases in our data. Whatever the reason, our inclination is to be very sceptical about attributing clear effects to any model with only one site. 5. All models are rated as showing average effectiveness on the PSI test. We had not expected this result since our preliminary analyses of the PSI indicated that it is probably our most reliable measure. In retrospect, however, we suspect that the reason for the lack of clear differences among models on the PSI is due to the nature of the test itself. The PSI was developed as a general test to assess the overall impact of preschools on children. As such, it attempts to measure a wide range of skills probably rendering it relatively insensitive to particular differences among curricula. Thus it is probably more appropriate to the tasks of assessing the overall average impact of preschools and of individual differences among children. Conclusions. A number of the findings of this study are not surprising. Numerous studies have shown that Head Start has a clear and strong short-term effect on cognitive outcomes (see, e.g., Datta 1969 or Stearns 1972). And many more studies of school effects indicate that it is unlikely that one curriculum will show a clear advantage over other curricula in the production of cognitive growth (see Jencks et al. 1972). What is surprising, however, is the seemingly clear relation between the goals of a preschool curriculum and the child outcomes 'produced' by the curriculum. The success of the academically oriented curricula on the test of letters and numerals and High Scope's success on the StanfordBinet are the two principal examples. These findings, along with the apparent insensitivity of the Preschool Inventory to different curricula, seem to indicate that evaluations of preschool (and school) curricula should attend more to the goals of the curricula than to the psychometric characteristics of the tests. If, for example, we had used only the most 'reliable' and 'valid' test in this study (the PSI), we would have had very little of interest to report.

566 Social organizations C. DIMENSIONAL ANALYSIS OF PRESCHOOL PROGRAMS by Louise B. Miller, University of Louisville, Kentucky

It was pointed out in the early 1960s by Medley and Mitzel (1963) that: 'Progress in the development of scientific knowledge about teaching behavior awaits the development of a set of ... dimensions, which are educationally and scientifically significant, so that a substantial portion of the differences between different classrooms can be described quantitatively and related to other variables' (p. 309). Most classroom observation systems have been tailored to the needs and purposes of a particular study. As long as this situation obtains, the results of various studies will be difficult to compare, and the body of research on instructional methods will remain fragmented. More importantly, describing classroom activity is only one part of a problem which confronts all investigators in the social sciences: How does one select categories for the description of human behavior? Is it possible to develop an observation system with sufficient flexibility to accommodate the multiple domains of inquiry - the numerous perspectives from which behavior can be viewed? Are there logical reasons why we must remain satisfied with a number of particularized approaches tailored to the specific concerns of different investigators? It happens that preschool programs differ in more numerous, or at least more obvious, ways from each other than is the case with instructional programs at more advanced levels. This makes the task of comparing such programs more difficult, but this very difficulty may be an adventitious circumstance if those who are concerned with preschool teaching can focus on the logical and theoretical issues involved. This paper is an attempt to follow that prescription. The observation method we developed was designed to solve a research problem. The primary purpose of the research was to examine the effects of four descriptively different programs on preschool children, and the four programs were selected to represent extremes in philosophy and method in as many respects as possible. It was our hope that we would be able by observation in classrooms not only to identify and describe the four programs, but also to relate the dimensions of programs to the development of the children in these classrooms. As will be discussed in more detail later, the second goal proved to be much more elusive than we

Comparison of model preschool programs

567

had anticipated, partly because we succeeded too well in the first! To carry out these research aims, two instruments were developed. One was designed for use in the classroom and provided for time-samples of teacher and child behavior and also for assessing a number of ecological variables - such as whether children worked individually or in small groups, whether children were occupied with the same or different activities, whether groupings were stable or unstable, number of episodes, etc. The second method, called 'DATA-TAPE', was designed for use with 10-minute video-tapes and focused entirely on the tallying of all teacher behavior into an exhaustive set of categories. These two instruments were to some extent redundant since we had no way of predicting which type of observation method would work best for us. As it turned out, the DATA-TAPE procedure, using only observation of teacher behavior, was quite adequate for discrimination of programs. This paper, therefore, focuses on the categorization of teacher behavior rather than on summative evaluation of programs per se. The important issues of reliability and analysis have also been by-passed in favor of an examination of the goals and nature of an observation system. Conceptual framework. Before describing the method, the logical framework should be explained. Behavior is typically conceptualized as a series of acts. Let us assume for the moment that agreement has been reached on what constitutes an act. An act may then be viewed from several perspectives. One perspective is the substance or nature of an act. If the act is a communicative act, substance would correspond to message. In philosophical terms, the substance of an act would be analogous to the meaning of a proposition which can be entertained in several different modes. For example, the same proposition can be asserted, questioned, doubted, negated, wondered about, approved of, etc. The second perspective corresponds to the mode or style of the act. An assertion, for example, can be made by a visual demonstration, a pantomime, a verbal statement, a physical gesture, etc. At more advanced educational levels, most teaching acts are verbal, but at the preschool level,' that is'not the case. a The third perspective is intentionality or intent. What did the actor mean or intend by the act? Because the assessment of intentionality involves interpretation of behavior and thus jeopardizes objectivity, the tendency is to attempt to eliminate or minimize it or, more often, to combine it with substance on the basis of interrater agreement. This may

568

Social organizations

not be wise. When the act is something like a verbal assertion, 'That is correct', it makes little difference whether one tallies the act in terms of the substance (i.e., reinforcement) or whether one tallies it in terms of the teacher's intent, which is to approve or confirm. For heuristic purposes, however, the two perspectives probably need to be distinguished. Consider the act of physically picking up a child and removing him from one place to another. This may be done to protect him from danger or to stop him from undesirable behavior. The substance (movement of child from A to B) and the mode (physical handling) are the same in both cases but the intention is not. It is certainly conceivable that the frequency with which a child is physically transported can have effects on him regardless of the reasons, and it is equally conceivable that the frequency of either being blocked from ongoing activity or being protected would have effects on the child regardless of the mode or means of doing these things. To put it more concretely, a child who is forced to recite the multiplication tables as a punishment may still learn them. Of course, he may also learn to hate mathematics, or memorizing, or teachers. To summarize in journalistic terms, the three perspectives described substance, intentionality, and mode - correspond to the 'what, why, and how' of behavior. When a system is to be used in tallying an instructional situation, these three perspectives may be defined as follows: One axis of the matrix is the nature of the lesson or the propositional content of the communication (substance). A second axis is the purpose of the act, typically the learning goal (intentionality). The third is the technique or method being used (mode). Thus a multiple coding system consisting of a complete 3-way matrix was the desired solution to the problem of the overall framework. Such a matrix would provide for tallying of each act in a cell corresponding to the intersection of these three axes. The system actually used is not so tidy. We began with Bales' Interaction Process Analysis (1950), a general observation system widely used by social psychologists. In terms of the previous analysis, the major dichotomy in the task or neutral area of Bales' system, 'Giving' vs. 'Asking' appears to correspond with intentionality. A similar distinction has been made in a number of systems for classroom observations, e.g., Klein (1971). 'Asking' is defined as any behavior on the part of the teacher which is intended to produce a response from the child - eliciting. By 'Giving' is meant an act on the part of the teacher intended to convey meaning to the child with no actual or implied expectation of a response - emitting.

Comparison of model preschool programs

569

These two categories of intentionality were totally separated and tallied at different times. Bales' system also makes a distinction between accepting and rejecting what another has emitted. These are emotional aspects of behavior for him, but they correspond in a loose way to the reinforcement variable which also has been included in a number of classroom observation systems. In terms of our conceptual analysis, reinforcement is a subcategory of 'Giving', but because of its peculiar importance in the teaching situation, we have treated it separately in tallying and often in analysis. Using the Bales IPA, alone and unmodified, it would be possible to monitor the classroom and still have no idea of the techniques used by the teachers, the nature of the activities in which children were engaged, or the specific content of any interaction. Even with some modification of the IPA, the original version of DATA-TAPE accounted for very little of the substance of activity. Substance is similar to what Bales (1950) calls 'topical' or 'idiosyncratic content' (p. 34) and his original system was not designed to recover it. We relied on the in-class monitoring for this. Later, we attempted to expand DATA-TAPE to include substance. At first, it was not entirely clear to us that Bales' categories, such as 'opinion' and 'orientation' were subdivisions of intentionality. Although DATA-TAPE has been modified many times, some of these still remain as row categories under 'Asking'. This reflects the fact that it is much easier to determine intentionality when the teacher attempts to elicit behavior than it is to do so when she is emitting. Applications and results. Successive modifications and applications of the instrument over four years, at prekindergarten, kindergarten, and firstgrade levels, uncovered conceptual as well as practical difficulties in constructing a three-way observation system. In the case of academic programs where explicit, overt teaching occurs, the intent and substance of a teacher's activity are frequently identical. On the other hand, in the more traditional, maturationally oriented preschool programs, it is not so easy to specify either one! For example, if the teacher is supervising the children who are engaged in free play, such as a pretend game of family life, what are they supposed to be learning? To use their imaginations, to assume appropriate social roles, the techniques of home management, or other things? The system thus represents a compromise. On the row axis are provided several categories which correspond to a lesson content - e.g., 'academic-

570 Social organizations verbal'. These categories actually represent both substance and intentionally. On the same axis are several categories which correspond to activities in which children might be engaged. In this way, even if no learning goal or substance could be specified, information would be obtained about what the children were doing. Several apparently onedimensional acts were placed outside the matrix, e.g., out-of-contact, maternal help such as tying shoes, and 'setting standards'. A more abstract and consistent system might make it possible to include these categories in the matrix. The 'how' axis translated easily to teaching technique. These column categories were derived entirely from program descriptions with one qualification: We selected so far as was possible those techniques corresponding to variables that have been experimentally studied in research on learning. The four programs which we were attempting to implement were Bereiter-Engelmann, essentially a skill-training, small-group program; DARCEE, also emphasizing cognitive skills and attitudes but somewhat less structured; Montessori, an academic but non-didactic program; and Traditional, a typical enrichment, child-centered program. Analysis of the program descriptions produced the following technique dimensions on which the four programs could be roughly rank-ordered. Non-Contingent Reinforcement (general praise or criticism); Contingent Reinforcement (related to standards for specific behavior); Knowledge-ofResults (dependent upon the correctness or incorrectness of a particular response); Modeling and Imitation (direct imitation of the teacher or matched-dependent behavior as described by Miller and Dollard 1941); Role-playing; Verbal Instruction; Exemplification (use of visual aids); and Manipulation of Materials. These technique dimensions were further subdivided to provide more specific information relevant to program styles, and the category of 'supervision' was added later to accommodate individualized programs in which children are doing academic work (e.g., with workbooks), without continual interaction with the teacher. Since the two classes of intentionality, Giving and Asking, were separated, the system effectively collapsed to a two-way matrix, with column categories representing techniques and row categories representing a mixture of learning goals and activities. Each act was tallied in a cell representing the intersection of a teaching technique and a learning goal, or in a cell representing the intersection of a teaching technique and an activity in which the children were engaged.

Comparison of model preschool programs

571

A number of predictions were made about the behavior of teachers in the four treatments or program styles. Most of these predictions were confirmed. Since these results have been presented in detail elsewhere (Miller and Dyer 1970), only a selected few are presented in Figure 1. Figure 1. Teaching techniques and outcome measures - Four preschool programs Bereiter-Engelmann DARCEE 10i

^

Montessori Traditional

50

8-

40-

6-

30-

4-

20

m

10-

1

xxX xx> • •

M

Variable A (Group Feedback) r A c = .55

Variable B (Imitation. Modeling. Req. Acad. Verb.) t b c = .55

Variable C (Ach. -Motiv. & Resis. to-Distraction)

572

Social organizations

Delighted with our success in identifying dimensions and differentiating program styles, we turned our attention to the problem of relating these dimensions of teaching to the outcome variables, consisting of measures of cognitive and motivational tests on children. At this point, we encountered some logical problems in the use of observational systems for this purpose. Problems encountered. First, even if the variables monitored are those which in fact do identify programs, these variables will always be associated with others which were not assessed. I could report, for example, that the variables shown in Figure 1 - feedback to the group and a combination of Modeling, Requests for Academic Verbal Performance, and Requests for Imitation - predicted children's scores on measures of Achievement Motivation and Resistance to Distraction; while this would be true, it would not provide any information different from the fact that the Bereiter-Engelmann classes had high scores on Achievement Motivation and Resistance to Distraction, since these dimensions of teacher behavior were quite characteristic of that program (Figure 1). And there is, of course, no way to determine whether the test scores were a function of those variables or resulted from some other aspect of that particular program such as the nature of the materials. The use of a larger number of classes or a greater variety of programs might alleviate this problem somewhat. But more classes would not solve a second problem in relating program dimensions to pupil change - the fact that some of the variables which best discriminate programs may not be the variables which are related to changes in children. Robert Soar (1972) has obtained some information of this kind. A third problem was that teaching techniques were recorded for the class as a whole and not directly related to the experiences of individual children. It may simply be the case that experimental studies in the classroom in which the variables not being studied can be controlled will provide the only avenue for obtaining unambiguous information about the relationship between teacher behavior and pupil learning. One troublesome question even in experimental work is this: 'Are certain dimensions, though logically distinct, linked behaviorally to each other?' The results of some of our subsequent experimental work in the modification of teacher behavior suggest that many dimensions may be in practice related to each other or to the nature of the activity or task.

Comparison of model preschool programs

573

For instance, a teacher who focused only on increasing the amount of language modeling succeeded in doing so but also she dramatically increased the amount of positive feedback. This result was inadvertent on her part and quite unexpected. Subsequently it was found that for all teachers studied, the amount of positive feedback varied directly with the number of all requests for academic performance across both baseline and experimental sessions. In short, there may be very close relationships between some program dimensions and others even though they are not logically related. According to Rosenshine (1973, p. 124), this problem is being experimentally studied at Canterbury University in New Zealand. A further problem is that there are program components which are probably important - in themselves or in their mediation of teacher behavior - but which are difficult to recover from a molecular analysis of interactive processes. For example, the four programs we studied can be roughly ordered along a dimension defined as 'sequencing'. What is referred to is a systematic change in program content over time. BereiterEngelmann, Montessori, and DARCEE all involve 'programed' activity, but with important differences: Bereiter-Engelmann is completely programed; in Montessori, the program is inherent in the nature of the materials and the ways in which they can be used; in DARCEE, sequencing is embedded in the content units around which tasks are organized. Traditional is not sequenced at all. Sequencing would be very difficult to assess by monitoring on a periodic basis, although if substance were specified in sufficient detail, it might be possible. 'Pace' is a dimension which is peculiarly related to other program dimensions and also to the method used in observation systems. The average amount of time required to complete an act will vary as a function of the mode of the act. For instance, manipulation of blocks in the teaching of arithmetic requires more time than a verbal statement of a mathematical fact. Role-playing usually requires more time than manipulation. Even different verbal acts may vary considerably in length depending upon the style of the teacher, the nature of the communication, and the subject being taught. Gump (1969) and Gordon (1973) have alluded indirectly to this problem. The solution which was adopted for DATA-TAPE is the provision that any act which persists or continues longer than five seconds is coded every five seconds thereafter. Thus, in the DATA-TAPE system no 'act' requires longer than five seconds' time, although many acts may require much less than five seconds. This procedure still allows for some imbalance

574

Social organizations

in the total number of acts tallied, depending upon the prevailing mode of teaching, and permits the indirect recovery of such dimensions as 'Demandingness' in terms of the total number of tallies under 'Asking'. But defined in this way, pace is a program variable, closely related to other techniques and to materials which are more characteristic of programs than teachers. For a number of reasons, therefore, it is difficult to construct an observation system which will discriminate programs and at the same time predict children's behavior. On the other hand, a monitoring system sufficiently well-designed to permit a solution to some of these problems would be highly useful in experimental work. The ideal observation system. What would an ideal observation system be like? One obvious characteristic, as indicated earlier, would be generalizability. The ideal system would be useful in all settings in which a description of behavior is needed. If such a system were available, one could not only compare educational programs successfully but it would also be possible to compare a classroom situation with a therapeutic interview, or the interaction in a child's home with those in his classroom, estimate the extent to which the classroom setting influences the teacher's behavior as compared with her behavior in other settings, etc. A system which was equally appropriate for a child and a teacher would permit us to examine the effects of children's behavior on teachers and the match between teacher and child behavior as well as the influence of teacher behavior on children. Caldwell (1969) has attempted to satisfy this criterion. Second, the ideal observation system would have scientific relevance. That is, it would contain a set of categories related to the dimensions of behavior which have been identified by various disciplinary research efforts as being explanatory in the predictive sense. Third, the ideal observation system would provide for expansion and contraction in accordance with various purposes, and for the transition from molecular to molar, from concrete to abstract dimensions in a systematic fashion. A suitable model might be that of a thin sheet of rubber which can be stretched in different areas and in different directions. Such a flexible system would enable us to identify heretofore unexamined dimensions of behavior and eliminate redundant information (linkages). What seems to be required is a matrix such that any behavior can be tallied at the intersection of all axes. Since a system with more than three axes and replication in all cells is probably not practical, perhaps some

Comparison of model preschool programs

575

precision for alternate subsets of categories would be useful. In situations in which behavior is primarily verbal, for example, various categories such as complexity, inflection, quality, etc., could be provided under verbal while other modes could be ignored. Finally, of course, the system should be relatively easy to learn and to use. These criteria do not exhaust the characteristics of the ideal system but they do serve to indicate the magnitude of the problem. Are there logical reasons why these criteria cannot be met? I have no answer to this question, but would like to indicate some of the questions which arose during our struggle to adapt a system to various settings. (1) Should a matrix system provide for classification of each single act in only one cell? The multi-dimensionality of behavior creates serious problems of classification, especially with respect to the mode axis. Most communicative acts take place in several modes simultaneously. In fact, if our hands were tied and our eyebrows locked, some of us could not communicate at all! In the teaching situation, even more complexity occurs because of the addition of demonstrative aids. When the teacher simultaneously shows the child a large 'A' and says, 'This is an A', it is clear that the substance is language and the intention is giving information. But should the act be classified as verbal or exemplary in mode? Should this be considered two acts, one verbal and one exemplary? Or should we be able to classify it as a single act of combined verbal and exemplary nature? The DATA-TAPE system required that every act be tallied in only one cell. We found that in preschool a great many techniques are frequently combined with verbalization and also that verbalization is a highly salient aspect of behavior for the observer. Therefore, in order to insure that other techniques would be adequately attended to, a decision tree was developed which established a hierarchy of categories. The use of this algorithm1 produced high interrater reliability and guaranteed adequate tallying of the more subtle techniques, but it did not permit us to distinguish between demonstrations with verbalizations and those without. A better solution to the multi-dimensional character of behavior should be sought in the nature of the matrix. (2) Should a category system be exhaustive of all acts during a specified time period? This characteristic is desirable because it provides flexibility in the summary and analysis of results. For example, one cannot determine what proportion of all behavior a certain category consists of unless all behavior has been accounted for somehow. Nor is it possible to recover

576

Social organizations

pace when tallying is done only at specified intervals. If all acts are to be tallied, the system must provide for location of each act in «-dimensions. Should n be the same for all categories? A system providing for all permutations of a set of exhaustive categories is designated by Rosenshine (1973, p. 134) as an analysis of variance design with all variables crossed (as contrasted with a design in which some variables are nested under others). Flanders (1970, p. 181) argues that such a system is cumbersome and unnecessarily complex in that it typically results in several empty cells either because the appropriate behavior for those cells is logically impossible or because it simply does not occur. Within a given setting, it is certainly correct that some combinations do not occur, regardless of what is provided for by the system, but it may be argued that the ideal system would provide for those combinations should they happen to occur and would be constructed in such a way that no combinations were logically impossible. (3) How can subsets of categories be designed so that they are independent of each other and independent of supraordinate categories? Consider the category 'evaluate' which would be applicable to any behavior involving judgment on the part of one individual regarding the behavior of another. This category could collect all such acts from a therapist's mild remark to a slap in the face. However, one of these acts is verbal and the other is physical. One may be neutral, the other hostile. If these latter categories are also part of the system, they are not independent of the evaluative category. If we must provide for these as sub-categories under evaluative, and also under all categories coordinate with 'evaluate', the number of 'category chains' (Flanders 1970, p. 180) will quickly become unwieldy. Systems which contain logically related categories cause grave difficulties in the use of techniques such as factor analysis as Klein (1971) has pointed out. (4) By what means should an observation system take into account the serial nature of human interactions? The fact that behavior is not simply a set of isolated acts causes at least two kinds of problems in observation. The first has been alluded to in the discussion of substance and intentionality. The meaning of a single act at a given point in time is not independent of those interactions which preceded it. In fact, no single act is comprehensible at all when considered in total isolation from context. 'No' may mean 'permission refused' or 'you were incorrect'. A system which attempts to preserve the serial nature of behavior by defining categories in terms of a mixture of substance and intentionality

Comparison of model preschool programs

SII

requires continual reinterpretation by the observer. Substance remains constant while intentionality varies, hence our attempt to tally these separately; but substance is often difficult to categoiize apart from context. This is the same dilemma of circularity encountered by experimental psychologists in defining variables such as reinforcement. The second problem connected with the serial nature of behavior is that of recovering through an observation system the order in which events occurred. How important is temporal order? Can it be recovered by observing only the behavior of one individual? Is immediate context sufficient or must the entire chain of events be traced throughout the observation period? Is sequence any more important within settings (e.g., in a one-hour class) than across settings (from class to bus to home, for example)? (5) Can an observation system adequately account for the important influences on learning which occur in an unobtrusive or ecologically structured environment? In many situations interpersonal interactions are so predominant that any other type may be disregarded, e.g., a therapeutic interview or a teacher-controlled didactic classroom. In other situations interpersonal interactions are of low frequency, but interactions of an individual with materials are the focus of attention, e.g., learning a motor skill. In a classroom, both of these types of interactions may be important, particularly in a preschool classroom. Most educational programs involve specified materials which are considered an essential portion of the program, and in some programs, interactions with materials constitute the major technique, e.g., computer-assisted instruction or the Montessori method. And what of the class in which the teacher has carefully prepared in advance an environment designed to arouse curiosity, discovery, manipulation, etc. ? (6) How can the affective component of behavior best be accounted for? It has been shown (Brooks et al. 1969) that the same reinforcers given in a negative, neutral, or positive tone of voice had dramatically different effects on kindergarten children of lower socioeconomic status, while the effects for middle-class children were more alike in all these conditions, presumably because the latter were responding primarily to substance. We need much more information of this sort in order to determine the salience of various aspects of a teacher's behavior for young children and the conditions under which they have relative importance. Can we recover such factors as teacher warmth or hostility by combining various categories'' Or must these things be assessed separately? Bales (1950)

578

Social organizations

points out (p. 52) that every act has an adaptive, an expressive, and an instrumental component. In the teaching situation, a system which combines these components does not permit a distinction between informative negative feedback and harsh criticism. (7) Interpretation of the actor's verbal behavior from the viewpoint of the 'generalized other' may work very well when both observer and observed are adults, but in preschool, where the 'others' are small children, this perspective is risky. For example, incongruities between substance and intentionality produce subtle style characteristics such as irony, sarcasm, and humor. To what extent do children recognize or attend to these? Unless we can answer such questions, how can we determine the relative importance of content, goals, or techniques for pupil development or achievements? (8) The hardest question of all, I think, is this: 'To what extent should observation systems to be used in preschool programs be designed to assess those dimensions of teacher behavior that are related to what we can already measure in children?' At present, our most accurate and reliable tests for young children do not tell us what we most urgently need to know about early development. Certainly description without prediction is barren. But I believe it would be unfortunate if we failed to profit by the history of psychometrics and allowed strictly pragmatic criteria to dominate the necessary work on the stimulus side of the street.

Comparison of model preschool programs

579

D. COMPARING MODEL PRESCHOOL PROGRAMS by John T. Neisworth and John E. Dopyera, Pennsylvania State University, University Park and Margaret Z. Lay, Syracuse University, Syracuse

Let us state initially that our major concern is the determination of how to provide programs which will have greatest benefits for children. We will, therefore, focus on the issues having to do with how programs for young children are implemented and how they may be most effectively examined to empirically determine their effects. There are two themes running through our discourse. One is that there is a delicate balance between formal and implicit theory in the implementation of programs for young children. The degree to which a formal model may be successfully implemented depends upon a number of factors. Included among these factors are explicitness of the model, characteristics of the supportive context of the program and characteristics of the persons who staff the program. The second theme is that more productive and differentiated means of analyzing program effects should be employed instead of the global comparisons of model-based programs typically used. In pursuing these themes we will examine several kinds of evidence and will attempt to further emphasize and elaborate some of the issues merely suggested by recent major reviews of early childhood programs, such as those done by Beller (1973) and by Bronfenbrenner (1973). We will proceed in the following way. First we will examine evidence regarding the extent to which and the conditions under which formal theories may be successfully translated into practice. Then we will address the question of whether the research models being used to generate the evidence warrant further use or whether there are alternative models which might have been, or would be, more productive. We will summarize the discussion with some criteria and recommendations for further research efforts. Our examination of recent research on programs suggests that there are at least three factors which may influence the extent to which theoretically based models may be successfully translated into practice. They are (a) the degree of explicitness of the model; (b) the characteristics of the supportive context; (c) the characteristics of program staff.

580

Social organizations

Model explicitness. Relevant to the issue of model explicitness, we might expect that, other things being equal, the greater the model explicitness, the greater the congruence between teacher behavior and model-intended behavior. Miller and colleagues (1970a, 1970b, 1970c, 1971) examined four model programs not only with respect to different outcomes for children, but also in detecting differences in actual teacher behavior. In looking at the Miller et al. (1971) results, it appears that model explicitness and actual practice isomorphic to model expectations are related. That is, for the Bereiter-Engelmann and the Demonstration and Research Center for Early Education programs, teacher behavior more fully met model expectations than the observed teacher behavior in the Montessori or traditional Head Start programs. Banta (1966), in a presentation entitled, 'Is there really a Montessori method?' points out that there appears to be great variability among Montessori programs and that a large sample of Montessori programs would have to be examined in order to determine common elements and outcomes. Bissell's (1971) findings across eight models also lend support to the notion that model explicitness is at least a partial determinant of implementation feasibility. Specifically, it was found that teachers in preacademic models were more frequently rated as being high or at least medium in meeting model expectations. Teachers in discovery models were more often judged as being low in meeting model standards.Verma (1973) has demonstrated that differential teacher behavior is reliably found in programs based on different models when the differences are made explicit through comparison and contrast. Generally, then, it appears that where models are ambiguous or noncommital on program dimensions, interpretations are made and gaps filled in as a function of the implicit theories held by staff members. Thus, it seems reasonable to assume that the slip between what one may perceive to be the intent of a model and actual program practice may result from inadequate model clarity along dimensions that open the door to idiosyncratic and variable interpretations. We concur with Rosenshine and Furst (1973) when they refer to '... the futility of treating all programs within a model as appropriate exemplars of that model.... Without data on implementation, comparative data on outcomes seem meaningless' (p. 28). Model clarity or explicitness has not only direct but also indirect implications for teacher behavior. Explicitness with regard to room arrangement, for example, may be an indirect but effective way to provide

Comparison of model preschool programs

581

prompts for or set limits on staff behavior. Generally, then, we are suggesting the hypothesis that model explicitness yields greater operationalizability of teacher behavior and, thus, greater apparent conguity between model and practice. While there is not a quantity of evidence that can be found which directly relates to this issue, the above stated arguments and studies seem supportive of the hypothesis. It would also appear logical that a person would be better able to enact a behavior if he knows what that behavior is than if it is unspecified or vague. The data and logic, however, do not obviate two other factors. The first is that models may vary on what is specified to be included as well as on the extent to which characteristics are emphasized. An example is the difference between what Piaget says about educational implications of his theory of development and the differential translations and emphasis characterized by Kamii (1972) and Lavatelli (1970). Also, it should be noted that explicitness need not be confused with structure of limited degrees of freedom for children. It is quite possible to be very explicit about programs which afford maximum richness, openness, and thus variety and options for both children and staff. Responsive Care, an open day-care model described by Lay (1972), is one such program. Supportive context. A second set of conditions which determines whether implementation can be successful concerns the supportive context within which the model is to be installed. The evidence on this issue is myriad, coming from direct practical experience of the writers, and from the empirical studies of attempted model replication. Every report of replication attempts, e.g., Head Start, Planned Variation (Bissell 1971), has highlighted the unforeseen problems encountered in field settings which may make the 'fidelity' of the replication questionable on some dimensions considered important to the model developer. Even when quality control measures assure that the necessary behaviors and conditions as prescribed by the respective model may be met, there may still be such inequalities between supportive contexts that comparisons between specific and essential similarities and differences are difficult to make. For example, in the Karnes (1969) comparison of ameliorative, academic preschool and Montessori programs, any differences between classes were difficult to attribute to model characteristics per se since there were inequities between programs in training and supervision of teachers, in ratios of adults to children, and in amount of supplementary education throughout

582

Social organizations

kindergarten. It may be of practical interest to know that one mode specifies lower ratios or necessitates more training from the view of the developer. However, the question of the influence of variations of these contextual factors within a model on what children in programs actually encounter, while recognized as a factor (Bissell 1971), has not been adequately and systematically explored. One would anticipate that a range of contextual factors such as space, equipment scheduling, administrative arrangement, size of center, training/ monitoring, sponsorship and funding, and community relationships are highly significant in determining whether a particular model can be successfully implemented. Weikart (1971), as a result of his experience with attempts to compare cognitively oriented, Bereiter-Engelmann, and unit-based programs, concludes that the operational conditions are far more potent in affecting outcomes than particular intended curricula. In view of this it may be relevant to consider the impact of contextual factors as reported by Prescott (1970) and by Barker and Gump (1964). In these latter studies size was apparently related to organizational structure: the larger, the more hierarchical. Larger preschool settings have been determined to be associated with, among other things, more teacher emphasis on rules and routine guidance, more teacher emphasis on restraint and control of children, less provision of opportunity for pleasure, wonder, and delight, less provision for children to participate in wide age-range groupings. In addition, ratings of teacher style were more often neutral or distant, but almost never were teachers rated as sensitive in large centers as contrasted with small centers. A second study by Prescott (1973) showed another pattern of organizational influence on program. In this study open and closed group daycare settings were compared with family home care and nursery school settings relative to a number of program characteristics. Programs were determined to be comparable in quality relative to characteristics such as stability, sponsorship, adequacy of physical plant, funding, etc. Nothing is reported relative to the size of the centers, although day-care homes would be presumed to be smaller. Programs were classified as open or closed on the basis of opportunity for and frequency of child initiated activities. The data in this study were determined through observation of teachers and children for a unit described as an 'activity segment'. Findings were that children's activities were initiated and terminated through adult pressure in closed day-care centers considerably more than in the other types of programs. Teacher approach to children was con-

Comparison of model preschool programs

583

siderably less sensitive. Teacher emphasis on social rules, control, and restraint was higher, and teacher-directed small group sessions and structured transitions between activities were encountered more frequently by children with free choice of activities less available. While the research report does not specify the size of all closed centers, case descriptions of the three most closed in the study indicate they were all large, well established in the community, and were recently moved into new (expensive) facilities. Clearly, contextual factors are of real significance in determining what a child encounters in a program and may well overshadow any differences that accrue to particular models or theoretical approaches. Staff characteristics. Finally, certain aspects of staff characteristics seem related to the feasibility of practicing what a model preaches. While there may be a variety of personality and other dimensions of relevance, teacher belief, attitude, or style has been examined and found to be associated with success of model implementation. For example, Head Start programs were judged to be more thoroughly implemented when the teachers reported they felt the curriculum was worthwhile (Bissell 1971). Prescott's (1967) finding that restrictive rather than guidance techniques are employed by teachers with little teaching experience has implications for model translatability. Where a model calls for non-restrictive, open procedures to be used by the teacher, apparently employment of relatively more experienced teachers will raise the probability that model expectations will be fulfilled. Studies by Harvey, White, Prather, Alter and Hoffmeister (1966") and Prather (1969) relate teacher behaviors to measured beliefs. Concrete as opposed to abstract teachers functioned differently in the classroom on several behavioral dimensions. Such belief or personality characteristics might be antagonistic with model expectations. Indeed, it has been shown that concrete VJ abstract teachers function differently in open VJ. closed environments (Tuckman, Forman and Hay 1971). That is, there is an interaction between teacher belief and the context in which one must operate. Here, again, success at actualizing a model will depend not only on the harmony between model expectations and teacher beliefs or values, but also on the congruity among model explicitness, teacher beliefs, and contextual arrangements. Verma (1973) studied the congruence between early childhood education, teachers' beliefs, and classroom practices in relationship to actualization of intended models. After reviewing literature relative to the philosophy and

584

Social organizations

theory of operant and Piaget-based early childhood education models, twenty-four items were developed concerning beliefs toward child development and learning; twelve of these reflected an operant, and the remaining twelve items a Piaget or cognitive-developmental, stance. The inventory was given to self-selected head and assistant teachers and faculty consultants for each of the experimental early childhood education units at the Pennsylvania State University. Each unit is intended to operationalize in as pure a way as possible a distinct model, specifically an operant, and cognitive-developmental model. Results suggested the validity of the inventory in that there was almost a perfect relationship between belief preference and program affiliation, with operant and Piaget participants showing somewhat dichotomous responses. In a second phase of the study, the belief items were behavioralized to yield twenty-four corresponding specific teacher behaviors. Continuing staff teachers and previously trained undergraduate student teachers were unobtrusively observed to detect frequency of belief related behaviors. Using a sign test, the results were as follows: Twenty-one out of twentyfour behaviors were differentially displayed in the predicted direction; two behaviors showed no difference, and one went contrary to prediction. Thus, it seems that extreme clarity of model guidelines and correlated staff attitudes conjoin to maximize congruence between theoretical model and practice. In a study we have just initiated at the Pennsylvania State University, eighty beginning student teachers were administered a battery of inventories prior to their random assignment to one of the three early childhood education (Operant, Piaget, Guided Discovery) model classrooms on campus. The battery includes several short personality assessment scales, the Verma (1973) belief inventory, a dogmatism scale, and questions relating to personal background and family. A later phase of the study will involve assessing the students on their satisfaction with the program to which they were randomly assigned. Further, systematic observations will be made to evaluate teacher behavioral congruence with model-expected behavior. Thus, we will be able to examine the match between personality and belief statements and actual teacher behavior. Furthermore, a cross-program comparison of child progress along seventeen dimensions is being completed and will be related to these findings. The importance of the congruence between staff beliefs and a particular model is, thus, illustrated by the studies cited. This influence can perhaps be further understood from the perspective of implicit theory. In a series

Comparison of model preschool programs

585

of studies (Dopyera 1969; Dopyera and Lay 1972; Howe 1971; and Lay 1972) teacher's sanctioning behavior has been analyzed in several contrasting settings. What is postulated in implicit theory (which stems from concepts of cognitive congruence in social psychology) is that all persons hold standards or assumptions about the way things ought to be. When facets of the external world are perceived as discrepant or incongruent from these standards, a tension results which activates an effort to either eliminate or modify the source of tension. A behavioral view of this would suggest that coping or adjustment behavior results when discriminative stimuli, reinforcers, and contingencies are discrepant from those in one's history. Whatever theoretical rationale is preferred, the observational procedures derived from this perspective on the amount and nature of teacher sanctions have been useful in establishing differences between kinds of programs and differences within programs relative to the amount of professional training, etc. Thus, from this perspective, it would follow that successful implementation of a program model is enhanced when there is a strong mesh between the theoretical model and what the staff members already implicitly believe (i.e., the current verbal behavior related to program issues). If this is not the case, a thoroughly explicit and convincing re-orientation along the lines of the model in question would be needed, i.e., 'brain washing'. In summary, then, the extent to which theoretical models can be successfully implemented in practice depends on at least three factors. There is most probably a complex interplay among model explicitness, teacher beliefs, and supportive context. An optimal chance for a model to be validly implemented for children would involve an alignment of model clarity, sympathetic staff beliefs, and a facilitative administrative and physical context. Whenever one of these factors is less than optimal, we should expect to find less than optimal congruence between model and practice. It is clear that these optimal conditions are difficult to establish and maintain in field settings. Pragmatic considerations often take precedence in the daily operation of programs. Guidelines from funding agencies or administrative personnel are ignored or modified if they are perceived as interfering with details of day to day program management. When these guidelines are unclear, when supportive materials and conditions are not available, when the staff basically finds their implementation discrepant from their own views of the best kinds of program practice, it is not surprising that practices depart from the model.

586

Social organizations

Methodological issues in program research. Let us assume for the moment that optimal conditions could be developed and maintained for the inpractice implementation of a theoretical perspective. That is, let us assume that the procedures for implementation were very explicit, administrative and other contextual conditions were supportive, staff were sympathetic and well-trained. Were these conditions present, would we then be more able to adequately address the questions of program influence than we are now with less than optimal conditions generally per vailing 9 Would we then be in a situation where we might learn how programs might be better constructed to provide benefits for participants? We think not! We believe the evidence supports the contention that even if most conditions were optimal, the assumptions inherent in our methodologies would preclude determination of the extent and influence of these programs. We are not here referring to the problems commonly seen in experimental design of inadequate randomization, controls, etc., although clearly if such a design is inadequately implemented interpretations of the results can involve little more than conjecture. Rather, we are speaking about the series of problems inherentin the use of classical experimentation for the determination of program influences. The first of these problems is conceptual. The use of classical experimental design implicitly places all of the effect in the treatment, something which anyone who has watched children knows to be fallacy. There are individual differences in children who participate in programs, and these differences in children are likely influences on results of program participation of these children. To view subject differences as contributions to 'error' in classical experimental design appears to us to be somewhat naive, or at best, experimentally unsophisticated. The use of experimental design for the global comparisons of two or more programs based on different theoretical models creates additional problems. Programs are potentially complex configurations of social and/ or physical events and phenomena, not single or simple treatment variables. This is not to say that these events or factors might not be dimensionalized and quantified. It is to say that this potential complexity should be acknowledged. A major problem associated with global comparison of programs is that there is no way to attribute cause to differences found or not found when comparisons are made. If differences are determined one doesn't know why. It could be due to any number of factors. It may furtherbe the

Comparison of model preschool programs

587

case that if no significant differences are found it is because no significant differences in the program itself were encountered by the participants. Relative to the first concern described, the problem may be more complex. Program encounters of a given kind may have one influence on one kind of participant and a different influence on another. These interactions would be classified as error variance. We have examined some of the evidence describing the conditions under which theories might be adequately translated in practice, as well as the problems in determining program influences on participants. We've suggested that use of classical experimental design is not effective for the task. Potential sources of influence are confounded and participant program interactions are obscured. Then where are we as professional researchers attempting to document program influences on participants? Should we conclude that programs have little or no influence at all - that differences are totally predetermined by generic or familial considerations? If we accept program comparison data at face value, this conclusion would seem reasonable. If there are program influences, research to date seems to indicate that it is more apt to be limited in scope, not very pronounced and of short duration rather than pervasive, strong, and lasting. That's the evidence. At this point we would like to tell you a dirty story. It concerns a point of view and a question about sex. It goes something like this - 'If you are one of those persons who finds sex to be a pain in the ass (pause), maybe you're doing it wrong'. It's our view that we've been doing the research on program effects the wrong way too much of the time. Maybe the differences we know to exist between learners do all stem from straight biology or family influences. But from our point of view, even if there are influences from programs, we won't find them if we continue with inadequate research models and analytic designs and unrealistic conceptualizations of processes, and naivety about the successful translation of theory into on-going programs. We feel that it is erroneous to conclude that research on program effects is impossible to conduct due to the many variables and complexities or to conclude that programs have no influence. We would like to suggest an alternative inductive approach to the determination of program influences on participants which addresses many of the problems discussed previously.

588

Social organizations

Suggestions for an alternative strategy. The alternative approach to research on program effects would require the following: (1) The actual encounters of individual children in programs would be directly assessed. There would not be the assumption that because a child was enrolled in a particular type of program that this program experience would be predictable. His actual encounters with material and persons would be fully documented by observers systematically sampling across time. This procedure has been demonstrated to be feasible and productive in studies by Prescott (1967, 1970, 1973), Brandt (1972), Dopyera (1969) and Lay (1972). (2) There would be a broad range of measures of child behavior, repertoire, and abilities at different points in time (including pre-data). These would be continued long enough to assess long-term as well as short-term effects. (3) There would be a distillation of both input and outcome data into configurations through factoring or clustering procedures. Soar's handling of Follow Through data from a number of classroom observational measures and from a number of outcome criteria measures is exemplary. Input data were similarly factored in studies conducted by Prescott (1967, 1970, 1973), Harvey et al. (1966) and Dopyera (1969). (4) There would be the expectation (and concomitant requirement for analysis) that similar program encounters may influence children of differing characteristics and histories in different ways. While too often ignored, this phenomena has been illustrated in a number of studies. An account by Beller et al. (1972) of the differing effects of nurturance and withdrawal on learning for dependent and independent children is exemplary of the advantage of looking for differing effects of the same treatment. Other examples of disordinal interactions have been presented by Bracht (1970). (5) The determination of effects would be accomplished in multivariate analyses or, preferable, aptitude-treatment-interaction analyses. (6) There would be the awareness that some kinds of influences may well have curvilinear functions rather than linear. In some areas the assumption that if a moderate degree of something has good effects, then more will be even better, just isn't justified. Soar's (1971) follow-through data illustrates need for this awareness in regard to such variables as effects of amount of freedom on achievement outcomes.

SECTION VI

Interaction in social groups and the development of the individual

URSULA LEHR, BERNICE L. NEUGARTEN AND JACQUELINE M. FALK

1

Social interaction and personality development

A. SOCIAL PERSPECTIVE-TAKING TRAINING: EMPATHY AND ROLE-TAKING ABILITY OF PRESCHOOL CHILDREN by Cornelis F. M. Van Lieshout*, Gerard Leckie and Betty Smits-Van Sonsbeek**, University of Nijmegen, The Netherlands

Empathy and role-taking can be considered as two aspects of a single developmental process. In the beginning of this process the child becomes gradually aware of the feelings and emotions of others, identifies with the feelings of others and can anticipate these (Borke 1971). In a later phase the child can take on the point of view of the other in concrete and imaginary social interactions, even when that viewpoint differs from his own perspective. In a further phase the child adjusts his behavior to that of the other while he accounts for the different standpoint of the other in his response (Flavell, Botkin, Fry, Wright and Jarvis 1968; Chandler and Greenspan 1972; Borke 1972). In this way the child changes his socially egocentric perspective either gradually or in a sequence of hierarchical stages in socially centered ways of thinking. The variation of tasks used in different studies to operationalize the * Now at the State University at Utrecht, the Netherlands. ** The authors wish to thank Cary Snellen, Annemieke vanBeelen,and Ellen Janssen for their assistance in executing the intervention program, and Amanda Op't Veld-Hesp and Miss M. Jansen for their assistance during testing. We are grateful to Peter G. Heijmans for his methodological advice and to Nan Stevens for translating the manuscript.

592

Interaction in social groups

role-taking concept suggests that role-taking is not a unitary concept. Selman (1971b) differentiates between 'perceptual role-taking' as the ability to take another's visual perspective (Piaget and Inhelder 1956; Flavell et al. 1968; Shantz and Watson 1971; Huttenlocher and Presson 1973) and 'conceptual role-taking' as the ability to take another's mental perspective, to impute another's knowledge or intentions. Perceptual roletaking requires a person to identify object locations from another's point of view. Conceptual role-taking requires him to identify the reasoning in a concept formation task of another. The verbal communication task used by Krauss and Glucksberg (1969) requiring a speaker to communicate new information to a listener can also be considered as a conceptual roletaking task. For efficient communication, the speaker should adapt the content of his information to the level of knowledge of the listener. Roletaking in binary choice social guessing games (Flavell et al. 1968; De Vries 1970) and in real and imaginary social interactions (Borke 1971; Chandler 1972) can be termed 'emotional-motivational role-taking'. In this task the subject is required to evaluate the emotional state or the motivation of another, who is involved in a social interaction from another motivational or emotional point of view. Another kind of role-taking is required in recursive thinking (Miller, Kessel and Flavell 1970). Recursive thinking refers to the ability of the child to understand the recursive or self-embedded nature of thinking, e.g., 'He is thinking that she is thinking that he is thinking ...'. The role-taking concept may seem complex. Rubin (1973), however, found in his study on the internal consistency of the decentering concept, that several aspects of role-taking, such as recursive thinking (Miller, Kessel and Flavell 1970), role-taking in verbal communication tasks (Krauss and Glucksberg 1969), and perceptual role-taking (Flavell et al. 1968) show an underlying close interrelation and likewise are related, to a great degree, to other decentering processes such as conservation (Goldschmidand Bentler 1968). Also Selman (1971b) described a highly correlated hierarchical sequence in the development of motivational (De Vries 1970), perceptual, and conceptual role-taking ability. In addition to studies on the internal structure of the role-taking concept, research was done on the relation of social decentering to other aspects of prosocial and antisocial behavior. Kohlberg (1969) assumed as a condition for higher levels of moral thinking that a child is capable of role-taking. This assumption was tested by Selman (1971a) who, with children aged 8-10 years, controlling for intelligence, ascertained that

Social interaction and personality development

593

role-taking skill must be considered as a necessary condition for the development of conventional moral thinking. Chandler (1972) assumed that antisocial behavior of delinquent boys was related to retardation in the development of the social decentering process. He found that delinquent boys indeed received lower role-taking scores on his role-taking test. Their scores corresponded with those of younger non-delinquent boys. According to the Piagetian tradition a child develops social decentration ability by repeated varied interpersonal interactions (Piaget 1950; Flavell 1963; Looft 1972). Thus Chandler (1972) could raise to a certain extent the considerable retardation in role-taking ability of the delinquent boys in his study by means of a social perspective-taking training program composed of dramatic role-playing and discussions of videotapes of this role-playing. Likewise the assessed delinquent behavior of the experimental group appeared to have significantly declined in a follow-up period of 18 months. In the first study, reported in this paper, a number of role-taking tasks weie developed for three- to six-year-olds. In view of the age level of the children, these tasks represented behavior patterns which can be considered as early childhood antecedents of emotional-motivational role-taking behavior, such as the recognition of social stereotypes, the recognition of the roles and role attributes of people in the environment of the child, the evaluation of needs and states of mind of people in the environment of the child, and antecedents of perceptual role-taking. In the composition of the tests, items are chosen which require the child to predict rolecharacteristics of others which are as different as possible from his own role characteristics. In the second study, the effect of a social perspectivetaking training program on the role-taking skill of three-, four-, and fiveyear-olds is examined. Study 1. The subjects weie 142 three-, foui-, and five-year-olds, 76 boys and 66 girls. The group consisted of 43 three-year-olds (24 boys and 19 girls), 52 four-year-olds (28 boys and 24 girls), and 47 five-year-olds (24 boys and 23 girls). The mean age in the three groups was respectively 42 months (37-47 months), 54 months (49-59 months) and 66 months (61-71 months). The three-year-olds attended a preschool 4 to 5 half days per week, and the four and five-year-olds were pupils in several kindergartens in Nijmegen. The subjects came primarily from the middle to upper social-economic classes. The mean intelligence was above average. Nine Role-Taking Tests (RTT) were individually administered during

594 Interaction in social groups school time by students in developmental psychology. The tests were administered in the fixed order stated below. The total number of items in the nine tests was 66. A person can score right or wrong on an item. On several tests (2, 3, 4, 5) the motivation of the answer was decisive in the right or wrong scoring. The RTT were the following: (1) Recognition of Role Attributes. Ten professions and ten associated role attributes are represented by pictures. The task of the subject is to match the role attribute with the profession, e.g., hammer with carpenter. The ten professions and the ten role attributes are simultaneously presented. (2) Recognition of Role Behavior of Family Members. This test is adapted from Emmerich (1959). The test administrator shows the subject four drawings, each representing, respectively, a man, a woman, a girl, and a boy. The test administrator introduces the pictures as follows: 'Here you see four pictures of a woman, a man, a boy, and a girl'. Next the subject must categorize ten statements according to these four roles, e.g., 'Who says "I am the son"?' (3) Choosing Gifts. On this test, which is adapted from Flavell et al. (1968), the child must choose a gift for a man, a woman, a girl, and a boy. The child can choose from eight different gifts. The gifts are represented by photographs. (4) Recognition of Needs. On five pictures a child is drawn, each time showing a specific need, being hungry, being warm, being cold, being in pain, and having shoes that pinch. For each the subject must choose from five alternatives how the portrayed child can best solve his need situation. (5) Empathy. This test is an adaptation of the test used by Feshbach and Roe (1968) and Borke (1971). In this task the tester shows the subject four drawings of faces depicting the emotional expressions of'happy', 'afraid', 'sad', and 'angry', which he is asked to identify. Then eight stories are told in which something happens to another child. Following the presentation of each story the subject is asked 'to select the picture that corresponds with the feeling of the child in the story. (6) Visual Perspective: Upright/Upside Down. This task and task 7 and 8 are derived from perceptual role-taking tasks, used by Flavell et al. (1968). In these tasks the child is asked to identify the visual perspective of another person. The subject must rotate a picture so that the test administrator who sits across from him sees the picture right-side-up/upside-down. This test has five items. (7) Visual Perspective: In Front/Behind. The material for this test consists of ten cards with a picture on both sides. After the child has seen the picture on both sides, the test administrator, who sits across from the subject, holds up the card between them. The subject must now say what the test

Social interaction and personality development 595 administrator sees on the card. (8) Visual Perspective: Three Dimensional. The material for this subtest consists of two identical sets of two threesided right prism-formed blocks with pictures on the three sides, two cubes with pictures on four sides, and two cubes with pictures on all six sides. The tester and subject, who sit across from each other, each have an identical prism or cube before them. The instruction reads: 'Turn the block so that you are looking at the same picture as I am.' The test consists often items. (9) Left/Right Identification. On this test, which is originally a task used by Piaget (Elkind 1961), the child must successively show his right and left hand, and right and left foot. Then he must do the same on the test administrator who sits across from him. Results. The mean difficulty level, expressed in the quotient of the number of correct scores per item divided by the total number of respondents (N = 142), was .61. Of the 66 items no item was too difficult (limit < .25) and three items were too easy (limit > .90). The mean difficulty level, determined for each of the nine tasks, ranged from .27 to .83. By means of the Chi-square test the frequency distribution of the total scores on all the tests was analyzed for normality. The population investigated appeared to be normally distributed (M = 40.32, Md = 43.61, SD = 11.62, df = 21, Chi-square = 47.28, p < .001). The internal consistency, determined for the items from all the subtests together according to Kuder-Richardson (K.R.20), was .91. In addition, the internal consistency was tested by means of the product-moment correlations between the scores on the different tests and the total sum of scores on the nine tests. The correlations from the first eight tests with the total score vary from .49 to .76 (p < .01). The correlation of test 9 with the total score is much lower (.32, p < .01). Table 1. Means of role-taking tests Role-taking test

3 yrs

4 yrs.

5 yrs.

Role attributes Role behavior of family members Gifts Recognition of needs Empathy VP: Right-side-up-upside-down VP: In front-behind VP: Three dimensional Left-right identification

3.47 6.40 1.74 2.19 2.81 2.02 7.47 3.23 .60

5.94 7.48 2.85 3.13 3.94 2.15 8.33 6.08 1.19

7.15 8.51 3.45 3.85 4.62 3.13 9.30 7.26 1.36

596

Interaction in social groups

The means of the separate RTT, specified for age level, are presented in Table 1. Two-way analyses of variance with age (3 levels) and sex as factors yielded a significant (all p < .01) age effect for the first eight tests. The F values were, for role attributes, F (2, 136) = 27.89; role behavior of family members, F (2, 136) = 13.76; gifts, F (2, 136) = 26.98; recognition of needs, F (2, 136) = 21.68; empathy, F (2, 136) = 14.36; visual perspective: right-side-up/ upside-down, F ( 2 , 136) = 4.41; visual perspective: in front/behind, F (2, 136) = 9.20; and visual perspective: three dimensional, F(2,136) = 20.14. For left/right identification the age effect failed to reach significance {F (2,136) = 2.60). The main sex effect and the age x sex interaction effect were not significant on any of the nine tests. Calculating t ratios (Kirk 1968) revealed most age means to differ significantly (p < .005) from one another, indicating that the older children are more socially decentered than the younger ones. Only the comparisons between means for ages 3 and 4 and 4 and 5 on visual perspective: right-side-up/upside-down and left/right identification failed to reach significance. To get a more accurate picture of the dimensions represented, the nine tests were submitted to a Kaiser and Caffrey Alpha factor analysis. Factor extraction was stopped when the eigenvalue was less than 1.00. Two factors fitted the data best. The interpretation of the factors was based on the unrotated factor matrix, since the Varimax criterion remained smaller than .30 after four iterations. Interpreting the factors, variables with loadings greater than or equal to .30 were considered significant (Nunnally 1967). Factor I explained 31.9 % of the total variance. The first eight subtests loaded significantly on this factor. The left/right identification test had no significant loading on either of the two factors. Factor I can be identified as the role-taking factor. Factor II explained 4.2 % of the total variance. The only variable which loaded significantly on this factor is recognition of needs. This test also loaded significantly on the first factor. Study 2. The subjects, a subgroup from the first study, were 72 children, the total population of two preschool classes and two kindergarten classes connected to the Psychological Laboratory of the University and the School for Kindergarten Teachers in Nijmegen. The subjects came from middle and upper social class. The experimental and control groups each consisted of one preschool class and one kindergarten class. The

Social interaction and personality development

597

experimental and control groups had no contact with each other before or during the experiment. Within these conditions the subjects were divided into three age groups. The experimental group consisted of 13 three-year-olds (37-47 months; 8 boys and 5 girls), 17 four-year-olds (49-59 months; 8 boys and 9 girls), and 9 five-year-olds (61-71 months; 4 boys and 5 girls). The control group consisted of 11 three-year-olds (6 boys and 5 girls), 11 four-year-olds (6 boys and 5 girls), and 11 fiveyear-olds (7 boys and 4 girls). The experiment consisted of three phases: pretest, treatment, posttest. (1) Pretest. The subjects of the experimental and control groups were individually pretested for empathy and role-taking ability, by means of the role-taking test (RTT) developed in the first study, and for intelligence (Stanford-Binet, Form L, 1960). (2) Treatment. After the pretest the experimental group was enrolled in a social perspective-taking training for a period of 18 weeks. The program was conducted by the class teachers during nursery school time, for 30 minutes a day. The social perspectivetaking training consisted of the following elements: (a) Predicting and interpreting feelings and behavior of others in standardized stories, doll play and role play. The prediction and interpretation of feelings and behavior was elicited by introducing in the stories new figures who either did not know what had happened in the preceding events, who interpreted the preceding events incorrectly, or reacted unusually to the feelings or behavior of other figures in the story, (b) Initiating role play during free play activities of the children. The teacher initiated role play by assigning the children specific roles. Once the role playing was proceeding well, she gradually withdrew from it. Role play was further elicited by offering role play materials, e.g., finger puppets, (c) Making feelings explicit. When children expressed specific feelings (e.g., crying, laughing, comforting each other), the teacher pursued this. She asked other children what the affected child felt when it cried, needed to be comforted, etc. (d) Exercising of social and cognitive perspective taking concepts (e.g., taking turns, sharing, playing honestly, left/right, in front/behind, above/ below) and attributes (e.g., attributes of professions). Altruism, helping each other, and comforting were modeled, induced, and reinforced when the possibility arose during the day. (e) Dominoes and slide material with professions and expressions of feelings. During this phase the control group had a nursery school program without special emphasis on social perspective taking training. (3) Posttest. Both groups were tested with the same RTT as during the pretest.

598 Interaction in social groups Results. The total scores of the RTT were submitted to two analyses of covariance. In the first analysis of covariance the total RTT pretest score was used as a covariable. In the second analysis of covariance, in addition to the RTT pretest score, the mental age (Stanford-Binet, Form L, 1960) was used as a covariable. As factors, age (3, 4, 5 yr.), sex, and condition were introduced in the analysis. When only the effect of the RTT score on the pretest was eliminated as a covariable the main age (F (2,59) = 9.94, p < .001) and condition (F (1,59) = 12.36, p < .001) effects were significant. The condition effect (F (1,58) = 13.83, p < .001) remained present, when, in addition to the RTT pretest score, mental age was eliminated as a covariable. In this analysis the age effect, however, disappeared. Understandably, mental age has a substantial proportion of common variance with age (r = .82). Table 2. Means of RTT pre- and posttest

RTT pretest RTT posttest

3 years Exper.

Contr.

4 years Exper.

Contr.

5 years Exper.

Contr.

30.62 41.15

28.36 33.55

39.59 50.65

42.09 42.73

46.89 56.22

48.18 54.82

On the pretest no significant differences between experimental and control groups appeared for the variables chronological age, mental age, and role-taking ability. Calculating t ratios (Kirk 1968) revealed that on the RTT posttest the 3- and 4-year-olds in the experimental group scored significantly higher than those in the control group (t values are 3.13, p < .01, and 5.24, p < .01). For the five-year-olds this difference failed to reach significance (t = .53, NS). In a comparison of the three age levels without specification for condition it appears that the four-year-olds attain a significantly higher RTT posttest score than the three-year-olds (t — 6.24, p < .01) and the five-year-olds attain a significantly higher RTT posttest score than the four-year-olds (t = 4.57, p < .01). Discussion. The results from the first study suggest that different aspects of early childhood role-taking behavior, such as recognition of social stereotypes and roles and role attributes of people in the environment of the child, evaluation of needs and feelings of people in the environment, like antecedents of perceptual role-taking, show a significant interrelation-

Social interaction and personality development

599

ship. Although the tasks used in this study for the most part were of a different nature than those used by Rubin (1973) the correlations assessed in this study with 3-5-year-olds can be considered as a supplement to the correlations found by Rubin for 5-11-year-olds. As with the tasks that Rubin presented to his subjects, a successful solution of the test used in this study, especially tests 6-9, required the child, in order to solve the problem, to approach it simultaneously or successively from different perspectives. Although there were significant intercorrelations found between 8 of the 9 role-taking tests, compared to those found by Rubin, the correlations were not very high. Characteristics related to the age of the subjects, such as limited language skills, less insight into the purpose of the test, and fluctuating motivation, presumably caused a relatively big portion of error in the total variance. The standard deviation of the test scores is smaller on most tests for the older children than for the younger children. As in other studies (Rubin 1973; Flavell et al. 1968; Miller, Kessel and Flavell 1970; DeYries 1970; Selman 1971b; Shantz and Watson 1971; Huttenlocher and Presson 1973), this study demonstrates a correlation between the age of the subjects and their role-taking skill. In the tests used in this study no so called phase thresholds were built in, so that we can only conclude that role-taking skills, for the group that we investigated, increase gradually with age and with increasing cognitive maturity. Before the experiment the role-taking ability of the experimental and control groups did not differ; however, the perspective-taking training increased the role-taking ability of the three- and four-year-old children considerably. With the five-year-olds this effect was not significant. The effect remained apparent when the differences in mental age of the children were controlled in addition to role-taking ability on the pretest. In accordance with the Piagetian view, we can conclude that a child develops his decentering ability especially as a result of repeated interpersonal interactions (Piaget 1950; Flavell 1963; Looft 1972). Thus not only can retarded role-taking ability be raised by a perspective-taking training (Chandler 1972), but the 'normal' development of role-taking ability in young children can also be stimulated by letting them experience specific social interactions. Although these findings may contribute to a better understanding of the development of role-taking ability, they should still be interpreted with caution. It is by no means clear which aspect of the total perspectivetaking training program brought about the effect. It involves both

600

Interaction in social groups

cognitive structuring of social interactions, such as induction, and behavior modification techniques derived from social learning theory, such as modeling, shaping, and social reinforcement. Furthermore, it can not be demonstrated whether the intervention program only extended the number of role-taking responses or at the same time contributed to a sequential restructuring of the development of role-taking ability. In this connection it is also important to know whether the intervention program exclusively influenced internal behavior structures or at the same time influenced the concrete social behavior, e.g., the amount of aggressive behavior, cooperation, altruism, etc. Due to these unsolved questions, further research on the possibility of modifying the decentering ability of children by directed interpersonal interactions is indicated. Specific insight into the nature of the variables which play a role in this behavior modification process is necessary.To assess the effect of such modification as precisely as possibly, conceptual role-taking and recursive thinking should be trained and evaluated, as well as emotional-motivational and perceptual role-taking.

Social interaction and personality development

601

B. LEVELS AND PATTERNS OF SOCIAL ENGAGEMENT AND DISENGAGEMENT FROM ADOLESCENCE TO MIDDLE ADULTHOOD by Joel Shanan, Hebrew University, Jerusalem, Israel

Common to all theories of psychological development, divergent as they may be in many respects, is the idea that a certain amount and/or a certain type of interaction between the individual and his social environment, at a certain point or points in time, constitutes a necessary condition for successful socialization or 'normal' or 'adjustive' personality development. In all these theories it is either explicitly stated or at least implied that lack of sufficient and/or satisfying social interaction leads to feelings of deprivation, rejection, or frustration in general, and consequently to the development of deviant patterns of personality and social and cognitive functioning in adult life, or arrest of development on an 'infantile' level. In this report a series of empirical studies shall be discussed which were organized around the general hypothesis that the developing individual's position in and his relations to social groups, i.e., the type and extent of his social engagement, affect certain aspects of his ego development and ego functioning. It was posited that this should hold true, though possibly with variations, at different stages of development, but particularly during the period from adolescence to middle adulthood, characterized by a constant broadening of social relations. Since ego development and ego functioning are rather broad terms, we intend to limit ourselves to a particular aspect of ego functioning, which has been studied by us over a number of years (Shanan et al. 1965; Shanan 1967), 1968, 1972; Lerner and Shanan 1972; Mor, Shanan and Levinson 1973) in different contexts and with different techniques: coping behavior, i.e., an individual's tendency or willingness (and capacity) to (1) orient his attention in a way which enables him to identify in his personal field complexity of goals and sources of conflict whether social, emotional, or cognitive and (2) articulate his field by coping affectively, cognitively, or by action in a way which safeguards optimal self-esteem and adjustment to changing environmental demands, that is, maximizes the potential for further integral development. This concept of active coping can be regarded as a parameter of ego strength which according to the assump-

602 Interaction in social groups tional framework of ego psychology (Hartmannef al. 1946; Rapaport 1967) is conceived as contingent upon the availability of free mental energy for the purpose of overcoming barriers, i.e., coping with challenge and stress in the service of self-realization. A basic feeling of being accepted by and a subsequent possibility to identify with significant others without major conflict is posited as one, though not the only one, of the major sources of free mental energy. Inversely, conflict with or isolation from such significant others or psyche groups should lead to an expenditure of effort to regain a feeling of acceptance, real or imagined. Such conditions are expected to drain subsequently the ego's resources for relating to and coping actively with novel situations which require restructuring of the individual's psychological field. Our thesis then may be stated more specifically as follows: The tendency of the individual to cope actively with novel and/or stressful situations whether perceived as self-selected, i.e., as challenge, or imposed by forces external to the ego, i.e., as stress-as opposed to the tendency to give in passively, will depend in its extent on the degree of present and/or prior engagement in significant social networks. Inversely, the greater an individual's past and/or present disengagement, real or imagined isolation, and concomitant social deprivation, the weaker will be his tendency to cope actively. Under such conditions the likelihood to give in passively or to see himself at the mercy of forces external to self and beyond its command is expected to rise. In the course of time we have developed various techniques of measuring coping behavior. In most studies a sentence completion technique was used (Shanan 1973) which yields measures of active coping behavior defined more specifically and operationally by: (1) Availability of free cathectic energy for directing attention to sources of potential difficulty, i.e., for identifying complexity; (2) articulation of the perceptual field, i.e., perception of the internal and external environment in a way that facilitates (3) coping with complexity or conflict, while (4) optimal balance is maintained between the demands of reality and the demands - developmental and integrative - of the self. The test contains four groups of items, ten items in each group, two referring to field articulation, which measure the extent of differentiation in a person's perception of aims and goals (item group 1), and in his perception of sources of fear and frustration (item group 2), that is, in his cognitive-motivational structure. The two remaining item groups measure the extent to which a person is ready to cope actively with novel

Social interaction and personality development

603

and/or complex situations (item group 3) and the extent to which he perceives himself in a positive manner (item group 4). Positive scores in the first two item groups are given to responses showing that the person's motivations are tied to objects other than the self. In the third group of items, positive scores are given to responses showing a tendency to cope actively with the situation (as opposed to responses indicating a passive giving in). In the fourth group of items, responses showing positive self-perception are scored plus. The sum of positive scores obtained in all four groups represents a global measure of active coping in the broader sense of level of adjustment. (For more detailed scoring instructions and their rationale see Shanan 1968, 1973.) Some illustrative research findings can now be presented. These studies refer to three age groups, adolescents from 14-18, young adults from 21-26, and middle aged parents during their thirties and forties. Varying degrees of engagement and disengagement from 'normal' to rather pathological types were studied. In some of our work only one specific age level and in other studies different age levels have been investigated simultaneously. Relevant data have also been gathered on later midadulthood. They are reported elsewhere (Shanan 1968, 1971, 1973). The effects of cultural background on gifted adolescents. The first study to be reported is one on normal adolescents. It was carried out by Talia Naftali and myself (1973) and supplied a good deal of support for our way of thinking, concerning the relationships between type of social interaction of which the individual had been or is part and personality development. In this study the sample consisted of 50 adolescent boys of oriental background and 68 boys of European background of above average intelligence. All boys had been applying for admission to diiferent types of state-owned boarding type high schools, when they were just about to complete their 8th year of grade school. They had been groupadministered a number of psychological tests for selection purposes. Among the tests administered was also Shanan's sentence completion form. The sentence completion test had been analyzed by the conventional quantitative technique. This analysis showed that there were no statistically significant differences on the total 'adjustment' score or on the two variables measuring the readiness to structure the field cognitively in a sufficiently complex and reality oriented way. However, on both selfesteem and particularly on the items measuring the tendency to cope actively statistically highly significant differences could be observed: The

604

Interaction in social groups

middle-class children of European origin were more likely to cope actively with problems and/or challenge, when perceived as such, and were also able to maintain a significantly more positive self-concept. A further interesting finding was that when the discrepancies between the four subscores on the SC were counted, it was found that the adolescents of lower class, eastern origin showed significantly more discrepancies. Briefly, the latter were characterized by a different coping style, in spite of similarity in general level of capacity to adjust. It should be recalled that both groups were 'normal' adolescents. From adolescence to young adulthood. As a next step we compared three consecutive age groups within one cultural group, middle class of European descent: 68 boys aged 14-15, 50 boys aged 17-18, and 50 young men aged 21-26. To begin with, a steady increase with age in the general level of the tendency to cope actively was observed. This increase was mostly accounted for by an increase in the capacity to differentiate and to identify more realistically and more successfully sources of conflict and difficulty as well as by an increased readiness to cope actively with such situations. It is also interesting to note that on the field structuring variable (no. 2) the major increase occurred in the group consisting of boy who returned from the army service, i.e., age 21-26. The major rise in the readiness to cope occurred during the transition from middle to peak or late adolescence which is in Israel the preenlisting period. Finally it should riot be overlooked that self-esteem tended to decrease slowly but consistently from age 14 to 26, suggesting that it might have been too high relative to the other aspects of coping at age 14-15. While these data open up a good many intriguing questions, to be discussed elsewhere, the general impression gained is that development during adolescence leads to an increase of the tendency to cope actively and to a more mature level of active adjustment. Last but not least such development correlated also with an emergence of qualitative differences in coping style, beyond a mere raise of level of activity. This development parallels an increasing social participation or active engagement, particularly in Israel with its youth movements and army services. In the light of these data, findings from the previous studies, in which age had been kept constant and only indices of engagement were varied, could possibly be interpreted in terms of different stages of personality development. Next we present illustrative findings from an unpublished study carried out in cooperation with Yael Lifson. Four groups of about 30 adolescents

Social interaction and personality development

605

were investigated in Israel: A group of juvenile delinquents, a group of orphans, a group of children referred to collective settlements (kibbutzim) to be educated there in order to relieve their parents from some of their socioeconomic stress, and finally a group of youngsters referred for education to the kibbutz as well, but for different reasons: To enable them to develop their above average intellectual potential, since it could be assumed that they could not achieve this while living within the framework of their families. All of these adolescents were between 13-15 years old and belonged to large size families (4 children or more) characterized by low income and 'eastern', mediterranean origin. Their background was similar to that of the 'eastern' group in the earlier mentioned study by Naftali. Applying the theoretical notions outlined above, one can consider the subcultural group to which all these adolescents belonged as relatively 'disengaged' from the Israeli middle class core population of European origin and consequently as more or less socially and culturally deprived. In fact, many adolescents with this type of background in Israel are prone to drop out of school, to develop delinquent and other socially deviant behavior. We had selected the four groups on the assumption that they represent marker groups on a continuum from disengagement to engagement in social frameworks. We assumed the deliquents to be the least engaged, since they not only h a d - t h r o u g h their delinquent behavior - given evidence of rejecting the rules of society, but in fact they were stigmatized and rejected on that account by society at large. The orphans on the other hand had suffered loss of at least one parent during early childhood and were thus deprived of the normal familial group in which a feeling of deep emotional engagement could be developed. The children from the under-privileged large families who had been sent to the kibbutz for education to alleviate the social situation of their parents were assumed to have viewed this situation at least partly as a rejection on the part of their parents, rather than something done in their personal favor only or predominantly. Finally, those children who had been sent to the kibbutz to develop their above-average intellectual potential, we knew, were quite well aware of the reasons for which they had been selected and therefore rightfully felt that they were 'chosen'. We could therefore assume that they were likely to feel the most engaged. To all these adolescents we had administered Shanan's sentence completion form as described above. Findings showed very clearly that the level of readiness to cope actively with stress was the lowest in the delinquent group, next to them came the orphans, then after a gap,

606 Interaction in social groups followed the children who had been transfered to the kibbutz for social reasons. Finally, the gifted children to whom society had reached out to provide extra opportunities had the highest scores. In fact, their average score was nearly four times as high, numerically, as that of the lowest group, the delinquents. Self-perception showed the same trend, though to a weaker extent. It should also be noted that while for self-perception the dividing line ran between the delinquents and all the others, for the readiness to cope actively the dividing line ran between the fostered, gifted group and the rest. These quantitative data can be complemented with some additional information derived from a content analysis of the sentence completion test, comparing the responses of the orphans with those of the other two non-delinquent groups. It turned out that the orphans felt the major source of happiness was the company of people. Achievement constituted for them less a source of happiness than for the other adolescents. The orphans also felt more than the others that they were ready to overcome personal difficulties and to work hard only when their needs for dependency had been satisfied. Loss of any kind was perceived by them as both a source of anger and depression, more so than either personal failure or rejection, absolutely and when compared with the other adolescents. These data supported very strongly indeed our hypothesis that the tendency to cope actively develops as a function of real or felt engagement in a narrower and/or broader social network. They also demonstrate that severe disengagement lowers dramatically, already during early adolescence, both readiness to cope actively and self-esteem. Mental illness and immigration. The next piece of empirical evidence for our argument stems from two studies in which age, social status, education, and culture were kept constant. Findings followed the trend observed in the preceding study and reached a peak in the case of mental illness. In one study two groups of students were compared: A psychologically healthy group, the 21-26 age group referred to earlier, and a matched group of University students who had turned to psychiatric help. They were preponderantly of European or 'Western' origin. In the other study a group of healthy Israeli soldiers of Moroccan origin had been compared shortly after their immigration to Israel, with a similar group who had sought psychiatric help.1 There is no need to present detailed data, since 1. Dr. Michaela Lifschitz has been helpful in gathering these data.

Social interaction and personality development

607

in both studies the patient groups obtained average coping scores about 50 % (!) lower than the healthy group. In addition to that, the new immigrants, the Moroccan group had mean scores about 25 % lower than that of the 'Western' oldtimers. That this latter effect was a combined effect of disengagement due to recent immigration, minority culture, and bad mental health and not only due to the fact that the group in case was an oriental group is born out by more recent findings from a study of more than 1,000 new immigrant students carried out in cooperation with Mirian Sharon. About 80% of the immigrant students were from the western hemisphere. Here again we witnessed depressed coping scores in the not yet 'engaged' newcomers as compared to the data on the students born in the country. Since we cannot automatically assume that all these immigrants or even most of them had been retarded in their ego development in general or intheir capacity to cope actively in comparison to the Israeli students, we suggest that disengagement from earlier social ties may bring about an at least temporary regression in ego development as observed in the lessened readiness to cope actively. Further support for this line of thought can be gained from the disproportionately high percentage of immigrant students who had been found in need of psychological or psychosocial support, as compared with students born in Israel. Engagement and perception of parental roles in mid-adulthood. Going up the developmental ladder we would like to present some data pertaining to peak adulthood, that is, people in their thirties and forties. These data pertain to a study on the perception of their parenthood by Israeli parents of two cultural origins, European and Iraqui, all of them belonging to the Israeli middle class.2 These pairs of parents had been presented with a series of questionnaires measuring the perception of their image of fatherhood and motherhood, distribution of educational responsibility between them, their readiness to serve as models for their children, and the degree of value congruence as perceived subjectively by each spouse. They had also been administered Shanan's sentence completion form. The highly interesting findings shall be reported in detail elsewhere, along with those on recent immigrants from Russia and Jewish and non-Jewish Greek and Austrian parents. Pertinent to the issue raised here are two sets of findings which were found to hold up over both cultures for fathers and 2. Mrs. Sylvia Marcus-Fenster was very helpful in the analysis of the data of this study.

608

Interaction in social groups

for mothers, though for each sex and culture and their combination to varying extents and with reference to varying contents. First there was a highly significant relationship between subjectively perceived value congruence and perception of parents as relatively child-centered or supportive as opposed to parent-centered or restrictive. Secondly there was a similar relationship between a tendency to cope actively and to perceive oneself positively and the perception of oneself as a childcentered parent. If we introduce value congruence as our index of social cohesiveness of the family one can say that those parents who perceived themselves as congruent with their spouse, i.e., perceived themselves as part of a whole, were more 'engaged' than those who perceived themselves as less congruent with their spouse. Engagement thus correlated here with a specific way of defining one's role in a social system, the family. If one feels part of a social group one also can feel more 'liberal' and 'accepting' toward the weaker members of the group - in this case the children. Here again engagement in a social network and readiness to cope were found to go together. Concluding comments. This paper concerned itself with the extent and nature of the individual's engagement in significant social frameworks or structures, such as family and peer-groups and their impact on a specific dimension of ego functioning, namely coping behavior. Summing up the findings reported here we observed that individuals of different age groups, adolescents during mid and late adolescence and adults during the peak and during middle to later adulthood varied in their level of readiness to cope actively and in their style of coping according to extent and type of engagement in social networks. In the adult groups we were also able to observe differences in the readiness to cope in a particular way with specific central tasks of adulthood, such as certain aspects of parenthood. We noticed that this readiness to get involved and to cope with complexity is not necessarily related to the general level of adjustment; neither does it relate in all cases to a readiness to perceive reality in all its complexity. While in general we get the impression that a lowered tendency to cope actively is correlated with lowered self-esteem, this does not hold for all groups studied. Differences in the patterning of the coping variables could be interpreted as differences in coping style. These correlated also with progressive phases of development. The more detached or disengaged a person was from social net-

Social interaction and personality development

609

works, the greater was the likelihood of his coping on a more passive level. Engagement on the whole was correlated with active coping and positive self esteem. It seems therefore legitimate to view the developing individuals and groups of individuals on a continuum of either having or lacking access material and/or psychological - to basic gratifications, offered normally by interaction within the social framework to which the individual or group belongs. A state of belonging in which an individual feels fully gratified psychologically and materially in terms of the norms of 'his' group, i.e., the group to which he feels legitimately that he belongs, we call (social) 'engagement'. Social disengagement (its extreme form is social stigma) on the other hand would be a state in which an individual is led, internally or externally, to give up satisfying ties to a group of people whom earlier he had perceived as 'significant others'. Thus the term 'engagement' refers to a perception or feeling on the part of the individual of belonging to a significant group, i.e., a group offering him basic gratification and privileges considered legitimate in a given society. The term 'disengagement' on the other hand refers to the perception of the individual of himself as lacking access to such basic emotional, interpersonal or material gratifications as a consequence of disruption of communication with a group of significant others, regardless of whether the disruption is real or imagined, voluntary or forced. It may be noted that the concept of disengagement was used here somewhat differently than by its original proponents (Cumming and Henry 1961) who coined it to denote mainly autonomously initiated disengagement, as internally perceived. At least theoretically it should become possible to classify states of social interaction both on a perceptual and on a behavioral level, according to the amount and type of engagement actually observed or perceived by the individual. Type and amount of legitimately expected gratifications made inaccessible or type and number of significant contacts severed could serve as markers or measures of psychological deprivation. Similarly, situations or given periods in development could be classified in this way and could be viewed as parts of a process during which the individual has to cope continuously with the necessity to engage in new and to disengage from earlier affective and social ties. Attempts to attain 'developmental tasks' (Havighurst 1952) or to resolve 'developmental crises' (Erikson 1959) could be then viewed as transitional situations in which the developing individual has to disengage himself from a given

610

Interaction in social groups

social, interpersonal or cognitive network to be able to engage subsequently in new types of social and cognitive interactions. Such transitional states of incomplete or overlapping, double 'belonging' lead to lack of security, anxiety, and stress characteristic of conflict situations in general and result in difficulty in developing and sensing a firm identity, self or ego. One could be tempted to proceed from here to the badly needed but maybe too ambitious task of developing a taxonomy of developmental phases in terms of levels and patterns of social engagement and disengagement. The more modest aim of this paper was to establish the potential usefulness of the suggested approach by some research findings indicating that different types or levels of social engagement do relate indeed differentially to the development of central personality characteristics such as type of coping style. In essence the present findings may be considered consistent with our hypothesis and with what we consider a major point of consensus in a good many classical theories of socialization and personality development. Still we have represented only a modest beginning. We are quite aware that many important questions remain open, both on the theoretical as well as on the applied side. To mention only one: What would be an optimal extent of engagement-disengagement balance at given periods of development and how to assess it in given cultures and for individuals with different genetic endowment. While our findings on the whole pointed to passive coping and low self-esteem as correlates of disengagement and isolation, there is also some provocative evidence that a certain amount of 'autonomous detachment' in childhood may be related to later outstanding achievement in science (Shanan and Stein 1956). Keeping such a possibility in mind, what would be alternative implications of the present findings for some of the current educational problems in the advancement of 'under-privileged', 'disadvantaged', and so-called 'minority groups' all of which could probably better be designated by the common denominator of psychological or psychosocial disengagement? If these and some other reservations stemming from data on overengagement (Shanan et al. 1965, Shanan 1965) are kept in mind, we can conclude by quoting Martin Buber (1957) who stated that 'growth of the self is not accomplished, as people like to suppose today, in man's relation to himself, but in the relation between the one and the other ...'.

Social interaction and personality development

611

C. EDUCATION AS AN AID TO ADAPTATION IN THE ADULT YEARS* by Rolf H. Monge and Eric F. Gardner, Syracuse University, Syracuse

A major aim of education, and particularly of education in the liberal arts and sciences, is to increase the number of options from which the student may make choices in pursuing his life's goals. It it especially important in a rapidly changing world for an indiviual to have a wide variety of basic skills and knowledge that can be mobilized to meet new challenges. The fundamental skill of geatest adaptive value, of course, is the ability to acquire new knowledge, or knowing how to learn. For many individuals the experiences of everyday living tend to become increasingly canalized. As Kuhlen (1964) noted, people tend to build comfortable niches in life, niches defined by the webs of mutual commitments each individual has with others. These mutual commitments include those with spouse, children, and other family members, with people in his world of work, with his friends, with the organizations in which he holds membership, and with sundry other people and institutions. These mutual commitments serve to satisfy the individual's needs and to protect himself from threat and anxiety, but as earlier noted they also tend to canalize his energies, interests, and activities. Many of the skills and knowledge he has acquired in the course of formal schooling decay with disuse, while skills and knowledge pertinent to current activities increase and stabilize. In short, much of the flexibility in decisionmaking the individual once had is lost, both through disuse and through his being 'locked-in' to the web of mutual commitments he has made. His ability to adapt to new situations or to alterations in familiar situations of long standing tends to be reduced. There are no entirely satisfactory short-term solutions for the individual locked in to his web of commitments who must face a threat to the security of his niche. Escape from the niche is not possible short of a drastic severing of social ties. It is clear that long-term planning is required if the individual is to maintain an optimal balance throughout life between the security provided by his niche and the flexibility he must have * Much of the research reported here was supported by the U.S. Office of Education through Grant No. OEG-1-7-061963-0149 to Syracuse University, Rolf H. Monge and Eric F. Gardner, Co-principal Investigators.

612

Interaction in social groups

in order to meet and adapt to threats posed by sudden changes in his social and physical environment. Education is surely one means of increasing flexibility through the keeping open of options. Many of the sudden changes in life circumstances to which people must adapt are latent in the lives of all of us and can, thus, in some sense be planned for. In the occupational realm, for example, the family breadwinners) should have contingency plans to cover temporary or long-term loss of job due to illness or economic reversals in the field of employment (e.g., the recent difficulties of engineers in the aerospace industry, of elementary school teachers, and of Ph.D's in many fields), as well as plans for the post-retirement years. In the area of family living, all must face up to the loss of parents, spouse, and other close relatives and friends through death, and to the 'empty nest' as children leave home and strike out on their own. Everyone must face the more or less inevitable physical and psychological changes that accompany the aging process, changes to which we must adapt by changing activity patterns and emotional investments. Recreational habits, for example, must be altered as physical capacities change, and self-regarding attitudes must respond to the realities of age changes in highly valued bodily characteristics that do not directly affect behavioral capacities but do alter the individual's social stimulus value, such as loss of (or graying of) hair, wrinkling of skin, and changes in body proportions. Education can help to prepare an individual for all of these changes in life circumstances, both directly in the provision and maintenance of knowledge and skills specific to the meeting of crises, and indirectly through the building of a broad and sound foundation of general knowledge and adaptive capacities. Education in this context includes not only the formal, institutionalized education for the young, but also life-long continuing education. It is appropriate, therefore, to inquire into the status of people of different ages in areas of general knowledge, their specific skills in subject matter areas that could be designated tools of learning, and on personality variables that are pedagogically relevant. The research project. Starting in 1966 the late Professor Raymond G. Kuhlen and the authors conducted at Syracuse University a five-and-onehalf year project with the support of the U.S. Office of Education that involved the determination of adult age differences in a variety of abilities and personal characteristics of presumed importance to learning. In addition

Social interaction and personality development

613

to the studies reported here the project also included a program of experimental studies, the results of which are not reported here. Details of the whole project may be seen in the final report on the project (Monge and Gardner 1972). Sample. In problems relating to adult education (whether involving general education or more specific types of retraining) the focus must be upon differences between existing age groups, since it is with these differences that the adult educator must cope. Thus the cross-sectional design was chosen for use in this set of studies. Intact groups of adults in social, church, PTA, local organizations, and industrial establishments were recruited by offering to pay the treasury of the organization a per capita fee for the participation of a least 80 per cent of their active membership. The sample was drawn from groups found in central New York State with one exception. Because of the need for a large sample of the adult population over sixty years of age several testing sessions were held at the Dade County, Florida Senior Centers. The responses of these people were comparable to those of subjects over age sixty obtained in central New York. At least 100 males and 100 females in each of the age decades from the twenties through the seventies participated in this study. In order to make comparisons across age decades more meaningful detailed information was obtained from a brief biographical survey on age, sex, amount of education, and occupational level (for married women not employed outside of the home the occupation of the husband was used). One of the major problems in interpreting data related to age trends arises from the fact that the amount of education acquired by people of different ages varies considerably. Since education is an especially important factor in dealing with data from the kinds of instruments used in this research raw scores adjusted for education by covariance are presented. Adult age differences in general knowledge. A set of twenty-eight paper-andpencil tests was devised aimed specifically at discerning differential age trends among adults. The most promising of these were refined and used to collect the basic data reported here. The tests were all five-alternative, multiple-choice tests: (a) Form D-2, a general vocabulary test suitable for use across the entire adult age range, 30 items; (b) Form TR-2, concerning modes of transportation used at various times in the past 75 years, 23 items; (c) Form DD-2, diseases and other medical matters

614

Interaction in social groups

specific to various adult ages and matters related to death and dying, 20 items; (d) Form SL-2, slang current among different generations, 25 items; (e) Form FI-2, financial matters including stocks and bonds, estate management, wills, and installment buying, 20 items. The results obtained with these five instruments were as follows: 1. Form D-2, general vocabulary. Little sex difference was noted on the mean education-adjusted scores. Females were slightly superior in the decades of the 20s, 30s, and 40s, the sexes were essentially equal in the 50s and 70s, and males scored slightly higher in the 60s. For males, performance increased from the 20s to a peak in the 60s and declined slightly in the 70s. For the females, performance increased from the 20s to the 50s and remained stable from that age decade through the 70s. 2. Form TR-2, transportation. The curves relating performance to age for the two sexes were essentially parallel. Performance increased from the 20s to a peak in the 50s and declined slightly thereafter (to a level between the 20s and 30s) for the decades of the 60s and 70s. 3. Form DD-2, matters concerning death and disease. Females scored considerably better at all ages than did males. Males increased from the 20s to the 70s. Females increased from the 20s to a peak in the 50s and declined thereafter, with the means for the 60s and 70s declining to the level of the 40s and 30s, respectively. 4. Form SL-2, slang. Females' adjusted means were slightly higher than males' in the 20s, 30s, and 40s. The sexes were equal in the 50s and 70s, but in the 60s the males scored considerably better than the females. Males increased from the 20s to a peak in the 60s, but the 70s scored considerably worse than the 20s. Females increased from the 20s to the 40s, with a consistent downward trend thereafter to the 60s with the mean slightly below the level of the 20s, with the 70s still further down. 5. Form FI-2, finance. Adjusted means for both sexes were together and increased from the 20s to a plateau for the 30s and older. Curiously, the sexes diverged in the decade of the 50s, with females increasing and males decreasing. The sexes converged in the 60s and then diverged in the 70s, with the males increasing to slightly above the plateau and the females decreasing to the level of the 20s. If these five tests can be taken as samples of the domain of general knowledge, then it may be concluded that older adults have a greater fund of general knowledge than do younger adults, at least through the normal working years. These data support a cumulative model, with age increasing the fund of experience and knowledge upon which the indi-

Social interaction and personality development

615

vidual can draw. It is not clear why this model generally breaks down in the age decade of the 70s. Examination of the raw scores, which generally paralleled the scores adjusted for covariance with educational level, indicated that the declines noted in the 70s were not an artifact of the adjustment process. Adult age differences in educational skills. As a means of assessing the 'tools of learning' possessed by adults of different ages the try-out form (Form T - l ) of the Adult Basic Learning Examination (ABLE), Level III, was administered to the sample earlier described. The tests used from ABLE were spelling, reading comprehension, reading retention, arithmetic computation, and arithmetic problem-solving, with the vocabulary test omitted as superfluous (Karlson et al. 1967). The trend with age of mean scores, adjusted for covariance with educational level, is reported here, as in the previous section. 1. Spelling. Females were superior to males at every age. There was little variation among females as a function of age. Males increased slightly from the 20s to the 60s (except for a dip in the 30s), with a decline to the 70s. 2. Reading comprehension. Except in the decade of the 20s, males were superior to females in every decade. Females declined continuously from the 20s to the 70s and, except for a rise from the 20s to the 30s, the males also declined across the age span. 3. Reading retention. This was a subpart of the reading test that involved reading of a simulated newspaper for a predetermined interval and then answering questions about what had been read without referring back to the material. The sexes were about equal in the 20s and 30s, and then diverged. The males improved slightly to the 60s and declined slightly to the 70s. The females declined from the 20s to the 50s, and remained at about that level through the 70s. 4. Arithmetic computation. Males were superior to females in every decade, with the difference decreasing with age to near equality in the 70s. There was a general and significant decline with age in performance in both sexes. 5. Arithmetic problem solving. Males scored higher than females in every decade, and both sexes declined significantly across the age range. With the exception of the scores on the reading retention test, which changed little, and on the spelling test, which changed little for females and was irregular for males, the general picture is one of

616

Interaction in social groups

poorer performance in these basic educational skills on the part of the older persons. A disuse model would appear to be appropriate to describe this trend. That is, the further an individual is in time from his early formal schooling, the poorer is his performance on school-learned skills. Adult age differences in learning orientation. The extent to which adults of different ages are positively oriented toward the seeking of new educational experiences or the degree to which they react negatively to learning opportunities because of uncertainty or apprehension may, in a practical sense, be of even greater importance than their abilities. The level of an individual's performance in learning tasks will likely be determined to no small degree by the extent to which he is 'achievement oriented' or threatened by the situation in which he finds himself. Many adults may in fact be sufficiently threatened that they never make attempts at continuing education or vocational retraining. It was necessary in this phase of the study to obtain information regarding four types of characteristics: Motivation for learning and achievement; learning apprehension or anxiety; personal rigidity; and extent of participation in educational activities. 1. Motivation for learning was measured by adaptations of the 'Demand for Achievement' items in Sarason's (1957) Autobiographical Survey. Thirteen of his 41 items were randomly selected and reworded to make them more applicable to the adult years. (These 13 items were interspersed with 13 of Sarason's 24 items on 'General Anxiety' and 13 of his 21 items on 'Test Anxiety' to form an integrated 39-item test form.) 2. Learning apprehension was measured in three steps. First, measures of manifest (or general) anxiety were obtained using Bendig's (1956) scale and an adaptation (as described in the preceding paragraph) of Sarason's (1957) general anxiety items. Second, test anxiety was measured with items adapted from Sarason's (1957) scale. Third, to assess the degree to which anxiety generalizes across situations as a function of adult age, subjects took a 'situational anxiety' test, which was developed along the lines suggested by Endler et al. (1962). Subjects were asked to rate each of ten situations on twelve five-point scales. The situations were these: 'You are getting up to give a speech before a large group; You are just starting off on a long automobile trip; You've received a notice from your bank that an important check has bounced; You have just received a call from a hospital emergency room saying that your child is there; You are going to your doctor for a routine physical examination; All the lights

Social interaction and personality development

617

in your home have just gone off unexpectedly; The check-out clerk at the supermarket has just rung up a large bill for you, and you find you don't have enough money; You are about to make a long-distance telephone call; You have just been nominated for President of your organization; You have just narrowly missed a dog that ran out in front of your car.' These were devised to represent plausible situations for adults of all ages, but situations having different connotations for older than for younger adults. A bounced check, for example, means something different to an aged pensioner than to a young businessman, and longdistance calls tend to be associated in the minds of older people with bad news or an emergency, where such calls tend to be more or less routine with younger adults. The twelve scales on which subjects rated their degree of reaction to each situation were: Heart beats faster, Get an 'uneasy feeling', Emotions disrupt action, Feel exhilarated and thrilled, Want to avoid situation, Perspire, Mouth gets dry, Become immobilized, Get full feeling in stomach, Seek experiences like this, Experience nausea, Dizzy feeling. 3. Rigidity was measured by self-ratings on the 19 'best' items from Troldahl and Powell's (1965) developmental work on shortening the Rokeach dogmatism scale. (Their twentiethitem wasinadvertently omitted when the form was produced.) 4. The long form (99 items) of Litchfield's (1965) Leisure Activity Scale was administered. The 46 items designated by Litchfield as 'educational activities' were weighted by 'degree of educativeness' and 'frequency of participation' (all as described by Litchfield, 1965) to obtain an 'Extent of Educational Participation Score', with the total score potentially ranging from zero to 615. A 2 (sexes) X 6 (age decades) unweighted means analysis of variance was conducted on the raw scores on each measure with the following results: 1. Demand for achievement. Only the age effect was significant. After a drop from the 20s to the 30s, there was a general increase in score across the remainder of the age range indicating that the older adults felt a greater need to achieve. 2. Learning apprehension. Deferring discussion for the moment on 'situational anxiety', there was a significant sex difference on all of the other three anxiety measures, with females rating themselves higher than males on all scales at all ages. The age effect was significant on all but the modified Sarason test anxiety measure. On both the modified Sarason

618

Interaction in social groups

general anxiety test and on Bendig's manifest anxiety scale the high points were in the decades of the 20s and the 50s for both sexes, and the low points were in the 30s and 60s for men and the 40s and 70s for women. On the 'situational anxiety' scale, a mean and a variance was calculated for each person across the ten situations. The mean of the individuals' means for each sex and age decade was calculated to reveal the level of anxiety, and the mean of the individuals' variances was calculated (and transformed to logarithms before analysis of variance) to determine whether there was less intraindividual variance - and thus greater generality of anxiety - as a function of increasing age. The main effects of sex and age and their interaction were significant on the analysis of the means. Females scored higher on level of anxiety than males. For males, the 60-69 age group was lower in anxiety than the other ages, while for females the decades of the 20s, 40s, and 50s were higher than the decades of the 30s, 60s, and 70s. In neither sex was age systematically related in a simple way to anxiety level. As for the variances, the only significant effect was the sex difference, with females having a higher variance than males, and thus showing less generalization of anxiety across situations than did males. 3. There was a remarkable increase in scores with age on the short form of Rokeach's dogmatism scale. Females were lower than males at earlier ages (20s, 30s, and 40s), but were higher in dogmatism than males in the later years (50s, 60s, and 70s). The overall sex effect was not significant, but both the age effect and the age by sex interaction were significant. 4. Inspection of the means of the 'extent of educational participation' scores, derived from the Leisure Activity Survey, suggested that men vary more from decade to decade in the degree to which they participate than women do, that participation for both sexes is higher among people in the 20s and in the 60s than in the other decades, and that better educated people participate to a greater extent and more consistently across the age span from 20-79, than do people with less education. Examination of these analyses leads to several suggestions. First, with respect to the various measures of anxiety, it does not appear that age is related in any straightfoward way to either the level of anxiety or its generality across situations as measured by the several approaches used here. There is, on the other hand, ample evidence on the difference between the sexes. The female participants rated themselves higher in every decade on every anxiety measure than did the males. Information derived from factor analyses indicated that the anxiety measures cluster

Social interaction and personality development

619

together, that they are positively associated with being female, and that they are independent of age, education, dogmatism, and vocabulary measures. Second, dogmatism is positively related to age and negatively to education. Since age and education were also negatively related, the question arises whether age, as such, is independently related to dogmatism. As one attempt to cast some light on this issue, the partial correlation of age and dogmatism score with the effect of education removed was calculated. The value was + .43 which, with 1285 degress of freedom, was clearly significant. (The unadjusted - zero-order - correlation of age and dogmatism was + .50, of age and education was - .48, and of education and dogmatism was - .32.) Thus the lower educational level of the older individual does not appear to be the primary explanation of the positive relationship between age and dogmatism. As noted earlier, dogmatism was separated in the factor analyses from the anxiety measures. Examination of the correlations of dogmatism with the anxiety scales revealed that they ranged from - .01 to + .18 for the ten situations, with the mean of the intraindividual means correlating + .14 with dogmatism. Dogmatism correlated + .17 with the Bendig scale, and + .18 and + .12 with the adaptations of Sarason's test anxiety and general anxiety scales, respectively. The highest of these several correlations ( + .18) accounted for only 3.2 percent of the common variance with the dogmatism score. Thus, Kuhlen's (1964) explanation of the increasing trend with age in dogmatism in terms of increasing anxiety seems not to have been supported by these data. And it cannot be too plausibly argued that dogmatism and anxiety should not co-exist if the former is supposed to be a way of controlling the latter, since Rokeach (1960) found highly significant correlations between dogmatism and anxiety in seven different samples. However, Kuhlen's hypothesis about the increasing generality of anxiety across situations with age fared somewhat better. There was in both sexes (but more so in the females) a declining trend in the intraindividual variances among the ten situations with age. Although the downward trend was not statistically significant it was consistent and regular except for a slight elevation among males in the 50-59 decade. This should encourage further examination of Kuhlen's hypothesis. One possible explanation for the increase with age in dogmatism is that ready-made, stereotyped responses are energy-conserving since little or no thought is required, and there is little wasted action. And, more often than not, such responses are adaptive. They have, after all, been learned through

620

Interaction in social groups

oft-repeated experience and in a variety of circumstances through a long life. Some of the older subjects in this study, for example, betrayed something of this in their arguments with the proctors over how to treat the task of rating the 'bounced check' and 'supermarket over-run' situations. Many of the older people assured us that they would never let such things occur, and how in the world could they imagine how they might feel in such a situation? This unwillingness to experience, even vicariously, might certainly be construed as dogmatic or rigid behavior but it is also symptomatic of an unremitting caution in important matters, caution that has been translated into an iron-clad routine that avoids both trouble and the expenditure of energy that would be required to put things right. Discussion. The need for viewing education as a life-long process aimed at the keeping open of options, or the maintaining of adaptive capacity, should be obvious. The data presented here, though limited in scope and from a limited sample, tend to suggest three things. First, the fund of general knowledge held by adults tends to be greater for older people, indicating that a broader base exists for them upon which to build any educational program. Second, some of the basic educational skills (reading comprehension and arithmetic manipulations) possessed by older people are not as polished as those of younger adults, suggesting that some review and refurbishing of these skills may be needed in educational and retraining programs. Finally, the personality data suggest that older people, especially women, may feel more pressure in the learning context, and that older people of both sexes have a considerably more dogmatic, less flexible outlook on life than younger people. Whether this implies that older people are significantly less open to new ideas or new approaches to problems cannot be stated with assurance; nevertheless, this information should be an important element in consideration of the design of educational programs.

Social interaction and personality development

621

D. INTERACTION O F PERSONALITY, SES, A N D SOCIAL PARTICIPATION IN OLD AGE by Reinhard Schmitz-Scherzer, University of Bonn, Germany, and Ursula Lehr, University of Cologne, Germany

The relevance of social contacts for personality development in all age groups is shown by many findings. For infancy this relevance has been demonstrated due to the influence of psychoanalytical theories especially regarding mother-child relationship. Although Yarrow (1961) and Casier (1969) were right in criticizing an overemphasis of the mother we still have to regard the mother as the most important source of stimulation for the child, and everybody agrees that stimulation is the most decisive variable for personality development. During childhood and adolescence social contacts, especially with peers, are very important factors in shaping personality. 'Significant others' give direction and orientation and transmit values. Failure and success in school are influenced by processes of social interaction. The same is true for the occupational career in young and middle adulthood. We could show that satisfaction with job as well as life satisfaction in women depend especially on social contacts. For old age this relationship was stressed in an impressive way by 'disengagement theory' (Cumming and Henry 1961). We may summarize its implications for our topic in the following way: Middle age social contacts have to be restricted in old age in order to further satisfaction and adjustment to the ageds' role. So far there exist many arguments for and against this theory and quite a few modifications of it. Especially the hypothetical correlation between high degree of life satisfaction and lower degree of social interaction in old age has been questioned (Lehr and Rudinger 1969). In this connection we especially pointed to personality variables which were introduced as intervening variables in this theory by Havighurst et al. (1964). Those individuals who reported a lower degree of social participation during younger and especially middle age got higher scores in life satisfaction (LS) if they scored lower in social participation in old age. Subjects with high degree of social participation during younger and middle age scored higher in LS in old age if they had a high degree of social participation. According to the findings of the Bonn Gerontological Study (Thomae 1968) related to the first three years of observation (1965-1968), theses of disengagement theory (DT) with

622

Interaction in social groups

regard to a negative correlation between age and degree of social participation, and of a negative correlation between LS and degree of social participation, have to be questioned (Lehr and Rudinger 1969). Furthermore we found that the relationship between LS and degree of social participation was different in different social roles. Finally it could be concluded from these findings that personality and SES variables were major determinants of the quality of relationships between LS and degree of social participation (Lehr and Rudinger 1969, 1970). Methodological problems. Many problems combined with the discussion of DT imply methodological problems (Maddox and Eisdorfer 1963, Maddox 1966). (1) So far we use data from cross-sectional studies only, which very often lead to the belief in a one-dimensional relationship between social activity and life satisfaction; this kind of study can never show the real process of disengagement. The actual nature of the sequence of engagement-disengagement (like any other sequence) can be studied longitudinally only, as for instance Maddox and Eisdorfer emphasized. (2) Furthermore, it seems to be necessary to approach problems of social change in a multi-dimensional way which takes account of as many variables as relevant for the individual life sphere. So we have to ask for the variables related to social change that lead to an increase or decrease in social role activity, and, furthermore, we have to ask for variables accounting for the inner experiences of satisfaction or dissatisfaction. These principles were applied in the Bonn-longitudinal study of 220 men and women (aged 60-65 and 70-75) started in 1965 at the University of Bonn (Thomae 1968). These subjects were interviewed and tested for a whole week each in 1965, 1967, 1969, and 1970. Especially we tried to find biographical details of the life of the persons interviewed. Of course we administered an intelligence test and rated from explorations which lasted all in all about 4 to 6 hours per investigation formal characteristics of behavior, life activity, mood, security, and adjustment among others. We got measures of self-esteem and attitudes toward others and towards the present situation and took many measures concerned with social role activity during this period of observation and satisfaction with changes in these roles. Each year approximately 1000 medical, psychological, and sociological measures were taken per individual. According to the usual classification we rated activity and satisfaction in the following family and non-family roles. Parent, grandparent, spouse, kin, friend, acquaintance, neighbour, clubmember, and citizen.

Social interaction and personality development

623

From the very detailed interview which could be tape-recorded we can confirm the thesis of the two-way-character of role activity: The concept of role activity primarily refers to subject-initiated activity. Our data, however, show that social activity can be determined as well by activity of the subject himself as by attitudes and behaviors of the society which can exclude the individual from some roles • and lower his social activity. Therefore, in this context we shall avoid the term 'social activity' and use that of 'social participation' instead. In this paper we should like to give some data which refer to the degree of change of social participation during five years of observation and of satisfaction with this change and the pattern of changes of social participation in different roles in their relation to different social, economic, and personality variables. These data are based on two one-week interview and test periods, conducted with 144 subjects in 1965 and 1970. The interview schedule and the rating-scales for the social roles were adapted from those developed by Havighurst and co-workers for the Kansas City Study, modified by our Cross National Research on 'Adjustment to retirement'. Full information is given in the book edited by Havighurst, Munnichs, Neugarten and Thomae in 1969. Methodological problems of computing these rating data are discussed by Frohlich et al. (1969) and Bengtson (1969) in the same publication. Consistency and change in socialparticipation and in satisfaction with change. A comparison of the median scores for activity in the five-year interval demonstrates a considerable degree of consistency. Almost no significant changes are observed in the family roles, whereas in the extra-family roles there exists some amount of change which is defined by increase in some subgroups and roles and decrease in other roles. In the group of the 60-69 year old persons, for instance, we saw a significant increase of activity in the acquaintance role connected with a likely significant increase of the degree of satisfaction in that role. In the friend role we found a significant decrease in the degree of satisfaction although the activity in that role showed a nonsignificant decrease. None of the above mentioned trends were found in the group of 70 years and older people. Here the neighbour and the clubmember role showed significant decreases in activity but in the neighbour role we found a significant trend in the reaction to that change: Here people became more satisfied (when being less active in that role). Women showed a significant increase of activity and satisfaction in the

624

Interaction in social groups

acquaintance role. The degrees of satisfaction in the other different roles were found to point to significant decreases in the kin, the friend, and the club-member role without a change in the amount of activity. Men were found to show significantly more activity in spouse and less activity in the neighbour role where the degree of satisfaction showed an increase. In all the groups of men we found some changes in role activity and the reactions to that change. That and other insignificant trends can lead to the hypotheses that the situation in the family and profession too and perhaps many other aspects of the present situation and the perception of the situation by the individual influence the changes in the social role activity. But at first, we have to consider that perhaps different trends in changing roles - e.g., one part to a decrease, the other to an increase did eliminate significant changes in one direction when we try to prove them by medians. Therefore we have to differentiate our view by asking which individual did show an increase, a decrease, or consistency. This analysis points only to a moderate amount of consistency. An overall assessment of changes in the different roles points to a complete consistency in 47.9 % of the cases, to an increase in 22.5 % and to a decrease of social participation in 29.6 %. The degree of consistency differs, however, within different roles. The lowest degree of consistency is observed in the kin- and acquaintance-role, the highest in the role of a club member and in family roles like the spouse and most of all in the parent role. The change in the spouse-role goes in the direction of increase (p < .01). In the grandparent-role, there is also a tendency towards increase, especially in men (p < .01). Within the parent-role, those 39.8 % who changed did so in the direction of increased and decreased participation in nearly the same amount. A similar relationship was found in the kin-role, in which the degree of consistency is lower than in the other family roles; in the older group and in men, there is a significant increase (p < .01) compared with the younger group and the group in women. The 55.6 % of cases, who changed in their neighbour-role, did so in the direction of decreased participation. Those who changed within the club-member-role show a significant tendency towards decrease (p < .01) especially in the older group (p < .05) and in women (p < .01). Approximately 55 % showed changes within the friend-role. Here we found as well a decrease and an increase. Also those 57 % who changed in the acquaintance-role showed a tendency in both directions but the younger group showed more cases of increase (p < .05). Contrary to our 1969 analysis we might say that consistency in social

Social interaction and personality development

625

participation is not always the dominant tendency in social behavior and life conduct during the time observed. As far as change was observed we found significant increases in two of the roles (spouse and grandparents) and decreases of participation in three of the roles (kin, neighbour, club memfber); these changes and tendencies are different in different age and sex groups, however. Our findings do not support however, the first statement of DT: There is a general decrease of social activity as people grow older. Analysis of change in terms of disengagement-theory and activity-theorypatterns of change. According to DT those aged who disengage are persons that have a sense of psychological well-being and a high degree of satisfaction. Corresponding to AT of aging increase in social participation will result in satisfaction, and a decrease in dissatisfaction. Analyzing the data of our whole sample, we cannot state any change of social participation in 47.9 %. Of the remaining 52.1 % there are 19.1 % who behave in the way expected by DT, and 33.0 % as expected by activity-theory (AT). This relationship remains about the same for both the older and the younger group, as well as for the males and females. As far as different roles are concerned we can state that only in the parent-role is there no difference in the direction proposed by AT or DT. The younger group showed more changes in the direction proposed by DT, women more in the direction proposed by AT(/> < .01). In the spouse role, women changed more in the direction proposed by AT (p < .01). Generally changes in spouse-role lead to reactions explained by AT (46.8 % of changes, 15.9 % reactions in the DT-way and 30.9 % reactions in the AT way). The kin role, acquaintance role, club-member, and friend role show the same kind of change in participation, but there changes in the kin role in the older group and in men showed changes according to AT; the older group in the friend role were found to show the same changes (p < .01), and also men in the neighbour role and women in the clubmember role. Changes in the AT-way we found in the younger group in the acquaintance role and the younger group in the friend role (p < .01 in both roles). Generally, we may state that among those 52.1 % of the cases who showed a change, the DT pattern of social participation could be found in the different groups within one-third of the reactions to changing role participation, while the AT pattern could be found in two-thirds of the cases in that groups. Most changes in the roles in the whole sample were observed to be in the AT-pattern.

626

Interaction in social groups

This analysis should be supplemented in a more differentiated way or in terms of changes as defined by both theories (Table 1). Most changers in our sample who follow the DT pattern point to an increase of activity combined with decrease in satisfaction scores. Only in the neighbour and in the club role is more satisfaction to be found with a lower amount of activity. The changers following the AT pattern can be described mainly in terms of a decrease in activity and in satisfaction with the activity in such roles. Only in the spouse role was there more often a change in terms of an increasing amount in activity as well as in satisfaction. This very complex pattern of changes demonstrates again that behavioral reactions to the social situation of the aged cannot be explained by one theory and that DT accounts for the smallest number of cases. The explanation for this complexity may be found in personality variables and social conditions. Therefore we can regard studies like this as a good example for evaluation of relationships between personality and social interaction. Patterns of changes insocial participation in different roles in their relations to social and personality variables. Whereas it has been known since Havighurst et al. (1964) that personality is a variable correlated to social engagement in old age generally little is known so far regarding these relationships in different social roles. The longitudinal analysis of our data demonstrates a very complex pattern of relationships which may be surveyed in the most economical way by defining the social participation aspects of these patterns in terms of D T vs. AT (see Table 1). No social or personality Table 1 Percentage changes according to the views of theory and activity theory

Role

Total amount of changes

parents 18.6 grandparents 23.4 spouse 15.9 kin 18.4 acquaintance 20.2 friend 17.3 neighbour 21.5 17.7 club

disengagement

Increasing activity, decreasing satisfaction

Decreasing activity, increasing satisfaction

Total Increasing Decreasing amount activity, activity, of increasing decreasing changes satisfaction satisfaction

10.2 17.0 11.7 10.1 12.5 11.6 4.9 5.7

8.4 6.4 4.2 9.3 7.7 5.7 16.7 12.0

21.2 32.0 30.1 38.1 36.7 38.1 34.1 30.7

7.6 14.9 17.0 15.1 15.9 14.0 11.2 5.6

13.5 17.1 13.9 23.0 20.8 14.1 22.9 25.1

Social interaction and personality development

627

variables discriminate changers in terms of DT and those in terms of AT in the spouse role. On the other hand these variables are rather relevant in the parent role. Those aged persons who are dissatisfied with higher activity in this role, or satisfied with lower activity, show a significantly higher intelligence score: Thus disengagement theory could be defined as a generalization from findings on more intelligent persons regarding the parent role. On the other hand subjects who followed the activity-theory pattern in this role reported more problems with children and spouse, of being needed, and more contacts with relatives. In this role problem situations apparently are effective as stimulations for social activity and affective reaction to it. In the grandparent role we can find an almost opposite pattern: The pattern of change in this role as defined by DT pattern is related to more problems regarding health and spouse relationships. On the other hand grandparents who followed the the AT patterns of social change report significantly more often better health and got higher scores in participation in the roles of club member, acquaintance, and relatives. We could not find any significant difference regarding social background and personality in the roles of relatives. As far as extra-familial roles are concerned the AT pattern of change was correlated with high degree of tension and stress regarding family life and/or health in the neighbour's and citizen's role, whereas in the roles of club member and friend DT conforming changers more often report higher degree of tension and stress, e.g., spouse, or living conditions. Conclusion. Generally we may state that Disengagement Theory as evaluated by longitudinal findings such as ours describes one of many patterns of change which occurs in correspondence with very complex changes in personality and in social conditions. Social interaction in terms of these findings is a dependent variable of very different personality and social variables which may have a unique character in many cases. On the other hand social interaction apparently is one of the main determinants of major personality dimensions like perceived stress, health, mood, responsiveness, etc. Therefore we may conclude by confirming Maddox' statement that disengagement theory was a premature generalization from cross-sectional data which had to be supplemented and/or controlled by longitudinal analysis. This analysis emphasizes individual patterns of change in different areas of social participations as well as individual relationships between these patterns and the rather individually changing worlds in which both the changers and the non-changers live.

628

Interaction in social groups

E. THE GENERATION GAP: IMAGINATION OR REALITY? by Bengt-Erik Andersson, Goteborg University, Sweden

The socialization process. Life means change but also continuity. Incessantly, one generation follows another - in static societies very similar to the preceding, in societies with a high rate of change often quite different from the preceding generation (cf'. Mead 1970). Members of a generation are constantly exposed to influences from different quarters - a socialization process with both conscious and unconscious components. Socialization, which can be defined in a vast number of way (see, e.g., Wernersson 1972), may be regarded from several perspectives. From society's point of view, socialization can be described as a process through which an individual learns his position in society and the norms society needs and requires {cf. Elder 1968), i.e., socialization is 'the attempt to ensure continuity of a social system through time' (Bengtson and Black 1972). But socialization can also be regarded from an individual perspective implying an individual's acquisition of the knowledge, skills and values he needs to function in society {cf. Brim 1966). The process is life-long (Cronbach 1963), even if most theories of socialization treat almost exclusively childhood and children's development to adults. In this paper I will focus on development between ages 14 to 20. From birth to maturity an individual has been exposed to a socialization process with several socialization agents. The process has taken place both in groups and individually and has been of both formal and informal nature (Brim 1966; Wernersson 1972). At the beginning, the parents members of another generation - and other family members have dominated the process. Rather soon peers, preschool and school, mass media and other agents also start to exercise influence on the individual, even if parents keep some sort of dominance. The transition from a subordinate position in society as a child and an adolescent to a status as an independent member of society of equal standing with other adults may - depending on the social system {cf. Eisenstadt 1956) - sometimes be connected with a prolonged emancipation period filled with conflicts and strivings for autonomy. The output of the socialization process at a given moment is the result of the accumulated influences of different agents and the interaction between them and the individual (Andersson 1969; Andersson and

Social interaction and personality development

629

Ekholm 1971). During adolescence and early adulthood it happens that members of the young generation exhibit behaviors and values which may seem frightening and strange for older generations. They fear failure in their socialization attempts and feel concern for the youngsters' behavior and future. This is not a new phenomenon but an historical one, well documented, for hundreds and thousands of years. But the rapid technical change in our own century with all its consequences has led many authors to state that today's differences between generations are of another and more serious nature than during earlier epochs and lead to more serious consequences for society and for relations between individual members of different generations. The generation gap. This view is highlighted in the contemporary concern over the generation gap. Today's mass media have, by their inclination for dramatic incidents, contributed to the popularization of this concept among the general public. However, the opinion among researchers who have paid attention to the phenomena is far from unanimous, as Bengtson has showed (1970). Some are convinced of the existence of a great gap, others think it is an illusion, and a third group maintains that differences may exist in some areas while the similarities are greater in others. The Stug project (Studies of the generation gap) is a research project interested in the conditions in Sweden. Is it possible to talk about a generation gap and, if so, how can we explain its genesis (Andersson and Ekholm 1971, 1972)? These problems can be studied partly on a macrolevel when you regard the generations as large age-based aggregates differing or being similar in different aspects, and partly on a microlevel, where the interest is focused on interpersonal relations between individual members belonging to different generations and 'where the older person bears some responsibility for the young person's successful attainment of functional adulthood' (Bengtson and Black 1972). In this paper I will focus on the microlevel. While the disagreement is large between the debaters concerning the existence or nonexistence of a generation gap (Bengtson 1970), empirical results have been more clear-cut. Very few results seem to give support to the idea of large differences between, for instance, parents and their children (Simmon 1971; Kandel and Lesser 1972). One reason for the disagreement in spite of this may be the fact that both generations have seldom been studied simultaneously and in comparable ways. Another reason may be the lack of a distinct definition,

630

Interaction in social groups

In Sweden, as in the United States, there exist expectations about differences, opposition, and conflict between generations, in spite of the lack of convincing scientific support. The studies on the macrolevel which we have done so far, in which different age groups have rated partly their own generation and partly other generations, don't indicate any real differences in the way of looking at each other. On the other hand, there seems to exist a perception gap, which means that members of the different age groups expected or perceived differences that didn't exist. Especially the younger age groups, from 13 to 20 years of age, expected the adult generation to value young people rather negatively and at the same time value adults very positively - something that according to the results was wrong. Also the adult groups showed the same type of faulty expectations, but the gap was much smaller (Andersson 1973a, 1973b). Similar discrepancy between 'self- and hetero-perceptions' has been described by Ahammer (1970) and Ahammer and Baltes (1970). Three dimensions. The above mentioned results indicate that when you talk about a generation gap or generation conflicts you need to distinguish between several dimensions. I also prefer another term for the concept discussed in this paper - a term that better stresses the kind of phenomena we are interested in in the Stug project, namely the negative relationships between generations' negative attitudes and emotional reactions to each other. As a descriptive word for these relations we use the term 'opposition' (Andersson and Ekholm 1971, 1972). This concept involves at least three dimensions: (a) The existence-nonexistence of an actual difference between two generations or members of two generations. The existence of a difference is for many people enough to talk about a generation gap. But a generation gap defined in this way makes the concept very insipid. No one will deny that differences in values and behaviourmayarise between individuals with different life experience, age, and so on, even if they may be smaller than expected. Davis (1940), for instance, maintained that such differences were unavoidable between parents and their children. (b) An individual's cognitive perception of the existence or nonexistence of a difference between generations. If an individual is unaware of an actual difference, this difference may be of less importance to him than if he perceives a difference which in fact does not exist. Since we react to our perceptions he can react to this fictitious difference both cognitively and emotionally, (c) The third dimension is this reaction. How does an individual react to a perceived difference - positively, negatively towards

Social interaction and personality development

631

the matter in question or perhaps negatively to the person or generation? In case of strong, negative reactions we could talk about a generation conflict, otherwise we prefer the term opposition. In the Stugproject we try to describe the relationships between a group of 20-year-olds and their parents in, among other things, these three dimensions. Dimensions (b) and (c), which are the most important, can be described for each generation separately dimension (a) only by a comparison between the generations. The Stug project. The project is designed in the following way. In 1965 all students in grade 8 in Goteborg who were present a certain day at school participated in a large research project aiming to study adolescents' attitudes towards different aspects of their school situation, school motivation, achievements, attitudes, values, and behavior related to peers and leisure time, the teenage culture, etc. (Andersson 1969). About 90 % of the total grade - i.e., about 4,500 14-year-olds - participated in the investigation. These adolescents have now been followed up in the Stug project. When they were 20 years old 74% answered a mail enquiry. By help of this enquiry we have data to elucidate dimensions (b) and (c) for the young people. A sample from this group have been interviewed together with the mothers. A comparison between the two generations will give us information about dimension (a) .The interviews also contain material to elucidate dimensions (b) and (c) for the mothers. In this paper I will concentrate on the mail enquiry, since the interview material is not analysed yet. The data concerns 3,388 young people. Political opinions. Much of the talk about a generation gap has been inspired by the many demonstrations and protests of different kinds that took place all over the world during the sixties. To a great extent it was the students who participated in these protests - hence the name given by some authors, 'student movements' (cf. Feuer 1969). Compared to other Western countries Sweden has had very little action of this kind, but there have been some incidents that attracted attention, for instance the stopping of the tennis tournament between Sweden and Rhodesia (Lindblom 1969), and the occupation of the Student Union in Stocholm (Mosskin and Mosskin 1960). A common ground for many of our protest movements has been antimilitarism and a stand against the use of force. This common ground was especially noticeable at the end of the 1950s and the early 1960s. Even later on, the strong antimilitaristic

632

Interaction in social groups

opinion among youth against the United States' war in Southeast Asia expressed itself in demonstrations. This last-named protest movement has been very successful since it now includes members of all ages. The vast majority of the Swedish people now sympathize with their opinions. Protests and demonstrations have also occurred against the government's policy for underdeveloped countries and against the capitalist system. The youth movement in Sweden, as in other Western countries, often has had a radical and left-oriented content, and the general picture in mass media as well as among the general public is that young people of today are very radical and in conflict with the adult society. Lipset (1970) has stated with respect to the United States that this picture is totally wrong. More young people sympathize with conservative and reactionary movements than with radical ones. Let's see how this statement fits my group of Swedish respondents. To get an idea of which direction the young people preferred for the development of society we gave them the following question: 'There is a lot you wish to change or strive for in the Swedish society. How important do you consider it is to ...'. For each of a number of items the respondents had to state how important it was to change society in a certain direction. They used a five-point rating scale: (1) most important, (2) very important, (3) rather important, (4) not very important, (5) not at all important. Twelve of the items could be arranged in six opposite pairs. One item in each pair could be classified as representing a conservative attitude and one a radical attitude. On all but one item, more respondents think it is important to change society in a conservative rather than in a radical direction. For example, about 70 % want to increase private enterprise, strengthen the resources of the police, and give the most competent the highest salary. The only item discussed in the general debate by members of radical groups and which a majority of the respondents think is important is to increase the aid to underdeveloped countries. These results really don't fit the popular picture presented in mass media. When they were published in Sweden they also attracted a lot of attention from the mass media. Perception of parents' opinions. But let's now see how the young people perceived the values of their parents. To measure this, the respondents had to state for each item how they believed their parents would answer. Over the twelve items an average percentage of 56 expected their parents to choose exactly the same degree of priority as they did themselves. If

Social interaction and personality development

633

we consider that the difference between alternatives two and three (very important and rather important) and between alternatives four and five (not very important and not at all important) is not very large, the agreement is even higher. We can get another picture of the perceived agreement between the respondents and their parents if we make an index giving one score for each item where there is agreement between the 20-year-olds' answers and the expected answerd from the parents. Two-thirds of the respondents perceive an agreement of 80% or higher, which means that they expect their parents to have the same values as themselves on all items but two or three. Not more than four individuals expect different answers on all items, and one in twenty perceive similarity on 40% or less of the items. Our respondents to a very little extent perceive differences between themselves and their parents in values concerning social and political matters. To study emotional reactions, the respondents had to answer a question about their feelings towards their parents' views. Not more than five to seven percent dislike the views of their parents so much that they want them to change opinions. Another ten percent don't like the parents' views but don't think they should change them. Fifty percent like them and another thirty percent don't care what opinions their parents have. Even if we divide the group of respondents according to the perceived agreement between them and their parents it is hard to find large negative reactions to the parents' views. Among those respondents who perceive a difference, not more than ten percent want their parents to change opinions and another fifteen percent dislike their opinions but without requiring a change. The mail enquiry contained several questions where the respondents had to state their own views together with the views of their parents and the emotional reactions on these views. The questions concerned: (a) work life - the best thing to strive for in work, the best way to have success in life and the most important thing in work; (b) family life - the most important reason to raise a family and whether you should marry to live together with a partner; and (c) society - the best way to influence the development of society. The perceived agreement was very high. On the average a good half of the respondents choose identical responses for themselves and their parents on the six questions. About 50-55 % like the parents' views and don't want them to change the views. Another 25-30 % don't care about the parents'views. Only five percent want a change. We can conclude

634 Interaction in social groups that Swedish 20-year-olds from a large city show very little sign of a far-reaching generation gap or opposition between generations. On the contrary, they perceive only small differences between themselves and their parents in views relating to family, work, and society and only a very small minority show discontent with their parents' views. From many other questions in the mail enquiry we can also conclude that the relationships between children and parents and the psychological contact seem to be very good. As examples, 73 % of the males and 84 % of the females stated that their mothers mean a lot or quite a lot to them just now. Corresponding figures for the fathers are 63 % and 74 %. And between 80 and 90 % had warm feelings for their parents. The girls seem to be somewhat closer to the parents than the boys, and the mothers are closer related to their children than the fathers. Such is the picture of our respondents considered as a group. Bad contact with the parents, hard feelings, deprecation, discussions, and quarrels don't characterize the relations between our 20-year-olds and their parents. But, of course, this type of relationship also exists, even if it is uncommon. Between six and nine percent of the respondents quarrel often with their parents. About ten percent would never turn to them to get advice. About ten percent don't feel very warm toward their parents and twice as many don't think that their parents mean much to them. Neither do they perceive their parents to be interested in or have warm feelings for them. Finally, about five percent meet their parents once or twice a year at the most. Interaction between generations. In the title to this paper I put the question of whether the generation gap is a reality or an imagination. The results presented so far make it safe to conclude that we can't find any support for the idea of a vast generation gap - at least not between young people on the threshold to adult status and their parents. This does not mean that there cannot exist something worth the name of a generation gap between certain subgroups of young and adult people. In our material, there is a small minority of young people who maybe feel some sort of opposition or conflict. Let us study this group a little further. In doing this I want to test some hypotheses developed elsewhere (Andersson and Ekholm 1971) about the rise and maintenance of an opposition. One important variable in this context is the psychological communication or contact or interaction between generations or between individuals belonging to different generations (cf. Newcomb, Turner and Converse

Social interaction and personality development

635

1966; Sherif and Sherif 1969). The contact between parents and their children is to a large extent dependent upon the kind of upbringing pattern that exists in the family (Elder 1968; Kandel and Lesser 1972). It is therefore of importance to study this variable in analysing the reasons for certain oppositions between generations. To understand such oppositions, we must also consider the relevance of the area in which differences exist or are perceived to exist. In order that the differences may be experienced emotionally in a negative direction, we think they must touch on basic values in the individual's view of his environment. We get an indication that this statement is valid from our results. The only question on which a relatively large proportion of the young people dislike the parents' views is whether or not one should marry to live together with a partner. About one-fourth of the young generation say yes, but between two-thirds and three-fourths believe their parents will say yes. Among those who perceive different views 16% of the males and 21 % of the females want the parents to change their views and another 30 % of the males and 37 % of the females dislike the views but don't require a change. Probably this matter touches many of the young men and women personally more than questions about worklife, policy, and society. Therefore, they react strongly to a perceived difference between themselves and their parents. A positive and a negative group. We constructed two indices to find a group to study where the members probably perceived opposition towards their parents. For the first we used the six questions on which the respondents had to give both their own and their parents' answers. Low agreement was defined as the case when the respondents gave identical answers on none or one question (in case they had answered all the six questions identical answers on not more than two questions were allowed). High agreement was defined as identical answers on all or all but one question. For the second index we used the six questions about the emotional reactions to the perceived parental views. A negative score was defined as an average choice of one of the three first alternatives indicating that the respondents dislike their parents' views but don't necessarily want them to change views. A positive score was defined as an average choice of alternative 4, indicating that the respondents like their parents' views and want them to keep the views. By combining the two indices, we get a group of young people with low perceived agreement in certain views with their parents and with at least mildly negative reactions towards

636

Interaction in social groups

this disagreement. For the sake of comparison we have also chosen a group with high similarity to their parents and with positive reaction to this similarity. In spite of our mild criteria the 'negative' group is very small - 223 young people - which is about seven percent of the total group; the 'positive' group is larger - 506 persons - which is about sixteen percent of the total group. The present situation. Let us start by looking upon the present situation of these two groups. The members of the negative group are much more dissatisfied both with different personal conditions and with conditions concerning their jobs or studies, than members of the positive group are. The largest differences concern the housing conditions and the relations to parents, where 32 % more people in the negative than in the positive group want a change. That the present relations to parents are not so good in the negative group could also be seen from other questions. Less young people in this group than in the positive group feel that parents treat them mostly as grown-ups and more have violent discussions and quarrels and think parents' ideas are old-fashioned. They would not turn to parents for advice so often, and the parents do not discuss their problems with them as often as in the positive group. On the whole, the negative group discusses with the parents less frequently than the positive one except for political matters. The negative group seems to contain more persons with political interests, but the group is not politically homogenous. On an average 37 % want changes in society in a conservative direction, against 46 % in a radical direction. The positive group is more homogenous, but in a conservative direction; 54 % want such changes against 21 % for radical changes. If we look upon the self-evaluation of the individuals we find that members of the negative group seem to have a lower self-evaluation. This can be traced back to grade 8, where we find a tendency for the same difference. At this time members of the negative group also showed a larger peer orientation, i.e., a tendency to seek popularity and to do what they thought peers required. This peer orientation seems to remain six years later. Forty-six percent in the negative group against 28 % in the positive state, reported, for instance, that they 'pretty often' or 'nearly always' turn to friends for advice. How can the differences between the two groups be explained? The only thing we can do in a short paper like this is to point to some factors that may be of importance. Let us first take a look at some background

Social interaction and personality development

637

data for the two groups. The negative group consists of more males than females, while the distribution is even in the positive group. The negative group is overrepresented by young people from social group 2, which is that part of the middle class where parents have more than elementary education. But there is also an overrepresentation from the working class and an underrepresentation from the highest socioeconomic strata. In grade 8 a somewhat higher proportion of this group was directed towards continuous theoretic studies, and the plans were realized. The average intelligence was somewhat higher in the negative group, as well as school achievements. About 50 % of each group were working when they answered the mail enquiry. From test results we find that the negative group is not a negatively selected group. On the contrary, the intellectual capacity is a little higher and the received education is also a little higher than in the positive group. On the other hand, it is not an especially homogenous group. It consists of people from all social classes and even if intellectual capacity and school achievement are above average, the variation is not restricted, so the group contains people with varying capacity. Both students and workers are also represented. Parent-child relations during adolescence. I now turn to the young people's family structure and the parent-child relations during adolescence. We asked the respondents some questions about these relations during their preadolescence and early adolescence (age 10-15). There are difficulties in answering such questions in retrospect, and one must be careful in interpreting the results. In spite of this it may be well to take a look at them. They are shown in Table 1. Only the mother-child relations are shown. The father-child relations show the same pattern, but the differences between the positive and negative group are somewhat larger. The table shows that the mother-father relations differ between the two groups. The negative group had more often a matriarchate or patriarchate, while in the positive group both parents took part in the decision process. In the negative group the parents used a more authoritarian or permissive authority than in the positive group, where a democratic style dominated. In the last group parents also explained their decisions more often than in the negative group. We can also see that reliance and affective relations during adolescence were better in the positive than in the negative group. The family structures also differ. More parents were separated in the negative group, and more mothers had a job outside the home during both preschool and school period in the negative group. The members of this

638

Interaction in social

groups

group also have a feeling that the parents were away from home more often than in the positive group. Table 1. Parent-child

relations

and family

structure

during

adolescence*

% Negative % Positive

Internal structure Father and mother decided as often

26

47

Parental authority Authoritarian Democratic Permissive

43 33 20

22 66 9

Communication Mother explained decisions 'nearly always' or 'pretty often' 61

86

Reliance Asked mother for advice when needed 'nearly always' or 'pretty often' 51

72

Affective relations Felt 'quite' or 'very' warm for mother 76 Did 'quite a lot' or 'nearly everything' that was fun together with mother 18

95 42

Family structure Parents living together (at present) Parents separated (at present) Mother worked during preschool period Mother worked during school period Mother 'very' or 'pretty' often away from home during adolescence

74 18 37 58

85 10 23 48

18

8

Contact (at present) Living together with parents Living together with partner

70 10

71 17

* All differences are significant at least at the 5%-level. Thus there seem to be large differences in family structure, upbringing pattern and parent-child relations between our two groups during childhood and adolescence, and these differences may very well be of importance for the origin of the perceived oppositions between parents and children which we can find in the negative group. Many members of this group have experienced an upbringing characterized by authority, one parent dominating over the other, bad emotional contact with parents, and also bad communication and little common interests with parents. As an

Social interaction and personality development

639

effect we get individuals with low self-evaluation who seek recognition and status where they can get it, namely, among the peers. They protest against the parents and strive for autonomy and the protests are so strong and the interactions so blocked that the effects can be found even when the individuals have become older and stand on the threshold of adulthood.

640 Interaction in social groups F. AFFLUENCE, RECIPROCITY, AND SOLIDARY BONDS* by Irving Rosow, University of California, San Francisco

Since the Second World War, growing productivity and affluence have raised living standards in the United States and other developed nations. They have also created severe difficulties in pollution, traffic, waste disposal, urban congestion and decay, ticky-tacky suburbs, and many elements of a plastic life. Such obvious problems blur other profound, but subtler, changes in people's relation to society. This paper examines one neglected process that affects social integration: The effects of affluence on people's mutual dependence and solidary ties. This also has implications for social personality and social problems. The basic thesis is reasonably clear: Prosperity stifles symbiosis. Affluence creates personal opportunities that reduce people's dependence on family, kin, and friends. Theirin dependence of the group weakens the constraints of mutuality. This undermines the solidarity that depends not only on sentiments, but on the web of affiliation that binds people into a system of reciprocity. In most societies, particularly the impoverished, economic limitations made the individual extremely dependent on the group for his survival and welfare, his integration, and the realization of his goals. His objectives both the humdrum and the unusual - were often unattainable without the active support and cooperation of others. This was usually organized through kinship units, collaborative work groups, and other institutions of mutual aid. Cooperation was governed by a system of mutual rights and obligations that could be withheld from those who did not observe reciprocity (Gouldner 1960) or other valued beliefs and conventions. The point can be repeatedly documented, almost ad lib, from the vast anthropological literature. In a sensitive study of a Philippine village, the authors emphasize local interdependence in various contexts (Nydegger 1966): 'The community of Tarong is functionally dependent on an extensive although not formally structured system of mutual obligations, very few of which a member of the community can avoid. "How can a man live if he does not have neighbors to help him?" ... The answer is, of course, "He cannot. Therefore, he must get along with and help his neighbors or * I am indebted to Corinne Nydegger for her valuable review of this paper.

Social interaction and personality development

641

they will not 'see' him when he is in need" ... The way to strengthen and extend kin-neighbor bonds is to create obligations by working on behalf of the others, whoever they may be ... Each member of the community is obligated to offer such help and support to all others when they have need of it. 'Within the kin-neighbor group, restraint must be exercised in the realization of all goals on pain of withdrawal of support, respect and affection ... The [older adolescent] boy himself is therefore required to express his anger in a suitably adult way out of respect for his parents and, probably more effective, fear of loss of their economic and social support... Such restraint, needless to say, is modified within and extended only as far beyond one's group as expediency suggests ... [As for those whose ambition makes them heedless of others,] "the one who runs fast when stepping on the thorn is pierced deepest." And it inevitably pierces most deeply the one who steps on other Tarongans as he runs, for thorny indeed they can be. 'The strong dependence of Tarongans on one another at the local level makes such social control not only possible but very effective. The threat of ostracism and the loss of local support are far more powerful than the fear of a jail sentence.' In other words, prudence was most necessary within the sphere of mutual dependence, local support, and immediate social control. Assistance is contingent upon conformity to and respect for the rights and wishes of others. Insofar as an individual is dependent on the group, his freedom is restricted. He not only has to consider his own needs, but also those of his associates and their views of the propriety of his action. If he wants their cooperation, he must cooperate in turn; if they disapprove of his conduct, he cannot count on their support when he needs it. Clearly such cooperation often involves control of property and other resources, sponsorship, and the access to opportunities. The person is bound into a symbiotic system that promotes his welfare and even survival. The security and support assure the conformity of most group members. At the same time, the group restricts his deviance and limits those options that would remove him or are otherwise deemed inappropriate for him. Insofar as the individual is dependent, control over his behavior is centered in the group, and can be invoked with power. Several caveats are immediately in order. First, this anaysis in no way implies a strong conflict between group and individual interests, an inevitable coercion of the member and bending him to the group against

642

Interaction in social groups

his will. For example, 'as is true throughout the mutual obligation system, these obligations are willingly accepted in no small part for the pleasure inherent in the friendly social relations they entail' (Nydegger 1966, p. 101). The typical pattern in most social life shows a reasonable accommodation between the person and his immediate affiliates.This tends to be true in cognitive, affective, and normative terms. Indeed, the more isolated and homogeneous the group, the more likely are its members to regard its power with little heed as a simple fact of social life, cognitively on a par with the climate or landscape. They may chafe under it on occasion, but it is part of the social environment. At the same time, this does not imply that the compliance of group members never entails any personal costs so dear to students of human development, culture and personality, or social psychiatry. Second, this is not a stark economic determinism. Exchange theory (Blau 1964) or some other symbiotic model would be more appropriate. We have spoken primarily of economic opportunities and resources, but the same principles govern the access to other social utilities, values, and goals that are central to the culture, for example, withholding consent to a disapproved marriage of one's child. The conditions of highest group dependence are clearly exemplified in Durkheim's concept of mechanical solidarity: tightly-knit, isolated communities with homogeneous values and virtual subsistence-level economies. But interdependence is not limited to impoverished, preindustrial cultures. It is contingent on more general variables that typify such societies, but are not peculiar to them. These factors tend to be governing and to have cumulative effects. While we cannot discuss them, we can list the major ones to illustrate their range. Dependence on the group is directly related to some and inversely to others (and we will indicate which). Accordingly, dependence is correlated with the following variables: (1) Available resources, both economic and social (inverse); (2) economic productivity (inverse); (3) life chances and opportunity (inverse); (4) power and control of resources (inverse); (5) age (curvilinear) and sex (inverse); (6) cooperative social organization (direct); (7) social identity based on ascription (vs. performance) (direct); (8) extreme deprivation (direct). The variables increase the individual's dependence on the group and its

Social interaction and personality development

643

power over him, whether benevolent or not. Such dependency limits personal autonomy and curbs individuation. These patterns are well understood for preliterate cultures with marginal economies. But they have not been analyzed for developed, prosperous nations. In this sense, the variables above specify an ideal type, the limiting case at the low end of the continuum of productivity. They provide a baseline for comparison with advanced societies for the effects of affluence on dependence and group integration. In other words, how do an expanding economy, high productivity, and reasonably stable prosperity affect mutual aid and solidarity? Does the proliferation of opportunity weaken the flow of reciprocity? Obviously, any economy expanding faster than the population generates many new positions. The faster the growth rate, the greater the likelihood of labor shortages, as in modern Germany and Japan. However, what is crucial for our problem is the particular form that these opportunities take, their development in the public sphere, whether individuals can take advantage of them without primary group support, and the effects on person-group solidarity when a member chooses to do so. Note that we are not concerned with collective action in which groups capitalize on new opportunities together, as is common in migration - viz., the immigration to America, the movement of Okies to California, southern Blacks to northern cities, or Indian villagers to new urban factories where village pioneers establish beachheads for the recruitment and reception of new workers. For under these conditions, group dependence is intensified. We are concerned solely with the problem of individuals breaking away from groups on an independent basis. Mutual aid, group control, and available alternatives may well be life-long issues. But they are particularly crucial for the point at which a young adult enters the labor force and the number of job levels he may eventually traverse (Blau and Duncan 1967). In modern society, this means how his adult occupational history starts, whether directly in full-time work or by first going to college. Access to both channels has opened up drastically in the United States during the last generation. We will document this briefly to cast into sharp relief the sources of these growing opportunities. (Unless otherwise indicated, all summary data on work and education in this section are drawn from the annual volumes of the Bureau of the Census, U.S. Statistical Abstract, with the year and table number of the volume given in parentheses.)

644

Interaction in social groups

Job opportunities stem from several sources, primarily the expansion of the economy and Federal spending. Immigration ceased to be a factor fifty years ago, and differential class fertility which is declining accounts for proportionally few vacancies in the occupational structure (Kahl 1957; Blau and Duncan 1967). Most circulation within the labor market reflects changes in the economy that absorb population growth in the process. In 1950, 67 % of the men in the labor force were in occupational categories higher than their fathers (Kahl 1957). From 1950-1970, 20 million new jobs, more than half of them for women, increased the labor force by one-third (32 %), but the proportion of all adults in the work force remained stable at 58 % (1970: 316). Population growth feeds this expansion in the provision of markets, but this would not be possible without corresponding growth in productivity, as India and other underdeveloped nations demonstrate. Yet American expansion has been sustained on this basis, with man-hour productivity between 1950-1970 increasing by 57 % (1970: 338) and the gross national product eventually exceeding one trillion dollars (1970: 506). The role of the Federal government in this growth cannot be overestimated, through tax and fiscal policy, subsidies and social programs, and direct consumption of goods and services, including vast defense expenditures. The primary forces at work since World War II also feature the impact of science and technology in the development of new enterprises, the consequent rise in skill requirements, the increasing demand for professional and managerial talent, and the trickle-down effect of vacancies as people leave positions to advance. The effect of science and technology in the direct creation of new jobs is tremendously significant. By 1950, 20 % of all men in the labor force were technologically mobile, occupying new jobs that had not even existed for their fathers (Kahl 1957). This innovation has not slowed in the past twenty years. Accordingly, economic growth, public spending, and technological innovation have increased job opportunities. For young people, the key to advancement lies in higher education. As we shall see, the Federal government has played an indispensable role in making higher education broadly accessible. Over the past forty years, higher education in the United States has undergone dramatic expansion. This is reflected in steadily rising enrollments. In 1940, there were 1.5 million persons working for college degrees, and this increased to 7.1 million by 1970 (1972: 154). The proportion of college-age youth, eighteen to twenty-one, who were attending school

Social interaction and personality development

645

grew from 16 % in 1940 to 43 % in 1968 (1970: 191). In the past thirty years, the number of college degrees each year has increased fivefold from 217,000 in 1940 to 1.1 million in 1970, with doctors' degrees alone rising from 3,000 to 30,000 annually (1972: 212). Thus, in the last generation, higher education has shifted from a privileged to a popular system - a movement also found in other developed nations. Much of this growth has been virtually underwritten by public funds. College and university income from public sources grew from 214 million dollars in 1940 to over eight billion by 1968 (1970:191), accounting for 45 % of total expenditures (1972: 157). The government has made massive investments in higher education which have met the growing needs of the country for trained personnel. The sheer growth, however, is not conclusive for our analysis. For higher personal incomes can be spent on education without reducing children's dependence on parents. The issue is whether public funding does increase their autonomy, whether the government has given youth independent access to college regardless of their families' circumstances or attitudes. Unfortunately, the evidence on this is not exhaustive. The statistics department of the Bureau of Higher Education in Washington informs me that they have tried to assemble accurate, complete figures on all forms of student support through the years, but this has been impossible. Programs are dispersed across many agencies, some of them such as the National Youth Administration of the 1930s being defunct. Records are highly variable, and those adequate for administration and disbursement are often unsuitable for research. For example, the Veterans' Administration bases its operation on personal files, and its summary statistics are rudimentary. Therefore, we cannot summarize total assistance or always distinguish among different types: NYA; veterans' benefits; stipends, traineeships, assistantships; student loans; positions concealed in research or teaching budgets; and so on. But despite these short-comings, some crucial data on overall trends are available. So we can at least examine evidence of opportunities for two key groups: (1) The disadvantaged and (2) veterans. The disadvantaged. Underprivileged youth in government programs pursued either vocational training or a college education. Between 19501970, the number of vocational trainees increased from 3.4 to 8.8 million annually (1970: 217). Surprisingly, this is even larger than the number of

646

Interaction in social groups

students in college, but obviously for much shorter, less intensive training programs. During this period, public funds for vocational training increased fifteen-fold from 129 million in 1950 to 1.8 billion dollars in 1970 (1970: 217), with a sharp rise after 1965 and the swelling stream of returning Vietnam veterans. The Federal share of funding fluctuated between 15-25 %, with state and local agencies accounting for the balance (1970: 217). This training was presumably consequential, for of those who completed programs between 1963-1970 under the Manpower Development and Training Act, 78 % found jobs (1970: 218). At the same time, minority group access to college has also improved. While their attendance has certainly not reached the level of whites, their rate of increase is currently faster. In the 1960-70 decade, white enrollment has about doubled ( + 117 %) while that of non-whites has increased more than twice as fast by + 259 % (1972: 162). From 19401970, the proportion of persons 25-29 with a college degree has grown from 5.9 % to 16.4 % among whites, but from 1.6 % to 7.3 % among Blacks (1971: 164). These trends paint no picture of social justice and equity, but do reflect opportunities arising from public sources. Veterans. The GI Bill (PL 346) and subsequent legislation on veterans' educational benefits have probably provided more entry to higher education than any other single factor. The annual figures for the last thirty years show that each war has been followed by a surge to the colleges as servicemen streamed home. After World War II, 1.8 million veterans were in school in 1947, dropping to 1.1 million by 1951 as many finished, bottoming out at less than half a million in 1953 before the Korean War furnished a second wave for the next five years. However, the immediacy of response to educational subsidies is dramatically shown by the expansion of the war in Vietnam after 1965. Just before this, in 1964, only 18,000 veterans were in college. Yet two years later, shortly after the expansion, this number had suddenly jumped to 294,000 by 1966 and climbed rapidly to 1.1 million vets in 1971 as more and more youth were pumped through the Vietnam horror. While details are not readily available specifically for Black veterans, the effects are quite apparent at predominantly Black colleges where total enrollment jumped from 106,000 in 1964 before the Vietnam expansion to 163,000 in 1968 and 180,000 in 1970 (1972: 204). Where Federal expenditure for veterans' education stood at 100 million dollars in 1965, it rose to 591 million in 1967 and 2.2 billion by 1972 (1972: 221).

Social interaction and personality development

647

All this, of course, was in addition to numerous other programs that were not specifically earmarked for veterans, such as training grants, fellowships, student loans, and private scholarships, many of which quadrupled after 1960. Federal benefits were supplemented by state and local agencies, and those of universities by activities of private foundations. However, the Federal government was the prime mover whose programs were complemented by the other institutions. An objection might be raised that higher education actually prolongs the period of youthful dependency. But this misses the point. Obviously, it ignores the fact that many students would not even be able to attend college without governmental programs. For others, educational benefits significantly increase their independence. Both groups are substantial in size. They can take advantage of opportunities that they would otherwise not have or on terms that increase their autonomy. But aside from these probabilities, there is also a conceptual confusion. Higher education may defer a student's reaching full adult responsibility or his social maturity. But this is quite different from sheer financial dependence on his family. In underdeveloped cultures, people grow to adulthood from a state of immaturity and dependence (-) to one of maturity, but still with group dependence ( + -). Our model injects a transitional pattern for college students who are not yet socially mature, but are independent of the family (—|-). Hence, the sheer fact that their social adulthood may be deferred does not mean that they are fully or even significantly dependent on their families. We have concentrated on training and educational opportunities because they are crucial, but they do not exhaust the full extent of public programs that significantly increase life chances. These are extremely diverse, ranging from social security and medicare to agricultural subsidies, small business and low-interest mortgage loans (or guarantees), unemployment insurance, food stamps, school lunch programs, VA hospitals and a gamut of other social services. Accordingly, then, the affluent society has created many new opportunities for its members, directly in the expansion of the economy and indirectly through various governmental services and programs, primarily those in education and training that raise people's occupational qualifications. Public agencies were the major instruments for reducing personal dependence on primary groups and they did so to a remarkable extent. (It is too early to tell whether the Nixon cutbacks are a temporary or stable reversal of recent Federal policy.)

648

Interaction in social groups

Now, what happens to the web of affiliation and control over the person when opportunities develop outside his group and his life chances are no longer governed by his compliance to group pressures? How does this affect the relationship of the individual to the group - particularly to family, kin and friends? This is the heart of the problem. In contemporary terms, the person-group relationship is commonly played out in the form of social mobility which remains a powerful value in American life. This generates strong pressures that attenuate the ties to one's origins and intensify the alienating forces. In the pursuit of his goals, the individual has significantly less reliance on the group. Their opinions are no longer particularly consequential for his fate and his complaisance is no longer so necessary. His self-interest and self-orientation grow at the expense of group concerns. Consequently, the obligations he feels are weakened and the network of reciprocity and mutual claims undermined. Accordingly, he is likely to be marginal to or outside the sphere of group solidarity. As the group's control over him wanes, so do its claims on his deference and the force of its norms and principles of authority. This sketches a pattern of tenuous group ties and constraints on its marginal members. Thus, social mobility which is rooted in expanding opportunities makes the person independent of the group and its controls, undermining reciprocity and solidarity. But the stress on social mobility also erodes group ties in other ways that intensify these effects. These are proportional to the social distance that the mobile person has moved from his group of origin. While social mobility often entails geographic mobility, geographic mobility alone is probably less alienating. A person's mobility removes him from the group. This results in less interaction and contact which, as Homans (1950) has shown, tends to reduce common sentiments. Mobility also exposes the person to new sets of significant others who may now influence his life chances. As they become salient, they signify a major shift in reference group whose strength depends on the range of new options open to the person and their control over these. In other words, an adaptation of the member-group problem in a new context. Mobile people are also exposed, sometimes with compelling force, to new values that may intensify any incipient alienation from their social origins. Indeed, it is axiomatic that the mobile person acquires beliefs that are intermediate to those of his initial and terminal classes (Blau 1956; Inkeles and Bauer 1959). The possibility of value conflict also contains the seeds of role conflict because the class values are

Social interaction and personality development

649

embodied in different sets of people who claim the person's obligation minimally the families of orientation and procreation and their surrounding groups. So he is commonly forced to choose between them. Such choices are often mutually exclusive and not easily balanced out in the long run. Any consistency of bias in choice tends to feed on itself, with commitment in one direction pulling the person steadily farther from the other group, but not necessarily reducing his conflict or strain. Further, mobility indicates the differential life chances that separate the mobile person from those remaining at his level of origin. His son may go to college while his nephew may not. He is significantly more advantaged than the others in almost all sociological respects. They no longer face and share a common fate, nor enjoy the empathy and solidarity of those that do. Their life experience tends to diverge sharply in direction and substance so that they come to share less and less. Consequently, they risk a genuine conflict of economic and political interests. This carries over beyond attitudinal and sheer cognitive differences, as important as they are, to spheres of awareness, levels of aspiration, value priorities, and the very conception of the good life. Ultimately they involve major, even profound, differences of life style. People simply come to live by different standards for different values. Thus, the opportunities that originally draw the person out of his group have incipient social class implications. But when there is sufficient mobility, what starts out as class differences develops into the social status differences that held Max Weber's (1946) interest. This creates significant pressures in people's alienation from one another. The individual's tie to his original group may become weak, shallow, and attenuated. Blau (1956) argues that the effects of mobility on the person depend on how he relates to the two groups. Contact with members of his class of origin may yield gratifying deference to his success (though, we may add, perhaps limiting social access to his new associates). Interaction in his terminal class requires a drastic change in life style. But if he is integrated in neither group, this implies strong strains and insecurity. The crucial factor is that the person may choose between those in his past and his present (or future), just as one selects between other reference groups. However, structural pressures that are anchored in the present are typically the more powerful, for one's future fate is mainly governed by current rather than past actions and associations (Rosow 1969). This is particularly true for prospective status claims usually so important to the mobile. Hence, past constraints tend to be weak and obsolete.

650

Interaction in social groups

This certainly does not mean that all former ties disintegrate and completely disappear. For even in modern industrial society, the family remains a relatively viable unit in meeting major life crises (Litwak 1960a, 1960b; Litwak and Szelenyi 1959; Rosow 1967; Sussman 1959; Shanas and Streib 1965). Extreme family needs are seldom shucked off without some supportive response. However, crises tend to be unusual, and mutual aid in this context is quite different from integration in a network of workaday cooperation and continuous reciprocity. The basic issue is now more sharply focused: When structural contraints are weakened, mutual aid and reciprocity ultimately depend on emotional bonds and the voluntary acceptance of obligations. The crux of the matter is whether, in the absence of strong structural pressures, emotional ties alone are strong enough to maintain routine solidarity and reciprocal relationships. In other words, do people help each other as much from pure sentiment as from the combination of sentiment and necessity? The evidence on this remains mixed because research has not examined the relative strength of these two factors. As against the viability of the family in crisis, there is other scattered evidence of a different kind. Successful Blacks who have 'made it' are under constant overt pressure to fight for their brothers in the ghetto, to use their new influence on their behalf. While adult children assume varying responsibility for the financial and other needs of older parents, this is more common among poorer groups in which mutual dependence is particuarly high (Dinkle 1944; Kosa, Rachiele and Schommer 1960; Shanas 1967). The steady shift of responsibility for aged parents from their children to the state has been stimulated by the courts in rulings that establish the prior claims of children's self-advancement over the needs of parents (Shorr 1960). This is epitomized in a recent study of a retirement apartment building tenanted primarily by widows on welfare (Hochschild 1973): 'The residents, like most Americans over 65, were poor ... All of them were living on welfare except for a few who received social security and pensions ... Their granddaughters and daughters wore bouffant hairdos and drove up to visit them in fairly new, large American cars.' Further, when older people can no longer perform vital functions for their children's family - i.e., sustain reciprocity - they may be excluded and cut off from further contact with them (Townsend 1957; Pagani n.d.). Other reports indicate positive animus between adult children and parents, with the younger generation virtually abandoning and ignoring the older (Shorr 1966). With each increase in social security benefits, an increment of

Social interaction and personality development

651

parents who live with grown children moves out to live alone. And Townsend (1957) reports in his London study that the relations between socially mobile children and their parents are fraught with strain and tension, the status differences clearly attenuating the personal relationships. Even some of the high divorce rate may be predicated on increased income that allows people to leave stituations that they previously had to abide. These data are not conclusive. But they show that the common view of filial devotion needs qualification, particularly of the effect of social distance on the relations between a group and its former members. More important, the nuclear family is a special case whose internal relationships cannot simply be generalized to kin, friends, and other groups (Rosow 1967). The question remains whether sheer emotional ties can maintain as much mutual aid and solidarity as emotional ties plus mutual dependence. We cannot settle this issue definitively. But we have outlined the alienating pressures that may be expected from affluence and the breakdown of mutual dependence. When they are not reinforced by structural imperatives, emotional bonds may well be tenuous, vulnerable, and unstable. They are contingent on goodwill and the assumption of personal responsiblity when many competing pressures may sap their strength. Therefore, in human reciprocity, love alone is probably not as compelling and stable as situational necessity. Nor are virtue and character reliable substitutes for structural constraints. Accordingly, as opportunities proliferate in affluent society, personal gain may well be inimical to group cohesion and mutuality. The remaining issues are quantitative. First, how much do independent opportunities erode symbiosis? What is the strength of the major variables involved and what is their relative effect on solidarity and reciprocity? Second, what is the optimal level of personal alternatives before additional ones become dysfunctional? For a plethora of choice (as in supermarket trauma) can become indigestible and drain off energy from other worthwhile claims. How much overload can a person absorb before cognitive complexity and social cohesion become undue strains? Third, what are the cost-benefits in the exchange of social values - to the individual, to his group, and to the larger society in the integration of its members and institutions? The general problem also has implications for social personality, not only in cognitive complexity or the premium on autonomy and initiative

652

Interaction in social groups

as functional capacities, but in subtler fashion, the reformulation of human rights and distributive values, whether personal aspirations and competition increase at accelerating rates, the changing forms and limitations set on competition, the sheer extent of residual dependency needs, the institutional outlets for their satisfaction, the degree and depth of affiliation and group embeddedness, the quality of personal relationships, the extent to which primary group ties may shrink in favor of more impersonal institutional linkages, and the effect of self-interest on diverse social allegiances and responsibilities - in other words, the social psychological problems and possible adaptations to them that affluence might generate in the wake of massive institutional change. These are also consequential for our deepening social problems whose solutions ultimately will require a balance between opportunity for individuals and their sense of social responsiblity. Opportunities always mediate the relations of a member to any group. But the pursuit of selfinterest at the expense of the group may become unbalanced and vitiate the moral claims of any collectivity on the responsibility of its members. What was formerly so clear with primary groups has now pervaded major American institutions and those of other modern nations as well. Indeed, our vaunted economic growth rate has exacted a heavy penalty in our weakened sense of social responsibility. What is at stake is the adjudication of private and public claims, the balancing of private and public allegiances, and the very integration of society. We are not simply playing Cassandra or delineating a crude specter of Orwell's 1984, but raising one major problem about the direction of social change and the correlates of prosperity - not as the absolute imperative of man's fate, but as a set of strains that are implicit in an affluent mass society.

A R N O L D J. S A M E R O F F

2

Variations in infant-caretaker interactions

A. FAMILY INTERACTION IN THE NEWBORN PERIOD: SOME FINDINGS, SOME OBSERVATIONS, AND SOME UNRESOLVED ISSUES by Ross D. Parke,* Fels Research Institute, Yellow Springs, and Sandra E. O'Leary, University of Wisconsin, Madison

Mother-newborn interaction has received increased attention in recent years (Thoman, Turner, Leiderman and Barnett 1970; Richards and Bernel 1971); research on the father's role in the newborn period, however, remains relatively unexplored. Although the father is often recognized in theoretical discussions (Nash 1965), little empirical data is available concerning the father's behavior toward his infant in the early days of life. In part, this is due to the unavailability of fathers for study, but the paucity of information of father-infant relationships is also due to the secondary position assigned fathers by both the culture and psychological theorists (Bowlby 1951). Research that has been executed has either been based on maternal reports (Pedersen and Robson 1969) or * The first study was carried out at University Hospital, Wisconsin; thanks to Steven West for assistance in data analysis. The second study was executed in Cincinnati, Ohio; thanks for assistance in various phases of the research to Elizabeth Bodde, Sue Dimiceli, Frances Hall, Keri Lord, Alice Rudolph, and Lynn Woodhouse. A special note of thanks is extended to Laurine L. Cochran, R.N., Nursing Supervisor, Newborn Services and James M. Sutherland, M.D., Director, Newborn Division, for their cooperation. Finally gratitude is expressed to F. Falkner, M.D., Director, The Fels Research Institute and The Fels Fund of Philadelphia for their support of this research project.

654

Interaction in social groups

on very limited samples of behaviors such as Rebelsky and Hanks (1970) who tracked only paternal verbalizations. While both studies suggest that father plays a qualitatively and quantitatively different role in early infancy, the exact nature of the interaction remains to be detailed. Moreover, no comparative data exists concerning maternal and paternal behavior toward their newborn infant. To explore the manner in which the father interacts with his newborn infant and to compare paternal and maternal interaction patterns are the principal aims of the present investigations. In contrast to earlier studies, a direct observational approach was employed that permitted a detailed specification of father behaviors in the presence of the newborn. Three sets of observations were made. Since the nuclear family setting is often the context of early infant-father encounters, both mother and father were present during one observation session; this triadicarrangement permitted a comparison of the manner in which the two parents differ in their interactions when they are together with their newborn child. Second, observations of mother-alone with the newborn infant were secured in the first study. A third set of observations of father-alone with his infant was included in the second investigation in this series. This set of three types of observations permitted an examination of 'second-order effects', which Bronfenbrenner (1973) has defined as the impact of the presence of a third party on the pattern of interaction between two individuals. Specifically, we examined the modifying impact of the father's presence on mother-infant interaction by comparing the behavior of the mother alone with her infant and her pattern of interaction in the presence of father. Conversely, by comparing the interaction of the father's behavior when he was alone and in the presence of the mother, the modifying impact of the mother on paternal interaction patterns could be investigated. A second aim is to explore the modifying impact of two factors on parent-infant interaction - sex and ordinal position of the infant. First what are the effects of the sex of the newborn infant on mother and father interaction patterns? While girls tend to smile more in the newborn period (Freedman 1971) and three-month-old boys tend to cry more and be more irritable (Moss 1967), there is a paucity of information concerning the impact of these behavioral differences on early parent-infant interaction. Thoman, Leiderman and Olson (1972) recently reported that primiparous mothers talk and smile more at female than male infants. Similarly, Rebelsky and Hanks (1971) report that fathers of female infants verbalized more than did fathers of male infants at two and four weeks

Variations in infant-caretaker interactions

655

of age. Moss (1967), on the other hand, found that mothers of threeweek-old male infants vocalized more than mothers of female infants. Research involving direct comparisons of maternal and paternal behaviors with male and female newborn infants is clearly necessary to determine the nature of parental responsiveness to babies of different sexes. The present study will examine this issue. The final issue concerned the effects of ordinal position of the infant on parent-newborn interaction patterns. Recent research (Thoman, Barnett and Leiderman 1971; Thoman, Leiderman and Olson 1972) has demonstrated that primiparous mothers stimulate, smile, and talk to their infants more than muciparous mothers. An aim of the present study was to examine the impact of infant ordinal position on both maternal and paternal behaviors. Study I. The first investigation was conducted at University Hospital Madison, Wisconsin; the sample was middle class and well educated. Nineteen Caucasian couples and their first-born infants served as subjects. Mothers ranged in age from 19 to 30 years, while fathers ranged between 20 and 38 years of age. With one exception, the fathers were present during both labor and delivery; half of the couples had attended Lamaze childbirth classes. The infant sample consisted of 19 full-term single-birth infants - 9 males and 10 females. All of the infants were judged by the attending physician to be normal newborns without complications. With the exception of one Caesarian section, all births were spontaneous delivery or by use of low forceps, vertex presentation. Thirteen infants were breast-fed and seven were bottle-fed. Two sets of observations were made: (a) mother-father-infant and (6) mother-infant alone. The purpose of the study was introduced as 'how babies develop social behavior - how they act with their mothers and fathers'. The reason for presenting the infant as the main target of observation was to reduce parental anxiety concerning the observations. All observations took place in the mother's hospital room between 6 and 48 hours after delivery. For the mother-only sessions, the infant was placed in the mother's arms prior to the commencement of the observations. In the case of the mother-father-infant sessions, the investigator was brought to the mother's room, and the observer asked 'Whom shall I give the baby to?' The infant was then handed to the parent who indicated a preference to hold the child. A time sampling observational procedure was used. A ten-minute

656 Interaction in social groups observation period was divided into 40, 15-second intervals, and for each 15-second interval the observer recorded the occurrence of a parental or infant behavior. The following infant behaviors were recorded: cry, vocalize, move, mouth movements with or without object, look at mother, look at father, and look around. For both mother and father, the following behaviors were recorded: looks, smiles, vocalize, holds, kisses, touch, imitate, explores, feeds, and hands over to the other parent. Each family was observed on a number of occasions across the two post-partum days. The number of observations varied slightly across family, with the average number of mother-infant observations being three; there were an average of two mother-father-infant observations. Inspection revealed that there were no day effects and, consequently, the means of all sessions were combined for mother-infant and mother-father-infant observations. This yielded two scores for each family. The main observer was a 30-year-old female nurse; for reliability estimates a male of the same age was present. For purposes of training the observers, 8 mm. films of mother-father-infant interaction were made in the same hospital room setting that was used for the regular observations. Prior to field observations, the two observers achieved a percentage agreement between 88 and 100 % on all categories of infant, maternal, and paternal behaviors. In the hospital room the two observers watched mother-father-infant groups until a level of reliability comparable to the level achieved in the film training period was reached. Both observers were present for six observation sessions of mother-father-infant interaction and for eight mother-infant sessions. Inter-rater reliability was assessed by calculating the percentage of occasions across the 40 15-second time intervals that the observers agreed with each other. The mean percentage reliabilities for the mother-infant and mother-father-infant sessions ranged between 86 and 100 % agreement between the two observers. Results. In Table 1, the frequency (mean number of 15 second intervals) of maternal and paternal behaviors directed toward the infant are presented. Of interest is the fact that the father is a very active participant in the family triad. Subsequent analyses of variance indicated only one significant effect: mothers smile more than fathers (F = 5.14; df = 1, 18; p < .05). Two other effects were of borderline significance. Fathers tended to hold the infant more than mothers (F = 3.27; p < .09) and rock the baby in arms more than his spouse (F=3.18 ;p < .09). On all other measures, father was just as likely to interact with the baby as the mother.

Variations in infant-caretaker interactions

657

Since observations were made on both the mother alone and in the presence of the father, the effect of father presence on maternal-infant interaction was evaluated. The means for mother's behavior alone and in the presence of the father are presented in Table 1. The presence of the father reduced the amount of interaction between mother and baby. Mother was less likely to hold (F = 22.22; df = 1,17; p < .001), change position {F = 18.33; p < .001), rock (F = 14.21; p < .001), touch (F = 49.36; p < .001) or vocalize (F= 32.12;/? < .001) when father was present. Analysis of variance was used to assess the effects of sex of infant on parental interaction. Only one sex difference was present: both mothers and fathers touched male babies significantly more than female infants (F = 8.24; df — 1,17, p < .01). However, this sex difference was not present during the sessions when mother and baby were alone. Table 1. Mean frequency of parent behaviour (middle-class sample) Father & mother present

Hold arms Hand over Change position Look Vocalize Smile Rock Touch Explore Kiss Imitate Feed

Father

Mother

23.1 .3 3.5 38.7 5.2 8.6 3.1 13.7 1.0 .1 .7 2.1

12.7 .4 1.7 37.0 3.7 14.3 .6 12.3 .7 0.0 .3 7.1

Mother alone

31.4 0.0 5.5 37.7 12.0 11.3 4.6 26.9 1.2 .1 .4 12.7

Study II. A more recent and extensive investigation of these issues extends the previous study in a variety of ways. First, observations of fatherinfant interaction, as well as mother-infant and mother-father-infant, were included in this second study. This permitted a comparison of father's interaction pattern with his infant under two conditions: alone with the infant and in the presence of the mother. Possibly, the high degree of paternal-infant interaction observed in the initial study was due to the supporting presence of the mother. Moreover, the fathers in the original study were unique in other ways

658

Interaction in social groups

that may have contributed to their high degree of interaction with their infant. Over half of the fathers had attended Lamaze childbirth classes and with one exception all fathers were present during the delivery of the child. Both of these factors are likely to have increased the father's later involvement with their infants. Finally, these fathers were well educated and middle class, and their high degree of involvement may be unique to middle-class groups; parental involvement may be less in lower-class samples due to a more rigid definition of parental roles among lower-class parents (Bronfenbrenner 1958; Bronson, Katten and Livson 1959). To overcome the sample limitations of the original study, a group of lower-class fathers who neither participated in childbirth classes nor were present during delivery were observed. This study permitted a much more stringent test of father-infant involvement and permitted wider generalization of the previous findings. A final purpose was to examine parental interaction patterns with first and later-born infants of both sexes. If fathers and mothers do, in fact, differ in their interactions with their newborn, it is likely that these parental differences will be marked with the first-born child. The sample was drawn from a large metropolitan general hospital in Cincinnati, Ohio. To date fifty-one white and thirty-one Black families of lower socioeconomic status have participated. The age range for mothers was 15 to 43 years (X 20.7), while fathers ranged in age from 17 to 47 years (X 23.6). The infant sample consisted of 48 girls and 34 boys; 17 boys and 17 girls were first born, while the remaining boys and girls were later born. Apgar scores for all infants were 8 or above at one minute; all infants were bottle fed. No fathers were present during delivery; fathers are routinely not permitted to handle their infants during the post-partum period. Special exemption from these restrictions made possible the father-infant observations. All contacts were made in the hospital, within the first 48 hours, and the study was introduced as 'how babies develop social behavior - that is how they act with their mothers and fathers'. Three types of observations were made: {a) mother-infant, (b) mother-father-infant, and(c) father-infant. The observations involving mother (mother-infant and mother-father-infant) took place in the mother's hospital room. The fatherinfant observations were made in a room near the newborn nursery, which was furnished with a comfortable chair. To increase the range of parental behaviors, the parents were informed that they could either pick up the baby or leave him in the crib. (In the earlier study, the baby was handed

Variations in infant-caretaker interactions

659

to the mother or whichever parent requested the baby in the case of the mother-father observations). A total of 82 families participated in the mother-father-infant session, while approximately half of these fathers and half of the mothers in each family were seen on a second occasion alone with their babies. In the case of repeat observations, an attempt was made to counter-balance order of the mother-father-infant observations and the observations in which the parent was alone with their infant; inspection of the data revealed no order effects. All observations were 10 minutes; this period was divided into 40, 15-second intervals. The same infant and parental behaviors were observed as in the first study. Inter-observer reliability was established prior to the commencement of the study and ranged between 85% and 99% for infant and parent behaviors. Reliability was assessed by a second observer on 12 occasions for the mother-infant sessions, on 8 occasions for the father-infant sissions, and on 14 occasions for the mother-father-infant sessions, during the progress of the study and prior reliability levels were maintained. Table 2. Mean frequency of parent behaviour (lower-class sample) Mother & father present

Hold arms Hold lap Hand over Change position Look Vocalize Smile Rock Walk Touch Explore Kiss Imitate Feed

Mother

Father

7.7 2.7 .7 3.6 38.7 4.0 9.2 1.6 .2 9.2 2.6 .2 .0 2.9

21.6 4.1 .71 8.1 39.3 11.9 7.7 5.6 .4 12.3 3.3 .3 .2 8.2

Mother alone

Father alone

25.8 5.9

25.8 7.2

—

—

13.3 38.0 10.4 3.7 7.4 .3 17.7 1.0 .4 .1 17.0

11.3 38.6 12.8 4.8 13.7 .9 17.6 1.8 .2 1.0 9.8

Results. First, the frequency (mean number of 15-second intervals) of maternal and paternal behaviors directed toward the infant when they are together and alone with the infant are depicted in Table 2 (N = 82). As in the earlier study, the father is a very active participant; analyses of variance revealed that father is significantly more likely than mother

660 Interaction in social groups to hold and visually attend to the infant and to provide physical and auditory stimulation (all p < .01). Only in smiling does the mother outdistance the father (p < .05). However, is the father's active involvement limited to the situation in which the mother is also present? To answer the question, we compared the father alone with the infant and the father in the presence of the mother (N = 44). The mother's presence was clearly an unnecessary support for the father's active involvement. In fact, he was significantly more likely to touch and rock his infant when alone than with the mother (all p < .01). In general, he was an equally active participant in both settings. The presence of the mother had one positive effect: The father smiled more in her presence than alone (p < .05). To examine how mother's behavior shifted across social contexts, we next examined her behavior alone and in the presence of the father (N = 55). Mother-infant interaction is much higher when the father is not present. Analyses of variance indicated that mother was significantly less likely to hold, touch, rock, vocalize to, imitate, and feed their ofspring when father was present (all p < .01). However, mother was more likely to explore the infant and smile at baby when the father was present than when she was alone with her newborn (p < .05). Perhaps the most important comparison involves mother and father alone with their infant (N = 42). As Table 2 clearly indicates, when they are alone, fathers and mothers differed only slightly in their patterns of interaction; mothers fed the baby more frequently than father (p < .01). Sex and ordinal position effects. However, there were some sex and ordinal position differences that merit consideration. When mother and father are together, parents tend to hold first-born infants in their arms (F = 2.80; df = 1,78; p < .10; first born X = 16.06 vs. later born X = 13.67), while they hold later borns on their laps (F = 4.41; p < .01; first born X = 2.04 vs. later born X = 4.31). Parents are more likely to walk with the first born {X = .43) than a later born (X = .20) infant (F = 7.39), particularly a first-born boy (X = .50 vs. 106 for first and later-born boys respectively). Parents walked girls equally, regardless of ordinal position (X = .35 and .27 for first and later-born females). Finally, fathers touched first borns (X = 14.06) more than later borns (X — 10.98), while mothers tended to touch later-born infants ( X = 9.92) slightly more than first-born infants (X = 8.15) (F = 5.19; p < .05). The analysis involving mother alone and in the presence of the father indicated that mothers rocked first-born babies {X — 4.85) more than

Variations in infant-caretaker interactions

661

later-born (X = 3.46) infants (F = 4.14; df = 1,47; p < .05) and boys (X = 5.05) more than girls (X = 3.24; F = 4.45;p < .05). However, these ordinal position differences in maternal behavior varied with the presence or absence of the father. When the father was present the mother rocked first-born (X = 1.40) and later-born (X =1.17) infants equally (F = 8.35; p < .01). On the other hand, mothers tended to touch first-born girls (X = 22.78) more than first-born boys (X = 14.45); again the presence of the father eliminated these differences (F = 3.81; p < .07). A similar trend was present for vocalization. The father analysis (father alone vs. in the presence of mother) revealed sex X ordinal position interactions for touching (F = 4.42; p < .05) and vocalizing (F = 4.60;p < .05). Father, regardless if alone or with mother, touched first-born boys (X = 19.87) more than either later-born boys (X = 12.38) or girls of either ordinal position (X = 15.63 and 16.00 for first- and later-born girls). Father vocalized more to first-born boys (X= 16.56) than to first-born girls (X = 8.94); while he vocalized equally to later-born infants irrespective of sex (X = 13.27 and X = 13.33 for boys and girls respectively). In summary, both sex and ordinal position are important modifying variables in early parent-infant interaction. Implications. First, the father plays a more active role in early social interaction than previous research suggests (Rebelsky and Hanks 1971; Pederson and Robson 1969). Moreover, this high degree of involvement on the part of fathers is not restricted to middle-class highly educated groups, nor is the presence of the mother a necessary setting condition. However, there are other contextual factors that must be taken into account. Most importantly, the restraints imposed by the hospital setting probably affected the amount of father interaction and may account, in part, for the level of father interaction. The home situation, on the other hand, offers more freedom concerning how and when a father will interact with his infant. Whether the frequency of father-child interaction in the hospital setting is predictive of father-infant behavior in other contexts remains to be determined. Follow-up studies are currently being executed in order to determine whether or not the amount and/or the patterns of parental-infant interaction during the newborn period are of predictive value for later behavior. Another related issue merits consideration. At present, there is considerable controversy concerning the importance of the opportunity for early contact between parents and infants for later parent-child relationships;

662 Interaction in social groups this controversy, of course, has important implications for hospital caretaking arrangements and visiting schedules. Recent research (Klaus, Jerauld, Kreger, McAlpine, Steffa and Kennell 1972) has indicated that mothers who were given extended contact with their infants over the first three post-delivery days engaged in more soothing, eye-to-eye contact and fondling at one month than mothers who were given only feeding contact with their infants during this period. Whether or not a similar 'early exposure' effect occurs for fathers as well merits examination. Opportunities for early interaction may, in fact, be particularly important for the father, who, unlike the mother in the post-partum period, may not be biologically or culturally primed to be responsive to infant cues. The father may require longer exposure to the infant and/or exposure to a larger number of infant cues to elicit the same degree of responsiveness as the mother. The threshold for paternal and maternal responsivity to infants may, in fact, be different. However, more attention should be paid to the differential effects of father-infant contact in the presence of the mother and father-infant contact alone. Possibly, the opportunity for interaction in the nuclear family triad is associated with increased sharing of responsibility for infant caretaking in the post-hospital period. This type of interaction will probably be particularly important for new fathers; the mother in the situation can serve as a model and as a direct instructor for teaching the father simple caretaking and handling skills. In addition, this provides a supportive context for the father to practice these behaviors. On another level, increased attention needs to be paid by hospital personnel to a family orientation to newborn care, whereby the father is routinely included in feeding and caretaking demonstrations. Particularly, in light of shifting parental roles that are redefining the appropriateness of father involvement in infant care, providing the father with opportunities to both learn and practice caretaking skills during the newborn period will not only make it more likely that he will share these responsibilities, but that he will execute these tasks effectively and view these behaviors as role-consistent. Both sex and ordinal position of the infant are important determinants of parent-infant interaction - even in the first days of life. While some of our findings are consistent with the results of Thoman et al. (1972) concerning maternal involvement with first-born girls, other results were inconsistent to this pattern. Moreover, fathers stimulated first-born boys more than children of other sex or ordinal position. In combination with

Variations in infant-caretaker interactions

663

prior research, this pattern of results lends some support to Rothbart's (1973) hypothesis, which suggests that parents provide maximal stimulation for their same-sexed offspring. However, situational parameters such as the presence of the other parent can significantly modify these ordinal position-sex relationships. Our findings indicate that mother's differential responsivity to first- and later-born infants is eliminated in the presence of the father; the father's involvement with his first-born son, however, tends to be less readily modifiable by the presence of his spouse. These results leave unanswered a central question: What is the contribution of prior culturally determined parental expectations and experiences and the impact of different infant behavior patterns associated with sex and ordinal position? Combinations of parental interviews and independent behavioral assessments (cf. Osofsky and Danzger 1973) of the infant outside the interaction context will be necessary in order to disentangle these effects. Moreover, it is not only stable infant characteristics, such as ordinal position and sex, that are important, but shortterm changes in infant behavior induced by maternal medication and labor may also affect parent-infant interaction patterns (Parke, O'Leary and West 1972). In any case, it is clear that an understanding of socialization in the newborn period requires a bi-directional model which gives explicit recognition to the infant's role in the mutual interaction sequence (Bell 1968). If we are to satisfactorily deal with these issues, our observational methodologies need to be improved so that we can provide a richer picture of the reciprocal interaction patterns that are now obscured by our reliance on frequency counts. Possibly, it is not the amount of social input that is unique, but the qualitative features of the social interaction. In other words, do fathers respond to the same kinds of infant social signals that mothers do? And, do they respond with similar kinds of input? It is in these complex matrices of reciprocal interaction that we are likely to define more adequately the subtle and unique differences associated with motherhood and fatherhood.

664 Interaction in social groups B. THE RELATION OF INFANT'S TEMPERAMENT AND MOTHER'S PSYCHOPATHOLOGY TO INTERACTIONS IN EARLY INFANCY* by Penelope Kelly, Syracuse University, Syracuse

Many developmental theorists approach the process of socialization during the first year of life by proposing that mothers and infants progress through a series of increasingly stable patterns of interaction (Sander 1964; Spitz 1965; Moss 1965; Escalona 1968; Bowlby 1969). Developmental researchers have attempted to delineate the important variables which determine the stability of these patterns and which may be used as predictors of subsequent behavior, e.g., separation anxiety and attachment. Very few researchers have characterized the actual mother and infant behaviors that constitute patterns of interaction, although isolated behaviors, e.g., crying and smiling, often have been investigated. Those studies dealing with conglomerates of mother and infant behaviors usually rely on abstractions and inferences (e.g., Escalona 1968) or they present lengthy lists of individual behaviors (e.g., Moss 1965). The former studies lack operational validity and reliability. The latter studies suffer from an overload of variables too numerous to be subjected to sophisticated analyses. Thus, some researchers (e.g., Stern et al. 1969) have used the factor analytic method to summarize the many potential variables entering into mother-infant interaction. This method also permits characterization of individual mother-infant pairs in terms of interactive patterns of behavior through derivation of factor scores for each dyad. It is reasonable to expect many individual differences in the extent to which mothers and infants engage in any particular pattern of behavior. Further, it should be possible to specify the determinants of these individual differences through prospective studies of mother-infant interaction. Potential determinants of individual differences consist of such variables as mother's education, attitudes towards child-rearing, expectations concerning the infant, unique personality traits and socioeconomic level. It is equally important to specify the infant's contribution to patterns of mother-infant interaction. Variables such as temperament, physical * This research was supported by a grant from the National Institute of Mental Health, No. 16544, awarded to Dr. Arnold J. Sameroff.

Variations in infant-caretaker interactions

665

appearance, signaling abilities, and sex contribute to the type of relationship established. Even more important, the very process of interaction, itself, will have an effect on the initial determinants of a particular type of interaction over time. This study attempted (a) to specify patterns of interaction typical of mothers and their four-month-old infants, and (b) to examine certain maternal and infant characteristics which may be associated with particular patterns. The study was exploratory in nature. The major goals were (a) to define those patterns most important for further investigations, and (b) to suggest possibilities for research dealing with determinants of specific patterns. Consequently, the study focused primarily on the extremes of maternal personality variability and infant temperament. It was decided that the most useful indices of personality variability were the traditional psychiatric pathology categories. Indices of infant temperament consisted of the nine 'reactivity patterns' defined by Thomas et al. (1968), and three additional dimensions suggested by other researchers (e.g., Birns et al. 1966, 1969). Seventy-three mothers and infants were recruited from a group participating in a high-risk longitudinal study undertaken at Strong Memorial Hospital in Rochester, New York. The mothers had undergone a psychiatric interview during their seventh month of pregnancy. On the basis of the interview and their psychiatric histories, they were classified as schizophrenic, neurotic-depressive, personality-disordered, or control. Mothers in the psychiatrically disturbed groups were matched with control mothers on the basis of sex of infant, mother's age, socioeconomic status, marital status, race, education, and gravida. The schizophrenic and depressive mothers were also classified into high and low pathology groups based on (a) psychiatric history (chronicity of disturbance) and (b) current psychiatric status. These groups were compared to a randomly selected control group of mothers and infants. All mothers were contacted by letter shortly before the infants reached four months of age. If they agreed to participate in the study, the observers made a preliminary home visit to acquaint the mothers with the procedure and equipment and to make arrangements for the observations. The home observation sessions were approximately two hours long and took place on two separate days. Two observers were present during one-third of the observations in order to obtain reliability data. Reliabilities averaged .94 (range = .78-1.00) across 65 behavior categories. Infants were observed only during waking or drowsy states. Observational

666

Interaction in social groups

data were collected by means of a time-sampling technique originally developed by Moss (1967) and modified for the purposes of the present study. The major modifications involved separating behaviors such as 'mother touches infant' and 'infant vocalizes' into spontaneous and responsive categories, in order that mutual reinforcement contingencies could be described. Each mother was requested to complete a multiplechoice questionnaire designed to assess infant temperament. The questionnaire was a modified version of that developed by Carey (1970). A number of new items were added in an attempt to describe three temperamental dimensions (soothability, alertness, and social responsiveness) not originally included. Mother-infant pairs received interaction scores based on proportion of time spent in each behavioral category and phi-coefficients which determined the contingency relationships between selected categories. The scores were subjected to four factor analyses. The first analysis yielded proportion factors which simply described the types of mother and infant behavior that occurred together. The second consisted of infant responsive factors. These factors described interactions in which the infant played a reinforcing role. The third, mother responsive factors, was based on interactions in which the mother played a reinforcing role. The fourth analysis, combined factors, was based on the factor scores from the three initial analyses and represented an attempt to define a smaller number of new variables which expressed the relationships between the original patterns with minimal information loss. However, nine of the initial factors were unrelated to any of the others, i.e., they did not load significantly on any of the combined factors. Consequently, it was decided that the information loss was too great for subsequent analyses to be carried out on the combined factors alone. Further data analyses were based on the scores derived from the 24 factors produced by the three initial analyses and the ten factors produced by the combined analysis. Analyses of variance were performed separately for the three psychiatric groups and their controls (Psychiatric Group x Sex of Infant) and the two extent of pathology groups and their controls (Extent of Pathology X Sex of Infant). Use of the factor analytic method for interactional analysis. The factor analytic method appears to be a valuable tool for describing patterns of mother-infant interaction. Patterns characteristic of mothers and their four-month-old infants were identified successfully. Positive social

Variations in infant-caretaker interactions

667

behaviors were distinguished from negative social behaviors. For example, proportion factor 1 involved mutual smiling and vocalizing responses and proportion factor 2 involved mutual negative responses including infant protest behavior. Maternal position was shown to correlate with the extent of motoric development exhibited by the infants, as well as the amount of social responsitivity they demonstrated. For instance, combined factor 2 indicated that mothers who remained within eyeshot or earshot often left their infants in a supine position, but mothers who held and supported their infants in standing positions were responded to with smiles, vocalizations, and eye contact. Furthermore, mothers who touched and held their infants received negative scores on proportion factor 2, while mothers who remained only within visual range of their infants received positive scores. Finally, animate interactions loaded on different factors than inanimate play behavior. A variety of distinct contingency relationships, involving both mothers and infants, emerged. The combined analysis suggested several patterns in which maternal reinforcements and infant reinforcements were related. For example, negative responsivity from the infant was positively related to negative responsivity from the mother, especially during the infant's protests. Such a relationship suggests a mutual negative reaction pattern. The discriminative activity contingencies factors indicated that the mothers may shape behaviors by differentially reinforcing various aspects of one dimension of infant behavior. Diffuse activity (kicking and thrashing) was positively reinforced with smiles and kinesthetic stimulation, but directed activity (reaching) was negatively reinforced. Other factors suggested that both mothers and infants were capable of differentially reinforcing behaviors in two or more dimensions. For example, the inhibition of activity factor describes mothers who physically supported infants in response to smiling, but who gave negative responses to diffuse activity. The discriminative social contingencies factor describes infants who smiled and vocalized in response to kinesthetic stimulation, but who gave negative responses to maternal smiling and caretaking behavior. Maternal psychopathology and mother-infant interaction. The question of whether or not different types of social interaction patterns characterize mothers and infants depending upon the psychiatric status of the mother must be answered tentatively in the affirmative. However, this answer can be given clearly only when the extent of pathology, rather than a unique disease entity, is the independent variable. Furthermore, it must be noted

668

Interaction in social groups

that 'different' types of social interaction does not necessarily imply that maladaptive behavior is characteristic of pathological groups. There was no support for the proposition that mothers with a specific diagnosis, such as schizophrenia, and their infants behave differently than a matched control group of mothers and infants in terms of inadequate mothering, inconsistent behavior, or withdrawal from each other. The small number of tests which were significant might have been so by chance. Although these results must be interpreted as inconclusive because of the small number of subjects in the present sample (schizophrenic mothers = 9), the few differences which were obtained may be suggestive of future hypotheses. Generally, these differences were shown to consist of interactions between psychiatric group of the mother and infant's sex. For example, only male infants of schizophrenics exhibited both negative responsiveness and a lack of responsiveness towards their mothers. Perhaps as a consequence, the schizophrenic mothers and male infants engaged in fewer positive interactions than the schizophrenic mothers and female infants or the control groups. The same lack of positive social interaction was seen among the depressive mothers and female infants. Less social interaction would suggest withdrawal from the infant, but not necessarily inconsistent behavior on the part of the mothers. Thus, this hypothesis must remain a possibility, pending more substantial research. The personality-disordered mothers and infants did not differ from their controls, except to show slightly more positive interactions. There is some evidence that this finding may be reversed as the infants grow older. Anagnostopoulou (1973) found that personality-disordered mothers engaged in fewer and less-prolonged social interactions with their infants than a control group of mothers. It is not known whether the reversal is related to changes in the mothers' or infants' personalities or is merely a function of age-related developments in infant behavior. However, both the present study and the Anagnostopoulou study suggest that maladaptive social patterns associated with psychiatric disturbance may not become apparent until some time around the first birthday. This hypothesis is compatible with the findings of Fish (1971) that social precursers of schizophrenia were very subtle in the first year, but became prominent in the second year. Patterns of mother-infant interaction were much more likely to vary as a function of the extent of pathology shown by the mother than as a function of her specific diagnosis. However, the pattern of differences that emerges depends, to some extent, on the definition of extent of

Variations in infant-caretaker interactions

669

pathology. If extent of pathology is defined in terms of the mother's psychiatric history, i.e., the chronicity dimension, both pathology groups appear more deviant than the control group, although only the low pathology group differed at significance levels of p < .05 or less. The infants in the low pathology group tended to be more often in a prone or headprone position rather than a sitting position. They gave more negative responses to kinesthetic stimulation, usually in the context of background visual and auditory stimulation. This pattern was noted by the observers to occur most frequently when the infants were placed in a swing-o-matic in front of a television. Their mothers often failed to adjust their position when the infants slid foward in their swings. Sometimes the infants were left in uncomfortable positions for long periods of time even though they protested vigorously. One possible inference, based on these findings and informal observations, is that low pathology mothers did not respond appropriately to their infants' requirements for postural and kinesthetic stimulation. Rather, they preferred to provide such stimulation without maintaining physical contact with the infants and, consequently, failed to modulate the stimulation according to the infants' demands and abilities for self-modulation. One further indication of the relative insensitivity of low pathology mothers is that they less often responded socially to protests than mothers in the other groups. Although the sex difference statistical tests for the high pathology group were significant only at the p < .10 level, it appears that they exhibited patterns opposite of those shown by the control group. The high pathology mothers tended to engage in fewer mutual negative interactions with their male infants and presented fewer inanimate reinforcements for their female infants' vocalizations. The opposite patterns, i.e., many negative interactions with male infants and positive reinforcements of females' vocalizations, were common interactions among control mothers. These results are far too specific to lend themselves to an interpretation of general insensitivity among low pathology mothers or distorted sex-related distinctions among high pathology mothers. However, they do point to the need for close examination of the large and subtle variety of patterns that may be related to maternal psychopathology and the necessity for careful control of both specific psychiatric diagnoses and severity of disturbance. The current status analyses, in contrast to the chronicity analyses, suggested that neither the high nor low pathology mothers and their infants could be described as particularly deviant in terms of inconsistency,

670 Interaction in social groups withdrawal, or lack of responsiveness. Consequently, current pathology, by itself, does not appear to be a contributing factor to maladaptive mother-infant relationships. Generally, high pathology and control groups differed only if the infants were male, and, except in one instance, the high pathology group engaged in more positive interactions. However, the present study cannot be interpreted as conclusive evidence for the nondebilitating effects of high maternal pathology. The number of subjects is very small (high pathology mothers = 8; low pathology mothers = 9) and multiple effects (e.g., extent of pathology and infant temperament) cannot be evaluated. Rather, the results lend further support for the proposition that failure to account for multiple determinants of interaction processes leads to simplistic and, frequently, erroneous conclusions. In spite of the deficiencies inherent in small-scale investigations, the present results suggest a number of avenues for further study. For example, there is some evidence that control mothers are more likely to make subtle distinctions in their interactions, both in terms of specific behaviors and differential responses to male and female infants. Control mothers are the only group to differentially reinforce various aspects of activity in male infants and to differentially respond to protests of male and female infants. The latter finding is particularly interesting because male infants of control mothers were more frequent protesters and negative responders. Their mothers, in turn, more often responded negatively. This mutual negative pattern was not shown by control mothers and female infants. Under normal circumstances, maternal behavior may be more extensively controlled by infant behavior, especially if it is negative, and mutual contingency patterns may more easily emerge than under conditions of maternal psychopathology. Sex differences in mother-infant interaction. Generally, sex of infant appeared to be a significant factor in the interaction patterns that mothers and infants established. However, caution must be exercised in generalizing sex differences across broadly defined population samples. Main sex effects were not significant for either schizophrenic or personalitydisordered sets of analyses. Although there were three independent sex differences in the depressive group, all other sex differences interacted with the mothers' psychiatric statuses. Both sets of analyses of extent of pathology suggested that sex differences interact with the mothers' degree of mental illness, although the interaction process was more prominent

Variations in infant-caretaker interactions

671

when current status alone was considered. The relationship between maternal variables and infants' sex is an important area for future research. It is indeed possible that the mother's personality influences how she chooses to differentiate sex-related behaviors in her infant. It is also likely that early sex differences will lead to differential responses, and these responses may, in turn, vary as a function of maternal personality. A number of sex differences in mother-infant interaction patterns were independent of maternal psychiatric status. Generally, male infants received more intensive proximal social stimulation from their mothers. Such maternal behaviors may have been determined, in part, by the fact that the male infants more frequently reinforced proximal and kinesthetic stimulation with positive social responses, such as smiling and vocalizing. Moss (1967) also observed more intense interactions between mothers and male infants. In the present study, female infants received more distal stimulation from their mothers than male infants, especially in the form of smiling, vocalizing, and imitative responses to their vocalizations. Lewis and Ban (1971) and Lewis (1972) noted a similar sex difference in maternal behavior. Female infants were talked to more often than male infants. Finally, female infants in the present sample received more negative responses for diffuse activity than male infants. Apparently, as early as four months of age, active behavior in males is considered more acceptable than active behavior in females. Most likely this attitude is based more upon societal expectations than an abnormal amount of activity among female infants. No sex differences were found in the activity level dimension of infant temperament as perceived by the mothers. Infant temperament and maternal psychopathology. Factor-analysis of the nine infant temperament scales revealed that an easy-difficult infant dimension could be identified. A threshold-intensity of response dimension also emerged. The easy-difficult dimension derived from the present factor analysis differs in several respects from that same dimension as described by Thomas et al. (1968). Rhythmicity and intensity of response did not load highly on the easy-difficult factor whereas persistence and the newly added dimension of alertness did. The intensity of response dimension loaded on a separate factor along with sensory threshold. The core temperament characteristics of the easy-difficult dimension common to both this and the Thomas et al. study were adaptability, approach/ withdrawal, and mood. These dimensions are probably the most pertinent when dealing with parental perceptions of the easy or difficult infant.

672

Interaction in social groups

The analysis of infant temperament and maternal psychiatric disorder indicated that there were no differences in temperament among infants of schizophrenic, depressive, personality-disordered, or control mothers. Consequently, there was no support for the common suggestion that mothers with a specific disorder produce infants with unusual temperaments. Although such results do not refute an hypothesis of a genetic component in the development of schizophrenia and other mental disorders, they do suggest that temperament, per se, is not the causal agent in 'vulnerability'. However, other studies (Fish and Alpert 1962, 1963; Fish 1971) have indicated that the mother's current psychological illness may aifect her infant's temperament, at least in the newborn period. The results of the extent of pathology (current status) and infant temperament analysis lend further support to this hypothesis. Infants of high pathology mothers were found to be less adaptable than control infants. Yarden and Nevo (1968) suggested that the intrauterine environment of acutely disturbed mothers may be a factor in later behavioral disorders in their children. Perhaps the intervening variable in the subsequent chain of events is infant temperament. Generally, the question of whether or not temperamental differences exist among infants of mothers with a history of psychiatric disturbance (chronicity dimension) and infants of control mothers must be answered in the negative. Thus, the finding by Sameroff and Zax (1973) that infants of chronically ill, high pathology mothers experience more delivery complications and abnormal EEGs apparently is not reflected in the temperament of these infants. Rather, the present study suggests that temperamental differences may exist among infants of mothers who are mentally disturbed during pregnancy, regardless of past history, and infants of normal mothers. However, the small number of subjects was not sufficient to adequately test this hypothesis, and a more extensive and controlled investigation is needed. One difficulty with the present study is that the infants' temperaments were determined by multiple-choice questionnaires completed by mothers who were psychiatrically abnormal. The effects of abnormal maternal perceptions and abnormal temperamental traits cannot be separated. Infant temperament and mother-infant interaction. The home observation data and the temperament data were subjected to a correlational analysis (N = 54) to determine the existence of possible relationships between infant temperament and mother-infant interaction patterns. The overall

Variations in infant-caretaker interactions

673

level of correlation coefficients was low, and the number reaching statistical significance (p < .05) might have occurred by chance. However, the small number of significant correlations between infant temperament and mother-infant interaction patterns suggested that temperament, per se, may be defined differently depending upon whether one relies on maternal perceptions or independent observation. For example, many of the infants would have been classified as highly active by the experimenter who based his conclusions about activity level on the amount of kicking and thrashing exhibited. Nevertheless, activity level, as derived from maternal reports, did not correlate with any of the factors involving active spontaneous or responsive infant behavior. Many of the temperament dimensions and observed patterns of interaction lacked such correlations. On the other hand, a few correlations were significant in the expected directions. For instance, approach/withdrawal correlated positively with reaching for and playing with objects. Persistence correlated positively with negative responses from the mother for reaching behavior. The more difficult infants received negative responses from their mothers during protests and, themselves, gave more negative responses. Overall, infant temperament, as defined by responses on the multiplechoice questionnaire, was most frequently associated with negative social interactions and interactions involving inanimate stimulation. The lack of significant correlations involving positive social interactions suggests that positive patterns may be more or less independent of perceived infant temperament. The same perceptions are likely to be much more compelling determinants in the cases of negative patterns and the infant's ability to react positively to inanimate stimuli. Implications and conclusions. Ideally, because schizophrenia includes so many varied types, a thorough investigation of its relationship both to patterns of mother-infant interaction and to infant temperament should include a distinction between categories of the illness, onset time of the ilness, and mixed diagnoses. Shakow (1971) believes such distinctions are extremely important if one is interested in the development and course of the disease. One of the major limitations of the present study was that the only requirement for classification in the schizophrenic or depressive groups was that the mother, at some point in her life, received the relevant diagnosis. Consequently, many of the subjects in the psychiatrically disturbed groups had mixed diagnoses, and several were found to be currently without psychiatric problems. Such general definitions of a

674

Interaction in social groups

specific mental disorder complicate the results even for a genetic hypothesis, assuming the mode of inheritance is polygenic and the genes, themselves, are distributed normally in the population-at-large. At this point, a closer examination of the effects of extent of pathology is in order. Apparently the effects differ somewhat depending upon whether one has a long history or a short history of mental disorder. The social interaction effects of having a chronically disturbed mother may not become apparent until the child can communicate verbally and interact with his peer group. The effects of having a psychologically disturbed mother during pregnancy may become apparent immediately at birth via temperament measures, and these effects may proceed to interact with social exchanges throughout the first year of life and thereafter. A longitudinal study which examines such interactions at close intervals during the first few years of life is needed. Perhaps, since the consequences of infant temperament are, to some extent, dependent upon parental responses (Thomas et al. 1968), recovery of the mother may alleviate developmental problems, while chronicity may intensify them. An extensive research effort is needed for the design of temperament measures which would be independent of maternal perceptions. The unreliability of maternal reports has been frequently documented (e.g., Bell 1958; Pease and Hawkes 1960). It may be expected that mentally disturbed mothers will exhibit even greater tendencies to distort reports of infant behaviors than normal mothers (Beisser et al. 1967). While some aspects of temperament would be quite difficult to measure in the laboratory (e.g., rhythmicity), others could easily be adapted to laboratory methods of assessment (e.g., approach/withdrawal). One further point, frequently made by other investigators (see Escalona 1968) should be mentioned. In view of the vast number of variables that determine mother-infant relationships, a particular maternal or infant characteristic should not be expected to predict a particular interaction pattern, unless that characteristic is very extreme, e.g., a severe maternal psychosis. On the other hand, it should be possible to specify combinations of maternal and infant variables that co-determine a particular pattern. The small number of subjects in the present study did not allow for this type of analysis. However, one would expect that a currently high pathology mother and a difficult infant would establish different interaction patterns than a control mother and an easy infant and that the patterns, themselves, would have a differential effect on each mother and infant over time. Generally, it should never be assumed that the organism

Variations in infant-caretaker interactions

675

is a static entity and, since social exchanges involve at least dyads, one organism is always part of the environment of the other. Hence, one should never assume a static environment. Consequently, long-range predictions are impossible for all but highly extreme situations, unless trends in intermediate, short-term changes are identified. Certainly one cannot produce causal explanations of developmental outcomes without knowledge of these changes.

676

Interaction in social groups

C. MOTHER-INFANT INTERACTION, ATTACHMENT, A N D MOTHER'S PSYCHOPATHOLOGY by Rena Anagnostopoulou, Vassar College, Poughkeepsie

In recent years the early social environment of the child has increasingly become the focus of study as a major determinant of affective and cognitive development. Early experiences have also been considered responsible for the development of subsequent mental disorders. Most of the studies using direct observational methods have focused on motherinfant interaction in the natural environment with normal samples (e.g., Ainsworth 1971; Moss 1965, 1967; Yarrow, Rubenstein and Pedersen 1971). Prospective 'follow-up' studies are only beginning to be utilized as a means of identifying factors contributing to increased risk for later mental disorder (Garmezy 1971). The present exploratory study attempted to examine the relationship between maternal psychopathology and mother-infant interaction. Attachment behavior was focused upon since attachment is a central developmental landmark for the one-year-old infant and is directly related to his interaction with his mother. Ainsworth (1971) proposed the dimension of 'security-insecurity' as useful in characterizing the quality of an infant's attachment to his mother. She defined security as being demonstrated by enjoyment and active maintenance of contact with the mother, minimal protest upon its termination and in response to short separations, minimal crying, following the mother cheerfully, and using her as a secure base from which to explore. Studies on the origins of schizophrenia have often hypothesized that children who later develop mental disorder have early interactions characterized by maternal inconsistency (Arieti 1967; Burnham et al. 1969; Lidz 1967), lack of responsiveness, and inappropriate responsiveness (Will 1967), which can be conceived of as environmental stress. In terms of mother-infant interaction with mothers suffering mental disorder there is a paucity of data. Sobel (1961) observed the early environment of the offsprin of eight schizophrenic mothers. During the first 18 months of life, three out of four infants living with their schizophrenic mothers developed signs of emotional disorders, whereas none of the four offsprings living in foster homes did so.

Variations in infant-caretaker interactions

677

The mothers of the disturbed infants played with them rarely, failed to provide stimulation, and did not show positive affect and playfulness. His observations, however, were unstructured, and he provided little systematic information on the specific interactions and their development during the first year and a half of life. Recently there has been an increasing recognition of the importance of the infant's behavior in shaping maternal behavior (Bell 1968; Moss 1967). Bell (1971) has reviewed studies that convincingly show the infant's active role in determining the nature of interactions with his mother. Despite such evidence most of the studies on attachment and interaction have primarily emphasized the mother's role in shaping the child's behavior and the quality of their interaction. Within this framework the present study was an attempt to explore mother-infant interaction in mothers with mental disorders, normal control mothers, and their one-year-old infants in their homes, keeping in mind the infant's active role in such interactions. Home observations. The subjects were 43 mothers and their one-year-old infants. They were part of a larger sample participating in the Rochester Developmental Psychopathology Project, a longitudinal 'high-risk' study of genetic and environmental influences in the development of serious mental disorder (Sameroff and Zax 1972). The sample consisted of six schizophrenic, seven neurotic-depressive, eight personality-disordered, and 22 matched control mothers and their one-year-old infants. Each mother was contacted when her infant approached 12 months of age and an appointment was made for a home visit by the observers. They explained the procedure to the mother and encouraged her to ignore them and behave as she normally would. The home observations took place on two different days, lasting for an average of two hours per day. Two observers were present during most of the observations for reliability purposes. Inter-observer reliability was in the nineties (X: .95; range: .93— .97). A modified time sampling technique was used for the observations. It was based on a technique originally developed by Moss (unpublished manuscript) for observing four-month-old infants. Moss's technique was modified in several ways: (1) Items were eliminated from the predetermined checklist of behaviors and new items were added, in order to comply with the behavioral repertoire and expanded environment of the 12month-old infant; (2) infant attachment behaviors were added to the list,

678 Interaction in social groups taken from Ainsworth's (1967) observations; (3) a code was developed to allow for recording of emitted, elicited, and sequential behaviors in the infant's interaction with his mother, and (4) the time sampling unit was reduced from one minute to 20 seconds, so that the sequencing of mother and infant behaviors could be better represented. The behavior checklist thus consisted of 73 mother and infant behaviors, 34 of which could be coded as emitted, elicited, or sequential (see Table 1). In this manner it was possible to record who initiated the interaction and in what manner, the nature of a response, and the following response of the initiator. All interactions that followed after the second response were recorded as sequential. Each mother-infant pair received scores consisting of the proportion of time each behavior category was scored during the total observation time. To assess the interaction between the mother and the infant phicoefficients were computed for the occurrence of combinations of behavior categories. These phi-coefficients reflected combinations such as the mother both vocalizing and smiling, and contingencies, such as, infant cries and mother picks him up. Proportions of maternal and infant behaviors and five groups of phi-coefficients were then subjected to separate factor analyses. The factor scores derived were submitted to a second factor analysis yielding 14 second order factors. The proportions and factor scores were subjected to t tests to compare each diagnostic group with its control. Analyses of variance were performed on both sets of data to compare one diagnostic group with another and to compare the mental disorder group as a whole to its control. Before the presentation of the results a few points should be made regarding the significant findings. Very few significant differences were found between the psychiatric groups, and given the high number of statistical tests applied to the data, they did not exceed the number expected to occur by chance. The results, however, showed patterns of consistency which may be suggestive for future research. Diagnostic groups and their controls. Schizophrenic mothers were no less responsive than their controls, and showed no more inappropriate or non-contingent responsiveness than their controls. There was a tendency for schizophrenic mothers and their infants to smile less at each other in prolonged interactions {p < .10), and a tendency for schizophrenic mothers to show less rich responsiveness to responsive infant vocalizations than their controls (p < .10). Their infants did not show any more of the

Variations in infant-caretaker

interactions

Table 1. List of variables General categories Background events Infant drowsy Mother categories Distant Earshot Eyeshot Close Watches infant* Holds infant* Caretaking activities Feeds Caretakes Background stimulation Auditory Visual Mother variables Instigates physical contact* Terminates physical contact* Interferes Restricts Lack of response Negative physical response Frowns at infant* Rejects infant Provides tactile stimulation* Smiles at infant* Sings to infant* Complies Vocalization: Negative expressive Negative restrictive* Encourages Gives information Social* Vis-à-vis Stresses musculature Presents toy* Manipulates toy* Retrieves toy* Kisses infant*

Infant movement variables Restricted setting Free setting Follows mother Reaches toward mother Moves to mother* Moves away from mother* Explores through locomotion Infant variables Out of contact Instigates physical contact* Terminates physical contact* Lack of response Continues to protest Non-compliant Complies Obeys mother's requests Frowns* Whimpers* Cries* Rejects toys* Rejects mother* Smiles* Vocalizes* Laughs* Watches mother* Gaze aversion* Watches toy* Tries to open the door Manipulates mother* Buries face in mother's lap Clings Kisses mother* Imitates mother's verbalizations Imitates mother's actions Shows toy to mother* Simple play* Complex play* Non-nutritive sucking Cuddles Complex imitation

* Items coded as emitted (1), elicited (2), or sequential (3).

679

680

Interaction in social groups

behaviors indicating insecurity in attachment than their controls. They only showed a tendency to cry when the mother left the house more, but also to show strong negative affect when following the mother in response to minor separations less than their controls (p < .10). They tended to smile and vocalize spontaneously more, to watch the mother less, and to be more object-oriented and independent than their controls (p < .10). Looking at these tentative differences there appears to be some consistency in the opposite direction of that expected in view of the literature for infants of schizophrenics. With the reservations mentioned above in mind, the findings for the schizophrenic group failed to confirm Sobel's (1961) findings on schizophrenic mothers as providing inadequate and impoverished stimulation leading to emotional disorders in their infants. They also failed to indirectly support contentions that mothers with mental disorder behave in an arbitrary manner (Arieti 1967, 1968). The depressive mothers also showed no more lack of responsiveness or contingency than their controls. They were found to respond to whimpering in a rich manner, to physically prolong interactions, and not to be restrictive significantly more than their controls. Their infants could not be characterized as having an insecure attachment, though they tended to show certain behaviors in the direction of insecurity more than their controls. They tended to whimper spontaneously more, to watch the mother less, and to move to her less than their controls (p < .10). This might be due to the tendency for them to be in a restricted setting more than their controls (p < .10), which might also be the reason for their mothers' lesser restrictiveness. They prolonged interactions with positive affect and vocalization significantly less, and approached the mother for physical contact with negative affect significantly less than their controls. This was also true for contact in response to the mother or in prolonged interactions. Mothers with personality disorders did not show any more overall lack of responsiveness than their controls. They responded tactually significantly more but encouraged their infants significantly less than their controls. They kissed the baby when initiating and terminating contact and did not verbally restrict during caretaking significantly more than their controls. They were, however, significantly less responsive to spontaneous crying than their controls. Their infants, like infants of depressives, tended to actively engage in physical contact when showing strong negative affect significantly less than their controls. Mothers with personality disorders showed a tendency to respond through toys and

Variations in infant-caretaker interactions

681

physical contact more, and to provide rich positive stimulation and be restrictive less than their controls (p < .10). There was a tendency for their infants to show less positive affective responsiveness and exploration, and to have mothers who prolonged interactions proximally and distally and were verbally responsive and compliant less than their controls (p < .10). Mothers with personality disorders, then, tended to appear more non-contingent than their controls not in overall responsiveness, but in prolonging of interactions. There was also a tendency for them to use toys as a response to continued protest, to hold responsively, not to talk when initiating and terminating contact, not to prolong playful interactions, and to have babies who showed negative affect and unsoothability more than their controls (p < .10). This tendency to respond through toys to continued protest might be taken as an indication of inappropriateness of response, in view of the infant's tendency to show negative affect, unsoothability, and rejection of toys. Finally, this tendency of infants of personality-disordered mothers to show less positive affective responsiveness and exploration and more negative affect than their controls seems to be suggestive of more insecurity in their attachment than for their controls. When the three diagnostic groups were treated as a whole and compared to the no-disorder group on the first order factors, contrary to expectations, mothers with mental disorder were found to provide proximal stimulation and show promptness of response significantly more than their controls. Mothers with mental disorder tended not to present toys when terminating contact, not to scold when terminating contact in response to the baby, to kiss during caretaking, and to smile when initiating contact in prolonged interactions significantly more than their controls. They tended not to discourage whimpering, to use toys as a response to continued protest, and to respond tactually to responsive crying significantly more than control mothers. The personality-disordered group was already reported as showing differences from its control in this direction. There were no results indicating non-contingency of maternal responsiveness in the mental disorder group as a whole. These findings failed to support notions in the literature of lack of maternal responsiveness in psychopathological environments (Will 1967), and of arbitrariness in maternal responsiveness (Arieti 1968). No significant differences were found between infants of mothers with mental disorder and their controls. There was only a tendency for them

682 Interaction in social groups to be less independent and toy-oriented than their controls, with the exception of the schizophrenic group which tended to show the opposite pattern, as reported earlier. In terms of attachment, then, infants of mothers with mental disorder were as a whole no less attached to their mothers, and appeared no less secure in their attachment than their controls. Differences among diagnostic groups. The comparisons of the three diagnostic groups showed, in general, that schizophrenic mothers showed significantly more promptness of response and encouragement than depressive mothers. Depressives also failed to respond significantly more than personality-disordered mothers. The rest of the differences were at thep < .10 level. Infants of schizophrenics vocalized spontaneously more than infants in the other two groups. The personality-disordered group tended to show a more maladaptive mother-infant unit than the schizophrenic and depressive groups, with the infant showing negative mood and unsoothability, and the mother responding through toys to his continued protest and not prolonging pleasant interactions. Patterns of interaction and attachment. Although most studies on attachment and interaction used intercorrelations, a few factors of the present study were compared to patterns reported in the literature. Of the infant factors one was clearly related to negative affect and unsoothability. Rejecting toys, all forms of whimpering and crying, and continued protest despite maternal intervention all loaded highly on this factor. Vocalization had no substantial loading. Factors consisting of vocalization had loadings of positive affect and exploration but no crying. These factors lend indirect support to Bell's (1971) finding of a negative relationship between crying and other communicative signals. Initiation and termination of contact with the mother loaded positively on the same infant factor, unlike Ainsworth's et al. (1971) finding that infants who initiated contact with the mother tended not to be the ones terminating the contact. Moving to and away from the mother spontaneously, following, showing toys to her, moving away in response to her, being in a free setting, and exploring, all had substantial positive loadings on this factor. Ainsworth et al. (1972) reported a lack of relationship between crying in general and following the mother. In the present study, indeed, following and crying loaded on separate factors. The factor analysis of

Variations in infant-caretaker interactions

683

phi-coefficients, however, did show a factor consisting of following with negative affect in response to minor separations. In terms of infant responses to termination of contact by the mother, there was a bipolar factor describing a mother-infant dyad in which the mother responds in a rich manner to the infant's clinging, and the infant happily accepts termination of contact by the mother and shows a mild negative response only to her departure from the home. Another bipolar factor consisted of whimpering and crying in response to the mother's departure from the room, but strong negative responses to termination of contact had a high negative loading on this factor. Punishing continued protest also had a positive loading. These loadings failed to agree with Stayton and Ainsworth's (1973) 'security-insecurity' factor, in which crying when the mother leaves the room and negative responses to being put down both had positive loadings. This disagreement might be due to differences in samples or procedure. The factor analyses, however, showed that negative responses to separation and termination of contact might not necessarily show a positive relationship that can clearly fit the security-insecurity dimension. In general, then, the factors revealed patterns of attachment and interaction that were at least in partial agreement with those reported in the literature, though they showed a diversity which did not clearly fit the dimension suggested by Ainsworth (e.g., Stayton and Ainsworth 1973). Furthermore, some of the tentative differences found could not easily fit this dimension. One suggestive finding was that infants of depressive and personality-disordered mothers, who were the ones to show some tendencies in the direction of insecurity, tended to engage in active contact behaviors with strong negative affect significantly less than their controls. On the other hand, initiating contact with the mother, which according to Ainsworth et al. (1971) is a measure of security of attachment, did not differentiate the groups. It is possible that initiation of contact and physical contact behaviors toward the mother might be of less relevance than the affect shown in such conditions, in describing the quality of attachment. Thus, these tentative findings may be an indication that contact with the mother is not being perceived as comforting for these infants, even though they appeared equally concerned with their mother's presence as their controls. Interpretation of findings. Conclusions about the results can only be tentative, in view of the paucity of significant findings and the small

684 Interaction in social groups number of subjects. The present study, within these conditions, was unable to identify specific environmental stress in mother-infant interactions, even for the schizophrenic group. The few tentative diiferences found can be used as a suggestion for a transactional interpretation of data (SamerofF 1972). Instead of simply observing the tendency of schizophrenic mothers not to prolong distal interactions when their babies were smiling and vocalizing, the infant's contribution may be emphasized. Infants of schizophrenics tended to appear more independent and to show more spontaneous affect and vocalization than their controls. These tendencies may be the cause of their mothers' lesser prolonging of distal interactions. It is difficult to see these infant behaviors as the effect of the mother's present behaviors. On the other hand, it is possible that they were influenced by maternal behaviors in the past. Kelly (1973) studied the same infants at four months of age. While she found no differences between the two groups of infants, schizophrenic mothers tended to initiate and terminate socialization and watch their infants more than their controls. By twelve months of age, tentative differences begin to appear between infants of schizophrenics and their controls, whereas the differences in maternal behavior found at four months have disappeared. These suggestive changes in maternal behavior from four to twelve months challenge the notion of a static environment. Incidentally, lack of stress in the early environment of infants of schizophrenics does not guarantee lack of stress in the future. M. Goldstein (personal communication) has been finding emotional disorders developing in children of schizophrenics during late childhood and adolescence, and has proposed that the mothers provide them with a stressful environment once the child's striving for independence becomes a threat to the integrity of the mother's ego. The personality-disordered group appeared suggestive of maladaptiveness when compared to its control, the mother showing a tendency to be ineffective in soothing the infant, the baby showing less positive affective responsiveness, and the mother again not prolonging proximal and distal interactions. In the case of the personality-disordered group attributing the cause of these suggestive tendencies to the mother appears appealing, in view of Bell's (1971) finding that prolonged crying by the end of the first year correlated positively with maternal unresponsiveness to crying in the early months of life. Given also the tendency of personalitydisordered mothers to appear less compliant than their controls, it may be that their behavior is less modifiable by their infants than for their

Variations in infant-caretaker interactions

685

controls. Knowledge of their past history is, of course, necessary before any causal inferences can be drawn. Methodological considerations. The time sampling method with the use of a detailed checklist of behaviors appeared useful for its objectivity and easy attainment of reliability among observers. Reducing the time sampling unit to 20 seconds allowed for a more representative recording of interactions. One of the limitations of time sampling as compared to continuous recording was that long sequences of interaction could not be specified. One attempt for overcoming this problem was to create separate categories for emitted and elicited behaviors, and behaviors that were responses to responses, i.e., prolonged interactions. In view of the results and factor loadings this seemed to capture some of the complexity of mother-infant interactions mostly absent in studies of attachment. On the other hand, all behavior sequences that followed the response of the initiator of interactions could not be characterized any further. The advantage of continuous recording appears to be this capacity to capture longer sequences. The problem with continuous recording, however, remains the lack of sophisticated methods of data analysis (Bell 1971). The results also suggested the usefulness of searching for patterns of behavior through factor analysis, in that differences began to appear on the factors which were not evident when behaviors were considered individually. The phi-coefficients also appeared useful in that they captured some of the varieties of mother and infant behavior patterns. On the side of limitations, the present study was unable to record infant interactions with persons other than the mother, which, though not frequent, nonetheless occurred and might have been of significance. Finally, having focused on the mother's behavior only in relation to the infant, some of the infant's influence on the mother was missed. In view of the transactional model it may be useful to close with an observation on a schizophrenic mother and her infant. The mother appeared very anxious in the home, constantly moving and manipulating things, and talking incessantly to herself. Her speech was very fast and became progressively more and more incomprehensible as she went on. Her one-year-old girl was a very independent child, actively exploring her environment, and moving from room to room playing with her toys. Every time she approached the mother and actively engaged in contact with her the mother slowed down, showed no apparent anxiety, and her speech to the infant became normally paced and very clear.

686 Interaction in social groups Finding ways of objectively assessing such changes in maternal behavior initiated by the infant seems imperative, especially for the study of psychopathological environments. Perhaps a combination of time sampling of infant behavior and interactions with continuous summarized recording of maternal behavior will better represent their mutual influence. This kind of approach would also offer the much needed comparison of various methods upon the same observations.

Variations in infant-caretaker interactions 687 D. COHORT EFFECTS AND APPARENT SECULAR TRENDS IN INFANT RESEARCH* by Stephen W. Porges, University of Illinois, Champaign-Urbana

At any stage of infant development the observation of behavior is dependent upon two cohort influences. The first, which is very obvious, is the infant himself. This category includes all the antecedent conditions that may reasonably be related to the birth and early environment of the infant. The second, and often neglected, is the observer. Experimenters often assume if inter-observer reliability is demonstrated the validity of measurement is assured. This technique merely results in a consistency of a tradition within a given style of research or laboratory. The category of the observer actually includes the experimenter's theoretical and philosophical view of the infant, often a model or set of general expectations of infant behavior. The changing prenatal, perinatal, and neonatal environment through the years, and the various experimenter expectations of infant behavior, may interact and result in apparent secular trends in the behavior of very young infants. This paper will emphasize two separate but interacting influences which may result in the appearance of secular trends in infant behavior: The first is the collective effect of specific environmental factors which may affect the infant's behavior and development; the second is the researcher's orientation or view of the infant. The cohort of the infant. Logically, if infant research is concerned with the ontogeny of psychological processes the research should take into account a multitude of the factors which may influence the development and limitation of early infant behavior. Infant research, generally, has not dealt with many of these factors. Experimenters have assumed that these factors are 'randomized' among subjects, populations, and across time. These assumptions, if not firmly based, may result in a limitation of the generalizability of previous infancy research. Factors contributing to the behavior of very young infants are not totally dependent upon the environmental influences after birth, but are also dependent upon an accumulation of events and factors which may * This work was supported in part by grant NE-G-00-3-0013 from the National Institute of Education.

688

Interaction in social groups

have occurred long before birth. All these factors contribute differentially to the ontogeny of the first observable behavior of the newborn. During the past one hundred years, the period often used as an observational window to define the secular trends of accelerated growth and maturation, it has become increasingly obvious that many factors which affect human development have their genesis during the prenatal period. Some of these factors which influence the intrauterine environment may be dependent upon historical events and trends characteristic of the time of conception and not random during history. This paper will select a few of the major influences and briefly describe their possible consequences on infancy research. In our culture many drugs containing sex hormones have been given to females. Some are prescribed, such as fertility and birth control pills; others are absorbed via the diet. This is partially the result of the administration of sex hormones to cattle and chicken to facilitate growth. Since the fetal period is the time during which many of the biological characteristics of sex are determined, the ingestion of sex hormones may affect the developing fetus. Beginning in the late 1940s synthetic progestins were administered to pregnant women with histories or threat of miscarriage. The progestins were used to augment the women's natural level of progesterone in order to maintain pregnancy. Undoubtedly, many pregnancies which would have been aborted were brought to term. However, cases were reported of masculinization of some of the newborn females, even to the extent of a diagnosis of hermaphroditism (Money and Ehrhardt 1972). Male hormones, which are often medically prescribed to women, may have undesirable side effects, such as the growth of facial hair in a masculine pattern. Although pregnant women are resistant to the virilization effects of male sex hormones, a female fetus is not. Phoenix, Goy, and Young (1967) have shown that injecting male hormones'^ into a pregnant monkey bearing a genetically female fetus will produce a masculinized infant that not only shows sex organs which appear to be those of a male, but also exhibits more masculine behavior than normal females. The hormone, therefore, modifies the development of the nervous system during the fetal stage, as well as modifying the external appearance of the individual. The thalidomide tragedy of the early 1960s (Nyhan and Lampert 1965) stimulated an awareness that drugs from heroin to aspirin taken during pregnancy may be determinants of birth defects. A recent symposium

Variations in infant-caretaker interactions

689

(Symposium on Drugs and the Unborn Child 3/15-16/73) sponsored by the National Foundation-March of Dimes discussed this topic. The symposium participants agreed that 'in view of the inadequacy of our current knowledge of the effects of drugs on the fetus, a woman would be wise to avoid all drugs - both prescribed and self-administered - unless she were certain she was not pregnant'. They also discussed the benefitrisk ratio of specific drugs that may determine whether medication should be used. The placental barrier is critical to the evaluation of the effects of drugs. The notion that the placenta serves as a protective barrier between the fetus' independent circulation and his mother was shattered to pieces in 1961 with the tragedy of thalidomide. A number of infants whose mothers took thalidomide as a mild sedative during the first weeks of gestation were born with limb malformations and other defects. Placental size also increases and the density of the placental barrier decreases throughout gestation (Eichorn 1970). Many psychologists studying newborn behavior are interested in specific response systems. In my laboratory (Porges, Arnold and Forbes 1973; Arnold and Porges 1972; Porges, 1974) we have noticed that an autonomic index of attentional responsivity, heart rate variability, is suppressed during the first day post-partum. This level increases significantly, regardless of experimental manipulations and environmental experiences, between the first and second day post-partum. Correlated with this increase in heart rate variability is an increase in the infant's responsiveness to external stimulation. The only plausible explanation we have is that there is an abatement of the influences of the deliverymedication treatment. Another example is reported by Kron, Stein, and Goddard (1966) who investigated the relationship between sucking behavior of neonates and the level of anesthesia given to the mothers during labor. The central nervous system depressants administered to the mother severely depressed the sucking rate which remained depressed for the first four days of postnatal life. Pressure of the sucks was also reduced and this combination resulted in a reduction of nutritive consumption. Barbituates often given as premedication prolong the disorganization of the infant's central nervous system by one or two days, the adaptation to breast feeding by 41-48 hours, and weight gain by 24 hours (Brazelton 1961). There are, in addition, the effects of cigarette smoking, coffee drinking, and virtually any manipulation which may affect the biochemistry of the mother during pregnancy, including basic nutrition.

690

Interaction in social groups

Some general examples of drugs passing through the placental barrier which suppress fetal movements have been general anesthetics, such as ether, barbituates, codeine and other opium derivatives, and tranquilizers (Klingberg, Abramovici and Chemke 1972). If these drugs have been administered preoperatively and a general anesthetic used, the fetus may be completely inactive and, therefore, the newborn at birth may exhibit relatively little responsiveness. This would result in a lower Apgar score (Apgar 1953) which is often used for the selection of subjects for psychological research. This bias is compounded by the fact that generally patients from lower socioeconomic classes and complicated deliveries receive general anesthesia. These subjects when tested on behavioral indices during the first few days of life may appear to exhibit 'retarded' behavior. According to Conway and Brackbill (1970) these effects may last for at least four weeks after delivery. Since much of the neonatal research is done during the first three days post-partum, concern should be directed to the influence of drugs on newborn behavior. A quick survey of the literature indicates that many investigators were not cognizant of this problem - and had enough problems related to securing newborn subjects. Perhaps the 'drugged' subjects were considered excellent material for experimentation, since their activity level was low and they seldom cried. It is also possible that the infant is capable of an enlarged response repertoire which has been greatly attenuated by the 'drug' manipulation. During the past 100 years there have been numerous influences which have affected the fetus and early newborn behavior. Striking examples have been the effects of radiation resulting in malformation and microcephaly in live-born children in Hiroshima and Nagasaki who were irradiated during fetal development (Plummer 1952; Yamazaki et al. 1954). There have been examples of multiple birth occurring following fertility drugs, pseudohermaphroditism following hormonal treatments, and the traumatic effects of maternal diseases such as rubella on the unborn child. Our interest in these antecedents seems only to develop as a function of irreversible tragedies. However, being interested observers of the ontogeny of behavior and early predictors of later behavior we have to be cognizant that the behavior we are observing during the first few days post-partum is, in a sense, 'contaminated'. The cohort of the experimenter. The theoretical orientation of the experimenter as a source of research variance has often been neglected.

Variations in infant-caretaker interactions

691

Although the dramatic 'experimenter bias effect' has been investigated, a more subtle influence, the experimenter's view of the human organism, has been ignored. In modern developmental psychology there are two common views of the human organism. These two views have resulted in two families of psychological theory. The first is a derivative of the reductionism propagated by the English empiricists, and the second is a derivative of the Continental philosophy which emphasized an unfolding and synthesis of the organism. Reese and Overton (1970) have described two general models which fit these theories. They classified developmental theories such as Bijou and Baer (1961) and Sears (Sears, Rau and Alpert 1965) in a family of theories formulated within the contexts of the mechanistic world hypothesis, which portrays the human as a reactive organism; they categorized Piaget (1960) and Erikson (1950) in a family of theories formulated within the contexts of the organismic world hypothesis, which views the human as an inherently and spontaneously active organism. Overton and Reese (1972) have explored the implications of the mechanistic and organismic models for developmental psychology. In general, a model may aid the experimenter with theoretical organization and interpretation of the data. However, in addition to being an aid for understanding the data, a model causes the experimenter to see the data with a specific focus. A model acts to define the problems that should legitimately be investigated. Reese and Overton (1970) recognized that the underlying model selected by the experimenter will influence both the selection and the interpretation of research problems. Since the model causes the experimenter to see data with a specific focus it may result in biasing the statement of research objectives and, subsequently, the interpretation of the research. For example, the mechanistic model allows psychologists to study classical conditioning at any age or stage of development. Classical conditioning procedures employed in the study of infant learning were published as early as 1909 (Krasnagorski 1909, as cited in Kessen, Haith and Salapatek). Several attempts to classically condition the fetus were attempted during the 1930s and 1940s (Ray 1932; Sontag and Wallace 1935, 1936; Spelt 1948). Although the results were not definitive, prenatal conditioning was a viable research question. During the 1930s and 1940s there were also many attempts to classically condition the human newborn (Marquis 1931, 1941; Morgan and Morgan 1944; Wenger 1936; Wickens and Wickens 1940) with varying degrees of success. During the next twenty years the condition-

692 Interaction in social groups ability of the newborn was not commonly investigated. Apparently during the late 1950s and early 1960s the Zeitgeist of infant research (primarily in the United States) was directed towards a developmental model which emphasized maturation and suggested that the level of neurological development in the newborn was not sufficient to allow the appearance of conditioned responses. Many psychologists also emphasized during this historical period the behavioral limitation or incompetence of the newborn. During the late 1960s and early 1970s a series of studies, emerging from the laboratories of Lipsitt and his colleagues (Kaye 1965; Lipsitt 1966; Lipsitt and Ambrose 1967; Lipsitt and Kaye 1964), Clifton (1971), and Porges (Forbes and Porges 1972; Porges, 1974; Stamps and Porges 1973), have demonstrated classically conditioned responses in neonates. Moreover, the Soviet psychologists have been using the classical conditioning paradigm with varying success (Bystroletova 1954; Dashkovskaia 1953; Denisova and Figurin 1929; Krachkovskaia 1959; Polikanina 1961). Three points may be extracted from this historical review: First, the acceptance of newborn classical conditioning by the psychological community appears to be linked to specific historical periods during which the mechanistic model was inherent in the weltanschaung. (During these periods even the question of prenatal conditioning was viable within this context.) Second, although research on newborn classical conditioning has been criticized on a methodological basis, it appears that the establishment of classical conditioned responses are possible during the first few days post-partum even though the topographic response characteristics, such as magnitude and stability, may differ from the more mature organisms (Lipsitt 1970; Fitzgerald and Porges 1971; Porges, 1974). Third, the Soviet psychologists who have not been influenced by the same philosophical traditions appear to have found little conflict in investigating newborn conditioned responses. According to Riegel (1973), there is another model which would partially explain why the Soviet psychologists were less affected by the dualistic traditions embodied in the mechanistic and organistic world views. Riegel has described this as a dialectic model based upon the work of Rubinstein, a Soviet psychologist, which incorporates both viewpoints and, thus, overcomes the dualistic conception outlined above. Rubinstein, by utilizing psychological concepts consistent with the philosophical views of Marx, Engels, and Lenin, has provided a synthesis which enables the researcher to deal with the totality of the development of man. Riegel

Variations in infant-caretaker interactions

693

identified four major theoretical viewpoints of man and his development, including models of passive environment-passive organism, passive environment-active organism, active environment-passive organism, and the dialectic of active environment-active organism. Within this framework the mechanistic model would conceive of development as an accumulation of external information by an essentially passive organism and the organismic model would conceive of development as 'the spontaneous emergence of new models of operation for which the environment merely provides information as necessary material in order to enable the organism to make his own selection' (Riegl 1973). Infant behavior from the organismic view may be understood only in terms of the level of organization of the infant's cognitive system. The elements of the cognitive system, schemas for Piaget, permit the infant to make sense of or assimilate the stimuli found in his environment. Sameroff (1971) has critically discussed the newborn conditioning literature and interpreted the findings in a Piagetian-cognitive context. Taking the organismic view, he rejected all the positive evidence of classical conditioning in newborns. This is understandable from the Piagetian view, since the evidence of classical conditioning necessitates the development of a long-term memory associated with the permanence or recognition of the unconditioned stimulus. The above discussion has been a demonstration of an experimenter cohort which may result in an apparent secular trend in the conditionability of the newborn. During the first fifty years of this century, the behavioristic approach with the mechanistic model supported the notion that the newborn could be conditioned; therefore, newborn conditionability was a viable research question and the data produced by these studies were readily accepted and organized in existing theories. Since the 1950s the influence of Piaget and the organismic model has stressed that conditionability necessitates the development of a type of memory which occurs later in life as a function of maturation. This conflict is an example of how the experimenter's model may generate a set of behavioral expectations and limit his interpretation of the data. Although these contrasting views of the newborn's response capacity may be interesting on a philosophical level, they may have limiting effects on the area of infant research. For example, I recently submitted a grant proposal for the study of neonatal conditioning to the National Institutes of Health. The grant proposal if approved might have been funded by the National Institute of Child and Human Development. The proposal was

694 Interaction in social groups not funded nor was it approved. It was considered 'atheoretical', perhaps since the assumptions in the proposed research appeared to reject a maturational model for conditioning. A summary of the review was forwarded to me with the following statement: '... the reviewers noted that if the existing published literature in the field of infant psychology can be used as any index, the outcome of most of your research would be negative, since there are no published records of successful conditioning of neonates. Another unsuccessful attempt would have little significance. Although you have claimed successful neonate conditioning in an unpublished study which you described in an abstract, but did not provide any details, the weight of this meager evidence must be weighted against the bulk of available negative evidence. According to one eminent researcher, no one has published any study showing that infants below two months show any evidence of even short term memory, no less the long term memory necessary for conditioning ...' It is obvious that this statement presents strong biases regarding the potentials of newborn infants. The reviewers obviously have rejected the previously published literature (cited earlier in this paper) because it did not fit their model. It is highly ironic that the same study section of the National Institutes of Health approved and funded a grant to study newborn conditioning a year earlier and that a similar grant of mine was later funded by the National Institute of Education. The reviewers' comments regarding short- and long-term memory are other examples of biases associated with an experimenter cohort, since many experimenters dealing with conditioning do not include this construct as a determinant of conditioning. The 'memory' notion is obviously more in the cognitive realm associated with Piaget. However, if one is to refute the claims regarding newborn memory, a survey of newborn habituation and operant conditioning literature would be helpful. Many infancy psychologists utilize a 'habituation' technique to study memory. They assume that if the stimulus no longer elicits a response, the components of the stimulus have been encoded (assuming that it is not receptor fatigue). There have been a series of studies which have demonstrated that newborns are capable of habituation (Bartoshuk 1962; Bridger 1961; Keen, Chase and Graham 1965; Lipsitt 1966; Engen and Lipsitt 1965; Friedman, Nagy and Carpenter 1970). Other psychologists interpret operant conditioning as the development of memory regarding the response-reinforcement contingency. There have also been recent successful demonstrations of operant conditioning in newborns (Brown

Variations in infant-caretaker interactions

695

1972; Clifton, Siqueland and Lipsitt 1972; Lipsitt, Kaye and Bosack 1966; Siqueland 1968; Siqueland and DeLucia 1969). Moreover, Papousek (1967, 1961) has demonstrated newborn conditioning utilizing a paradigm which has classical and instrumental components. These studies may also be used to support the contention that the newborn is capable of conditioning and exhibiting at least a short-term memory. Conclusion. Research investigating early infant behavior may be influenced by many factors. This paper has described two categories of cohort effects which may influence both the data collected during the first few days post-partum and its interpretation. Within these two categories I have tried to select environmental conditions and theoretical orientations which may be easily seen to be related to historical or secular trends in infancy research. A survey of research investigating early infant behavior during the past century exhibits a total inconsistency. The historical changes affecting prenatal, neonatal, and perinatal care may be related to the changing trends in the expectancy of early infant behavior. Although these factors may influence newborn research, they do not exist in a vacuum independent of the theoretical orientation of the experimenter. If the response potential of the newborn is to be understood, the infant cohort and the experimenter cohort must be identified and their associated influences clearly stated. In the process of identifying these factors, the dominant theoretical orientation in developmental psychology may change to accomodate a more complete picture of the infant's behavioral repertoire.

V I C T O R G. C I C I R E L L I

3

Sibling influences on the development of the individual

A. SIBLING POSITION, SEX OF CHILD, A N D M A T E R N A L INVOLVEMENT by Mary K. Rothbart, University of Oregon, Eugene

The first and often most enduring setting for social development is the family. In the home, a child adopts and begins to perfect ways of relating to others and to the physical world, and we expect that a child's early learning will be a function of the family structure into which he or she is born. For this reason, a great deal of research has investigated personality differences between children of differing birth order, and a smaller number of studies have investigated patterns of early family interaction as a function of the child's ordinal position. We are only beginning to explore differences in early family interaction related to sibling position, and to date most completed research has involved the study of parentchild interaction rather than child-sibling interaction. Moreover, fathers have rarely been the focus of study; until very recently only the mother's interactions with her children have been studied in any detail. The present paper will attempt to review studies indicating that a child's sibling status (birth order and sex) is predictive of the mother's behavior toward the child in an interaction situation. It will be suggested that the mother's behavior reflects both her experience as a parent and the extent of her involvement with the child's performance, and that the mother's differential treatment of her children may affect sibling interactions in important ways.

698 Interaction in social groups Studies of ordinal position and parent-child interaction. The first major study of mother-child interaction as a function of birth order was performed by Lasko (1954). Using longitudinal data from the Fels studies, she was able to compare visitors' home ratings of a mother's behavior toward her first two children. In addition, because of the longitudinal nature of the data, Lasko was able to compare mother's behavior toward the first-born with her behavior toward the second-born at any given age from fifteen months to ten years. Comparing first and second children from 46 families, only two of twenty-one Fels scales showed overall significant differences: Second-boms were given both more approval and more babying. However, in terms of direction of differences, most 'warmth' variables favored the second child, while differences on 'control' variables were in the direction of the first child being more controlled by the mother. Although Lasko predicted overall greater maternal interference and anxiety over the first-born than the second child, these differences were not found. Lasko was also able to examine birth order differences in terms of age trends. She noted that the first child's original environment was rated as very warm and child-centered, but that these qualities declined during preschool years. While warmth scores started at a lower point for the second child, they maintained the same level for a longer period, finally exceeding ratings of warmth for the first child. At the age of five only (when the child was entering school), greater disciplinary friction, criticism, and attempts at acceleration were noted in the mother's behavior toward the first child. Lasko's study gives strong confirmation to the view that the treatment of the first child is less stable over time than the treatment of the second child. Unfortunately, she was unable to analyze the specific sex composition of her families. Extensive research cited by Sutton-Smith and Rosenberg (1970) has indicated that not only sex of the child but the sex of his or her siblings are related to birth order differences. Stout (1960) studied the behavior of both parents toward children of differing birth order in one to four child families, using ratings of parental behavior from the Berkeley Guidance study. She found that parents showed more directiveness toward the first-born, and that parents tended to jointly show or withhold affection from the first-born girl, while at least one parent showed greater warmth to the second daughter. Stout's findings on authority generally agree with Lasko's findings on 'control'; parents seem to be exerting more authority in connection with the first-born. However, there are problems in interpretation: Sibling spacing, sex

Sibling influences on the development of the individual 699 composition, and size of family were not controlled; ratings were summed over an 18-year period, a procedure that may have obscured possible early differences in parent-child interaction. Hilton reported an important study of birth order and mother-child interaction in 1967. In order to compare mother's interactions with firstborn and second-born children as a function of the child's success or failure, she gave mothers different information about their child's success or failure on a puzzle task. Hilton reported greater interference on the part of the mother to first-born and only children (grouped together) on several measures: Mothers of first-born and only children were rated as 'more involved', were more likely to initiate work on the puzzle during an unstructured play period, and gave more task-oriented suggestions and support to the first-born, especially if the child was reported as doing well; if the mother of the first-born had been told her child was failing, she was more likely to decrease her demonstrations of love and expressions of verbal support. Hilton concludes that the mother is both more interfering and inconsistent in her behavior toward her first child, although the differential warmth exhibited toward the first-born might also be viewed as a more contingent reaction on the part of the mother to the reported success or failure of her child. In a second structured mother-child interaction study, the present author (Rothbart 1971) observed mothers supervising their 5-year-old boys and girls from same-sex, two-child families in five achievementrelated tasks: An open conversation, interpretation of two cartoons, explanation of how a water tap works, completion of a memory task, and supervision of the child's performance on a puzzle. Findings from this study generally agreed with Hilton's results, with one major exception. Hilton failed to find consistent differences in the mother's behavior as a function of the sex of the child, while Rothbart's study indicated that a mother's behavior toward her first- or second-born child was also a function of the child's sex. Mothers gave first-born children more complex explanations than second-born children. Mothers were rated as more anxiously instrusive into the performance of the first-borns, and also exerted more pressure for achievement on first-borns in a memory task, but these last findings were strongly accentuated in the mother's behavior toward the first girl. The mother gave a higher estimate for her first-born girl's performance on the memory task, and showed significantly greater intrusiveness toward daughters than toward sons. Both the sex and birth order differences on the intrusiveness variable were contributed chiefly by

700 Interaction in social groups the mother's intrusive behavior toward the first-born girl. There is a strong similarity between these findings and those of Cushna (1966), reported by Sutton-Smith and Rosenberg (1970). Cushna observed 16- to 19-month-old children with their mothers. When mothers were asked to direct their child's performance on several tasks, the mothers were found to be more active in influencing the performance of their firstborn children, but in different ways for boys and girls. Mothers were reported to have been more supportive, strategic, and cautious in directing first-born boys, but more demanding, exacting, and intrusive toward first-born girls. Effects relating to birth order and sex have thus been found both in mother's behavior toward 4- and 5-year-olds and in mothers' behavior toward children under the age of two. Can birth order effects on maternal behavior be found with still younger children? Thoman and her collaborators (Thoman et al. 1970, 1971, 1972) have reported important data on this question. In their first study of neonate-mother interaction (Thoman, Turner, Leiderman and Barnett 1970) these investigators studied bottle-feeding patterns of mothers with their 2-day-old infants. They found that the total time the mother spent using the bottle, including both feeding and non-feeding activities, was longer for mothers of firstborn infants, even though their infants consumed less than the laterborn infants. Sex differences were not examined in this study. In a second study (Thoman, Barnett and Leiderman 1971), motherfeedings were compared with nurse-feedings of the same infants. While no differences were found due to birth order or child when the infant was fed by the nurse, time measures were greater for the mothers of first-born infants than for mothers of later-born infants on all measures (although not significant for this smaller sample). On one variable, number of feeding intervals, the direction of differences was reversed. Mothers used more intervals while feeding first-born infants, while nurses used more intervals while feeding later-born infants. The authors concluded that the differences in behavior between mothers of first-borns and later-borns were due more to the mother than to characteristics of the infant. Again, sex differences were not investigated. In a third study (Thoman, Leiderman and Olson 1972), primiparous and multiparous mothers were observed during breast-feeding. In this study, mothers of first-borns, in comparison with mothers of later-borns, spent more time in non-feeding activities, fed male infants for a greater percentage of time, changed activities more frequently, stimulated their

Sibling influences on the development of the individual 701 infants more, talked to the first-born infant more, especially to the firstborn female, and smiled more at the first-born female. Talking to the child at the 2-day feeding correlated significantly with the length of time in weeks that the mother continued to breast-feed her child. Since mothers also talked more to their first-born daughters, this correlation may have resulted from mothers both nursing and talking to their first-born daughters for longer periods. Thoman et al. (1972) note parallels between the mother's behavior with the neonate and findings with older children. They suggest that the greater number of feeding intervals and frequent changes of activity during feeding indicate greater inconsistency or variability in the behavior of mothers of first-born children. They also conclude that: 'The results of the feeding studies... provide evidence that the primiparous mother interferes with or dominates the feeding interaction to a greater degree than the multiparous mother' and that 'the parity of the mother differentially affects interaction with male and female infants during the earliest days'(p. 117). Summary of research findings. On the basis of these studies of motherchild interaction, it is possible to point to some consistencies in findings, and to speculate about mechanisms that may contribute to the results. First, there has been a general finding of greater intrusiveness and control of the mother in the performance of her first-born (Hilton 1967; Stout 1960; Lasko 1954) with this intrusiveness heightened in the case of the daughter (Cushna 1966; Rothbart 1971; Thoman et al. 1972V Second, more variability has been noted in the behavior of the mother toward the first-born child, both in a single interaction sequence (Hilton 1967; Thoman et al. 1972) and over time (Lasko 1954). There has also been some evidence to indicate that the mother tends to more often criticize the first daughter and tell her when she is incorrect in an achievement task (Cushna 1966; Rothbart 1971). There are of course limitations in the generality of these findings: Mothers in these studies have been for the most part middle-class Americans, the ages of children studied have not exceededfiveyears, and father-child interaction has not yet been adequately observed. The situations in which mother-child interaction were observed have also for the most part been limited to feeding and achievement tasks. The age range does, however, include the period during which motherchild interaction is probably most frequent and intense, and interactions during feeding and problem-solving are undoubtedly important ones for the parent and the child.

702 Interaction in social groups As Sampson (1965) has pointed out, finding differences as a function of birth order is not in itself nearly as interesting as trying to determine the psychological experiences and family experience that may lead to such differences. Sampson was referring to birth order differences reflected in characteristics of the child or young adult, but his argument may also be related to differences in the parent's behavior as a function of the birth order and sex of his or her child. Bell (1971) and Harper (1971) have given extensive support to the view that characteristics of the infant are significant contributors to the caretaker's behavior, and birth order and sex of child appear to be two such characteristics, at least for the mother. Maternal experience. In what ways might her parity influence differential behavior of the mother to her child? The most obvious intervening variable is the experience of the mother. Harper (1971) has reviewed research on several mammalian species indicating that prior experience with an infant enhances the mother's sensitivity to her infant's care-eliciting cues. Thoman et al. (1970, 1972) comparisons of primiparous and multiparous mothers during infant feedings provide additional evidence on this point: Experienced mothers were more likely than mothers of first-borns to respond appropriately to sucking cues of the infant. The mother's experience is not sufficient, however, to account for some additional findings. Although mothers of first-born infants tended to use more feeding intervals than mothers of later-born infants in the study by Thoman et al. (1971), both sets of mothers used more feeding intervals than did experienced nurses, and the mothers spent a far greater percentage of time in non-feeding activities than did the nurses. Differential maternal behavior as a function of sex of the first-born also could not be explained in terms of parental experience in mothering. Maternal involvement. The argument to be proposed here is that an additional influence on the mother's behavior in these interaction studies may be the degree to which the mother identifies with or is involved with her child's performance. In developing this proposal, I will first suggest that a parent is more concerned and involved with the adequacy of performance of his or her own child than with the performance of a strange child. There is support for this prediction in a study by Halverson and Waldrop (1970), where mothers of nursery school children helped administer simple tasks to their own and to another child. Mothers used a greater percentage of positive and encouraging remarks to the other child, but more negative and controlling statements to their own child. Halverson and Waldrop

Sibling influences on the development of the individual 703 suggest this interpretation of the mother's behavior with her own child: 'First, the interaction situation was, no doubt, perceived by a majority of the mothers as an achievement situation in which it was important for their own children to do well. Second, mothers when out "in public" with their own children generally tend to assume responsibility for their children's performance. As a result of these pressures, mothers might have been more task-oriented with their own children, as evidenced by the more negative and controlling environments which they provided. With other children, maternal involvement in child performance was probably lower...'(p. 844). To take the involvement hypothesis one step further, we might predict that parents will be more involved with the performance of the first-born than the second-born child for at least three reasons: (a) The first-born child is continually confronting situations that are new to the parent, and many of them are public ones (for example, starting school for the first time). With a later-born child, the parent will be less concerned and involved with his child's performance due in part to greater familiarity with possible demands of each situation; (b) to the extent that the parent of the later-born has a wider range of child behaviors from which to make judgments about his or her child, the parent may be more tolerant about the later child's behavior, and less likely to intercede or try to control it; and (c) at least initially, the parent simply has more time to spend with the first-born child. Halverson and Waldrop's interpretation might thus be extended to say that degree of maternal involvement is reflected in the mother's exerting greater control and directiveness toward the first-born. The involvement variable may also be extended to the greater intrusiveness of the mother during the performance of her first daughter. It is possible that of all ordinal positions, the mother identifies herself most closely with the first-born girl, since the first-born girl of all her children would most resemble the mother at any given time. This could lead to both greater involvement in the success of her daughter, and a tendency to be critical of mistakes in this child's performance (Cushna 1966; Rothbart 1971). The mother's identification would be less for a son of any birth order, while a second daughter with an older sister would also receive less involvement from the mother. The involvement-identification hypothesis predicts that studies of father-child interaction would show the greatest intrusiveness and control toward the first-born son. The hypothesis also predicts that in families with a first-born son and a second-born daughter, birth order effects on the mother's behavior would not be as strong as those

704 Interaction in social groups found in same-sex, two-child families. A second reason for studying fatherchild interaction relates to an alternative explanation for results reported here - the possibility that the mother's differential behavior to her sons and daughters may only reflect her reaction to possible early sex differences in child behavior. This interpretation would not predict differential father behavior in the opposite direction from that of the mother. The author is pleased to add that the first extensive study of fatherchild interaction with infant boys and girls has been completed and is reported at this conference by Parke. In consistency with the parental involvement hypothesis, Parke reports that fathers of newborns (alone or in the presence of the mother) touched first-born boys more than either laterborn boys or girls of any ordinal position. Fathers also vocalized more to first-born boys than to first-born girls, while vocalizing equally to laterborn infants of both sexes. Direct comparison of father's behavior with mothers' behavior would be necessary to detect whether cross-parent interactions between sex and ordinal position of child were present. Evidence has been reported suggesting that sex differences exist in parental responses to their children, with parents being more controlling with a child of the same sex (Emmerich 1962; Kagan 1965; Osofsky and O'Connell 1972) and more permissive with the child of the opposite sex (Rothbart and Maccoby 1966), but more work needs to be done. Crosscultural research would be of definite interest: In cultures where achievement is not highly valued, parental involvement would likely be evidenced in different ways than through intrusiveness into the child's performance. Sibling relations. We may now speculate about possible effects of differential maternal involvement on sibling social development. One possible consequence is that while the first-born child (especially the daughter) has a more intimate experience with the mother in her socializing role, the later-born sibling, always encountering new social situations at a later point than the first-born, will have learned a good deal by observing parent interactions with the first-born child. In any socializing setting, the later-born child will have both previous observational experience and first-hand experience, and as a result might be expected to handle social situations with greater skill than the first-born. This advantage would of course vary with the degree of age spacing between siblings, with the closely spaced second sibling having the greater social advantage. On the other hand, if the first-born's experience is with a more powerful and controlling parent, modeling studies (e.g., Bandura, Ross and Ross

Sibling influences on the development of the individual

705

1963) suggest that the first-born child would be more likely to model the behavior of the parent than would the later-born child, leading to greater similarity in first-born and parental characteristics and values (Kammeyer 1967; Sutton-Smith and Rosenberg 1970). If the later-born child has the advantage of a more objective and independent view of interaction with the parent, he or she must also continually cope with the usually greater size and verbal skill of the first-born sibling, as well as the possibility that the parents are more involved in the exploits of the first-born. Differences in these possible sources of sibling power seem to be reflected in the results of a study by Sutton-Smith and Rosenberg (1968). They asked fifth- and sixth-grade children how they controlled the behavior of their siblings, and how their siblings controlled the child's own behavior. Both first- and second-born children agreed in describing the first-born as more bossy, while the second-born (with the exception of the physically assertive second-born male with an older sister) was described as more likely to ask other children or the parents for help. Possible modeling effects (or differences in actual role assignment by parents) were also found for the first-born girl, who was described as more likely to explain, give reasons, or to ask the sibling to do what she wanted. These results are intriguing and should promote additional research on home and structured laboratory interaction between siblings. Possible effects of greater parental involvement with the first-born may also be seen in different personality ratings of siblings within families. Dean (1947) found that mothers of same-sexed pairs of children more often described their first-born child as dependent, fearful, sensitive to having his or her feelings hurt, and less demonstrably affectionate, while the second-born was more often described as stubborn, physically aggressive, happy, and good natured. McArthur (1956) reported a study in which two generations were asked to describe their children and compare one sibling with another. He also obtained psychiatric descriptions of his Harvard Ss and asked them for self descriptions. Across all of these data, the first child tended to be described as sensitive, adult-oriented, serious, and conscientious, while the second child was more often described as peeroriented, easy-going, friendly, and independent. Finally, I would like to comment on the special relation of the firstborn daughter with the mother, and to move from considerations of research to more practical applications to child-rearing. If the position of the first-born daughter is especially susceptible to modeling influences, it is likely that the middle-class first-born girl has probably experienced intense

706

Interaction in social groups

value conflict in recent years between her attitudes toward achievement and values of feminity. Kammeyer (1966) in fact found that the first-born girls in his sample of college coeds tended to be more traditional in orientation toward the feminine role and more traditional in beliefs about psychological traits in women than did later-born women. Changes in society's expectations and attitudes about sex roles may have definite liberating tendencies for first-born girls, especially those from all female families. While the later-born daughter, less closely tied to the mother, probably has possibilities for greater independent development than her older sister, the social skills she may learn as a second child will also especially suit her to adopting an expressive feminine role. Perhaps the sensitive parent today should be careful not to let the later-born daughter rely too heavily on charm and attractiveness in getting what she wants; at the same time the parent might attempt to encourage instrumentally competent behavior in both first- and later-born daughters, and welcome the expression of emotion in both first- and later-born sons.

Sibling influences on the development of the individual 707 B. ON THE EXTENT OF SIBLING INFLUENCE by Walter Toman, University Erlangen-Niirnberg, Germany

Sibling positions are roles imposed upon a person by the sibling configuration into which he happens to have been born. These roles can amount to relatively pervasive behavior patterns, particularly with respect to a person's social behavior and social preferences. I have outlined perhaps more explicitly and systematically than others before me (Adler 1920, 1929; Koch 1956) what the basic sibling positions are in the first place and how they tend to form at least certain portions of a person's personality (Toman 1959, 1962). I may even have overdone it at that time, but since then ample evidence has suggested that there can be no doubt about the influence of such enduring environmental characteristics as the persons living with oneself under the same roof, eating 'from the same table', interacting with oneself steadily, day after day, from earliest life on. We have proven some of these characteristics to be effective 20 and 30 years after they began to exercise an influence, also 5, 10 and 15 years after a person had left his original family and thus in most instances strongly reduced the physical and even the psychological presence of his family members in his own life (Toman 1969, 1971; Toman and Preiser 1973). We distinguished eight basic types of sibling position. All other and more complicated sibling positions are composed of one or more of these eight types. These types are oldest brother of sisters, oldest brother of brothers, youngest brother of sisters, youngest brother of brothers, oldest sister of brothers, oldest sister of sisters, youngest sister of brothers and youngest sister of sisters. Empirical studies of families. The social behavior to which a sibling position predisposes a person vis a vis peers can be and has been characterised, among other things, along the dimensions of dominance of others and responsibility for others versus dependence upon others and carefree behavior. Oldest siblings tend to show the former, youngest siblings rather the latter. Moreover, social behavior is likely to display a predilection for contact with peers of the same sex or of the opposite sex depending on whether the person had same-sex or opposite-sex siblings, respectively, or both. We have found that persons with opposite-sex siblings tend to be more at ease with peers of the opposite sex than of the same sex, persons

708

Interaction in social groups

with same-sex siblings with peers of the same sex than of the opposite sex (Toman 1962, 1971, 1972b). We also found that single children who, by definition, have no sibling position, are more eager to associate with parent-like persons than with peers, or with peers willing to assume the roles of parents or seniors. We have shown too that the relationship, of parents to their children are not unaffected by the sibling roles of parents and those of their children. It is easier for a child to identify with his (same-sex) parent and for the parent to identify with his (same-sex) child if they have similar or identical sibling positions rather than opposite or complementary sibling positions (Toman 1962; Gasch 1967; Kurz 1969; Erzigkeit 1970). On the other hand, it is easier for a child to interact and interrelate with his (opposite-sex) parent and for a parent to interact and interrelate with his (opposite-sex) child, if they have opposite or complementary sibling positions rather than similar or identical sibling positions. There is also evidence that the problem child in a family, more often than chance, is a child who has either an identification-conflict with the same-sex parent or an interaction- and interrelation-conflict with the opposite-sex parent, or both (Toman 1969). There are other reasons, to be sure, for a child to become the problem child of the family, but compatibility of sibling positions of parent and child appears to be among them. We have also demonstrated that characteristics such as age distance of parents from each other as well as from their children, age range of siblings within a configuration of children, age distance from next-older and nextyounger sibling, the balance or imbalance of sexes in a sibling configuration, all tend to modify the trends that I have outlined. Even a person's uncles and aunts and their children, i.e., a person's cousins, are not irrelevant to his own family constellation and its effects upon him. We have found, e.g., that special groups of the population such as families of delinquents, of criminals, and of persons with neurotic problems differ significantly from the population of families at large. Step-siblings and half-siblings were more frequent with those special groups than in the population. The parents of neurotic youths had fewer siblings, those of delinquent and criminal youths more siblings than the parents of youths of the population at large. More siblings of parents of the special groups remained unmarried or had no children than expected. In the population an average of two out of three siblings of the parents are married, 85% of those married siblings have children, on average two of them (Toman, Preiser, Gasch and Plattig 1967; Toman and Preiser 1973).

Sibling influences on the development of the individual 709 Extraneous factors. Besides all these characteristics of family constellation, which should not be disregarded in any study of sibling influences, there are extraneous sources of variation. More or less hereditary features such as looks, height, physical strength, intelligence, vitality, or overall frustration tolerance manifest themselves independently of sibling positions. To give an example, we found that, by psychological tests, oldest siblings appear somewhat less intelligent and youngest siblings somewhat more intelligent than their school grades seem to indicate. On the other hand, oldest siblings get somewhat better grades, on average, than do youngest siblings. The net balance is no true difference of intelligence among different sibling positions. Apparent differences may be attributed to different motivations rather than different intelligence (see also Toman 1972a). In other words, there is little doubt in my mind that heredity works independently of sibling position. While the intelligence quotients of siblings are positively correlated with each other and with those of their parents (e.g., Newman, Freeman and Holzinger 1937; McNemar 1938; Skodak and Skeels 1949), variations of intelligence within a given sibling configuration appear to be random. Losses of family members. Consideration of characteristics that may affect a person's social behavior besides sibling position itself should always include losses of siblings and/or parents that have occurred in a given family constellation. Trivial as it sounds, there is reason to remind researchers in the field that a dead sibling is not the same as a live sibling. A sibling position may change after another sibling has dropped out for good, or if a parent has died or left the family permanently. We were able to show that age of person lost, age of oldest person in the nuclear family, age of subject at time of loss, time lived together with person lost, change of balance of sexes due to loss, size of family, time that has passed since loss occurred, and number of preceding losses suffered by subject, all bear on the overall psychological effect of loss. As for the effects themselves, we were able to establish, e.g., that early losses suffered by one or both partners tend to delay marriage by about one year and to reduce the number of children that they want or at least that they get (Toman, Preiser, Gasch and Plattig 1967). We were surprised to find even that, judged by year-end school-grades, those children of gradeschool who had lost a parent were significantly poorer performers than the average, whereas those children who had not suffered a loss themselves, but one

710

Interaction in social groups

of whose parents had suffered an early loss of a parent were significantly better performers than the average (Zielinski 1966; for other effects see Toman 1969; Toman and Preiser 1973). Temporary losses of family members (due to military calls of duty, extended travel, prison terms, illness etc.) turned out to follow the same rules as permanent losses. Partial losses, on the other hand, i.e., losses of certain (positive) aspects of personality of a family member (e.g., discovery that father is a drunkard or a coward, that mother had had an illegitimate child, etc.) are more difficult to appraise and compare. Sibling position and nursery school behavior. In order to illustrate how complicated the effects are that may be expected when children enter a new context, a little field study shall be reported. Nursery school behavior of children was observed and rated once a week by psychologists and the nursery school teachers themselves over a period of 3 months with classes of 3-, 4-, and 5-year-old children. Afterwards family characteristics including sibling positions had been established for all children by questionnaire. We found that oldest siblings tended to lead others and initiate activities in the 3-year-old group, whereas among the 4- and 5-year-olds youngest siblings appeared to be leaders and initiators more frequently than chance. We interpreted this to mean that oldest siblings succeeded at first with the accustomed role of leadership and responsibility, but other children eventually caught up with them and (after one or two years in nursery school) came into their own. Youngest siblings are perhaps particularly prone to sense their chances, since the oldest siblings in nursery school, their rivals for leadership, are their own age, whereas at home they are 2, 3 and more years older. This, together with identifications with their oldest siblings at home, may act to the advantage of youngest siblings in nursery school. Furthermore we found that by ratings and observer counts of actual contacts with other children those subjects who had no siblings of the opposite sex preferred contacts with children of the opposite sex, whereas children who had no siblings of the same sex tended to prefer contacts with children of the same sex. In other words, in both respects our simple prediction that children would do in a new context what they were used to at home did not quite bear out. On the contrary, children seemed to vie for those experiences in nursery school that were not available to them at home.

Sibling influences on the development of the individual 711 We suspected that this result may have been due to this particular nursery school. It was an experimental school on a college campus with relatively liberal teachers and progressive curricula. So we checked with another nursery school known to be stricter and more conservatively run. Only teachers' ratings were available of the children's behavior in that school. Checking the sibling positions of those children by separate inquiry we learned that oldest siblings of both age groups (4- and 5-yearsolds) showed a relatively higher frequency of leading and initiating activity, but also of obedient behavior, than did other sibling positions, especially youngest siblings. In both groups, moreover, children showed a preference for the sex they were used to at home in their contacts with other children. Thus we may say that in a more restrictive new environment children tended to apply the patterns of social behavior practiced at home. They did not try out new patterns and new experiences. Sibling position and individual friendship patterns. We have evidence, however, that even much later in life and in a most liberal context such as college the old patterns of behavior, in particular the sibling roles do still prevail to an extent and display their effects. College students were asked to list their friends and acquaintances as they came to their mind. Afterwards they had to provide six different types of information about the persons they had listed. Among them were the sibling positions of those friends and acquaintances. Checking for compatibility of sibling positions of subject and friends listed we found that males had selected more frequently than chance friends of both sexes whose age ranks were complementary to their own, and that they had listed relatively more friends and acquaintances of that sex which they themselves had been exposed to in their original families. When they had had no siblings of the opposite sex, they had listed relatively fewer friends of the opposite sex, when they had had no siblings of the same sex, they had listed relatively fewer friends of the same sex. In absolute terms all had tended to list more persons of the same sex than of the opposite sex. Female subjects displayed only random choices of friends (Toman 1972b). This trend observed with male subjects was more clearly apparent with male and female counselees and patients at student counseling centers and at child guidance clinics. Among those persons whom a subject mentioned spontaneously and whom he or she considered friends during 1 to 20 interviews persons with complementary sibling positions were significantly more frequent than chance.

712

Interaction in social groups

Clinical significance of family characteristics. Such findings have led me and associates to check routinely and unobtrusively for family characteristics of friends that a person mentions in clinical interviews and psychotherapy sessions. Patients don't mind. If done properly, this registers with them as an indication of the therapist's interest. These family characteristics of friends are part of what we consider necessary in any case history: A reconstruction of the most basic objective data about all those life contexts in which the subject has participated. We think we learn from those friendships how a subject has interpreted his family constellation, which of the options that it oifered him he seems to have preferred. A middle brother of two sisters, e.g., will indicate by the choice of his girl friends and by their sibling positions (of which he may be unaware consciously), whether he got along better with his older or with his younger sister, or perhaps best with his mother (and her sibling position). Theoretical foundations. All these effects of sibling positions and family constellations do not come about miraculously. They are the result of complicated, global learning processes. People tend to transfer experiences made in familiar social contexts to new social contexts. We could also say, they generalize from old contexts to current ones. They view any new situation automatically in the light of the most similar and familiar situations of the past. They learn before long, of course, in what ways the new situation differs from similar and familiar situations encountered in the past. That way the current situation becomes familiar in turn and may serve as a sample along with older situations when another new situation comes up. All this suggests conglomerate effects. We should expect, however, that older situations, especially those that have been encountered more frequently, more regularly, and earlier in a person's life, are likely to exert a greater influence upon the views, attitudes, and preferences shown in new situations than would those situations that have occurred less frequently, less regularly and/or only later in a person's life. I can think of several reasons for that. One is time which 'moves' in one direction only. It is logically impossible that recent events affect earlier events. It is psychologically possible only in the sense that experiences with recent events can change a person's views of earlier events. They cannot change those earlier events themselves, to be sure, but a new view of an early event might constitute a different kind of influence of that event on a new situation.

Sibling influences on the development of the individual

713

Generally speaking, however, this does not happen too often. It is one of the objectives of psychotherapy. Another reason is learning. Those situations that have occurred most frequently, regularly, and early in life are more likely to have been practiced to perfection than are less frequent, less regular and/or more recent situations. Better practiced patterns of behavior or roles are likely, in turn, to be more readily available in new situations than behavior patterns or roles that have been practiced less well. A third reason may be that people who are party to a given social situation that happens to be new to the person in question tend to respond to the initial behavior that this person shows. A boy who has learned to boss the girls in his family is likely to try the same in school. Some girls there won't mind, whereas others might quickly learn how to avoid him or even to do him in. He is likely to stick with those who don't mind and to skirt the others. Thus the others are perhaps as prone to perpetuate a mode of social interaction as a given person is himself. This holds even for family life itself. A person does not chose his role in the family. It is more or less assigned to him by the others. If he tries to change it, the others will object at first, or even forever. Quantitative estimates of the effects of social contexts. Let us assume that learning occurs in any situation that a person finds himself in regardless of what there is to learn (see Toman 1968). Let us assume, moreover, that duration of exposure to a situation is the variable that determines, perhaps more than others, the effect of any situation upon a person's choices of new situations. If we do that we could appraise the effects of a situation by estimating the time of a person's exposure to it in relation to his life-time (or age) and in relation to other situations that he has been exposed to. This applies to persons who have lived with a subject for different lengths of time and in different 'densities'. A person who has spent an average of five hours per week with a child would have a negligible effect compared to a person who spent 50 hours a week with the child. The time spent by a baby nurse and that spent by the baby's working or madly socializing mother may be in a relation of 10 to 1, and we should conclude that the baby nurse is likely to impress herself and her ways on the child much more than the mother can. Ordinarily, it is the mother who spends that much more time with her child than any other person. If a child's waking time increases and the time that the (ordinary)

714 Interaction in social groups mother spends with him decreases, other persons may become more influential, mother's influence relatively smaller. The sum-total of her influence over the child's lifetime, however, will remain the largest of all persons in the child's environment, at least up to a certain age and as reflected by such time counts. We could also appraise the effects of entire contexts such as home, school, playground, open country, church, etc., as to their temporal distribution in a growing person's life. The effect of the context home (E h ), if it were the only context to which the child has been exposed during his first 3 years of life would be time spent at home (TH) divided by time lived so far (T): EH = TH/T = 3/3 = 1. The effect of home has been maximal. If, during the next 3 years the child spends 20 % of his waking time in nursery school and kindergarden (Ps = 0.2) and, during his 4th to 6th year of life, perhaps 10 % on the playground (P p = 0.1), but the rest of his time at home, we could appraise the effect of home during the years 4 to 6 as E H = (TH/T) P h = (3/6) 0.7 = 0.35. The effect of nursery school plus kindergarden would be ES = (T s /T)Ps = (3/6) 0.2 = 0.1, and the effect of the playground E P = (T P /T) P p = (2/6) 0.1 = 0.03. The cumulative effect of home at the end of the first 6 years of life could be expressed as E H = (1/2) (EHL + E H2 ) = (0.5) (1 + 0.35) = 0.68. The actual relationships between exposure to situations and their effects may be much more complicated, but even these simple ratios seem to reflect psychological realities. The effects themselves can be observed, say, in a person's choices of permanent friends or of a marriage partner, in his decisions to have children, in his choice of a professional position, of a geographical location, even of a hospital in case of illness, etc. The persons to whom he has been exposed the most are likely to exert the strongest influences on his choices. The contexts that have surrounded him the longest are likely to guide him in the selection of his new contexts, in shaping them in line with his old contexts, even in literally returning to the old contexts when he has a chance. Conclusions. I would like to repeat that sibling position is but one of many characteristics of a persons's psychological environment that bear on his social behavior. Wherever possible those other characteristics should be heeded too in any study of the effects of sibling position, even if the study gets more cumbersome that way and larger samples of subjects might be needed to reach significant results.

Sibling influences on the development of the individual 715 C. SIBLING INTERACTION AND COGNITIVE DEVELOPMENT by Victor G. Cicirelli, Purdue University, Lafayette

Previous research, in particular the pioneering studies of Helen Koch and the more recent work of Rosenberg and Sutton-Smith, has amply demonstrated that the cognitive development of a child is related to his position in the sibling structure1 of the family. However, just how such effects are produced is not yet clear. Many investigators explain the relationship of sibling structure to cognition, at least partially, in terms of differing parental treatment, or differing parent-child interactions, depending on the position of the child in the sibling structure. But sibling-sibling interactions may also be an important factor. (Within the family, interactions between family members may be grouped into three subsystems: parent-parent interactions, parent-child interactions, and sibling-sibling interactions. Certainly the latter must be considered in understanding the development of a child with siblings.) Sibling interaction means the reciprocal interchange of verbal and nonverbal communication between siblings. If siblings indeed have an influence on each other, independent of parental influences, then this influence should be manifested in the everyday interactions between siblings in the home. Assuming that a sibling pair has developed a moreor-less stable and characteristic mode of interaction, then by observing what goes on when this sibling pair interacts in a controlled situation, one should be able to draw some conclusions about whether there is direct sibling influence and how such influence works. At least, this was my hope at the start of these investigations. 1. Two children have a sibling relationship when they share the same parents. The sibling structure of a family is the network of positions for children in the family, defined by the number of children, birth order, sex of the children, their ages, and the age spacing between the children. The sibling status of two or more children is their poisition within the network. For some researchers, sibling status is identified with one dimension of the sibling structure, e.g., birth order. Research is then carried out to establish the relationship between that dimension and cognitive-intellectual or social-emotional variables (both short- and longrange effects). Other researchers attempt to establish the relative strengths of different dimensions of the sibling structures (e.g., birth order vs. number of children) as predictors of certain outcomes. In my own work, sibling status is defined by age, sex, birth order, and age spacing taken simultaneously (rather than by one dimension acting alone).

716

Interaction in social groups

Up until this time, there had been little or no direct study of sibling interaction, although the negative effects of sibling rivalry have been discussed for many years. In one study of siblings, Sutton-Smith (1966) asked fifth-grade children what games they played with their sibs and with their nonsib playmates, and who was usually 'boss' in the play. He found that first-born children took high power roles with their younger siblings, and equal or lower power roles with their friends, while later-born children took low power roles with their older siblings but high power roles with their friends. Sutton-Smith and Rosenberg (1970) also reported the results of an interview-questionnaire administered to upper-grade elementary school children in regard to tactics to get their sibling to do what they wanted him to do. Boys used attack and offense more often, while girls used reasoning, defense, and making the sibling feel obligated; certain tactics were more typical of first-born than of later-born children. Also, same-sex siblings used more powerful tactics than did opposite-sex siblings. Thus, there is some evidence that children in different status positions interact differently with their siblings; these studies, however, do not attempt to determine what the outcome of the interactions is in terms of the other sib's behavior. Irish (1964) has said: 'Interactions with siblings function as one avenue for the socialization of children ... Sometimes siblings are more effective teachers than adults, particularly if youthful skills are involved ... Siblings may serve as role models for one another; particularly may the younger observe the older siblings of the same sex. They can serve as challengers and stimulators' (p. 282). In studying sibling influences on cognitive development, I have looked at the effect of the older sibling upon the younger (recognizing, of course, that there may well be effects in the other direction). However, since much of cognitive development is cumulative in nature, the older sibling occupies a position in which he possesses greater knowledge and experience about most topics. With this in mind, in the two studies I have completed thus far, I have observed the interactions between a pair of siblings as the older child attempted to teach or help his younger sibling in a cognitive task situation, and then measured the younger child's subsequent task performance. In looking at the data, I tried to answer the following questions: (1) Do older siblings influence their younger sibling's performance on a cognitive task? (2) If so, does this influence depend upon sibling status? (3) Do the observed behaviors of the sibling pair as they interact with each other in the task situation depend upon

Sibling influences on the development of the individual

717

sibling status? (4) Do different observed interaction behaviors lead to differences in younger siblings' task performance? Concept learning study. The first study (Cicirelli 1972) investigated sibling interaction as the older child taught a concept to the younger child. Each of the 120 first-grade children in the study had an older sibling in third grade. (Number of children was not held constant in this study: The mean number was four.) Equal samples of 30 sibling pairs were drawn from populations of boys with older sisters, boys with older brothers, girls with older sisters, and girls with older brothers. For half the children in each group, the older sib served as the teacher of his or her younger sib; the remaining half were re-paired so that the older child taught an unrelated first-grade child. The experimenter trained the older child in the trapezoid concept to a given learning criterion in a 30-45 minute session. A standardized teaching procedure involving a variety of teaching techniques was used. Then the older child was asked to teach the trapezoid concept to the younger child. The interaction between the two children was recorded on an observation schedule by the experimenter as well as on magnetic tape. Following this 10-minute teaching-interaction session, the younger child was given a concept attainment test to determine his mastery of the trapezoid concept. The results of an analysis of variance in the 2 x 2 x 2 factorial design carried out for the learner's concept attainment scores provided an answer to the first two of the general questions. Older siblings did influence their younger sibling's performance on this concept attainment task, but this influence depended on sibling status. To be more specific, older sisters were significantly more effective teachers of younger siblings (boys or girls) than older girls were of unrelated younger children (boys or girls). On the other hand, older boys tended to be more effective teachers of unrelated younger children (boys or girls) than older brothers were of younger siblings. Viewed in another way, older sisters were significantly more effective teachers of younger siblings than were older brothers. Older boys and girls were not significantly different in effectiveness as teachers of unrelated younger chidren, however, even though those unrelated children occupied the same sibling status positions as the older children's own siblings. The highest-scoring subgroup of the study was the one in which older sisters were teachers of younger brothers, while the lowest-scoring subgroup was the one in which older brothers were teachers of younger sisters.

718 Interaction in social groups Chi-square tests of association were used to determine whether observed interaction behaviors led to differences in the younger sibling's concept attainment score (classified as high or low), and whether these differences depended on sibling status. It was found that girls teaching their sibs used a deductive teaching method and its associated teaching behaviors (explaining and describing, demonstrating and illustrating attributes, and selection of concept examples) more than did other groups. When the two extreme subgroups were compared, older sisters teaching younger brothers tended to use the deductive approach, while older brothers teaching younger sisters tended to use an inductive approach. (In an inductive approach, the teacher gives many examples of the concept and the learner must abstract the rules with varying degrees of guidance by the teachers, while in the deductive approach, the teacher gives and explains rules followed by examples.) There were, however, no specific interaction behaviors which were related to task performance in a general way. Object sorting study. Because the concept teaching task used in the first study was highly structured and allowed only a rather narrow range of behaviors to be displayed in the teaching-interaction session, a second study (Cicirelli 1973a, 1973b) was carried out using an object sorting task in which there were no 'right' answers. It was hoped that this would provide a better vehicle for studying sibling interaction, in that the children would be freer to react to the task and to each other than in the concept learning study. In contrast to the earlier study, this study selected only children from two-child families; also, while age of the siblings and age spacing between them were held constant in the first study, they were varied in this study since they were found to be important factors in research on sibling structure (Sutton-Smith and Rosenberg 1970). From this research, one would expect a child's siblings to have the greatest effect on him in the preschool and early elementary years, with the effect diminishing as he gets older. Similarly, a close age spacing is held to be beneficial for girls, but detrimental for boys. However, when the older sib is helping the younger on a cognitive task, the younger sib might well view the widespaced older sibling as a more powerful model or a more effective source of aid. A 2 x 2 x 2 x 2 x 2 factorial design was used, with four sibling structure factors and one treatment factor. There were 160 sibling pairs

Sibling influences on the development of the individual

719

in all, 40 from each of the following groups: kindergarten children with second-grade siblings, kindergarten children with fourth-grade siblings, second-grade children with fourth-grade siblings, and second-grade children with sixth-grade siblings. Each of these groups of 40 contained 10 sibling pairs from each of the four sex combinations (boys with brothers, boys with sisters, girls with brothers, and girls with sisters). The cognitive task was an object-sorting task, in which the children were asked to form groupings of 50 pictures of familiar objects. Half the children in the study were assigned to an 'alone' condition, where the child made an initial or practice categorization of the pictures of the object sorting task while alone; the remaining half of the children were assigned to a 'sib' condition, where the child made the initial categorization with the help of his sib. The experimenter observed and recorded the interaction between the sibling pairs in the 'sib' condition as they worked on the sorting task. Measures of the younger sib's task performance were obtained from a subsequent 'test' administration of an alternative form of the object sorting task. These included measures of categorization style (number of groups formed, size of groups, number of ungrouped items) and measures of conceptual style (type of category formed, such as descriptive, relational, and inferential). The results of the study were complex, involving a number of interactions between variables in the analysis of variance. I will bring up only those results with bearing on the question of sibling influence here. First, the results provide evidence that older siblings do influence their younger sibling's performance on an object sorting task, and that this performance depends, at least in part, on sibling status. In relation to categorization style, children in the 'sib' condition made more groups than did children in the 'alone' condition. Similarly, on the average, children in the 'sib' condition left fewer items ungrouped than children in the 'alone' condition; however, this finding depends on the age spacing between the siblings. Among children with sibs four years older, those in the 'sib' condition left fewer items ungrouped than those in the 'alone' condition, while among children with sibs two years older, there was little difference between those in the 'sib' and 'alone' conditions. In regard to group size, among children with sibs four years older, those in the 'sib' condition made larger groups than those in the 'alone' condition, but among children with siblings two years older, those in the 'sib' condition made smaller groups than those in the 'alone' condition. Looking at the conceptual style variables, the older sib had no direct

720 Interaction in social groups influence on the younger sib's use of relational categories, but there was an influence on the use of descriptive and inferential categories which depended on sibling status. Among kindergarten children, those in the 'alone' condition used a higher percentage of descriptive categories than those in the 'sib' condition, while among second-grade children there was little difference between the two conditions. Among children in the 'sib' condition, girls used a higher percentage of inferential categories than did boys when their siblings were four years older, but boys used a higher percentage than did girls when the sibs were two years older; among children in the 'alone' condition, these relationships were reversed. To answer the remaining two major questions (whether differences in observed interaction behaviors lead to differences in the younger sibling's object sorting performance, and whether these differences depend on sibling status), an analysis of behaviors observed while the sibling were interacting in the object sorting task situation was carried out. The interaction behaviors were coded into 29 categories; 18 of these were significantly correlated with the object sorting measures (correlations were weak to moderate in strength). These behaviors were: Older sib explaining, telling younger sib what to do, giving cues, questioning, giving category name, encouraging, criticizing, giving confirmation of a group younger sib makes, adding to younger sibs group, making a group himself, nonverbally encouraging and criticizing, and rearranging groups made by the younger sib; younger sib verbalizing his actions as he sorts, accepting direction from the older sib (verbally and nonverbally), rejecting help, showing objects to the older sib, and working independently. Point-biserial correlations were computed between the sibling status variables and the frequency of the sibling interaction behaviors. Sex of the younger sibling was significantly related to two of the older sib's behaviors such that the older sib showed a greater tendency to give cues or hints and to verbalize his actions when the younger sib was a girl. Sex of the older sib was related in such a way that older sisters showed a greater tendency to point to objects for grouping, add objects to the younger sib's grouping, and to observe the younger sib working, while older brothers showed a greater tendency to verbalize their actions. The younger sibling was more likely to accept direction (nonverbal) when the older sib was a girl, and was more likely to work independently of the older sib when the older sib was a boy. The older sib was more likely to make groups himself when the younger sib was of kindergarten age rather than a second grader, while the second-grade child showed a

Sibling influences on the development of the individual 721 greater tendency to verbalize his actions than did the kindergarten child. Finally, the younger sib showed a greater tendency to accept direction when the older sib was four years older rather than only two years older. I could go on to list the interaction behaviors that are correlated with the more complex sibling status positions, but they would do little more to illuminate the situation. In an attempt to simplify this detailed mass of data, I carried out a principal components factor analysis of the sib interaction behavior data (admittedly a somewhat dubious procedure, considering the small number of cases and other characteristics of the data). The first five factors account for 80 % of the variance, and suggest meaningful interaction patterns. In the first, the older sibling is highly directive of the younger sib's sorting activities without being explanatory. In the second, there is some explanation along with a good deal of verbal interaction between the sibs. The third and fifth factors involve nonverbal interaction between the sibs, the third being a directive pattern, and the fifth a criticizing, interfering pattern of behavior. The fourth factor, reminiscent of the deductive teaching style in the concept learning study, is an explanatory teaching pattern with the older sib giving category labels, giving cues, asking questions, and so on, and the younger sib accepting this help and verbalizing his own sorting actions. (These behaviors were associated with the younger sib's use of inferential categories.) Some conclusions and interpretations. Looking back over the two studies, it seems rather clear that older siblings do have an effect on the younger siblings' performance on cognitive tasks, and that such effects depend on sibling status positions (although there is less consistency in these results than one would like). The analysis of the sibling pairs' behaviors during the interaction, however, failed to provide an explanation of how the older sib's effect was produced. One theme which is common to the two studies is that of using verbal labels, giving explanation, and asking questions. In concept learning, this provides the child with labels that might not be in his repertoire and helps him to become aware of attributes which he might be unable to abstract for himself from examples presented to him. In the object sorting task, if one perceives relational, descriptive, and inferential cognitive styles on a continuum going from simple to complex, then the data suggest that the use of more advanced inferential categories is facilitated by the older sib's verbal explanatory behaviors. On the other hand, nonverbal activities tend to be less helpful to the

722

Interaction in social groups

younger sib, and these are associated with the greater use of relational categories. Sutton-Smith and Rosenberg have theorized that the interaction between siblings is characterized by both modeling and reactive components, with second-born children relying more heavily on modeling of the older sib. Direct modeling of the older sibling's object sorting performance does not seem to have been important in this study, since in only a few cases was the older sib observed to make groups himself. On the other hand, such reactive behaviors as dependency on the older sib and conformity to his suggestions also appear to be of minor value in explaining the outcome, as evidenced by the weak correlations between behaviors and outcomes. (The fact that the younger sibs accepted help and direction more readily from a sib who was four years older does provide some evidence for a reactive component of the interaction, however.) An alternative possibility, suggested by social facilitation theory, is that the mere presence of the older sib in the task situation serves to arouse or energize the younger sib, leading to greater task involvement on his part. Another possibility is that subtle modeling of the older sib's task-appropriate behaviors occurred, even though the older sib did not actually model the task. If some sort of modeling is responsible for the effect, it is quite reasonable that the widely spaced older sib would tend to be viewed as a more powerful and competent model. Although children close in age spacing tend to interact more in the home situation, this study suggests that one might distinguish between amount of interaction and quality of interaction. Widely spaced older sibs might be more effective as models on cognitive tasks even though the actual amount of interaction between sibs might be less than if the children were closely spaced. At the present time, I am in the process of analyzing data from a third study in which the relative effects of mother-child interaction and sib-sib interaction are compared in the object sorting situation. Data collection and analysis in this type of study is far from easy, but I am still hopeful that real insights into the dynamics of the child's cognitive development lie at the end of this difficult path.

Sibling influences on the development of the individual

723

D. SOCIAL CLASS, FAMILY SIZE, A N D COGNITIVE PERFORMANCE by Kevin Marjoribanks, University of Oxford, England, and Herbert J. Walberg, University of Illinois at Chicago Circle, Chicago

Investigations of the relations between sibling constellation variables and cognitive measures have usually found that sibsize is negatively related to performance on verbal-type tests (e.g., Nisbet 1953; Maxwell 1969; Eysenck and Cookson 1970; Oldman, Bytheway and Horobin 1971; Murray 1971; Kellaghan and Macnamarra 1972). Also, support has often been provided for Anastasi's (1956) proposition that in the highest socioeconomic levels the negative relations may disappear (e.g., Cicirelli 1967; Nisbet and Entwistle 1967; Record and McKeown 1969; Douglas 1968; Kennett and Cropley 1970; Poole and Kuhn 1973). The present study is a further analysis of the relations between social status characteristics, sibsize and cognitive measures. But the study departs from much of previous research in two ways: First, by analyzing longitudinal data and second, by using a potentially more sensitive sibling constellation variable, the inverse of sibsize. The inverse of sibsize was introduced as it was reasoned that since children share adult resources of intellectual stimulation in the home, the mathematical relationship between sibsize and parental stimulation is not linear but is of a hyperbolic form involving the term, one divided by the number of children in the family. That is, the amount of parental attention which each child receives decreases as the number of children in the family increases in such a way that with each additional child the successive decrements in shared attention become smaller. Therefore, the expected percentages of parental attention given children in 1-, 2-, 3-, 4-, and 5-child families would be 100, 50, 33, 25 and 20 respectively. Thus a single child in a family may score higher on cognitive tests because he receives all available parental stimulation, whereas a child with four siblings may have lower ability scores because he receives more like one-fifth of the available parental stimulation. The present analysis is, in part, a test of a model which was developed during an examination of the mental ability development of a sample of children from Southern Ontario, Canada (Marjoribanks, Walberg and Bargen, in press). In that study, it was found that a three-term equation

724 Interaction in social groups which included father's occupation, inverse of sibsize, and the product of the two variables accounted for as much significant variance in verbal and number ability scores as complex many-term equations which also included father-education, mother-education, sibsize, ordinal position, crowding ratio of the home, the squares of the variables (to test for non-linearity) and the products of the variables (to test for interactions). The results indicated that while children with more siblings tended to score lower on verbal and number ability tests, the relation was attenuated and minimal in families in which the father had high occupational status. As the study emphasized the parsimony of using father occupation and the inverse of sibsize when accounting for variance in cognitive scores, it was decided to test the model again for a different set of cognitive measures with a different sample. Method. The data used in the study were collected in 1964 and 1968 as part of a national survey of English primary school children and their parents. The results of the first survey, which was commissioned by the Central Advisory Council for Education in England, were published in what is commonly known as the Plowden Report (1967). The sampling procedure in 1964 had two stages. First, a random sample was taken from all types of maintained primary (elementary) schools in England. Then a random selection was made of children within these schools. The total sample, which included 3,092 children from 173 primary schools, was divided into three age-cohorts, each of approximately 1,000 children. The average age of the senior children in 1964 was approximately 11, of the middle group 8, and of the youngest group 7. In 1968 when the three cohorts were surveyed again, the senior cohort was in the fourth year of secondary school, the middle cohort in the first year of secondary school, and the youngest group in the last year of the primary school (Peaker 1971). The number of children with complete records for both surveys is 2,350 with approximately 800, equally divided between boys and girls, in each of the cohorts. During the surveys the children completed a series of cognitive tests. In 1968 the three groups were assessed using the Watts-Vernon English comprehension test, the Vernon graded arithmetic''test, and the AliceHeim non-verbal intelligence test (AH4). In 1964 the senior'group was assessed using the Watts-Vernon reading test, and the other two groups were assessed on reading tests developed by the National Foundation for Educational Research in England. A structured home-interview schedule

Sibling influences on the development of the individual 725 was used in the surveys to gain a measure of the social-psychological environment of the family. During the interviews, data were also collected on father occupation, father education, sibsize, oldest-only child, and income of the family. Social class was coded according to the Registrar General's classification based on occupations (1960 edition) with unskilled occupations classified as category 1, manual semi-skilled occupations as category 2, non-manual semi-skilled as category 3, manual skilled as 4, non-manual skilled as 5, managerial-supervisory as 6, and professional-managerial occupations as category 7. Results and discussion. A set of multiple regression models was used to examine the relations between the social status indicators, the sibling variables, and the cognitive scores. The 1968 measures of reading, arithmetic, and non-verbal ability formed the criterion vectors while the predictors included the 1964 reading scores, the 1964 social status characteristics, and the 1964 sibling variables. Also, the 1964 and 1968 measures of father occupation and sibsize were averaged to form two composite variables. It was assumed that the two new variables, father occupation and sibsize, would reflect the cumulative nature of the environment between the two time periods. In general, two-term equations containing the 1964 reading scores and the composite father occupation variable accounted for as much significant variance in the 1968 scores as complex many-term equations, with the inverse of sibsize explaining no additional variance after taking into account prior reading scores and father occupation. As an example, the results for the 1968 reading scores are presented in Table 1. Since the results indicated that the most potent predictor of the 1968 cognitive measures was the 1964 reading scores, it was decided to examine the relation between the 1964 assessments of father occupation, inverse of sibsize, the product of the two variables, and the 1964 reading scores. The results of the regression analyses showed that social class accounted for a significant percentage of the variance in the reading scores for each group, while the inverse of sibsize explained significant additional variance in the two senior groups and for the middle boys. After accounting for father occupation and the inverse of sibsize, no additional variance was explained by the product of the two variables, in any of the groups. A set of regression surfaces was constructed to show the regressionfitted sibsize influence on reading achievement scores at different socioeconomic levels. The curvature of the surfaces in the figures for the

726 Interaction in social groups Table 1. Multiple regression of 1968 reading scores on 1964 reading, father's occupation, and inverse of sibsize 1968 reading Senior boys

Senior girls

1964 reading

Father occupation

.99 a .03 b .72**°

.56 .32 .00* .85 .37 .01

1.06 .04 .68**

Sibsize inverse

Product

Multiple R

2.32 2.65 .00

— .55 .64 .01

.85**

4.34 3.38 .00

— .94 .79 .01

.83**

Middle boys

.64 .30 .62**

1.22 .40 .02**

5.51 3.22 .00

—1.09 .81 .00

.80**

Middle girls

.63 .03 .59**

.82 .39 .02**

3.51 3.80 .00

— .45 .86 .00

.78**

Junior boys

.41 .03 .38**

.95 .39 .03**

4.06 3.34 .01**

— .24 .84 .00

.66**

Junior girls

.41 .03 .36**

— .16 .42 .02**

1.65 .88 .01

.62**

a. raw regression weights b. standard errors of weights c. variance accounted for by variable

—5.47 3.83 .00 * P < 0.05 ** p < 0.01

junior groups and for the middle girls indicated the overall lack of influence of sibsize on the reading scores, after accounting for the influence of social class. In the middle boys group there is a tendency in large families for the inverse of sibsize to have an increasing influence on the reading scores as the socioeconomic level of the family decreases. For the two senior groups, eleven-year-olds, the curvature of the surfaces reflects the potentially deleterious influence of large family size on reading achievement. Even at high socioeconomic levels, after the number of siblings increases beyond five, the reading scores start to decline rapidly. For each age-cohort the results indicate the powerful influence of social class on academic achievement. In the previously mentioned study of Canadian children, it was found that if children came from small families, their performances on mental ability tests were quite similar

Sibling influences on the development of the individual

727

regardless of their social class background. It was not until sibsize increased beyond two, in families of low social status, that family size appeared to have a deleterious influence on academic performance. An analysis of the social-psychological environments in the homes of the Canadian families indicated that in spite of low occupational status, parents with small families often provided a stimulating learning environment (Marjoribanks and Walberg, in press). The present study suggests that perhaps working-class parents of even small families in England do not, or possibly are unable to, provide a learning environment within the home for their children that might be related to high academic achievement. What is now required is a longitudinal study of adolescent multivariate development which includes not only social and sibling correlates of both cognitive and affective performance, but also measures of the environmental process variables operating within the family. Such a study is at present being undertaken.

728 Interaction in social groups E. THE ASSESSMENT OF STABILITY AND CHANGE IN PEER INTERACTION OF NORMAL HEARING AND DEAF PRESCHOOL CHILDREN by Cornelis F. M. van Lieshout,* University of Nijmegen, The Netherlands

The development of behavior during the life span is characterized by change and stability. Change and stability can be conceptualized in different ways: (a) As the variability of behavior, i.e., the degree to which the behavior ranges around the mean of the individual's behavior. Low variability is called stability, high variability is considered as change; (ib) as quantitative increase or decrease vs. absolute or relative constancy of the frequency of responses, schemes, or behavior units. When the amount of a behavior is the same over time we talk about absolute stability, e.g., the level of cognitive functioning of an adult mentally retarded person. When the intensity of the behavior of a person remains the same in relation to the intensity of the same behavior of another person or a population, this is called relative stability (Emmerich 1964, 1968); (c) as qualitative restructuring of the behavior or stage discontinuity. The stage models in developmental psychology are examples of the discontinuity of behavior over the life span. Quantitative increase or decrease of behavior and qualitative restructuring or discontinuity can go together with relative stability of a behavior, e.g., over a certain period the intensity of altruistic behavior of two persons A and B increases or decreases, while the ratio of the behavior of A and B remains the same. Discontinuity goes together with stability, when there is a functional relation between the behavior before and after a stage transformation and when the relative amount of the behavior in the individuals of a group remains the same (cf. Emmerich 1964, 1966, 1968; Kagan 1969, 1971; Maccoby and Feldman 1972). * Now at the State University at Utrecht, The Netherlands. The author wishes to thank Franz J. Moenks and Willard W. Hartup for their guidance in conducting this study. He is further indebted to the Rev. Fathers A. van Uden and J. van Eindhoven of the Institute for the Deaf, St-Michielsgestel, for their cooperation as well as to Mr. A. G. A. Krijnen, Sr. Cecilia, Sr. Alma and the other teachers of the children. The observations were made by Joke Vermeulen-Maan, Jan Bekhuis, Henk van der Hoeven, Herman van den Munckhof, Frans IJseldijk and Koos Zaad. The author is grateful to Peter G. Heijmans and Henk van der Hoeven for their help in conducting the statistical analyses.

Sibling influences on the development of the individual 729 Behavior change and stability can be the result of internal personal factors or of situational factors (Van Lieshout 1972a, 1972b). Kagan (1971) recently termed this endogenous and exogenous stability. Personal factors such as the developmental level or the sensitivity and exposure of a person to external stimulation can influence behavioral change or stability. Bloom (1964) argued that the individual is more sensitive to external influences to the extent that development progresses faster. The degree to which behavioral change and stability can be assessed depends upon the theoretical model and the research strategy used (Mischel 1968, 1970; Van Lieshout 1972a). This point can be made clear by contrasting the psychodynamic and trait theoretical theories against social learning theory. The points of difference may be grouped under three headings: (a) Localization of stability and change in behavior. The stability and change of behavior is localized by trait theorists primarily within the individual. According to social behavior theorists stability in behavior is maintained by behavior contingent reinforcement and cognition, while change results from changes in behavior contingent reinforcement and informative feedback; (b) strategies of investigation of stability and change. Trait theories prefer to assess stability and tend to use correlational studies often extending over long periods of time. Broad definition by trait theorists of the investigated behavior, however, often leads to overestimation of stability. In experimental studies social behavior theorists mainly emphasize the assessment of change. The use of posttests some time after experimental manipulation to assess generalization over time is inappropriate to assess stability; (c) research techniques. Techniques such as questionnaires and behavior ratings yield relative scores and generally give an overestimation of behavioral stability and an underestimation of change. Detailed definition and direct observation of the investigated behavior results in absolute scores, which are required in stability and change research (Emmerich 1971). The aim of this study is to investigate the particular effects of two different research techniques, behavior rating and direct observation, on assessing stability and change in social interactions of two divergent groups of preschool children, prelingually deaf and normal hearing children. These divergent groups were chosen to find out whether behavior ratings and direct observations each reflect expected differential changes in the social interactions of the children of the two groups. This study further investigates differences in the stability in social interactions of hearing and deaf children. Until the age of four, verbal communication

730 Interaction in social groups of the deaf children is very limited compared with the hearing children. The limited verbal communication of deaf preschool children reduces the amount of environmental stimuli for deaf children. Because of this limited exposure of the deaf to environmental stimuli their social interactions are expected to be more stable than those of the hearing children. Method. The subjects were 34 prelingually deaf and 34 normal hearing preschool children. The mean impairment of hearing of the deaf children was 97 dB (SD = 10,7 dB; ISO-Fletcher index). The mean age at the beginning of the study was 60 months (48-72 months) in the hearing group and 61 months (47-72 months) in the deaf group. In each group were 18 boys and 16 girls. Social interactions in both groups were observed according to a time sample scheme (5 weeks x 3 observations per week X 3 minutes per observation = 45 minutes) during free play. The observations were conducted in matched subgroups of 5 to7 children each. After six months the observations were repeated using the same procedure. Observational categories were derived from Skinner (1953) and Charlesworth and Hartup (1967) as delineating behavior possessing some level of generalized reinforcing value. The observations were coded on the spot and recorded on tape. The total time in which a child performed a certain behavior was used as a score. The following categories were distinguished: Positive Interaction - (a) verbal attention; (b) visual attention; (c) physical contact; (d) approval; (e) mutual attention; ( / ) No Interaction; and (g) Negative Interaction. During the same period the social interactions of the children were rated on a 76 item checklist. Iterative cluster analysis on the ratings of a pilot group of 188 preschool children showed the checklist to contain three clusters, viz., (a) sociability vs. social inhibition, shyness; {b) dominance vs. submissiveness; and (c) attention seeking. The behavior ratings were made by 14 teachers; the observations were carried out by seven observers, especially trained for the purpose. The mean interrater agreement per item of the checklist between 14 teachers in the pilot study was .63. The mean interobserver agreement between each pair of observers was .81 and fluctuated from .76 to .85 per observer pair and from .57 to .88 per observational category. Interobserver agreement was computed by dividing the number of seconds observers A and B agreed in assigning a category by the total time A and B observed together.

Sibling influences on the development of the individual

731

Results. In each group and at each time of measurement product moment correlations are calculated between the total scores of the subjects on the three clusters of the checklist and on the observational categories. Only in one of four occasions did the cluster sociability correlate significantly with one of the other clusters (sociability - attention seeking, r = —.36), while in three of four occasions the correlations between the clusters dominance and attention seeking were significant (r varies from .31 to .67). The clusters dominance and attention seeking, therefore, should be considered as aspects of one single dimension. In both groups a number of correlation coefficients between the observational categories were significant at both times of measurement; another number of correlation coefficients were significant at one time of measurement and not at the other. Significant correlations at both times of measurement can be considered as stable interrelations between the observational categories. In both groups and at both times of measurement mutual attention is negatively correlated with No Interaction (r varies from —.69 to —.78). In the hearing group also verbal attention (resp. r — —.62 and —.41) and Negative Interaction (r =—.36 and —.37) are negatively correlated with No Interaction. Visual attention is negatively correlated with verbal (r — —.49 and —.44) and mutual attention r = —.46 and —.63). Negative Interaction is positively correlated with No Interaction (both rs = .33). In the deaf group touching is positively correlated with mutual attention (r = .52 and .59) and Negative with No Interaction (r = —.65 and —.60). In the hearing group the stable relation between some categories suggests the presence of one stable dimension determined by the categories mutual attention, verbal attention, and Negative Interaction, contrasted with No Interaction and visual attention. In the deaf group this dimension is determined by the categories mutual attention and touching, against No Interaction. The interrelation between the clusters of the checklist and the observational categories was assessed by canonical correlations with the clusters as predictors and the observational categories as criteria, computed separately for both groups at both times of measurement. The total percentage of common variance, covered by the first canonical factor, was in the hearing group 11.72 % (ns) and 16.87 % (F = 2.17, df - 21/69.46, p < .001) at the respective periods of measurement. In the deaf group the respective percentages of common variance were 13.64 % (F = 1.81, df= 21/69.46,/» < .05) and 12.69 % ( F = 1.79, df= 21/69.46,/» < .05). The stability coefficients of the clusters and observational categories,

732 Interaction in social groups expressed in the product moment correlations between the scores on the clusters and the observational categories at the first and second time of measurement, are presented in Table 1. Table 1. Stability coefficients Hearing children Total Boys Girls (n = 18) (n = 16) (n = 34)

Variable Cluster: Sociability Dominance Attention seeking Observational category: Verbal attention Visual attention Physical contact Approval Mutual attention N o interaction Negative interaction * p < .05;

Deaf children Boys Total Girls (n = 18) (n = 16) (n =

.53* .61** .30

.61** .57** .77**

.57** .68** .56**

.70** .85** .76**

.86** .87** .54*

.78** .87** .65**

.23 .74** .12 -.02 .52* .72** .47

.56** .77** .26 .00 .50* .21 -.06

.39* .68** .18 -.01 .54** .45** .52**

.34 .32 .29 .32 .63** .86** .52*

.53* .33 .85** .63** .77** .81** .90**

.33* .34* .53** .46** .70** .83** .65**

** p < .01.

In both groups the stability coefficients of the clusters of the checklist are generally higher than those of the observational categories. Except in two of ten occasions the stability coefficients of the deaf group were higher than those of the hearing children. This difference is significant in the cluster dominance (p < .05) and the observational category No Interaction (p < .01). The stability coefficient of the category visual attention is significantly lower (p < .05) for the deaf than for the hearing children. The social interactions of the deaf girls tend to be more stable than those of deaf boys. In the hearing group this sex difference is less consistent. The clusters of the checklist and the observation categories were submitted to analyses of covariance according to a split plot factorial design with repeated measures on the last factor (Kirk 1968). Group (hearing, deaf), sex (male, female), and time of measurement were introduced as factors. Age was used as a covariable because several earlier studies (e.g., Parten 1932; Martin 1964; Emmerich 1964, 1966; Charles-

Sibling influences on the development of the individual

733

worth and Hartup 1967) showed age changes in the social interactions of preschoolers between the ages 4-6. The analyses of covariance on the three clusters of the checklist showed two main group effects and one main sex effect. Eliminating age, hearing children are rated as more sociable than deaf children (F(l,63) = 9.96, p < .001). The deaf children are rated as more dominant than the hearing children (F (1,63) = 15.77, p < .001). Boys in both groups are rated as more dominant than girls (1.63) = 12.01, p < .001; means are 11.9 and 7.5). The analyses of covariance on the observational categories yielded significant main group effects for all observational categories except visual attention (all ps < .001). When age differences were statistically eliminated, the hearing children had more social interactions than the deaf children, especially verbal ( F (1,63) = 68.09) and mutual attention (F(l,63) = 26.19). The deaf children, on the other hand, exhibited more social interactions of an expressive kind, e.g., physical contact (F (1,63) = 26.22), approval ( F (1,63) = 15.57), and negative interactions (F (1,63) = 11.34) (see Table 2). Both groups spent an equal amount of time with onlooker behavior. The main sex effect for the category Negative Interaction was significant {F (1,63) = 4.64, p < .01). Boys had more'negative interactions than girls (means are 39.5 and 18.5 seconds).

Table 2. Mean scores on clusters and observational categories at two times of measurement

Clusters: Sociability Dominance Attention seeking Observational categories: Verbal attention Visual attention Physical contact Approval Mutual attention No interaction Negative interaction

Hearing children First Second

Deaf children First Second

18.2 7.5 3.8

14.2 12.0 3.4

488 470 53 47 565 986 10

19.1 7.3 3.4 399 443 20 23 748 1004 18

223 425 77 106 248 1498 44

15.1 12.5 3.2 209 556 89 45 476 1211 7

734 Interaction in social groups The significant main time of measurement effect revealed an increase over time in both groups of mutual attention (i7 (1,63) = 40.15, p < .001). The interaction effect group x sex x time of measurement was significant in the categories verbal attention {F (1,63) = 11.70, p < .001) and No Interaction (F (1,63) — 5.59; p < .001). During the first period of measurement the hearing children had more verbal interactions than the deaf and the boys had more verbal interactions than the girls. These differences decreased during the second period of measurement. During this period the greatest difference exists between both groups. During both periods of measurement deaf children spent more time non-interacting than hearing children. During the first period of measurement this difference was greatest between the boys in each group. These differences decreased during the second period of measurement, especially the sex differences. The group x time of measurement interaction effect was significant in the categories visual attention (F (1,63) = 9.44, p < .001), physical contact (F(l,63) = 13.93, p < .001), and approval (F (1,63) = 6.61, p < .01). Over time the amount of approval as well as the difference between both groups decreased. The deaf children had more physical contacts than the hearing children at both periods of measurement. While physical contacts of the deaf increased slightly over time, the physical contacts of the hearing children decreased sharply. Over time the visual attention in deaf children increased, while it decreased for hearing children. Discussion. The results of this study suggest that change over time can be assessed easier by direct observations than by rating techniques. While most observational categories revealed behavioral change over a halfyear period, the behavior ratings did not reveal any change. Presumably at the two times of measurement raters adapted their frame of reference to the age of the children. In the same way it can be explained that a number of correlations between age and observational categories were significant and none of the correlations between age and the clusters of the checklist. The age of the subjects - both groups and both periods of measurement taken together - was positively correlated with the observational categories verbal attention (r = .27) and mutual attention (r = .25) and negatively with visual attention (r = —.34). A further finding of this study is that the rating technique used yields generally higher stability coefficients than direct observation. The higher stability of social interactions of deaf children may presumably be considered

Sibling influences on the development of the individual 735 a consequence of the reduced variability of stimuli in their environment. The kind and frequency of social interactive behavior differs considerably in the hearing and deaf group. The hearing children exhibited more social interaction, especially verbal and mutual interaction, and were rated as more sociable than the deaf children. The deaf children, on the other hand, exhibited more social interactions of a more expressive kind, as touching, approval, and negative interaction, and their interactions were rated as more dominant than those of the hearing children. Also the changes over time, assessed with the observational categories, were different in both groups. The increasing total amount of social interactions in the deaf group shows that the deaf in the period between the two measurements approach the level of social interaction of the hearing children, presumably because of their extended introduction in the peer group. The amount of mutual attention increases in both groups (Parten 1932). Visual attention for the behavior of others increases between the two times of measurement in the deaf group and decreases in the hearing group. This probably is the result of a different history of both groups during the interval between the times of measurement. Parten (1932) stated that the onlooker behavior gradually decreased over the ages 2-5. This was reflected in our data by the negative overall correlation between age and visual attention (r = —.34). The increased visual attention in the deaf group between the two times of measurement is presumably caused by the intensive language training during nursery school age. Their language training was based on lip-reading and the children were trained to look at the faces of their interaction partners. In the deaf group, however, the onlooker behavior was also negatively correlated with age during the second period of measurement (r = —.62). This negative correlation in the deaf group reflected the relation ageonlooker behavior found by Parten in normal hearing children. The increase in physical contact in the deaf group is in agreement with the increase of the total amount of social interaction of the deaf. To start or maintain interaction the deaf often raise each other's attention by touching. At both times of measurement in the deaf group physical contact is significantly correlated with mutual attention (resp. r = .52 and .59). At the same time the hearing children show a decrease in their physical contacts. The differences in approval, when measured first, between both groups can be explained as follows. During the first time of measurement the deaf children direct the attention of others to their behavior by explicitly demonstrating what they are doing or give ex-

736 Interaction in social groups pressive approval to the behavior of others. When other ways of interaction become available approval asking and giving decreasesinthis group. On the whole, direct observations seem to be more sensitive in assessing differential forms of behavioral changes over time whereas ratings refer to gross behavior clusters and seem to be relatively insensitive in assessing behavioral changes.

Bibliography

Aakre, B., Strobel, D.A., Zimmerman, R.R., & Geist, C.R. Reactions to intrinsic and extrinsic reward in protein malnourished monkeys. Perceptual and Motor Skills, 1973,36,787-790. Abelson, R.P., & Lesser, G.S. A developmental theory of persuasibility. In C.I. Hovland & I.L. Janis (Eds.), Personality and persuasibility. New Haven, Yale University Press, 1959. Ademowore, A.S., Courey, N.G., & Kime, J.S. Relationships of maternal nutrition and weight gain to newborn birthweight. Obstetrics and Gynecology, 1972, 39, 460-464. Adler, A. The practice and theory of individual psychology (1920). New York, Harcourt Brace, 1929. Adler, A. Understanding human nature (1929). New York, Greenberg, 1946. Ahammer, I.E. Alternative strategies for the investigation of age differences. Paper presented at the Southeastern Conference on Research in Child Development, Athens, Georgia, April 1970. Ahammer, I.E., & Baltes, P.B. Desirability judgments of adolescents, adults and older people: Self- and heteroperceptions. Paper presented at the 78th Annual Convention of the American Psychological Association, Miami, September 1970. Ahr, P.R., & Youniss, J. Reasons for failure on the class inclusion problem. Child Development, 1970, 41, 131-143. Ainsworth, M.D. Infancy in Uganda. Baltimore, The Johns Hopkins Press, 1967. Ainsworth, M.D. Developmental changes in some attachment behaviors in the first year of life. Introduction; Discussion. Papers presented at the Biennial meeting of the Society for Research in Child Development, Minneapolis, Minnesota, April 1971. Ainsworth, M.D., & Bell, S.M. Attachment, exploration, and separation: Illustrated by the behavior of one-year-olds in a strange situation. Child Development, 1970, 41, 49-67. Ainsworth, M.D., Bell, S.M. Blehar, M.P., & Main, N.B. Physical contact: A study of infant responsiveness and its relation to maternal handling. Paper presented at the Biennial Meeting of the Society for Research in Child Development, Minneapolis, April 1971. Ainsworth, M.D.,Bell, S. M. & Stayton, D.J. Individual differences in the development of some attachment behaviors. Merrill-Palmer Quarterly, 1972, 18, 123-143. Allen, (Lady) M. Planning for play. Cambridge, M.I.T. Press, 1968.

ii

Bibliography

Allen, W. H . Instructional media research: Past, present and future. A V Communication Review, 1971, 19, 5-19. Alper, S.W., & Leidy, T . R . The impact of information transmission through television. Public Opinion Quarterly, 1970, 33, 556-562. Ambrosino, L., & Fleiss, D. An international comparison of children's television programming. National Citizens Committee for Broadcasting, July 1971. Anagnostopoulou, R. Mother-infant interaction and attachment behavior in one-yearold infants of mothers with mental disorders and normal mothers. Unpublished dissertation, University of Rochester, 1973. Anastasi, A. Intelligence and family size. Psychological Bulletin, 1956, 53, 187-209. Andersen, R. From an educational point of view. In A. Bengtsson (Ed.), Adventure playgrounds. New York, Praeger, 1972. Anderson, R. Quick child-care center. Family Circle, May 1973. Andersson, B-E. Studies in adolescent behavior. Stockholm, Almqvist & Wiksell, 1969. Andersson, B-E. Misunderstandings between generations - a general phenomenon? Scandinavian Journal of Educational Research, 1973, 17, l-10.(a) Andersson, B-E. Older and younger generations' views on each other. A study in misunderstanding. Unpublished manuscript, 1973.(b). Andersson, B-E., & Ekholm, M. The generation gap: Frame of reference and design of the Stug-project. Report from the Institute of Education, Goteborg University, no 13, 1971. Andersson, B-E., & Ekholm, M. The Stug-project (Studies of the generation gap). In F . K . Monks, W . H . Hartup & J. deWit (Eds.), Determinants of behavioral development. New York, Academic Press, 1972. Apgar, V. A proposal for a new method of evaluation of the newborn infant. Current Researches in Anesthesia and Analgesia, 1953, 32, 260-267. Aries, P. Centuries of childhood. New York, Knopf, 1962. Arieti, S. New views on the psychodynamics of schizophrenia. American Journal of Psychiatry, 1967, 124, 39-44. Arieti, S. The psychodynamics of schizophrenia: A reconsideration. American Journal of Psychotherapy, 1968, 22, 366-381. Arnold, W.R., & Porges, S. W. Heart rate components of orientation in newborns as a function of age and experience. Paper presented at the Meeting of the Eastern Psychological Association, Boston, April 1972. Baker, R . K . . & Ball, S.J. (Eds.), Mass media and violence (Vol. 11 of the Report to the National Commission on the Causes and Prevention of Violence). 1969. Bales, R . F . Interaction process analysis. Cambridge, Mass., Addison-Wesley, 1950. Ball, S., & Bogatz, G. A. The first year of Sesame Street: An evaluation. Princeton, N. J., ETS, 1970. Bandura, A. Principles of behavior modification. New York, Holt, Rinehart and Winston, 1969. Bandura, A., Ross, D., & Ross, S.A. A comparative test of the status envy, social power, and secondary reinforcement theories of identificatory learning. Journal of Abnormal and Social Psychology, 1963, 67, 527-534. Bandura, A., Ross, D., & Ross, S.A. Vicarious reinforcement and imitative learning. Journal of Abnormal and Social Psychology, 1963, 67, 601-607. Banta, T.J. Is there really a Montessori method? Paper presented at the Joint Meeting of the Ohio Psychological and Ohio Psychiatric Associations, February, 1966. Barcus, F. E. Saturday children's television: A report of television programming and advertising on Boston commercial television. Boston, Action for Children's Television, July 1971.

Bibliography

iii

Barcus, F. E. Network programming and advertising in the Saturday children's hours: A June and November comparison. Boston, Action for Children's Television, 1972. Barker, R.G., & Gump, P.V. Big school; small school. High school size. Stanford, Stanford University Press, 1964. Barnes, R.H., Cunnold, S.R., Zimmermann, R.R., Simmons, H. Macleod, R.B., & Krook, L. Influences of nutritional deprivations in early life on learning behavior of rats as measured by performance in a water maze. Journal of Nutrition, 1966, 89, 399-410. Barnes, R.H., Moore, A.U., Reid, I.M., & Pond, W.G. Learning behavior following nutritional deprivations in early life. Journal of American Dietetic Association, 1967, 51, 34-39. Barnes, R.H., Moore, A.U., Reid, I.M., & Pond, W.G. Effect of food deprivation on behavioral patterns. In N.S. Scrimshaw & J. Gordon (Eds.), Malnutrition, learning, and behavior. Cambridge, Massachusetts, M.I.T. Press, 1968. Barnes, R.H., Reid, I.M., Pond, W.G., & Moore, A.U. The use of experimental animals in studying behavioral abnormalities following recovery from early malnutrition. In R.A. McCance & E.M. Widdowson (Eds.), Calorie deficiencies and protein deficiencies. Boston, Little, Brown, and Co., 1968. Pp. 277-285. Bartoshuk, A.K. Human neonatal cardiac acceleration to sound: Habituation and dishabituation. Perceptual and Motor Skills, 15, 1962, 15-27. Battle, H. Relation between personal values and scholastic achievements. Journal of Experimental Education, 1957, 26, 27-41. Baumrind, D. Current patterns of parental authority. Developmental Psychology, 1971, 4, 302-303. Baumrind, D., & Black, A.E. Socialization practices associated with dimensions of competence in preschool boys and girls. Child Development, 1967, 38, 291-328. Beilen, H. The training and acquisition of logical operations. In M. Rosskopf, L. Steffe & S. Taback (Eds.), Piagetian cognitive-development research and mathematical education. Washington, D.C., National Council of Teachers of Mathematics, 1971. Pp. 81-124. Beisser, A., Glasser, N., & Grant, M. Psychosocial adjustment in children of schizophrenic mothers. Journal of Nervous and Mental Disease, 1967, 145, 429-440. Bell, R.Q. Retrospective attitude studies of parent-child relations. Child Development, 1958, 29, 323-338. Bell, R.Q. A reinterpretation of the direction of effects in studies of socialization. Psychological Review, 1968 , 75, 81-95. Bell, R. Q. Stimulus control of parent or caretaker behavior by offspring. Developmental Psychology, 1971, 4, 63-72. Bell, S.M. The effectiveness of various maternal responses as terminators of crying: Some theoretical implications. Paper presented at the Biennial Meeting of the Society for Research in Child Development, Minneapolis, Minnesota, April 1971. Bell, S.M., & Ainsworth, M.D.S. Infant crying and material responsiveness. Child Development, 1972, 43, 1171-1190. Beller, E.K. The evaluation of effects of early educational intervention on intellectual and social development of lower-class, disadvantaged children. In E. Grotberg (Ed.), Critical issues in research related to disadvantaged children. Princeton, New Jersey, Educational Testing Service, 1969. Beller, E.K. Research on teaching: Organized program in early education. In R. Travers (Ed.), Handbook for research on teaching. New York, Rand McNally, 1973. (a) Beller, E.K. Research on organized programs of early education In R. Travers (Ed.), Second handbook of research on teaching. Chicago, Rand McNally, 1973.(b)

iv

Bibliography

Beller, E.K., Adler, P., Newcomer, A., & Young, A. Motivation reinforcement and problem solving in children. In A. Pick (Ed.), Minnesota symposium on child psychology. Minneapolis, University of Minnesota Press, 1972. Bendig, A.W. The development of a short form of the Manifest Anxiety Scale. Journal of Consulting Psychology, 1956, 20, 384. Bengtson, V.L. Differences between sub-samples in level of present role activity. In R. Havighurst, J. Munnichs, B. Neugarten & H. Thomae (Eds.), Adjustment to retirement - a cross national study. Assen, van Gorcum, 1969. Bengtsson, V.L. The generation gap: A review and typology of social psychological perspectives. Youth and Society, 1970, 2, 7-32. Bengtsson, V.L., & Black, K.D. Intergenerational relations and continuities in socialization. Paper presented at the Third West Virginia Conference on Life-Span Developmental Psychology, Morgantown, 1972. Berndt, T.J., Collins, W.A., & Hess, V.L. Social inferences about motives and consequences for televised aggression: A developmental study. Paper presented at the Biennial Meeting of the Society of Research in Child Development, Minneapolis, March 1973. Bernstein, B. Primary socialization, language and education. London, Routledge & Kegan Paul, 1970. Bijou, S.W., & Baer, D.M. Child Development, vol. 1: A systematic and empirical theory. New York, Appleton, 1961. Birch, H.G. Malnutrition, learning and intelligence. American Journal of Public Health 1972, 62, 773-784. Birns, B., Barton, S., & Bridger, W. Individual differences in temperamental characteristics of infants. Transactions of the New York Academy of Science. December, 1969. Birns, B., Blank, M., & Bridger, W. The effectiveness of various soothing techniques on human neonates. Psychosomatic Medicine, 1966, 28 (4, Part. 1), 313-322. Bisseil, J.S. The cognitive effects of preschool programs for disadvantaged children. Unpublished doctoral dissertation, Harvard University, 1970. Bissell, J.S. Implementation of planned variation in Head Start. Washington, D.C., U.S. Department of Health, Education and Welfare, Office of Child Development, 1971. Blau, P. Social mobility and interpersonal relations. American Sociological Review, 1956, 21, 290-295. Blau, P. Exchange and power in social life. New York, Wiley, 1964. Blau, P., & Duncan, O. American occupational structure. New York, Wiley, 1967. Bloom, B.S. Stability and change in human characteristics. New York, Wiley, 1964. Borke, H. Interpersonal perception of young children: Egocentrism or empathy? Developmental Psychology, 1971, 5, 263-269. Borke, H. Chandler and Greenspan's 'Ersatz egocentrism': A rejoinder. Developmental Psychology, 1972, 7, 107-109. Bower, G.H., et al. Hierarchical retrieval schemes in recall of categorized word lists. Journal of Verbal Learning and Verbal Behavior, 1969, 8, 323-343. Bowlby, J. Maternal care and mental health. Geneva, World Health Organization, 1951. Bowlby, J. Attachment and loss, vol. 1: Attachment. New York, Basic Books, 1969. Bracht, G.H. Experimental factors related to aptitude treatment interactions. Review of Educational Research, 1970, 40, 627-645. Brandt, P. An observational investigation on instruction and behavior in an informal British infant school. Paper presented at the Annual Meeting of American Educational Research Association, Chicago, 1972. Brazelton, L.B. Psychophysiological reactions in the neonate: Effect of maternal

Bibliography

v

medication on the neonate and his behavior. Journal of Pediatrics, 1961, 58, 513-518. Bridger, W. H. Sensory habituation and discrimination in the human neonate. American Journal of Psychiatry, 1961, 117, 991-996. Brim, O.G., jr. Socialization through the life cycle. In O. Brim & Wheeler (Eds.), Socialization after childhood: Two essays. New York, Wiley, 1966. Brockman, L.M., & Ricciuti, H.N. Severe protein-calories malnutrition and cognitive development in infancy and early childhood. Developmental Psychology, 1972, 4, 312-319. Brodzinsky, D.M., & Jackson, J.F. Effects of stimulus complexity and perceptual shielding in the development of spatial perspectives. Paper presented at the Biennial Meeting of the Society for Research in Child Development, Philadelphia, 1973. Brodzinsky, D.M., Jackson, J.P., & Overton, W.F. Effects of perceptual shielding in the development of spatial perspectives. Child Development, 1972,43,1041-1046. Bronfenbrenner, U. Socialization and social class through time and space. In E.E Maccoby, T. Newcomb & E. Hartley (Eds.), Readings in social psychology. New York, Holt, 1958. Bronfenbrenner, U. An emerging theoretical perspective for research in human development. Paper presented at the Biennial Meeting of Society for Research in Child Development, Philadelphia, April 1973. Bronson, W.C., Katten, E.S., & Livson, N. Patterns of authority and affection in two generations. Journal of Abnormal and Social Psychology, 1959, 58, 143-152. Brooks, L.R. Spatial and verbal components of the act of recall. Canadian Journal of Psychology, 1968, 22, 349-368. Brooks, R., Brandt, L., & Wiener, M. Differential response to two communication channels: Socioeconomic class differences in response to verbal reinforcers communicated with and without tonal inflection. Child Development, 1969, 40, 453-470. Brown, J. Instrumental control of the sucking response in human newborns. Journal of Experimental Child Psychology, 1972,14, 66-80. Bruner, J. A. The nature and uses of immaturity. American Psychologist, 1972,27,1-22. Bruner, J.S. The act of discovery. Harvard Educational Review, 1961, 31, 21-32. Bruner, J.S. The course of cognitive growth. American Psychologist, 1964, 19, 1-16. Buber, M. Distance and relation. Psychiatry, 1957, 20, 97-104. Burke, B.S., Beal, V.A., Kirkwood, S.B., & Stuart, H.C. The influence of nutrition during pregnancy upon the condition of the infant at birth. Journal of Nutrition, 1943, 26, 569-583. Burnham, D.L., Gladstone, A.I., & Gibson, R.W. Schizophrenia and the need-fear dilemma. New York, International University Press, 1969. Butler, G.D. Recreation areas: Their design and equipment. New York, Barnes, 1947. Butter, C.M. Detection of hidden figures by rhesus monkeys. Perceptual and Motor Skills, 1966, 23, 979-986. Bystroletova, G.N. The formation in neonates of a conditioned reflex to time in connection with daily feeding rhythm. Zhurnal Vysshei Nervnoi Deiatelnosti, 1954, 4, 601-609. Caldwell, Bettye M. A new 'approach' to behavioral ecology. In J. P. Hill (Ed.), Minnesota symposia on child psychology, vol. 2. Minneapolis, University of Minnesota Press, 1969. Pp. 74-107. Canosa, C. A., Salomon, J. B., & Klein, R. E. The intervention approach: The Guatemala Study. In W. M. Moore, M. M. Silverberg, & M. S. Read (Eds.), Nutrition, growth

vi

Bibliography

and development of North American Indian children. DHEW Puplication [NIH] 72-26, Washington, D. C., U.S. Government Printing Office, 1972. Carey, W.B. A simplified method for measuring infant temperament. Journal of Pediatrics, 1970, 77,188-194. Carmichael, L. The onset and early development of behavior. In P.H. Mussen (Ed.), Carmichael's manual of child psychology. New York, Wiley, 1970. Carpenter, E. The new languages. In E. Carpenter & M. Mcluhan (Eds.), Explorations in communication. Boston, Beacon Press, 1960. Casler, L. Perceptual deprivation in institutional settings. In G. Newton & S. Levine (Eds.), Early experience and behavior. Springfield, 111., Charles C. Thomas, 1968. Chadwick, G.F. The park and town: Public landscape in the 19th and 20th centuries. London, The Architectural Press, 1966. Champakam, S., Srikantia, S.G., & Gopalan, C. Kwashiorkor and mental development. American Journal of Clinical Nutrition, 1968,21, 844-852. Chandler, M.J. Egocentrism and antisocial behavior: The assessment and training of social perspective-taking skills. Developmental Psychology, 9, 326-332. Chandler, M.J., & Greenspan, S. Ersatz egocentrism: A reply to H. Borke. Developmental Psychology, 1972, 7,104-106. Charlesworth, R., & Hartup, W. W. Positive social reinforcement in the nursery school peer group. Child Development, 1967,38,993-1002. Chase, H.P., & Martin, H.P. Undernutrition and child development. New England Journal of Medicine, 1970,282,933-939. Chow, B.F., & Stephan, J.K. Fetal undernourishment and growth potential. Nutrition Reports International, 1971,4,245-255. Cicirelli, V.G. Sibling configuration, creativity, IQ and academic achievement. Child Development, 1967,38,481-490. Cicirelli, V.G. The effect of sibling relationships on concept learning of young children taught by child teachers. Child Development, 1972,43,282-287. Cicirelli, V.G. Effects of sibling structure and interaction on children's categorization style. Developmental Psychology, 1973, 9,132-139. Cicirelli, V.G. Relationship of sibling structure and interaction on younger sib's conceptual style. Journal of Genetic Psychology, 1973, in press. Clarke-Stewart, A. Interactions between mothers and their young children: Characteristics and consequences. Monographs of the Society for Research in Child Development, 1973,38 (Serial No. 153). Clifton, R . K . Heart rate conditioning in the newborn infant. Paper presented at the Meeting of the Society for Psychophysiological Research, St. Louis, October 1971. Clifton, R.K., Siqueland, E.R., & Lipsitt, L.P. Conditioned head-turning in human newborns as a function of conditioned response requirements and states of wakefulness. Journal of Experimental Child Psychology, 1972, 13, 43-57. Cohen, J. Multiple regression as a general data-analytic system. Psychological Bulletin, 1968 , 70, 426-443. Collins, W. A. Learning of media content: A developmental study. Child Development, 1970,47,1133-1142. Collins, W. A. Effects of temporal spacing on children's comprehension and behavior following exposure to media violence. Unpublished doctoral dissertation, Stanford University, 1971. Committee on Maternal Nutrition, Food and Nutrition Board, National Research Council. Maternal nutrition and the course of pregnancy. Washington, D.C., National Academy of Sciences, 1970. Committee on Maternal Nutrition, Food and Nutrition Board, National Research Council. Nutritional supplementation and the outcome of pregnancy. Washington,

Bibliography

vii

D. C., National Academy of Sciences, 1973. Conway, E., & Brackbill, Y. Delivery medication and infant outcome: An empirical study. Monographs of the Society for Research in child Development, 1970, 35 24-34. Cooper, D. The application of technology in education: A proposal. Toronto, Ontario Institute for Studies in Education, 1971. Coopersmith, S. Antecedent of self-esteem. San Francisco, Freeman, 1967. Coursin, D.B. Relationship of nutrition to central nervous system development and function: Overview. Federal Proceedings, 1967,25,134. Cravioto, J. Application of newer knowledge of nutrition on physical and mental growth and development. American Journal of Public Health, 1963,53,1803-1809. Cravioto, J. The ecological approach: The Mexico study. In W. M. Moore, M. M. Silverberg, & M.S. Read (Eds.), Nutrition, growth and development of North American Indian children. D H E W Publication [NIH] 72-76, Washington, D. C., U.S. Government Printing Office, 1972. Cravioto, J., & DeLicardie, E.R. The long-term consequences of protein-calorie malnutrition. Nutrition Reviews, 1971,29,107-111. Cravioto, J., DeLicardie, E.R., & Birch, H . G . Nutrition, growth, and neurointegrative development: An experimental and ecologic study. Pediatrics, 1966,38,319. Cravioto, J., & Robles, B. Evolution of adaptive and motor behavior during rehabilitation from kwashiorkor. American Journal of Orthopsychiatry, 1965, 35, 449-464. Cronbach, L. Educational psychology. New York, Harcourt, Brace & World, 1963. Cumming, E., & Henry, W.E. Growing old: The process of disengagement. New York, Basic Books, 1961. Curtis, H. The school playgrounds of America. U.S. Bureau of Education, Report, 1967. Cushna, B. Agency and birth order differences in very early childhood. Paper presented at the Meeting of the American Psychological Association, New York, September, 1966. Dashkovskaia, V.S. The first conditioned reactions in newborn infants under normal and pathological conditions. Zhurnal Vysshei Nervnoi Deiatetnosti, 1953, 3, 247-259. Datta, Lois-ellin. A report on evaluation studies of project Head Start. Washington, D. C. U. S. Dept. of Health, Education and Welfare, Project Head Start, 1969. Davis, K. The sociology of parent-youth conflict. American Sociological Review, 1940, 5, 523-535. Dean, D. A. The relations of ordinal position to personality in young children. Unpublished master's thesis, State University of Iowa, 1947. DeGroat, A. F., & Thompson, G. G. A study of the distribution of teacher approval and disapproval among sixth grade children. Journal of Experimental Education, 1949, 18, 57-75. DeLicardie, E.R., & Cravioto, J. Language development in survivors of clinical severe malnutrition (a longitudinal study). Paper presented at the Ninth International Congress of Nutrition in Mexico City, September 1972. Denisova, M.P., & Figurin, N . L . The question of the first associated appetitional reflexes in infants. Voprosy Geneticheskoi Refleksologii i Pedologii Mladenchestva, 1929,7,81-88. DeVries, R. The development of role-taking as reflected by behavior of bright, average and retarded children in a social guessing game. Child Development, 1970, 41, 759 -770. Dieuzeide, H. Educational technology for developing countries. In D . R . Olson (Ed.),

viii

Bibliography

Media and symbols: The forms of expression, communication and education. 73rd Yearbook of the National Society for the Study of Education. Chicago, University of Chicago Press, 1974. Dinkle, R. Attitudes of children toward supporting aged parents. American Sociological Review, 1944,9, 370-379. Dobbing, J. Effects of experimental undernutrition on development of the nervous system. In N.S. Scrimshaw & J.E. Gordon (Eds.), Malnutrition, learning, and behavior. Cambridge, Mass., M.I.T. Press, 1968. Doell, C.E., & Fritzgerald, G.B. A brief history of parks and recreation in the United States. Chicago, The Athletic Institute, 1954. Dopyera, J.E. Assessing the micro-environments of individual preschool children. Final report, OEO Head Start, Research Division, OEO 4120,1969. Dopyera, J.E., & Lay, M.Z. Assessing the micro-environments of individual preschool children: Some further issues and results. Paper presented at the Annual Meeting of American Educational Research Association, Chicago, 1972. Douglas, J. W. B. All our future. London, Peter Davies, 1968. Dysinger, W.S., & Ruckmick, C.A. The emotional responses of children to the motion picture situation. New York, Macmillan, 1933. Eichorn, D.H. Physiological development. In P.H. Müssen (Ed.), Carmichael's manual of childpsychology. New York, Wiley, 1970. Eisenhower, M. (Chairman) Commission on the causes and prevention of violence, U.S. Government, 1969. Eisenstadt, S.N. From generation to generation. New York, Free Press, 1956. Elder, G. Adolescent socialization and development. In Bargatta & Lambert (Eds.), Handbook ofpersonality theory and research. Chicago, Rand McNally, 1968. Electric Heating Association, Inc. The learning place. 437 Madison Avenue, New York, N.Y., 10022. Elkind, D. Children's conceptions of right and left: Piaget replication study IV. Journal of Genetic Psychology, 1961,99,269-276. Ellis, M. Why people play. New York, Prentice-Hall, 1973. Emmerich, W. Young children's discriminations of parent and child roles. Child Development, 1959,30,403-419. Emmerich, W. Variations in the parent role as a function of the parents' sex and the child's sex and age. Merrill-Palmer Quarterly, 1962, 8,3-11. Emmerich, W. Continuity and stability in early social development. Child Development, 1964,55,311-332. Emmerich, W. Continuity and stability in early social development: II. Teachers' ratings. Child Development, 1966,57,17-27. Emmerich, W. Personality development and concepts of structure. Child Development, 1968,59,671-686. Emmerich, W. Structure and development of personal-social behaviors in preschool settings. Report PR-71-20. Princeton, N.J., Educational Testing Service, 1971. Endlei, N.S., Hunt, J. McV., & Rosenstein, A.J. An S-R inventory of anxiousness. Psychological Monographs, 1962, 76 (17, Whole No. 536). Engen, T., & Lipsitt, L.P. Decrement and recovery of responses to olfactory stimuli in the human neonate. Journal of Comparative and Physiological Psychology, 1965, 59,312-316. Erikson, E. Childhood and society. New York, Norton, 1950. Erikson, H. E. Identity and the life cycle. Psychological Issues, 1959,1. Erzigkeit, H., Über den Zusammenhang faktoriell aufgefaßter Persönlichkeitsdispositionen mit Identifikation, beschrieben durch den Ähnlichkeitsvergleich des se-

Bibliography

ix

mantischen Differentials. Diplomarbeit. Nürnberg, Universität Erlangen, 1970. Escalona, S. The roots of individuality. Chicago, Aldine, 1968. Evans, D.E., Moodie, A.D., & Hansen, J.D.L. Kwashiorkor and intellectual development. South African Medical Journal, 1971, 71,1413-1426. Eysenck, H.J., & Cookson, D. Personality in primary school children: 3-family background. British Journal of Educational Psychology, 1969,40,117-131. Featherston, H. Cognitive effects of preschool programs on different types of children. Cambridge, Huron Institute, 1973. Feldman, D. Crystallizing experiences in cognitive development. Grant Application, Yale University, 1972. Feshbach, N.D., & Roe, K. Empathy in six and seven year olds. Child Development, 1968,39,133-145. Feshbach, S., & Singer R.D. Television and aggression. San Francisco, Jossey-Bass Inc., 1971. Feuer, L. S. The conflicts of generations. New York, Basic Books, 1969. Fish, B. Contributions of developmental research to a theory of schizophrenia. In Exceptional Infant, vol. 2. New York, Brunner/Mazel, 1971. Pp. 473-482. Fish, B., & Alpert, M. Abnormal states of consciousness and muscle tone in infants born to schizophrenic mothers. American Journal of Psychiatry, 1962, 119, 439445. Fish, B., & Alpert, M. Patterns of neurological development in infants born to schizophrenic mothers. In J. Wortis (Ed.), Recent advances in biological psychiatry, vol. 5. New York, Plenum Press, 1963. Pp. 24-37. Fitzgerald, H. E., & Porges, S. W. A decade of infant conditioning and learning research. Merrill-Palmer Quarterly, 1971,17,79-117. Flanders, N.A. Analyzing teaching behavior. Reading, Mass., Addison-Wesley Publishing Co., 1970. Flavell, J.H. The developmental psychology of Jean Piaget. Princeton, N.J., Van Nostrand, 1963. Flavell, J. H., Botkin, P. T., Fry, C. L., Wright, J. W., & Jarvis, P. E. The development of role-taking and communication skills in children. New York, Wiley, 1968. Flavell, J. F., & Wohlwill, J. Formal and functional aspects of cognitive development. In D. Elkind & J.F. Flavell (Eds.), Studies in cognitive development. New York, Oxford University Press, 1969. Fleming, M.L., &Sheikhian, M. Influence of pictorial attributes on recognition memory. AVCommunication Review, 1972,20,423-441. Foman, S.J., & Egan, M.C. Infants, children, and adolescents. In Jean Mayer (Ed.), U.S. nutrition policies in the seventies. San Francisco, Freeman, 1972 Forbes, E. J., & Porges, S.W. Heart rate classical conditioning with a noxious auditory stimulus in human newborns. Paper presented at the Meeting of the Society for Psychophysiological Research, Boston, October 1972. Frank, F. On the conservation of liquids. In J.S. Bruner, R. Olver & P. Greenfield (Eds.), Studies in cognitive growth. New York, Wiley, 1966. Pp. 183-207. Freedman, D.G. An evolutionary approach to research on the life cycle. Human Development, 1971,14,87-99. Freire, P. Pedagogy of the oppressed. New York, Herder and Herder, 1972. Friedman, S., Nagy, A., & Carpenter, G. Newborn attention: Differential response decrement to visual stimuli. Journal of Experimental Child Psychology, 1970, 10, 44-51. Frisch, R.E. Present status of the supposition that malnutrition causes permanent mental retardation. American Journal of Clinical Nutrition, 1970,23 189-195.

x

Bibliography

Frisch, R.E. Does malnutrition cause permanent mental retardation in human beings? Psychiatria, Neurologia, Neurochirugia 1971, 74,463-479. Fröhlich, W.D., Becker, J., Bengtson, V.L., & Bigor, A. The problem of cross-national reliability of ratings, some tentative approaches and findings. In R.J. Havighurst, J. M.A. Munnichs, B.L. Neugarten, & H. Thomae (Eds.), Adjustment to retirement. Assen, The Netherlands, Van Gorcum, 1970. Furu, T. et al. The function of TV for children and adolescents. Tokyo, Monumenta Nipporica, 1971. Gardner, H. The arts human development. New York, Wiley, 1973. Garner, H., Howard, V., & Perkins, D. Symbol systems: A philosophical, psychological, and educational investigation. In D.R. Olson (Ed.), Media and symbols. Ch icago, University of Chicago Press, 1974. Garmezy, N. Vulnerability research and the issue of primary prevention. American Journal of Orthopsychiatry, 1971,41,101-116. Garn, S.M. Biological correlates of malnutrition in man. In Nutrition, growth and development of North American Indian children. Washington, D.C., U.S. Dept. of Health, Education and Welfare, 1972. Gain, S. M., & Frisancho, A. R. Growth retardation: Effects of malnutrition on size and skeletal development. In R.G. Hansen & H.N. Munro (Eds.), Problems of assessment and alleviation of malnutrition in the United States. Nashville, Vanderbilt University, 1970. Gasch, U. Über den Zusammenhang von Identifikation and Familienkonstellation. Diplomarbeit. Nürnberg, Universität Erlangen, 1967. Geist, C. R., Zimmermann, R. R. ,& Strobel, D. A. Effect of proteincalories malnutrition on food consumption, weight gain, serum proteins, and activity in developing rhesus monkeys (Macaca mulatto). Laboratory Animal Science, 1972,22, 369-377. Gerbner, G. Dimensions of violence in television drama. Report for the Mass Media Task Force: National Commission on the Causes and Prevention of Violence, 1969. Gerbner, G. Violence in television drama: Trends and symbolic functions. Television and social behavior, vol. 1. Washington, D.C., U.S. Government Printing Office, 1972. (a) Gerbner, G. Communication and social environment. Scientific American, 1972, 227, 152-162. (b) Gerth, H.H., & Mills, C.W. (Eds.), From Max Weber:Essays in sociology. New York, Oxford University Press, 1946. Giroud, A. The nutrition of the embryo. Springfield, 111., Charles C. Thomas, 1970. Goldschmid, M.L., & Bentler, P. M. Concept assessment kit-Conservation. San Diego, Educational and Industrial Testing Service, 1968. Goodman, N. The languages of art. Indianapolis, Bobbs-Merrill, 1968. Goranson, R. Media violence and aggressive behavior: A review of experimental research. In L. Berkowitz (Ed.), Advances in experimental social psychology, vol. 5. New York, Academic Press, 1970. Gordon, I. J., & Jester, R.F. Techniques of observing teaching in early childhood and outcome of particular projects. In R.M. W. Travers (Ed.), Second handbook of researchon teaching. Chicago, Rand McNally, 1973. Pp. 184-217. Gouldner, A. The norm of reciprocity. American Sociological Review, 1960,25,161-177. Greenberger, M. (Ed.), Computers, communications and the public interest. Baltimore, Johns Hopkins Press, 1971. Gregg, W.H. Playground. In R.O. Zeleny (Ed.), The world book, vol. 15. Chicago, Field Enterprises Educational Corporation, 1968. Griffiths, W., & Senter, R. The effect of protein deficiency on maze performance of

Bibliography

xi

domestic Norway rats. Journal of Comparative Physiological Psychology, 1954, 47,41-43. Gross, L. Modes of communication and the acquisition of symbolic competence. In D.R. Olson (Ed.), Media and symbols. Chicago, University of Chicago Press, 1974. Gross, N., Giacquinta, J.B., & Bernstein, M. Implementing organizational innovations: A sociological analysis of planned education change. New York, Basic Books, 1971. Gump, P.V. Naturalistic viewpoints in psychology research. New York, Holt, Rinehart, & Winston, 1969. Hale, G.A., Miller, L.K., & Stevenson, H.W. Incidental learning of film content: A developmental study. Child Development, 1968,39,69-78. Halliday, M. The state of language. Educational Review, 1969,22. Halverson, C.F., & Waldrop, M.F. Maternal behavior toward own and other preschool children: The problem of 'ownness'. Child Development. 1970, 41, 839-845. Hansen, J.D.L., Freesemann, C., Moodie, A.D., & Evans, M. A. What does nutritional growth imply ? Pediatrics, 1911,47,299-313. Harlow, H.F., Davis, R.T., Settlag, P.H., & Meyer, D.R. Analysis of frontal and posterior association syndromes in brain damaged monkeys. Journal of Comparative Physiological Psychology, 1952,45,419-437. Harver, L. V. The young as a source of stimuli controlling caretaker behavior. Developmental Psychology, 1971,4,73-88. Hartmann, H., Kris, E., & Loewenstein, R. M. Comments on the formation of psychic structures. In The psychoanalytic study of the child, vol. 2. New York, International University Press, 1946. Harvey, O. J., White, B.J., Prather, M.S., Alter, R.D., & Hoffmeister, J.K. Teachers' belief systems and preschool atmospheres. Journal of Educational Psychology, 1966, 57,373-381. Havighurst, R. Developmental tasks and education. New York, Longmans, Green & Co., 1952. Havighurst, R. J., Neugarten, B., & Tobin, S.H. Disengagement and patterns of aging. Gerontologist, 1964,4,217-228. Havighurst, R.J., Munnichs, J. M.A., Neugarten, B.L.,& Thomae, H. (Eds.), Adjustment to retirement. Assen, The Netherlands, Van Gorcum, 1970. Hawkins, R.P. Learning of peripheral content in films: A developmental study. Child Development, 1973, 44,214-217. Hein, G. E. Piaget, materials and open education. EDC News, (Educational Development Center), Winter, 1973. Heinstein, M. Behavior correlates of breast-bottle regimes under varying parent-infant relationships. Monographs of the Society for Research in Child Development, 1963, 4,133. Henry, W. The theory of intrinsic disengagement. In From Hansen (Ed.), Age with a future. Munksgaard, Copenhagen, 1964. Hetherington, E.M., & Brackbill, Y. Etiology and covariation of obstinacy, orderliness, and parsimony in young children. Child Development, 1963,34,919-943. Hilton, I. Differences in the behavior of mothers toward first- and later-born children. Journal of Personality and Social Psychology, 1967, 7, 282-290. Hochschild, A. The unexpected community. Englewood Cliffs, N. J., Prentice-Hall, 1973. Hoehn, A.H. A study of social class differentiation in the classroom behavior of nineteen third-grade teachers. Journal of Social Psychology, 1954,39,292-296. Hoffman, M.L., & Saltzstein, H.D. Parent discipline and the child's moral development. Journal of Personality and Social Psychology, 1967,5,45-57.

xii

Bibliography

Homans, G. The human group. New York, Harcourt, Brace, 1959. Howe, A. Unpublished manuscript. Syracuse, New York, Syracuse University Reading and Language Arts Center, 1971. Hunt, J.McV. Intelligence and experience. New York, Ronald Press, 1961. Huttenlocher, J. Constructing spatial images: A strategy in reasoning. Psychological Review, 1968,75, 550-560. Huttenlocher, J., & Presson, C.C. Mental rotation and the perspective problem. Cognitive Psychology, 1973.4,277-299. Illich, I. The breakdown of schools: A pioblem or a symptom? Interchange, 1971,11, 4-5. Inkeles, A., & Bauer, R. The Soviet citizen. Cambridge, Harvard University Press, 1959. Irish, D.P. Sibling interaction: A neglected aspect in family life research. Social Forces, 1964, «,279-288. Jacobson, H. N. Nutrition and pregnancy. Journal c,/the American Dietetic Association. 1972,69,26-29. Jencks, C., with Smith, M., Acland, H., Bane, M., Cohen, D., Gintis, H., Heyns, B., & Michelson, S. Inequality: A reassessment of the effect of family and schooling in America. New York, Basic Books, 1972. Kagan, J. The child's perception of the parent. Journal of Abnormal and Social Psychology, 1956,53,257-258. Kagan, J. The three faces of continuity in human development. In D. A. Goslin (Ed.), Handbook of socialization theory and research. Chicago, Rand McNally, 1969. Pp. 983-1002. Kagan, J. Change and continuity in infancy. New York, Wiley, 1971. Kahl, J. American class structure. New York, Holt, Rinehart and Winston, 1957. Kallen, D.J. (Ed.), Nutrition, development and social behavior, Washington, D.C., DHEW Publication [NIH] 73-242,1973. Kamii, C. An application of Piaget's theory to the conceptualization of a preschool curriculum. In R. Parker (Ed.), The preschool in action: Exploring early childhood program. Boston, Allyn & Bacon, 1972. Kammeyer, K. Birth order and the feminine sex role among college women. American Sociological Review, 1966,31,508-515. Kammeyer, K. Birth order as a research variable. Social Forces, 1967,46,71-80. Kandel, D.B., & Lesser, G. S. Youth in two worlds. San Fransisco/Washington/London, Jossey-Bass, 1972. Kaplan, B.J. Malnutrition and mental deficiency. Psychological Bulletin, 1972, 78, 321-334. Karlson, B., Madden, R., & Garner, E.F. Manual for adult basic learning examination. New York, Harcourt, Brace, Jovanovich, 1967. Karnes, M.B. Investigations of classroom and at home interventions: Research and development program on preschool disadvantaged children. Final report, Bureau No. 5-1181, Bureau ofResearch, Office of Education, U.S. Department of Health, Education, and Welfare, ED 036 633,1969. Kaye, H. The conditioned babkin reflex in human newborns. Psychonomic Science 1965,2,287-288. Keen, R.E., Chase, H. H., & Gi aham, F. K. Twenty-four hour retention by neonates of an habituated heart rate response. Psychoncmic Science, 1965,2,265-266. Kallaghan, T., & Macnamarra, J. Family correlates of verbal reasoning ability. Develop-

Bibliography

xiii

mental Psychology. 1972, 7,49-53. Kennett, K.F., & Cropley, A.J. Intelligence, family size and socioeconomic status. Journal of Biosocial Science, 1970,2,227-236. Kessen, W., Haith, M.M., & Salapatek, P.H. Human infancy: A bibliography and guide. In P.H. Müssen (Ed.), Carmichael's manual of child psychology. New York, Wiley, 1970. Pp. 287-445. Kirk, R.E. Experimental design: Procedures for the behavioral sciences. Belmont, Wadswoith, 1968. Klaus, M.H., Jerauld, R., Kreger, N.C., McAlpine, W., Steffa, M., & Kennell, J.H. Maternal attachment: Importance of the first post-partum days. New England Journal of Medicine, 1972,286,460-463. Klein, R.E., Freeman, H.E., Kagan, J., Yarbrough, C., & Habicht, J.P. Is big smart? The relation of growth to cognition. Journal of Health and Social Behavior, 1972, 13,219-225. Klein,R.E.,Gilbert,0.,Canosa, C.,&DeLeon,R. Performance of malnourished children in comparison with adequately nourished children. Paper presented to the Annual Meeting of the American Association for the Advancement of Science, Boston, 1969. Klein, R.E., Habicht, J. P., & Yarbrough, C. Effect of protein-calorie malnutrition on mental development. Advances in Pediatrics, vol. 18. Yearbook Medical Publishers, Inc., 1971. Klein, R.P. An approach to the measurement of preschool environments. Paper presented at the Annual Meeting of the American Psychological Association, Washington, D. C., September 1971. Klingberg, M.A., Abramovici, A., & Chemke, J. (Eds.), Drugs and fetal development. New York, Plenum Press, 1972. Knittle, J. Statement made in Nutritional supplementation and the outcome of pregnancy. Washington, D.C., Committee on Maternal Nutrition. Food and Nutrition Board. National Research Council, National Academy of Sciences, 1973. Koch, H.L., Children's attitudes toward their peers as related to certain characteristics of their siblings. Psychological Monographs, 1956, 70 (19), No. 426. Kohlberg, L. Stage and sequence: The cognitive-developmental approach to socialization. In D. Goslin (Ed.), Handbook of socialization theory and research. Chicago, Rand McNally, 1969. Kosa, J., Rachiele, L., & Schommer, C. Sharing the home with relatives. Marriage and Family Living, 1960,22,129-131. Krachkovskaia, M. V. Reflex changes in leukocyte count of newborn infants in relation to food intake. Pavlov Journal of Higher Nervous Activity, 1959,9,193-199. Krauss, R. M., & Glucksberg, S. The development of communication: Competence as a function of age. Child Development, 1969,40,255-267. Krön, R.E., Stein, M., & Goddard, K.E. Newborn sucking behavior affected by obstetric sedation. Pediatrics, 1966,57,1012-1016. Kuenzlen, M., in collaboration with Cashman, T., & Happ, H. Playing urban games. New York, Braziller, Inc., 1972. Kuhlen, R.G. Personality change with age. In P. Worchel & D. Byrne (Eds.), Personality change. New York, Wiley, 1964. Pp. 524-555. Kuhn, D. Mechanism of change in the development of cognitive structures. Child Development, 1972,43,833-844. Kurz, C. Über die Identifikation zwischen Eltern und Kindern auf Grund ihrer Familienkonstellation und ihre Auswirkungen auf die Erziehbarkeit. Diplomarbiet, Nürnberg, Universität Erlangen, 1969.

xiv

Bibliography

Lasch, C., Inequality and education. New York Review of Books, May 17,1973. Lasko, J.K. Parent behavior toward first and second children. Genetic Psychology Monographs, 1954,49,96-137. Lavatelli, C. S. Piaget 's theory applied to an early childhood curriculum. Washington, D.C. American Engineering & Science, Inc., 1970. Lay, M.Z. The responsive care model: A program manual and report of implementation. Syracuse, New York, Syracuse University Early Childhood Education Center, 1972. Lehr, U., & Dreher, G. Determinants of attitudes towards retirement. In R. Havighurst, J. Munnichs, B. Neugarten & H. Thomae (Eds.), Adjustment to retirement. Assen, The Netherlands, Van Gorkum, 1970. Lehr, U., & Rudinger, G. Consistency and change of social participation. Human Development, 1969,12,255-267. Lehr, U., & Rudinger, G. Strukturen der sozialen Teilhabe im höheren Lebensalter. Veröffentlichungen der Deutschen Gesellschaft für Gerontologie, 1970,4, 81-88. Leifer, A.D. The relationship between cognitive awareness in selected areas and differential imitation of a same-sex model. Unpublished master's thesis, Stanford University, 1966. Leifer, A.D. Effective use of television in encouraging social competence. Final report to the Children's Television Workshop, 1973. Leifer, A.D. et al., Developmental aspects of variables relevant to observational learning. Child Development, 1911,42,1509-1516. Leifer, A.D., & Roberts, D.F. Children's responses to television violence. Television and social behavior, vol. 2. Washington, D.C., U.S. Government Printing Office, 1972. Leifer, A.D., & Roberts, D.F. Vicarious reinforcement and aggressive behavior: Rocky and Johnny revisited. Unpublished manuscript, Harvard University, 1973. Lerner, J., & Shanan, J. Coping style in psychiatric patients with somatic complaints. Journal of Personality Assessment, 1972, 36, 18-32. Levin, H. Permissive child-rearing and adult role behavior. In D.E. Dulany, R.L. DeValois, D.C. Beardsley & M.R. Winterbottom (Eds.), Contributions to modern psychology. New York, Oxford University Press, 1958. Pp. 307-312. Lévi-Strauss, C. Structural anthropology. New York, Anchor Books, 1967. Levitsky, D. A., & Barnes, R.H. Effects of early protein-calorie malnutrition on animal behavior. Nature, 225,468-469. Levitsky, D.A. & Barnes, R.H. Nutritional and environmental interaction in the behavioral development of the rat: Long term effects. Science, 1972,176,68-71. Levitsky, D. A., & Barnes, R.H. Malnutrition and animal behavior. Nutrition development and social behavior. Washington, D. C., U.S. Dept. of Health, Education and Welfare, 1973. Lewin, K., Lippit, R., & White, R. Patterns of aggression in experimentally created social climates. Journal of Social Psychology, 1939,10,271-299. Lewis, M. State as an infant environment interaction: An analysis of mother-infant interaction as a function of sex. Merrill-Palmer Quarterly, 1972,18,95-121. Lewis, M., & Ban , P. Stability of attachment behavior: A transformational analysis. Paper presented at the Biennial Meetings of the Society for Research in Child Development, Minneapolis, April 1971. Lidz, T. The family, personality development, and schizophrenia. In J. Romano (Ed.), The origins of schizophrenia. New York, Excerpta Medica Foundation, 1967. Pp. 131-138. Lindblom, B. (Ed.), Fallet Bastad: En Studie i svensk opinionsbildning. Stockholm, Wahlström & Widstrand, 1969.

Bibliography

xv

Lindvig, E. K., & Zimmermann, R. R. Conditional learning in rehabilitated malnourshed monkeys. Paper presented at the Montana Psychological Association Meetings, Missoula, 1973. Lipset, S. M., & Raab, F. The non-generation gap. Commentary, 1970,50, 35-39. Lipsitt, L.P. Learning processes of human newborns. Merrill-Palmer Quarterly, 1966, 72,45-71. Lipsitt, L. P. The experimental origins of human behavior. In L. R. Goulet & P. B. Baltes (Eds.), Life-span developmental psychology: Research and theory. New York, Academic Press, 1970. Pp. 285-305. Lipsitt, L.P.,& Ambrose, J. A. A preliminary report of temporal conditioning of three types of neonatal stimulation. Paper presented at the Meeting of the Society for Research in Child Development, New York, March 1967. Lipsitt, L.P., & Kaye, H. Conditioned sucking in the human newborn. Psychonomic Science, 1964,I,29-30. Lipstitt, L.P., Kaye, H., & Bosack, T.N. Enhancement of neonatal sucking through reinforcement. Journal of Experimental Child Psychology, 1966,4,163-168. Litchfield, Ann. The nature and pattern of participation in adult education activities. Unpublished dissertation, University of Chicago, 1965. Litwak, E. Geographic mobility and extended family cohesion. American Sociological Review, 1960,25, 385-394. (a) Litwak, E. Occupational mobility and extended family cohesion. American Sociological Review, 1960,25,9-21.(b) Litwak, E., & Szelenyi, I. Primary group structures and their functions: Kin, neighbors, and friends. American Sociological Review, 1969,24,465-481. Lloyd, B.B. Perception and cognition: A cross-cultural perspective. London, Penguin Books, 1972. Loban, W. Problems in oral English: Kindergarten through grade nine. Champaign, 111., National Council of Teachers of English, 1966. Looft, W.R. Egocentrism and social interaction across the life-span. Psychological Bulletin, 1972, 78,73-92. Lowenberg, M.E., Todhunter, E.N., Wilson, E.D., Feeney, M.C., & Savage, J.R. Food and man. New York, Wiley, 1968. Lucas, C., & Wohlleb, C. Implementation of Head Start planned variation, 1970. Cambridge, Huron Institute, 1972. Lyle, J. Television in daily life: Patterns of use (overview). Television and social behavior, vol. 4. Washington, D.C., U.S. Government Printing Office, 1972. Macy, I.C., & Hunscher, H. A. Calories - a limiting factor in the growth of children. Journal of Nutrition, 1951, 45, 189-99. Maccoby, E. E., & Feldman, S. S. Mother-attachment and stranger-reactions in the third year of life. Monographs of the Society for Research in Child Development, 1972,37, (1, Whole No. 146). Maddox, G.L., &Eisdorfer C. Some correlates of activity and morale among the elderly. Social Forces, 1963,40,254-260. Maddox, G.L. Persistence of life style among the elderly: A longitudinal study of patterns of social activity in relation to life satisfaction. Proceedings of the Seventh International Congress of Gerontologists, Wien, 1966,6. Marjoribanks, K., &Walberg,H. J. Ordinal position, environment, and mental abilities. Journal of Social Psychology, in press. Marjoribanks, K., Walberg, H. J., & Bargen, M. Mental abilities: Sibling constellation and social class correlates. British Journal of Social and Clinical Psychology, in press.

xvi

Bibliography

Marquis, D.P. Can conditioned responses be established in the newborn infant. Journal of Genetic Psychology, 1931,39,419-492. Marquis, D.P. Learning in the neonate: The modification of behavior under three feeding schedules. Journal of Experimental Psychology, 1941,29,263-282. Martin, W.E. Singularity and stability of profiles of social behavior. In C.B. Stendler (Ed.), Readings in child behavior and development. New York, Harcourt Brace, 1964. Pp. 448-466. Maxwell, J. Intelligence, education and fertility: A comparison between the 1932 and 1947 Scottish surveys. Journal of Biosocial Science, 1969,1,247-271. McArthur, C. Personalities of first and second children. Psychiatry, 1956,19,47-54. McCay, C. M., Crowell, M. F., & Maynard, L. A. The effect of retarded growth upon the length of life span and upon the ultimate body size. Journal of Nutrition, 1935,10, 63-79. McKay, H.E., McKay, A., & Sinisterra, L. Behavioral intervention studies with malnourished children: A review of experiences. In D.J. Kallen (Ed.), Nutrition, development and social behavior. DEHW Publication [NIH] 73-242, Washington, D.C., U.S. Department of Health, Education, and Welfare, 1973. McLuhan, M. Understanding media: The extension of man. New York, McGraw Hill, 1964. McMeekin, R. Costs analysis of planned variation Head Start models. Cambridge, Huron Institute, 1973. McNemar, Q. Special review of Newman, Freemand & Holzinger's 'Twins: A study of heredity and environment'. Psychological Bulletin, 1938,35,237-249. Mead, M. Culture and commitment: A study of the generation gap. New York, Basic Books, 1970. Medley, D.M., & Mitzel, H.E. Measuring classroom behavior by systematic observation. In N.L.Gage (Ed.), Handbook of research on teaching. Chicago, Rand McNally, 1963. Pp. 247-328. Mendel, G. Children's references for differing degrees of novelty. Child Development, 1958,29,453-465. Mendelson, G., & Young, M. A content analysis of black and minority treatment on children's television. Boston, Action for Children's Television, 1972. Mero, E.B. American playgrounds. Boston, American Gymnasia Co., 1908. Mickelsen, O. Adolescent nutrition. Journal of Periodontology, 1971,42,460-466. Millen, J.W. The nutritional basis of reproduction. Springfield, Illinois, Charles C. Thomas, 1962. Miller, L. B., Dyer, J. L., et al. Experimental variation of Head Start curricula: A comparison of current approaches. Progress Report No. 9, Research Grant CG 8199, Office of Economic Opportunity. Louisville, Ky., University of Louisville, 1971. Miller, L.B., Dyer, J.L., et al. Experimental variations of Head Start curricula: A comparison of current approaches. Progress Report No. 7, Research Grant CG 8199, Office of Economic Opportunity. Louisville, Ky., University of Louisville, 1970. Miller, L.B., Dyer, J.L. et al. Experiment variation of Head Start curricula: A comparison of current approaches. Progress Report No. 5, Research Grant CG 8199, Office of Economic Opportunity. Luisville, Ky., University of Louisville, 1969-1970. Miller, L.B., & Dyer, J.L. Two kinds of kindergarten after four types of Head Start. Louisville, Ky., University of Louisville Press, 1970. Miller, N.E., & Dollard, J. Social learning and imitation. New Haven, Yale University Press, 1941. Miller, P. H., Kessel, F. S., & Flavell, J. H. Thinking about people thinking about people thinking about...: A study of social-cognitive development. Child Development,

Bibliography

xvii

1970,41, 613-623. Mischel, W. Personality and assessment. New York, Wiley, 1968. Mischel, W. Sex-typing and socialization. In P.H. Müssen (Ed.), Carmichael's manual of childpsychology, vol. 2. New York, Wiley, 1970. Pp. 3-60. Mitchell, G. Abnormal behavior in primates. In L.S. Rosenblum (Ed.), Primate behavior, developments infieldand laboratory research. New York, Academic Press, 1970. Mitchell, H.S. Protein limitation and human growth. Journal of the American Dietetic Association, 1964,44,165-171. Money, J., & Erhardt, A. A. Man & Woman - Boy & Girl: Differentiation and dimorphism of gender identity from conception to maturity. Baltimore, John Hopkins Univ. Press, 1972. Monge, R.H., & Gardner, E.F. A program of research in adult differences in cognitive performance and learning: Backgrounds for adult education and vocational retraining. Final report to the U.S. Office of Education on Grant No. OEG-1-7061963-0149, Project No. 6-1963, January 1972. Mönkeberg, F. Effect of early marasmus malnutrition on subsequent physical and psychological development. In N. S. Scrimshaw & J. E. Gordon (Eds.), Malnutrition, learning, and behavior. Cambridge, Massachusetts, M.I.T. Press, 1968. Pp. 266-277. Moore, O.K., & Anderson, A.R. The responsive environment project. In R.D. Hess & R. M. Bear (Eds.), Early education: Current theory, research and action. Chicago, Aldine, 1968. Mor, E., Shanan, J., & Levinson, A. Motivation and coping behavior in adaptation to contact lenses. Journal of Personality Assessment, 1973,37,136-143. Mora, J.O., et al. Nutrition, health, and social factors related to intellectual performance. World Review of Nutrition and Dietetics, 1974,19,205-236. Morgan, J. J.B., & Morgan, S.S. Infant learning as a developmental index. Journal of Genetic Psychology, 1944,65,281-289. Moss, H.A. Coping behavior, the need for stimulation and normal development. Merrill-Palmer Quarterly, 1965,11,171-179. Moss, H.A. Methodological issues in studying mother-infant interaction. American Journal of Orthopsychiatry, 1965,35,482-486. Moss, H.A. Sex, age, and state as determinants of mother-infant interaction. MerrillPalmer Quarterly, 1967,13,19-36. Mosskin, L., & Mosskin, P. Ungdomsupproret. Stockholm, Bonniers 1969. Munsterberg, H. The film: A psychological study. New York, Dover, 1972. Murray, C. The effects of ordinal position on measured intelligence and peer acceptance in adolescence. British Journal of Social and Clinical Psychology, 1967,10,221-227. Naftali, T. Coping patterns in adolescents. Unpublished master's thesis, Hebrew University of Jerusalem, 1973. (Hebrew). Nash, J. The father in contemporary culture and current psychological literature. Psychological Bulletin, 1965,36,261-297. Nelson, K. Structure and strategy in learning to talk. Monographs of the Society for Research in Child Development, 1973,38, (Serial No. 149). Neugarten, B.L., Havighurst, R.J., & Tobin, S. Personality and patterns of aging. Gawein. Tijdschrift Psychologische Kring van de Nijmeegse Univ., 1965,13,249-256. Newcomb, Th.M., Turner, R.H., & Converse, P.E. Social psychology. London, Tavistock, 1966. Newman, H.H., Freeman, F.N. & Holzinger, K.H., Twins: A study of heredity and environment. Chicago, University of Chicago Press, 1937. Nicholson, S. How not to cheat children: The theory of loose parts. Landscape Ar-

xviii

Bibliography

chitecture, October 1971. Nisbet, J . D . Family environment and intelligence. Eugenics Review, 1953,45,31-42. Nisbet, J.D., & Entwistle, N.J. Intelligence and family size, 1949-1965. British Journal of Educational Psychology. 1967.37,188-193. Nunnally, J. C. Psychometric theory. New York, McGraw-Hill, 1967. Nydegger, W., & Nydegger, C. Tarong: An Ilocos barrio in the Phillipines. New York, Wiley, 1966. Nyhan, W.L., & Lampert, F. Response of the fetus and newborn to drugs. Anesthesiology, 1965,26,487-500. Oldman, D., Bytheway, B., & Horobin, G. Family structure and educational achievement. Journal of Biosocial Science, Supplement 3,1971,81-91. Olson, D. Cognitive development. New York, Academic Press, 1970. Olson, D. Is instructional media the message? Abstract guide of the 20th International Congress of Psychology, Tokyo, 1972.(a) Olson, D. On a theory of instruction: Why different forms of instruction result in similar knowledge. Interchange, 1972,3,9-25.(b) Olson, D. On a theory of educational experience: What is worth knowing and what can be taught. School Review, 1973,82,27-43. Olson, D. R., & Bruner, J. S. Learning through experience and learning through media. In Olson, D. R. (Ed.), Media and symbols. Chicago, Univ. of Chicago Press,1974. Opie, I., & Opie, P. Children's games in street and playground. Oxford, Clarendon Press, 1969. Orth, R. The acquisition of attitudes as an instrumental conditioned response with consideration of the developmental aspects of the effects of delayed reinforcement. Unpublished doctoral dissertation, Stanford University, 1969. Osofsky, J. D „ & Danzger, B. Relationships between neonatal characteristics and mother-infant interaction. Papei presented at Biennial Meeting of the Society for Research in Child Development, Philadelphia, April 1973. Osofsky, J.D., & O'Connell, E.J. Parent-child interaction: Daughters' effects upon mothers' and fathers' behaviors. Developmental Psychology, 1972,2,157-168. Overton, W. F. Piaget's theory of intellectual development and progressive education. In J . R . Squire (Ed.), A new look at progressive education. Washington, Association for Supervision and Curriculum Development, 1972. Pp. 88-115. Overton, W . F . On the assumptive base of the nature-nurture controversy: Additive versus interactive conceptions. Human Development, 1973,16, 74-89. Overton, W.F., & Brodzinsky, D . Perceptual and logical factors in the development of multiplicative classification. Developmental Psychology, 1972,6,104-109. Overton, W.F., & Reese, H.W. Models of development: Methodological implications. In J . R . Nesselroade & H.W. Reese (Eds.), Life-span developmental psychology: Methodological issues. New York, Academic Press, 1972. Pp. 65-86. Overton, W. F., Wagner, J., & Dolinsky, H. Social class differences and task variables in the development of multiplicative classification. Child Development, 1971,42,19511958. Owen, G . M . , Nelson, C.E., & Garry, P.J. Nutritional status of pre-school children: Hemoglobin, hematocrit, and plasma iron values. Journal of Pediatrics, 1970, 76, 761-763. Pagani, A. Causes of institutionalization. Cross-national surveys cf old age. Ann Arbor, University of Michigan, Division of Gerontology, No date. Pp. 53-58. Paivio, A., & Yarmey, A . D . Pictures versus words as stimuli and responses in pairedassociate learning. Psychonomic Science, 1966,5,235-236.

Bibliography

xix

Palmer, E.L. Formative research in the production of television for children. In D.R. Olson (Ed.), Media and symbols. Chicago, University of Chicago Press, 1974. Papousek, H. Conditioned head rotation reflexes in infants in the first months of life. Acta Paediatrica, 1961,50,565-576. Papousek, H. Conditioning during early postnatal development. In Y. Brackbill & G.G. Thompson (Eds.), Behavior in infancy and early childhood: A book of readings. New York, Free Press, 1967. Parke, R.D., O'Leary, S.E., & West, S. Mother-father newborn interaction: Effects of maternal medication, labor and sex of infant. Proceedings of the American Psychological Association, 1972, 85-86. Parker, R., Sperr, S., & Rieff, M. Multiple classification: A training approach. Developmental Psychology, 1972,7,188-194. Parten, M.B. Social participation among preschool children. Journal of Abnormal and Social Psychology, 1932-1933,27,243-269. Paulson, F.L., McDonald, D.L., & Whittemore, S.L. An evaluation of Sesame Street programming designed to teach cooperative behavior. Monmouth, Oregon, Teaching Research, 1972. Paeker, G.F. The Plowden children four years later. London, National Foundation for Educational Research, 1971. Pease, D., & Hawkes, G. Observation of parent-child interaction. American Journal of Orthopsychiatry, 1969,30,453^59. Pedersen, F. A., & Robson, K.S. Father participation in infancy. American Journal of Orthopsychiatry, 1969,39,466-472. Peregoy, P.L., Zimmermann, R.R., & Strobel, D.A. Protein preference in protein malnourished monkeys. Perceptual and Motor Skills, 1972,35,495-503. Pettus, J. P. Recovery from malnutrition: Food preference and neophobia. Unpublished master's thesis, University of Montana, 1972. Phoenix, C.H., Goy, R.W., & Young, W.C. Sexual behavior: General aspects. In L. Martini & W.F. Ganong (Eds.), Neuroendocrinology, vol. 2. New York, Academic Press, 1967. Piaget, J. The psychology of intelligence. New York, Harcourt, Brace, 1950. Piaget, J. The general problems of the psychobiological development of the child. In J. M. Tanner & B. Inhelder (Eds.), Discussions on child development: Proceedings of the World Health Organization study group on the psychobiological development of the child, vol. 4. New York, International Universities Press, 1960. Pp. 3-27. Piaget, J. Science of education and the psychology of the child. New York, Viking Press. 1969. Piaget, J. Piaget's theory. In P.H. Müssen (Ed.), Carmichael's manual of child psychology. New York, Wiley, 1970.(a) Piaget, J. Science of education and the psychology of the child, New York, Orion Press, 1970.(b) Piaget, J. Structuralism. New York, Basic Books, 1970.(c) Piaget, J. Biology and knowledge. Chicago, University of Chicago Press, 1971. Piaget, J., & Inhelder, B. The child's conception of space. London, Routledge & Kegan Paul, 1956. Piaget, J., & Inhelder, B. The psychology of the child. New York, Basic Books, 1969. Plowden, B., et al. Children and their primary schools. London, H. M. Stationery Office, 1967. Plummer, G.W. Anomalies occurring in children exposed in utero to atomic bomb in Hiroshima. Pediatrics, 1952,10, 687-692. Polanyi, M. Personal knowledge. Chicago, University of Chicago Press, 1958. Polanyi, M. The tacit dimension. New York, Doubleday, 1966.

xx

Bibliography

Polikanina, R.I. The relationship between autonomic and somatic components of a defensive conditioned reflex in premature children. Pavlov Journal of Higher Nervous Activity, 1961,11,72-82. Pollitt, E. Ecology, malnutrition, and mental development. Psychosomatic Medicine 1969,31,193-200. Pollitt, E. Desnutrición, antecedentes bio-sociales y desarrollo cognoscitivo. Revista de Neuro-psiquiatria, 1972,35,21-37. Poole, A., & Kuhn, A. Family size and ordinal position: Correlates of academic success. Journal of Biosocial Science, 1973,5,51-59. Porges, S.W. Heart rate indices of newborn attentional responsivity. Merrill-Palmer Quarterly, 1974,20, 231-254. Porges, S.W., Arnold, W.R., & Forbes, E.J. Heart rate variability: An index of attentional responsivity in human newborns. Developmental Psychology, 1973,8, 85-92. Prather, M., Project Head Start teacher-pupil-parent interaction study. In E. Grotberg (Ed.), Review of research 1965 to 1969. Washington, D.C., Project Head Start (OEO), 1969. Prescott, E. et al. Group day care as a child-rearing environment: An observational study of day care programs. Pasadena, Calif., Pacific Oaks College, 1967. Prescott, E. The large day care center as a child-rearing environment. Pasadena, California, Pacific Oaks College, 1970. Prescott, E. A comparison of three types of day care and nursery school-home care. Paper presented at Biennial Meeting of society for Research in Child Development, March 1973. Pritchett, V. S. Street actions. New York Review of Books, July 23,1970. Ramalingaswami, V., & Deo, M. Experimental protein-calorie malnutrition in the rhesus monkey. In R. A. McCance & E.M. Widdowson (Eds.), Calorie deficiencies andprotein deficiencies. London, Churchill, 1968. Rapaport, D. Some metapsychological considerations concerning activity and passivity. In M.M. Gill (Ed.), Collected papers of David Rapaport. New York, Basic Books, 1967. Ray, W.S. A preliminary report on a study of fetal conditioning. Child Development, 1932,3,175-177. Rebelsky, F., & Hanks, C. Fathers' verbal interaction with infants in the first three months of life. Child Development, 1911,42,63-68. Record, R.G., McKeown, T., & Edwards, J.H. The relation of measured intelligence to birth order and maternal age. Annals of Human Genetics, 1969,33,61-69. Reese, H.W., & Overton, W.F. Models of development and theories of development. In L.R. Goulet & P.B. Baltes (Eds.), Life-span developmental psychology: Theory andresearch. New York, Academic Press, 1970. Pp. 115-145. Ricciuti, N.H. Malnutrition, learning and intellectual development: Research and remediation. In F.F. Korten, S.W. Cook & J.I. Lacey (Eds.), Psychology and the problems of society. Washington, D.C., American Psychological Association, 1970. Richards, M.P.M., & Bernal, J.F. Social interaction in the first days of life. In H.R. SchafFer (Ed.), The origins of human relations. New York, Academic Press, 1971. Pp. 3-13. Riegel, K.F. Developmental psychology and society: Some historical and ethical considerations. In J.R. Nesselroade & H.W. Reese (Eds.), Life-span developmental psychology: Methodological issues. New York, Academic Press, 1973. Pp. 1-24. Roberts, D.F. A developmental study of opinion change: Source-orientation versus content-orientation at three age levels. Unpublished doctoral dissertation, Stanford University, 1968.

Bibliography

xxi

Rohwer, W. Prime time for education: Early childhood or adolescence? Harvard Educational Review, 1971,41,316-341. Rokeach, M. The open and closed mind. New York, Basic Books, 1960. Rosenshine, B., & Fürst, N. The use of direct observation to study teaching. In R. Travers (Ed.), Second handbook of research on teaching. Chicago, Rand McNally, 1973. Rosow, I. Social integration of the aged. New York, Free Press, 1967. Rosow, I. Retirement leisure and social status. In Duke University Council on Aging and Human Development, Proceedings of seminars, 1965-1969. Durham, Duke University Center for the Study of Aging and Human Development, 1969. Pp. 249-257. Ross, M.H. Aging, nutrition and hepatic enzyme activity patterns in the rat. Journal of Nutrition, 1969,97,563-601. Rothbart, M.K. Birth order and mother-child interaction in an achievement situation. Journal of Personality and Social Psychology, 1971,17,113-120. Rothbart, M. K., & Maccoby, E. E. Parents' differential reactions to sons and daughters. Journal of Personality and Social Psychology, 1966,4,237-243. Rovet, J. Can spatial skills be acquired through visual media. Unpublished doctoral dissertation, University of Toronto, 1974. Rubin, K.H. Egocentrism in childhood: A unitary construct? Child Development, 1973, 44,102-110. Rule, B.G., & Duker, P. Effects of intentions and consequences on children's evaluations of aggressors. Journal of Personality and Social Psychology, 1973, 27, 184— 189. Ryle, G. The concept of mind. London, Peraguine Books, 1949. Salomon, G. Can we affect cognitive skills through visual media? An hypothesis and initial findings. A VCommunication Review, 1972,20,401-423. Salomon, G. Sesame Street in Israel: Its instructional and psychological effects on children. New York, Children's T.V. Workshop, 1974. Sameroff, A.J. Can conditioned responses be established in the newborn infant: 1971? Developmental Psychology 1971,4,1-12. Sameroff, A.J. Schizotaxia revisited: Model issues in the etiology of schizophrenia. Paper presented at the Dorado Beach Conference in Risk Research, Dorado Beach, Puerto Rico, October 1972. Sameroff, A. J., & Zax, M. Perinatal characteristics of the offspring of schizophrenic women. Paper presented at the Meeting of the American Psychological Association, Honolulu, September 1972. Sampson, E.E. The study of ordinal position: Antecedents and outcomes. In B. Maher (Ed.), Progress in experimental personality research, vol. 2. New York, Academic Press, 1965. Samuels, J.S. Effects of pictures on learning to read, comprehension and attitudes. Review of Educational Research, 1970,40,397-409. Sander, L. Adaptive relationships in early mother-child interaction. Journal of the American Academy of Child Psychiatry, 1964,3,231-264. Sarason, I. G. Test anxiety, general anxiety, and intellectual performance. Journal of Consulting Psychology, 1957,21,485-590. Schorr, A. Filial responsibility in the modern American family. Washington, Social Security Administration, 1960. Schorr, A. On selfish children and lonely parents. Public Interest, 1966,4,8-12. Schramm, W.,Lyle, J., & Parker, E.B. Television in the lives of our children. Stanford, Stanford University Press, 1961. Schuck, J.R. Pattern discrimination and visual sampling by the monkey. Journal of

xxii

Bibliography

Comparative Physiological Psychology, 1960,53,251-255. Schuck, J.R., Polidora, V.J., McConnell, D.C., & Meyer, D.R. Response location as factor in primate pattern discrimination. Journal of Comparative Physiological Psychology, 1961,54,543-545. Sears, R.R., Maccoby, E.E., & Levin, H. Patterns of child rearing. Evanston, Row, Peterson, 1957. Sears, R.R., Rau, L., & Alpert, R. Identification and child rearing. Stanford, Calif., Stanford University Press, 1965. Selman, R.L. The relation of role-taking to the development of moral judgment in children. Child Development, 1971,42,79-91.(a) Selman, R.L. Taking another's perspective: Role-taking development in early childhood. Child Development, 1971,42,1721-1734.(b) Shakow, D. On doing research in schizophrenia. In R. Cancro (Ed.), The schizophrenic syndrome: An annual review. New York, Brunner/Mazel, 1971. Pp. 189-232. Shanan, J. Family structure as a predictor of transient amenorrhea in adolescent girls of two cultures. Proceedings of the Second International Congress of Psychosomatic Medicine and Childbirth. Vienna, 1965. Pp. 337-380. Shanan, J. Active coping behavior. Megamot., 1967,16,188-196. (Hebrew). Shanan, J. Psychological changes during the middle years. Jerusalem, Gon and Grafica, 1968. (Hebrew). Shanan, J. Cross-cultural design in the study of structural aspects of development during middle and later adulthood. In F.K. Monks, W.H. Hartup, & J. de Wit, (Eds.), Determinants of behavioral development. New York, Academic Press, 1972. Shanan, J. Coping behavior in the prediction of complex tasks. Proceedings of the Seventeenth International Congress of Applied Psychology. Bruxelles, Editest, 1973. Shanan, J., Brzezinsky, A., Sulman, F., & Sharon, M. Active coping behavior, anxiety and cortical stereoid excretion in the prediction of transient amenorrhea. Behavioral Science, 1965,70,461-^65. Shanan, J., & Stein, M.I. Early detachment and independence in a study of creativity. Proceedings of the Thirty-third Annual Convention of the Society for Social Research. University of Chicago, 1956. Shanas, E. Family help patterns and social class in three countries. Journal of Marriage and the Family, 1967,29,257-266. Shanas, E. Family kin networks and aging in cross-cultural perspective. Journal of Marriage and the Family, 1973, 3, 505-554. Shanas, E., & Streib, G. (Eds.), Social structure and the family: Generational relations. Englewood Cliffs, N.J., Prentice-Hall, 1965. Shantz, C.U., & Watson, J.S. Spatial abilities and spatial egocentrism in the young child. Child Development, 1971,42,171-181. Sherif, M., & Sherif, C. W. Social Psychology, New York, Harper & Row, 1969. Sherman, H.C. Chemistry of food and nutrition. New York, The MacMillan Company, 1952. Sigel, I.E. The distancing hypothesis: A Neo-Piagetian view. Paper presented at the Annual Meetings of the American Psychological Association, Miami, 1970. Simmons, L. The real generation gap: A speculation on the meaning and implication of the generation gap. Youth and Society, 1971,3,119-135. Siqueland, E.R. Reinforcement patterns and extinction in human newborns. Journal of Experimental Child Psychology, 1968,6,431^442. Siqueland,E.R., & DeLucia, C.A. Visual reinforcement of non-nutritive sucking in human infants. Science, 1969,165,1144-1146. Skinner, B. F. Science and human behavior. New York, Macmillan, 1953. Skodak, M., & Skeels, H.M. A final follow-up of one hundred adopted children.

Bibliography

xxiii

Journal of Genetic Psychology, 1949, 75, 3-19. Smilansky, S. The effects of sociodramatic play on disadvantaged preschool children. New York, Wiley, 1968. Smith, M.S. Some short-term effects ofproject Head Start: A report on the second year of planned variation, 1970-71. Cambridge, Huron Institute, 1973. Soar, R.S. A measure of teacher classroom management. Paper presented at Annual Meeting of American Psychological Association, Washington, 1971. Soar, R.S., & Soar, R. M. An empirical analysis of selected Follow-Through programs: An example of a process approach to evaluation. In I.J. Gordon (Ed.), Early childhood education. Chicago, The University of Chicago Press, 1972. Soar, R. S., Soar, R. M., Ragosta, M., & Blekking, E. Factors affecting pupil cognitive growth in disadvantaged children. Paper presented at Annual Meeting of American Education Research Association, Chicago, 1972. Sobel, D. E. Children of schizophrenic patients: Preliminary observation on early development. American Journal of Psychiatry, 1961,118,512-517. Sontag, L.W. Effect of noise during pregnancy upon foetal and subsequent behavior. Paper presented at the Meeting of the American Association for the Advancement of Science, Boston, December 1969. Sontag, L.W., & Wallace, R.F. The movement response of the human fetus to sound stimuli. Child Development, 1935,6,253-258. Sontag, L. W., & Wallace, R. F. Changes in the rate of the human fetal heart in response to vibratory stimuli. American Journal of Diseases of Children, 1936,51,583-589. Spelt, D.K. The conditioning of the human fetus in utero. Journal of Experimental Psychology, 1948,38,375-376. Spicer, E. (Ed.), Human problems in technological change. New York, Russell Sage, 1952. Spitz, R. A. The first year of life. New York, International Universities Press, 1965. Stamps, L.E., & Porges, S.W. Heart rate variability and heart rate conditioning in human newborns. Paper presented at the Meeting of the Society for Psychophysiological Research, Galveston, October 1973. Stanford Research Institute. Implementation of planned variation in Head Start: Preliminary evaluations of planned variation in Head Start according to followthrough approaches (1969-70). Menlo Park, Calif., 1971. Stayton, D. J., & Ainsworth, M.D. S. Individual differences in infant responses to brief, everyday separations as related to other infant and maternal behaviors. Developmental Psychology, 1973,9,226-235. Stearns, Mariam S. Report on preschool programs: The effect of preschool programs on disadvantaged children and their families. Washington, D.C., DHEW, Office of Child Development, 1971. Steiner, G. The people look at television. New York, Knopf, 1963. Stern, G., Caldwell, B., Hersher, L., Lipton, E., & Richmond, J. A factor analytic study of the mother-infant dyad. Child Development, 1969,40,163-181. Sternglanz, S.H., & Serbin, L. An analysis of the sex roles presented on children's television programs. Paper presented at the Biennial Meeting of the Society for Research in Child Development, Philadelphia, March 1973. Steuer, F.B., Applefield, J.M., &Smith, R. Televised aggression and the interpersonal aggression of preschool children. Journal of Experimental Child Psychology, 1971, 11,442-447. Stoch, M. B., & Smythe, P. M. Does undernutrition during infancy inhibit brain growth and subsequent intellectual development? Archives of Diseases in Childhood, 1963, 38, 546. Stoch, M.B., & Smythe, P.M. Undernutrition during infancy and subsequent brain

xxiv

Bibliography

growth and development. In N. S. Scrimshaw & J. E. Gordon (Eds.), Malnutrition, learning, and behavior. Cambridge, Mass., M.I.T. Press, 1968. Pp. 278-288. Stout, A.M. Parent behavior toward children of differing ordinal position and sibling status. Unpublished doctoral dissertation, University of California at Berkeley, 1960. Strauss, S., & Langer, J. Operational thought inducement. Child Development, 1970,41, 163-175. Strobel, D.A., & Zimmermann, R.R. Manipulatory responsiveness in protein malnourished monkeys. Psychonomic Science, 1971,24,19-20. Strobel, D. A., & Zimmermann, R.R. Responsiveness of protein deficient monkeys to manipulative stimuli. Developmental Psychobiology, 1972,5,291-296. Surgeon General's Scientific Advisory Committee on Television and Social Behavior. Television and growing up: The impact of televised violence. Washington, D. C., U.S. Government Printing Office, 1972. Sussman, M. Isolated nuclear family: Fact or fiction. Social Problems, 1959,6,333-340. Sutton-Smith, B. Role replication and reversal in play. Merrill-Palmer Quarterly, 1966, 12,285-298. Sutton-Smith, B., & Rosenberg, B.G. The sibling. New York, Holt, Rinehart, & Winston, 1970. Sutton-Smith, B., & Rosenberg, B.G. Sibling consensus on power tactics. Journal of Genetic Psychology, 1968,112,63-72. Swedish Information Service, New York. Children in our environment. 1971. Tanner, J. M. Growth at adolescence. Oxford, Blackwell Publications, 1962. Tanner, J.M., &Taylor, G.R. Growth. New York, Time, Inc. 1965. Tappel, A. L. Where old age begins. Nutrition Today, 1967,2,2-7. Thomae, H. Persönlichkeit und Altern. Veröffentlichungen der Deutschen Gesellschaft für Gerontologie, 1968,1,191-203. Thoman, E. B., Barnett, C. R., & Leiderman, P. H. Feeding behaviors of newborn infants asafunction of parity of the mother. Child Development, 1971,42,1471-1483. Thoman, E.B., Leiderman, P.H., & Olsen, J.P. Neonate-mother interaction during breast-feeding. Developmental Psychology, 1972,6,110-118. Thoman, E.B., Turner, A.M., Leiderman, P.H., & Barnett, C.R. Neonate-mother interaction: Effects of parity on feeding behavior. Child Development, 1970, 41, 1103-1111. Thomas, A., Chess, S., & Birch, H. Temperament and behavior disorders in children. New York, New York Universities Press, 1968. Thomas, S.A. Violent content in television: Effect of cognitive style and age in mediating children's aggressive responses. Proceedings of the Eightieth Annual Convention, American Psychological Association, 1972. Thompson, W. R. Influence of prenatal maternal anxiety on emotionality in young rats. Science, 1957,125,698-699. Toman, W., Family constellation as a basic personality determinant. Journal for Individual Psychology, 1959,15,199-211. Toman, W. Family constellation. New York, Springer, 1962. (rev. ed., 1969; (German translation, 1965). Toman, W. The 'duplication theorem' of social relationships as tested in the general population. Psychological Review, 1971, 78,380-390. Toman, W. Geschwisterkonstellation und Leistungsmotivation. Schule und Psychologie, 1972,19,65-72. (a) Toman, W. Complementarity of sibling roles in friendship systems. Paper presented at the 20th International Congress of Psychology, Tokyo 1972.(b)

Bibliography

xxv

Toman, W., Preiser, S., Gasch. B., & Plattig, G. Die Wirkungen von Familienkonstellationen auf die Person, ihre sozialen Beziehungen und die nachfolgende Generation. Projekt der Deutschen Forschungsgemeinschaft, Bad Godesberg, 1967. Toman, W., & Preiser, S., Familienkonstellationen und ihre Störungen. Stuttgart, Enke, 1973. Townsend, P. Family life of oldpeople. London, Routledge & Kegan Paul, 1957. Troldahl, V.C., & Powell, F. A. A short-form dogmatism scale for use in field studies. Social Forces, 1965, 211-215. Tukey, J. M. Discussion on temporal changes in treatment-effect correlation. In Glass (Ed.), Proceedings of the 1971 invitational conference on testing problems. Princeton, Educational Testing Service, 1972. Turiel, E. An experimental test of the sequentiality of developmental stages in the child's moral judgment. Journal of Personality and Social Psychology 1966,3,611-618. U.S. Bureau of the Census. Statistical abstract of the United States. Washington Department of Commerce, 1970,1971,1972. Valadian, I. The adolescent - his growth and development. Paper presented at the National Education Conference, Washington, D.C., 1971. Van Lieshout, C. F. M. Het vaststellen van stabiliteit in de sociale interacties van kleuters. Een onderzoek bij hörende en dove kleuters. Nijmegen, Berkhout, 1972.(a) Van Lieshout, C.F.M. Structure of'social interaction in preschool children: Stability over time and relation to concurrent and contingently antecendent and consequent behavior. In F. R. Mönks, W. W. Hartup, & J. de Wit (Eds.), Determinants of behavioral development. New York, Academic Press, 1972.(b) Verma, S. Relationships between a day-care teacher's beliefs and practices. Unpublished doctoral dissertation, Pennsylvania State University, 1973. Walberg, M.J. Generalized regression models in educational research. American Educational Research Journal, 1971,5,71-93. Walker, D., Bane, M.-J., & Bryk, T. The quality of Head Start planned variation data. Cambridge, Huron Institute, 1973. Wallach, M.A., & Kogan, N. Modes of thinking in young children. New York, Holt, Rinehart & Winston, 1965. Walters, R.H., & Thomas, E.L. Enhancement of punitiveness by audiovisual displays. Canadian Journal of Psychology, 1963,17,244-255. Watson, G. Some personality differences in children related to strict or permissive parental discipline. Journal of Psychology, 1957,44,227-249. Watts, J.C., Barnett, I.C., & Halfar, C. Environment, experience and development, Report submitted to the Office of Economic Opportunity, 1973. Weikart, D.P. Relationship of curriculum, teaching, and learning in preschool education. Paper presented at the Hyman Blumberg Memorial Symposium on Research in Early Childhood Education, Baltimore, February 1971. Wenger, M. A. An investigation of conditioned responses in human infants. University of Iowa Studies in Child Welfare, 1936,12,7-90. Werner, H., & Kaplan, B. Symbol formation. New York, Wiley, 1967. Wernersson, I. Socialisation och skola. Nägra teoretiska utgängspunkter. Rapporter frán pedagogiska institutionen, Göteborgs universitet, no. 82,1972. White, S.H. Evidence for a hierarchical arrangement of learning processes. In L.P. Lipsitt and C.C. Spiker (Eds.), Advances in child development and behavior, vol. 2. New York, Academic Press, 1965. White, S., with Day, M.C., Freeman, P., Hantman, S., & Messenger, K. Federal pro-

xxvi

Bibliography

grams for young children. Cambridge, Huron Institute, 1972. Wickens, D.D., & Wickens, C.D. A study of conditioning in the neonate. Journal of Experimental Psychology, 1940,26,94-102. Wilinsky, H. Observational techniques in preschool classrooms. Institute for Developmental Studies, School of Education, New York University. (In ERIC Bibliography No. 3. Urbana, 111., University of Illinois, 1968,15-23.) Will, O.A. Schizophrenia: The problem of origins. In J. Romano (Ed.), The origins of schizophrenia. New York, Excerpta Medica Foundation, 1967. Pp. 214-227. Winick, M. Cellular growth of the placenta as an indication of abnormal fetal growth. In K. Adamsons (Ed.), Diagnosis and treatment of fetal disorders. New York, Springer 1967. Winterbottom, M. In D.C. McClelland, J.W. Atkinson, R.A. Clark, & E.L. Lowell (Eds.), The achievement motive. New York, Appleton-Century-Crofts, 1953. Pp. 297-306. Wise, L. The effects of protein deprivation on dominance measured by shock avoidance competition and food competition. Unpublished master's thesis, University of Montana, 1972. Wise, L. A., & Zimmermann, R.R. Shock thresholds of low and high protein reared rhesus monkeys. Perceptual and Motor Skills, 1973,36,674. Wise, L.A., Zimmermann, R.R., & Strobel, D. A. Dominance measurements of low and high protein reared rhesus macaques. Behavioral Biology, 1973,9,77-84. Wohlwill, J. F. Responses to class-inclusion questions for verbally and pictorially presented items. Child Development, 1968,59,449^*65. Wohlwill, J. F. The concept of experience: S or R? Human Development, 1973, 16, 90-107. World Health Organization. World Health Organization Technical Reprint Service, 1965,7,302. Wozniak, R.H. Dialectical materialism, and cognitive developmental theory: A reexamination of certain basic assumptions of Piagetian theory. Paper presented at the Biennial Meeting of the Society for Research in Child Development, Philadelphia, 1973. Yamazaki, J.N., Wright, S.W., & Wright, P.M. Outcome of pregnancy in women exposed to atomic bomb in Nagasaki. American Journal of Diseases of Children, 1954, 57,448-463. Yarden, P., & Nevo, B. The differential effect of the schizophrenic mother's stages of illness on her children. British Journal of Psychiatry, 1968,114,1089-1096. Yarrow, L. J. Maternal deprivation: Toward an empirical and conceptual re-evaluation. Psychological Bulletin, 1961,5«, 459^90. Yarrow, L.J., Rubenstein, J.L., Pedersen, F.A., & Jankowski, J.J. Dimensions of early stimulation and their differential effects on infant development. MerrillPalmer Quarterly, 1972,18,205-218. Youniss, J. Classificatory schemes in relation to class inclusion before and after training. Human Development, 1971,14,171-183. Zielinski, W. Schulleistung, Intelligenz und Familienkonstellation. Inauguraldissertation, Nürnberg, Universität Erlangen, 1966. Zimmermann, R.R. Effects of age, experience and malnourishment on object retention in learning set. Perceptual and Motor Skills, 1969,28,867-876. Zimmermann, R.R. Retention in learning set and latent memory as a function of malnutrition and deprivation. Annual Meeting of the Rocky Mountain Psychological Association, Albuquerque, New Mexico, 1970.

Bibliography

xxvii

Zimmermann, R. R. Reversal learning in the developing malnourished rhesus monkey. Behavioral Biology, 1973, 8,381-390. Zimmermann, R.R., Steere, P.O., Strobel, D.A., & Horn, H.L. Abnormal social development of protein malnourished rhesus monkeys. Journal of Abnormal Psychology, 1972,50,125-131. Zimmermann, R.R., Strobel, D.A. & Maguire, D. Neophobic reactions in protein malnourished infant monkeys. Proceedings of the Annual Convention of the American Psychological Association, 1970. Pp. 187-198. Zimmermann, R.R., & Wells, A.M. Performance of malnourished rats on the HebbWilliams closed field maze learning task. Perceptual and Motor Skills, 1971, 33, 1043-1050. Zunich, M. Child behavior and parental attitudes. Journal of Psychology, 1966, 62, 41-46.

Index to Part II

Activity theory, 625-627 Activation, 417-420 Adaptation, 611-612, 620, 601-606, 608 Adolescence, 628-629, 631, 637-639 Adult education, 611-613, 620 Affluence, 640-652 Aggression, 495-503 Aging, 445-456 Anxiety, 611, 616-620 Assessment, 522, 529-530, 539-540 Attachment, 676-686 Attention, 469-473, 524

Deafness, 728-736 Deficiencies, 445-456 Dependence, mutual, 640-652 Depressive mothers, 677, 679-682 Deprivation, 457-464 Development, 445-456, 499-503 Developmental models, 690-693 dialectic, 692-693 Discipline, 422-423 Discrimination, 467-473 Disengagement, 601-610, 621-622, 625627 Dogmatism, 616-620

birth, order, 654, 658, 660-663, 696-705 Categorization, 719 Cognitive development, 413-420, 457-464 475-485, 487-494, 723-727 Cognitive tests, 560-565 Communication medium, 475-485 Concept learning, 717-721 Conceptual style, 719-722 Consistency, 623-625 Coping, behavior, 601-610 style, 601-610 Culture, 663 Curriculum, preschool, 556-565 Curriculum implementation, 547-555 Curriculum strategies, 518-522, 525-528, 534-539, 542

Education, 475-485 home based, 507-514 Educational, skills, 611-612, 615-616, 619-620 Ego development, 601-610 Empathy, 591-600 Engagement, 601-610 Environment, 413-420, 507-511, 541-442 Environmental reciprocity, 421-430 Family, loss of member, 709-710 Father-child interaction, 696, 702-703 Follow Through, 547, 548, 552 Friendship patterns, 711 Generation gap, 628-639 Geriatrics, 443

xxx

Index

Group control, 640-652 Growth, 445-451 Head Start, 543,547-555,556-565 Identification, 422 Imitation, 495-503 Independence, 640-652 Infancy, 457-464, 653-663, 664-675 cohorts, 687-695 Information, televised, 499-500 Instruction, 475-485 Intellectual development, 457-464, 515518 Interaction, family, 653-663 social, 728-736 Israel, 490-494 Knowledge, 475-485 general, 611-615 Language, 523-530 functions, 523-524, 529 Life satisfaction, 621-622 Malnutrition, 445-456, 457-464, 465-473 Maternal elaboration, 516, 520-522, 528 Media, 475-485,487-494 Model explicitness, 579-580 Monkeys, 465-473 Mother-child interaction, 524-525, 531534,664-675, 676-687,696-703 Nursery school behavior, 710-711 Nutrition, 445-456 Nutritional rehabilitation, 471-472, 457, 461-464 Observation, 729-734 Observation system, 566-578 Opportunities, 640-652 Opposition, 630-631, 634-635 Organismic, 413-420 Patent-child interactions, 634-635, 637639, 696-705 Personality-disordered mothers, 677, 679683 Perspective-taking, 591-600 Persuasibility, 500-502 Planted Variation, 547-555,436-443, 556565

Play, 515-522 Playgrounds, 431-444 adventure, 438-441 equipment, 438-443 Political opinions, 631-633 Pregnancy, 446-448 Prenatal influences, 687-690, 695 Preschool children, 457-464 Preschool programs, 556-565, 566-574, 579-588 evaluation, 586-588 implementation, 579-585 models, 579-588 Psychopathology, 664-675, 676-684 Reciprocity, 640-652 environmental, 421-430 parental, 425-427 teacher-child, 427-428 Recreation movement, 433-438 Recursive thinking, 592, 600, 569-570, 573, 577 Role-taking, conceptual, 591-593, 595-600 emotional-motivational, 592-600 perceptual, 591-600 Research design, 513-514 Satisfaction, 621-627 Schizophrenic mothers, 676-683 Secular trends, 687, 688, 693, 695 Self-interest, 641, 648-652 Sesame Street, 490-494, 498 Sex differences, 654, 656, 660-663, 696698,699-705 Siblings, 707-709 interaction, 715-722 relations, 703-705 sibsize, 723-727 theory, 712-713 Skill, cognitive, 478-485, 487-494 Social, class, 723-727 contacts, 621-627 contexts, 711-712, 713-714 development, 531-540 participation, 621-627 role, 622-627 Social deprivation, 457-464, 601-610 Socialization, 421-430, 495-503, 628-629, 653-663 Solidarity, 640-652

Index Staff characteristics, 583-585 Strategies, 509-412 Stress, 453-454 Structuralism, 413-415

Teacher role, 552, 553 Team sports, 436-437 Television, 487-494, 495-503 Trait theory, 729

xxxi