Fertility and Family Planning in the United States 9781400877447

Citation preview

FERTILITY AND FAMILY PLANNING IN THE UNITED STATES

Fertility and Family Planning in the United States BY

PASCAL K. WHELPTON ARTHUR A. CAMPBELL JOHN E. PATTERSON

PRINCETON UNIVERSITY PRESS PRINCETON, NEW JERSEY

1966

Copyright © 1966 by Princeton University Press ALL RIGHTS RESERVED

L.C. Card: 65-12995

Printed in the United States of America by The Maple Press Company

PREFACE Whether or not to have another child is an important decision, both from a personal and a national point of view. If married couples in the United States were to have an average of 2.5 births by the time they completed their families, the population would eventually level off and nearly stop growing. However, if couples had only one more child apiece, or 3.5 altogether, the population would double within 40 years. We have reached a point in our demographic history where either of these developments is clearly possible. The lower alternative has already been demonstrated by married women who reached age 50 in the period 1955-1965. The higher alternative may be approached, if not attained, by younger women who are still in the reproductive years of life. It has become increasingly important, therefore, to try to understand the variety of factors that influence family growth in the United States. The first nationwide effort to find out about couples' family planning attitudes and practices was an interview survey conducted in the spring of 1955 under the direction of the Survey Research Center of the University of Michigan and the Scripps Foundation for Research in Population Problems, Miami University, Oxford, Ohio. The findings of this survey are reported in Family Planning, Sterility, and Population Growth, by Ronald Freedman, Pascal K. Whelpton, and Arthur A. Campbell (McGraw-Hill Book Company, 1959) and in other publications listed in the Selected Bibliography, which appears in the latter part of the present book. In the summer of 1960, the same organizations sponsored a second survey, which is the subject of this book. One important purpose of the second survey was to see how well the wives interviewed in 1955 had predicted the number of children that women Uke themselves would have in the 1955—1960 period. In addition, the second study was designed to get more information on certain subjects, such as couples' ability to have children and their success in using contracep­ tion, than was obtained in the first study. Also, the second study provides, for the first time, some data on the family planning attitudes and practices of nonwhite couples. Both the 1955 and the 1960 surveys were supported largely by grants from the Rockefeller Foundation. Additional funds were ν

PREFACE

provided by the Population Council. Neither of these organizations is responsible in any way for the findings or interpretations of the authors. Two sponsoring committees endorsed the purposes and auspices of the 1960 survey in letters that interviewers showed to respondents in order to reassure them about the scientific usefulness of the study and about the confidentiality of their individual contributions to it. Although the committee members endorsed the purposes of the survey, it must be clearly understood that they have not been consulted about the findings or interpretations of the authors and are not responsible in any way for them. Their names and affiliations, as shown on the sponsoring letters, are presented below. The members of the general sponsoring committee were as follows: Dr. John D. Millett, President, Miami University, Oxford, Ohio Mr. Frederick Osborn, Former Presi­ dent, The Population Council, New York City Dr. Thomas Parran, President, Avalon Foundation, New York City Dr. Lowell J. Reed, President Emeri­ tus, The Johns Hopkins Univer­ sity, Baltimore, Maryland Miss Anna Lord Strauss, Former President, League of Women Voters of the United States Mr. Charles P. Taft, Member of City Council of Cincinnati Dr. R. B. Von Kleinsmid, Chancellor, University of Southern California

Mr. Samuel W. Anderson, Former Assistant Secretary of Commerce for International Affairs Dr. Detlev W. Bronk, President, The Rockefeller Institute, New York City Dr. Harry J. Carman, Member, Board of Higher Education, New York City Rev. Lawrence J. Cross, S.J., Chair­ man, Dept. of Sociology and Social Work, University of De­ troit, Detroit, Michigan Rev. Harry Emerson Fosdick, Pastor Emeritus, Riverside Church, New York City Mr. Ralph McGxll, Editor, The Atlanta Constitution

The members of the medical sponsoring committee were as follows: Dr. Nicholas J. Eastman, The Johns Hopkins Hospital, Baltimore, Maryland Dr. John E. Gordon, School of Pub­ lic Health, Harvard University, Boston, Massachusetts Dr. Alan F. Guttmacher, The Mount Sinai Hospital, New York City

Dr. Russell R. De Alvarez, School of Medicine, University of Wash­ ington Dr. Baynard Carter, Duke University Medical Center, Durham, North Carolina Dr. Frederick A. Coller, St. Joseph Mercy Hospital, Ann Arbor, Michigan Vt

PREFACE

Dr. Frank R. Lock, The Bowman Gray School of Medicine, Wake Forest College, Winston-Salem, North Carolina Dr. John Dale Owen, St. Mary's Hos­ pital, Milwaukee, Wisconsin Dr. John Rock, Clinical Professor of Gynecology, Emeritus, Harvard Medical School, Boston, Massa­ chusetts

Dr. Howard C. Taylor, Jr., College of Physicians and Sturgeons, Columbia University, New York City Dr. Herbert F. Traut, University of California Hospital

The interview schedule for the second study was developed in late 1959 and early 1960 in collaboration with Richard F. Tomasson, who was on the staff of the Scripps Foundation at that time, Harry P. Sharp, and Harold Organic, who were with the Survey Research Center. Two preliminary versions of the schedule were pretested: one in December 1959 (in Detroit, Tallahassee, and a southern rural area), and the other in February 1960 (in Paterson, Little Rock, and another southern rural area). The sample was designed and selected under the direction of Leslie Kish and Bernard Lazerwitz of the Survey Research Center. The interviewing was done in the summer of 1960 under the direction of Morris Axelrod of the Survey Research Center. His staff of supervisors and professional interviewers did a truly remarkable job of questioning women about subjects of a highly personal and delicate nature. Their success is attested to by the fact that fewer than 1 per cent of the wives interviewed refused to tell about their use of con­ traception. The editing and coding of the interviews was supervised by Doris Muehl of the Survey Research Center. Throughout the study, both in the early phases and during the analysis of the data, the authors benefited greatly from the generous help and advice of the senior author of the first study, Ronald Freedman of the University of Michigan. During the analytical phase of the study, many people were asked to read preliminary versions of the chapters and to offer their com­ ments. In the process, they made many helpful suggestions, which are gratefully acknowledged. In addition to Ronald Freedman they include Thomas K. Burch, AnsIey Coale, Jerry W. Combs, Jr., John V. Grauman, Frederick S. Jaffee, Clyde V. Kiser, Norman Lawrence, Everett S. Lee, Frank Lorimer, Edward W. Pohlman, Steven Polgar, Robert Potter, Jr., Lee Rainwater, Philip C. Sagi, Warren S. Thompson, Christopher Tietze, and Charles F. Westoff. Ruth W. Smith of the Scripps Foundation gave valuable assistance

PREFACE

in the statistical analysis of the survey data, which was essential to the success of this study. Mrs. Paul Demeny drafted the figures for publication. Mrs. Annice Cottrell prepared the index. During the final stages of the work on the manuscript of this book, the senior author, Pascal K. Whelpton, died. The junior authors wish to acknowledge with great thanks his stimulating and kindly direction of all phases of the study. The extent to which this book may be regarded as useful is due largely to him. ARTHUR A. CAMPBELL JOHN E. PATTERSON

via

CONTENTS Preface List of Abbreviated References 1. Background of the Study and the Reliability of Birth Expectations 2. Ideal, Desired, and Expected Family Size 3. Family Size by Religion and Socioeconomic Status 4. Fecundity Impairments 5. The Control of Fertility 6. Family Planning 7. Methods and Effectiveness of Contraception 8. Trends and Differentials in the Timing and Spacing of Births 9. White—Nonwhite Differences in Fertility 10. Family Size and Population Growth in Future Years

ν xxxi

1 32 69 125 174 221 276 300 334 371

APPENDICES

A. Sampling B. Technical Notes

405 416

Selected Bibliography

433

Index

439

TABLES 1. Number of Women Interviewed in 1960, by Characteristics Defining the 1960 Sample, by Eligibility for the 1955 Sample (Including Glossary of Symbols used) 2. Per Cent Distribution by Births by 1955 for Wives 18-39 Years Old in 1955, According to Sample for 1955 and 1960 3. Per Cent Distribution by Number of Births in 1955-1960, Expected by Wives in the 1955 Sample and Reported by Wives in the 1960 Sample, for Wives 18-39 Years Old in 1955, by Fecundity in 1955 4. Per Cent Distribution by Most Likely Total Number of Births Expected, for Wives 18-39 Years Old in 1955, According to Samples for 1955 and 1960 5. Consistency of 1955 and 1960 Statements about Numbers or Births, by Wife's Religion, by Cohort 6. Consistency of 1955 and 1960 Statements about Numbers of Births, by Wife's Education, by Cohort 7. Consistency of 1955 and I960 Statements about Numbers of Births, by Wife's Religion, by Wife's Education 8. Consistency of 1955 and 1960 Statements about Numbers of Births, by Wife's Education, by Number of Births in 1955 9. Consistency of 1955 and 1960 Statements about Numbers of Births, by Use of Contraception, by Cohort 10. Consistency of 1955 and 1960 Statements about Numbers of Births, by Fecundity, Duration of Marriage, Husband's Occupation, Couple's Farm Experience, and Region of Residence 11. Number of Children Considered Ideal for the Average American Family and Per Cent Distribution by Number Considered Ideal, for Wives, by Color 12. Per Cent Distribution by Number of Children Considered Ideal, 1941, 1945, 1955, and 1960 13. Average Number of Children Considered Ideal for the Average American Family, a High Income Family, and a Low Income Family, for Wives, by Color 14. Number of Children Wanted if Life Could Be Relived and

4

12

14

16 17 19 20

22 23

24

33 34

35

TABLES

15. 16. 17. 18.

19. 20.

21. 22. 23. 24. 25.

26.

27. 28.

29.

30.

Per Cent Distribution by Number Wanted, for Wives, by Color Number of Children Wanted at Interview and Per Cent Distribution by Number Wanted, for Wives, by Color Number of Children Wanted by Wife and by Husband, by Color Percentage of Wives Who Wanted Fewer Children than Their Husbands, the Same Number or More, by Color Average Number of Births by Interview, Children Living at Interview, Additional Births Expected, Total Births Expected, and Total Children Expected, for All Wives (1960), White Wives (1955 and 1960), and Nonwhite Wives (1960) Number of Births Expected and Per Cent Distribution by Number Expected, for Wives, by Color Number of Children Considered Ideal for the Average American Family, Wanted If Life Could Be Relived, Wanted at Interview, and Expected, for Wives, by Color Percentage of Wives Who Expected Fewer Children than They Wanted, the Same Number or More, by Color Per Cent Distribution by First Reason for Not Wanting a Smaller Family, for Wives, by Color Per Cent Distribution by First Reason for Not Expecting a Larger Family, for Wives, by Color Number of Births Expected and Per Cent Distribution by Number Expected, for White Wives, by Age Number of Births by 1960 and Additional and Total Number Expected, for White Wives, by Age in 1960, by Age at First Marriage Number of Births by 1960 and Additional and Total Number Expected, for White Wives, by Age, by Years Since First Marriage Additional and Total Number of Births Expected, for White Wives, by Age, by Number of Births by 1960 Number of Births by Interview and Additional and Total Number Expected, for White Wives, by Age in 1955 and 1960 Number of Children Considered Ideal for the Average American Family, Wanted If Life Could Be Relived, Wanted at Time of Interview, and Most Likely Expected, for White Wives, by Age, Years Since First Marriage, Age at First Marriage, and Number of Births by 1960 Percentage Who Expected Fewer Children than They xii

37 38 41 41

46 47

52 53 54 55 58

60 61

62

63

65

TABLES

31. 32. 33.

34. 35. 36. 37.

38.

39.

40.

41.

42.

43.

44.

45. 46.

Wanted, the Same Number or More, for White Wives, by Age, Years Since First Marriage, Age at First Marriage, and Number of Births by 1960 Per Cent Distribution by First Reason for Not Wanting a Smaller Family, for White Wives, by Age Per Cent Distribution by First Reason for Not Expecting a Larger Family, for White Wives, by Age Births by 1960 and Total Number Expected, for Wives, by Religion of Wife, Husband, and Couple, and by Who Performed the Marriage Per Cent Distribution by Total Number of Births Ex­ pected, for Wives, by Religion Births by 1960 and Total Number Expected, for Protestant Wives, by Type of Denomination Births by 1960 and Total Number Expected, for Wives, by Religion, by Age in 1960, by Age at First Marriage Births by 1960 and Additional and Total Number Ex­ pected, for Wives, by Religion, by Years Since First Marriage Births by Interview and Additional and Total Number Expected, for Wives, by Religion, by Age, in 1955 and 1960 Births by 1960 and Total Number Expected by Wife, for Catholic Wives and Husbands, by Frequency with Which They Received the Sacraments Births by 1960 and Total Number Expected by Wife, for Catholic Wives and Husbands, by Amount of Elementary and Secondary Education in CathoKc Schools Most Likely Expected Total Number of Births, for Catholic Wives, by Amount of Elementary and Secondary Educa­ tion in Catholic Schools, by Age Most Likely Expected Total Number of Births, for Catholic Wives, by Amount of Elementary and Secondary Educa­ tion in Catholic Schools, by Education Total Number of Births Expected, for Wives, by Religion, by Importance of Religious Activities in Daily Life of Family When Wife Was a Child Total Number of Births Expected, for Wives, by Religion, by Importance of Religious Activities in Daily Life of Family at Present Total Number of Births Expected, for Wives, by Religion, by Change in Wife's Interest in Religion in Recent Years Number of Children Considered Ideal for the Average

66

67 68

71 72 74 76

77

78

83

84

86

87

88

89 90

TABLES

American Family, Wanted if Life Could Be Relived, Wanted at Time of Interview, and Most Likely Expected, for Wives, by Religion 47. Percentage Who Expected Fewer Children than They Wanted, the Same Number or More, for Wives, by Religion 48. Per Cent Distribution by First Reason for Not Wanting a Smaller Family, for Wives, by Religion 49. Per Cent Distribution by First Reason for Not Expecting a Larger Family, for Wives, by Religion 50. Births by 1960 and Total Number Expected, for Wives, by Education of Wife, Husband, and Couple 51. Births by 1960 and Total Number Expected, for Wives, by Religion and Education 52. Total Number of Births Expected, for Wives, by Age, by Education, by Religion 53. Total Number of Births Expected, for Wives, by Religion, by Education, 1955 and 1960 54. Number of Children Considered Ideal for the Average American Family, Wanted if Life Could Be Relived, Wanted at Time of Interview, and Most Likely Expected, for Wives, by Religion, by Education 55. Percentage Who Expected Fewer Children than They Wanted, the Same Number or More, for Wives, by Religion and Education 56. Per Cent Distribution by First Reason for Not Wanting a Smaller Family, for Wives, by Education 57. Per Cent Distribution by First Reason for Not Expecting a Larger Family, for Wives, by Education 58. Births by 1960 and Total Number Expected, for Wives, by Husband's Income 59. Total Number of Births Expected, for Wives, by Religion, by Husband's Income 60. Number of Children Considered Ideal for the Average American Family, Wanted if Life Could Be Relived, Wanted at Time of Interview and Most Likely Expected, for Wives, by Husband's Income 61. Percentage Who Expected Fewer Children than They Wanted, the Same Number or More, for Wives, by Husband's Income 62. Births by 1960 and Total Number Expected, for Wives, by Family Income 63. Births by 1960 and Total Number Expected, for Wives, by

90

91 92 93 95 96 97 98

99

100 101 102 104 105

105

106 107

TABLES

64.

65.

66.

67.

68. 69. 70. 71.

72.

73. 74. 75.

76.

77.

Years Wife Worked Since First Marriage, by Reason for Working Total Number of Births Expected, for Wives, by Years Wife Worked Since First Marriage and Reason for Work­ ing, by Fecundity of Couple Total Number of Births Expected, for Wives, by Years Wife Worked Since First Marriage and Reasons for Work­ ing, by Wife's Religion Number of Children Considered Ideal for the Average American Family, Wanted if Life Could be Relived, Wanted at Time of Interview, and Most Likely Expected, for Wives, by Year's Wife Worked Since First Marriage and Reason for Working Percentage Who Expected Fewer Children than They Wanted, the Same Number or More, for Wives, by Years Wife Worked Since First Marriage and Reason for Working Births by 1960 and Total Number Expected, for Wives, by Husband's Occupation Total Number of Births Expected, for Wives, by Hus­ band's Occupation, by Wife's Religion Total Number of Births Expected, for Wives, by Hus­ band's Occupation, 1955 and 1960 Number of Children Considered Ideal for the Average American Family, Wanted if Life Could Be Relived, Wanted at Time of Interview, and Most Likely Expected, for Wives, by Husband's Occupation Percentage Who Expected Fewer Children than They Wanted, the Same Number or More, for Wives, by Hus­ band's Occupation Births by 1960 and Total Number Expected, for Wives, by Size of Place of Residence Total Number of Births Expected, for Wives, by Size of Place of Residence, by Religion Number of Children Considered Ideal for the Average American Family, Wanted if Life Could Be Relived, Wanted at Time of Interview, and Most Likely Expected, for Wives, by Size of Place of Residence Percentage Who Expected Fewer Children than They Wanted, the Same Number or More, for Wives, by Size of Place of Residence Births by 1960 and Total Number Expected, for Wives, by Farm Residence of Wife and/or Husband

109

109

110

110

111 113 113 114

115

115 118 118

119

120 121

TABLES

78. Births by 1960 and Total Number Expected, for Wives, by Region of Residence 79. Number and Per Cent Distribution by Fecundity, for White and Nonwhite Couples 80. Percentage of Couples with Contraceptive and Remedial Operations, by Wife's Age, Duration of Marriage, and Number of Births 81. Per Cent Distribution by Reason for Operation, for Couples with Operations that Prevent Conception, by Type of Operation 82. Percentage of Couples with Contraceptive and Remedial Operations, by Use of Contraception, whether Couple Had an Accidental Conception, and Pregnancy Planning Status 83. Percentage of Couples with Contraceptive and Remedial Operations, by Wife's Religion, Education, Region of Residence, and Size of Place of Residence 84. Percentage of Couples with Contraceptive Operations, by Region of Residence, by Spouse on Whom Operation was Performed, by Reason for Operation 85. Percentage of Couples with Contraceptive Operations, by Region of Residence, by Wife's Religion and Education, and Size of Place of Residence 86. Percentage of Couples with Contraceptive and Remedial Operations, by Wife's Age, 1955 and 1960 87. Percentage of Couples Classified as Subfecund because of Operation Preventing Pregnancy, Recognized Pathology, or Long Period of Noncontraceptive Exposure, by Wife's Age, by Duration of Marriage 88. Percentage of Couples Classified as Subfecund, by Dura­ tion of Marriage, by Use of Contraception 89. Percentage of Couples Classified as Subfecund, by Wife's Age, by Number of Conceptions before Contraception was Begun 90. Per Cent Distribution by Fecundity, for Couples by Wife's Religion and Education 91. Per Cent Distribution by Fecundity, for Non-Sterile Couples, by Wife's Religion, by Use of Contraception 92. Per Cent Distribution by Fecundity, for Non-Sterile Couples, by Wife's Education, by Use of Contraception 93. Per Cent Distribution by Fecundity, for Couples by Wife's Age, 1955 and 1960 XVt

122 132

135

136

141

142

145

147 150

156 157

157 158 158 159 161

TABLES

94. Per Cent Distribution by Evidence of Subfecundity, for Couples by Wife's Age, 1955 and 1960 95. Percentage of Couples Who Are Childless and in Specified Fecundity Categories, by Wife's Age 96. Percentage of Couples Who Are Childless, by Fecundity, by Wife's Age 97. Percentage of Wives Who Expect Fewer Children than Wanted and Are in Specified Fecundity Categories, by Wife's Age 98. Percentage of Wives Who Expect Fewer Children than Wanted, by Fecundity, by Wife's Age 99. Average Number of Births by Interview, Average Number of Births and Children Expected, Average Number of Children Wanted by Wife, and Per Cent that Number of Children Expected Is Above or Below Number Wanted, for Couples by Fecundity 100. Average Number of Children Expected and Wanted by Wife, for Couples by Fecundity, by Wife's Age 101. Average Number of Births by Interview and Average Total Number Expected, for Fecund and Subfecund Cou­ ples, and Estimates of These Averages for Subfecund Couples and All Couples Assuming No Fecundity Impair­ ments, by Wife's Age 102. Per Cent Distribution by Wife's Attitudes toward Fertility Control, for Wives, by Religion and Education 103. Per Cent Distribution by Wife's Attitude toward Fertility Control, for Wives by Religion, by Frequency of Church Attendance 104. Per Cent Distribution by Wife's Attitude toward Fertility Control, for Protestant Wives by Type of Denomination, and for Catholic Wives by Frequency of Sacraments 105. Per Cent Distribution by Wife's Attitude toward Fertility Control, for Catholic Wives by Amount of Education in Catholic Schools 106. Per Cent Distribution by Wife's Attitude toward Fertility Control, for Wives, by Religion of Wife and Husband 107. Per Cent Distribution by Wife's Attitude toward Fertility Control, for Wives by Cohort, from 1916 to 1942 108. Percentage of Couples Who Have Used or Expect to Use Contraception, by Fecundity, Wife's Age, Duration of Marriage, and Number of Births 109. Percentage of Couples Who Have Used or Expect' to Use Contraception, by Wife's Religion and Education, HusXVll

162 164 165

167 167

169 171

171 177

178

179

180 180 181

184

TABLES

110.

111.

112.

113.

114.

115.

116.

117.

118. 119.

120.

band's Income and Occupation, Region of Residence, and Size of Place of Residence Percentage of Couples Who Expect Not to Use Contra­ ception, by Fecundity by Use of ODFC for Fecund Cou­ ples, by Wife's Age, Duration of Marriage, and Number of Births Percentage of Couples Who Expect Not to Use Contra­ ception, by Fecundity, by Use of ODFC for Fecund Cou­ ples, by Wife's Religion and Education, Husband's Income and Occupation, Region of Residence, and Size of Place of Residence Per Cent Distribution by Degree of Subfecundity, and Average Number of Births, for Subfecund Nonusers Who Expect Not to Use Contraception Cumulative Percentage of Couples Who Used Contracep­ tion before Specified Pregnancy, and Percentage Who Are Fecund or Subfecund Nonusers, by Number of Preg­ nancies, by Wife's Education Cumulative Percentage of Couples Who Used Contra­ ception before Specified Pregnancy, and Percentage Who Are Fecund or Subfecund Nonusers, by Number of Preg­ nancies, by Husband's Income Cumulative Percentage of Couples Who Used Contracep­ tion before Specified Pregnancy, and Percentage Who Are Fecund or Subfecund Nonusers, by Number of Preg­ nancies by Wife's Religion Percentage of Users Who Delayed Using Contraception because They Wanted One or More Children As Soon As Possible, for Users Who Did Not Use Contraception before the First Pregnancy, by Wife's Religion and Education, and Husband's Income Percentage of Couples Who Used Contraception before the First Pregnancy, by Wife's Attitude toward Fertility Control, by Wife's Religion and Education Percentage of Couples Who Have Used or Expect to Use Contraception, by Wife's Religion and Education Percentage of Couples Who Used Contraception before the First Pregnancy, by Wife's Religion and Education, by Husband's Income and Occupation, Region of Residence, Size of Place of Residence, and Number of Years Wife Worked, by Reason for Working Percentage of Couples Who Have Used Contraception, by Fecundity, by Wife's Religion and Education, Husband's XVtll

185

188

189

190

194

196

197

198

200 201

202

TABLES

121. 122.

123.

124.

125.

126.

127.

128.

129. 130.

131.

132.

Income and Occupation, and Number of Years Wife Worked Per Cent Distribution by Use of Contraception and Inten­ tion to Use, for Couples, by Cohort, 1916-1942 Per Cent Distribution by Number of Pregnancies before Use of Contraception Began and by Fecundity for Nonusers, for Couples by Number of Pregnancies, 1955 and 1960 Percentage of Couples Who Had Used Contraception by 1955 and 1960, for Couples in 1960 Sample, by Cohort by Eligibility for Study in 1955 and 1960 Percentage of Couples Who Had Used Contraception by 1955 and 1960, Percentage Who Expected to Use as of 1955, and Percentage Who Began Use during 1955-1960, for Couples Eligible for the 1955 Sample by Cohort Percentage of Couples Who Have Used or Expect to Use Contraception, by Fecundity, Wife's Age, Duration of Marriage, and Number of Births, 1955 and 1960 Percentage of Couples Who Have Used or Expect to Use Contraception, by Wife's Religion and Education, Hus­ band's Income and Occupation, and Number of Years Wife Worked, 1955 and 1960 Percentage of Couples Who Have Used Contraception and Percentage Who Have Used or Expect to Use, by Wife's Age, by Wife's Education, 1955 and 1960 Percentage of Couples Who Have Used Contraception and Percentage Who Have Used or Expect to Use, by Wife's Age, by Wife's Religion, 1955 and 1960 Per Cent Distribution by Contraceptive Status, for All Pregnancies, by Order of Pregnancy Number of Births by Interview, Most Likely Total Number of Births Expected, Number of Children Wife Wants, Age at Marriage, Number of Years Since First Marriage, Per Cent Subfecund, and Per Cent Users, for All Couples, by Pregnancy Planning Status Average Number of Months in Inter-Birth Intervals, Average Number of Years from Marriage to Births of Specified Order, and Average Age at Births of Specified Order, for All Wives, by Pregnancy Planning Status Percentage with Completely Planned Fertility, for Couples Who Began Using Contraception before the First Preg­ nancy, by Wife's Age, Duration of Marriage, and Number of Births

205 207

210

211

213

214

216

217

218 223

229

230

231

TABLES

133. Average Number of Births Expected and Children Wanted by Wife, for All Couples, by Pregnancy Planning Status, by Husband's Income 134. Average Number of Births Expected and Children Wanted by Wife, for All Couples, by Pregnancy Planning Status, by Wife's Religion 135. Average Number of Births Expected and Children Wanted by Wife, for All Couples, by Pregnancy Planning Status, by Wife's Education 136. Per Cent Distribution by Pregnancy Planning Status, for All Couples and Fecund Couples, by Wife's Age, Dura­ tion of Marriage, and Number of Births 137. Per Cent Distribution by Pregnancy Planning Status, for All Couples, by Wife's Education and Religion 138. Percentage Whose Last Conception Was Unwanted and of Specified Contraceptive Status, for All Couples, by Wife's Education and Religion, and Husband's Income 139. Per Cent Distribution by Pregnancy Planning Status, for Couples with Protestant Wives by Type of Denomination, and for Couples with Catholic Wives by Frequency of Sacraments 140. Per Cent Distribution by Pregnancy Planning Status, for All Couples, by Wife's Religion and Education 141. Percentage with Excess Fertility, for Couples with Protes­ tant and Catholic Wives, by Number of Births 142. Per Cent Distribution by Pregnancy Planning Status, for All Couples, by Husband's Income and Ocupation, and Number of Years Wife Worked by Reason for Working 143. Per Cent Distribution by Pregnancy Planning Status for All Wives, by Wife's Education, by Husband's Income 144. Per Cent Distribution by Pregnancy Planning Status, for All Couples, by Size of Place of Residence and Region of Residence 145. Per Cent Distribution by Pregnancy Planning Status, Using 1955 Planning Status Categories, for All Couples, Fecund Couples, and Fecund Users, 1955 and 1960 146. Percentage of Couples with Completely Planned or Excess Fertility, for Couples Eligible for the 1955 Sample, by Cohort, 1955 and 1960 147. Percentage of Couples with Completely Planned Fertility, Observed for 1955 and 1960, and Hypothetical Percentage for 1960, by Wife's Age

240

241

242

245 247

248

253 254 255

257 258

260

261

262

263

TABLES

148. Percentage of Couples with Completely Planned Fertility, by Number of Births, 1955 and 1960 149. Percentage of Couples with Completely Planned Fertility, by Eligibility for 1955 and 1960 Samples, 1955 and 1960 150. Percentage of Couples with Completely Planned Fertility, for Couples Eligible for the 1955 Sample, by Cohort by Number of Births by 1955, 1955 and 1960 151. Percentage of Couples with Completely Planned Fertility, by Wife's Religion and Education, Number of Years Wife Worked Since Marriage, Husband's Income and Occupa­ tion, Size of Place of Residence, and Region of Residence 152. Percentage of Couples with Excess Fertility, by Number of Births, Wife's Age, and Duration of Marriage, 1955 and 1960 153. Percentage of Couples Who Expect No More Births, by Number of Births, 1955 and 1960 154. Percentage of Couples with Excess Fertility, by Wife's Religion and Education, 1955 and 1960 155. Percentage of Couples with Excess Fertility, by Number of Years Wife Worked Since Marriage, Husband's Income and Occupation, Size of Place of Residence, and Region of Residence 156. Percentage of Users Who Have Used Specified Method of Contraception, by Wife's Religion, 1955 and 1960 157. Percentage of Users Who Have Ever Used Specified Method of Contraception, for Couples with Protestant and Catholic Wives, by Wife's Education, 1955 and 1960 158. Percentage Distribution by Type of Contraception Ever Used, for Users, by Wife's Religion, 1955 and 1960 159. Percentage Distribution by Type of Contraception Used Most Recently, for Users, by Wife's Religion, 1955 and 1960 160. Percentage Who Conform to Catholic Teachings on Methods of Contraception, for All Couples with Catholic Wives, by Wife's Frequency of Church Attendance, 1955 and 1960 161. Per Cent Distribution by Wife's Attitude toward Use of the Contraceptive Pill, by Wife's Religion, by Use of Con­ traception 162. Percentage Distribution of Wives by Whether They Thought They Would Use "The Pill," for Protestant and Catholic Couples Who Had Used Contraception, by Selected Characteristics

263 264

265

267

268 270 272

273 278

281 283

284

285

287

288

TABLES

163. Number of Contraceptive Failures per 100 Years of Con­ traceptive Exposure during the First 12 Months of the Specified Interpregnancy Interval, by Type of Failure, for Couples Uang Contraception during the Specified Interval, by Interval of Use, by Number of Children Wife Wants 164. Number of Contraceptive Failures per 100 Years of Con­ traceptive Exposure during the First 12 Months of the Specified Interpregnancy Interval, by Type of Failure, for Couples Using Condom Only or Rhythm Only during Specified Interval, by Interval of Use, by Number of Children Wife Wants 165. Median Age at First Marriage and at First, Second, and Third Birth, for White Women in the Cohorts of 18801889 to 1935-1939 166. Median Length of Interval between Marriage and First Birth, Second Birth, and Third Birth, for White Women Who First Married in 1900-1909 to 1955-1959 167. Age at First Marriage and Length of Birth Intervals, for White Wives, by Cohort, by Years Since First Marriage 168. Average Length of Birth Intervals and Per Cent of Wives Having Births of Given Orders, for Cohorts, by Age at Marriage 169. Age at Marriage and Length of Birth Interval, for White Wives When at Specified Ages, by Cohort 170. Age at Marriage, Age-Specific Birth Rates, and These Rates as a Percentage of Most Likely Expected Total Births, for White Wives Married by Specified Ages, by Cohort 171. Final Birth Rate per 1,000 Women Living to Age 50, and Percentage Distribution by Age, for All Women in Cohorts of 1891-1895 to 1951-1955 172. Estimated Increase in Births and Population Growth dur­ ing Five-Year Periods Due to Different Ages of Childbearing: Timing Patterns of 1911-1915 Cohorts Compared with Those of Later Cohorts 173. Age at Marriage, Average Years Married, Length of Birth Intervals, Per Cent Having Specified Number of Births, and Number of Births by Interview and Total Expected for Selected Characteristics 174. Per Cent Distribution by Months between Marriage and First Birth, for Wives, by Religion and Education 175. Age at Marriage, Average Years Married, Length of Birth Intervals, Per Cent Having Specified Number of Births, XXll

294

296

303

304 306

307 309

313

317

318

321 324

TABLES

176.

177.

178.

179.

180.

181. 182.

183.

184.

185.

186.

187.

188.

and Number of Births by Interview and Total Expected for Protestant and Catholic Wives, by Wife's Education Age at Marriage, Average Years Married, Length of Birth Intervals, Per Cent Having Specified Number of Births, and Number of Births by Interview and Total Ex­ pected, by Wife's Education by Interval When Contracep­ tion First Used Age at Marriage and Per Cent Distribution of Most Likely Expected Total Births by Age for Protestant and Catholic Wives Marrying before Age 25, by Cohort Age at Marriage and Per Cent Distribution of Most Likely Expected Total Births by Age, by Wife's Education, by Cohort Age at Marriage and Per Cent Distribution of Most Likely Expected Total Births by Agej by Husband's Occupation, by Cohort Age at Marriage and Per Cent Distribution of Most Likely Expected Total Births by Age, by Size of Place of Resi­ dence, by Cohort Age at Marriage and Per Cent Distribution of Most Likely Expected Total Births by Age, by Region, by Cohort Per Cent Distribution by Number of Births by 1960 and Most Likely Expected Total Births, for Nonwhites, Matched Whites, and All Whites Average Number of Births by 1960 and of Most Likely Expected Total Births, for Nonwhites, Matched Whites, and All Whites, by Region of Residence, Southern Farm Residence, and Wife's Education Number of Children Ever Born per 1,000 Ever-Married Women 18-39 Years Old, for Nonwhites and Whites, by Region of Residence, by Rural-Urban Residence in South, 1960 Census Number of Children Ever Born per 1,000 Ever-Married Women 15-39 Years Old, for Nonwhites and Whites, by Wife's Education, 1960 Census Average Number of Births by 1960, for Nonwhites, Matched Whites, and All Whites, by Wife's Education, by Couple's Southern Farm Residence Average Most Likely Expected Total Number of Births, for Nonwhites, Matched Whites, and All Whites, by Wife's Education, by Couple's Southern Farm Residence Average Number of Births by 1960 and of Most Likely Expected Total Births, for Nonwhites, Matched Whites,

325

326

328

329

330

331 332

337

339

340 344

345

346

TABLES

189.

190. 191. 192.

193.

194.

195.

196.

197.

198.

199.

200.

201.

and All Whites, by Cohort, Use of Contraception, and Pregnancy Planning Status Average Number of Children Wife Wants, for Nonwhites, Matched Whites, and All Whites, by Region of Residence, Southern Farm Residence, and Wife's Education Per Cent Distribution by Fecundity, for Nonwhites, Matched Whites, and All Whites Per Cent Distribution by Evidence of Subfecundity, for Nonwhites, Matched Whites, and All Whites Percentage of Couples Who Have Used or Expect to Use Contraception and Percentage Who Do Not Expect to Use Contraception, by Fecundity, for Nonwhites, Matched Whites, and All Whites, by Southern Farm Residence Percentage of Couples Who Have Used or Expect to Use Contraception and Percentage Who Do Not Expect to Use Contraception, By Fecundity, for Nonwhites, Matched Whites, and All Whites, by Region of Residence Percentage of Couples Who Have Used or Expect to Use Contraception and Percentage Who Do Not Expect to Use Contraception, By Fecundity, for Nonwhites, Matched Whites, and All Whites, By Wife's Education Percentage of Couples Who Have Used or Expect to Use Contraception and Percentage Who Do Not Expect to Use Contraception, By Fecundity, for Nonwhites, Matched Whites, and All Whites, by Cohort Percentage of Users Who Have Used Specified Method of Contraception, for Nonwhites, Matched Whites, and All Whites Per Cent Distribution by Pregnancy Planning Status, for Nonwhites, Matched Whites, and All Whites, by Southern Farm Residence Per Cent Distribution by Pregnancy Planning Status, for Nonwhites, Matched Whites, and All Whites, by Wife's Education Per Cent Distribution by Pregnancy Planning Status, for Nonwhites, Matched Whites, and All Whites, by Region of Residence Percentage of Couples with Excess Fertility, for Nonwhites, Matched Whites, and All Whites, by Number of Births by 1960 Percentage Whose Last Conception Was Unwanted and of Specified Contraceptive Status, for Nonwhites, Matched Whites, and All Whites, by Wife's Education XXlV

348

350 352 353

354

358

359

359

360

362

364

365

365 367

TABLES

202. Cumulative Birth Rate by Age in 1955 or 1960, for Wives in GAF Studies, and for Ever-Married Women, According to Cohort Fertility Tables and the 1960 Census 203. Expected Final Birth Rate of GAF Wives Marrying at Specified Ages as a Percentage of the Rate for Those Marrying by Younger Ages, and Projected Percentages for Women Who Will Marry after 1960, by Cohort 204. Most Likely Final Birth Rate Expected by Wives, and Estimated for Single Women Who Will Marry before Ages 45-49, by Cohort, by Age at Marriage 205. Increase in Most Likely Expected Final Birth Rate as White Wives Grew Older from 1955 to 1960, and Pro­ jected Increase for Years after 1960, by Cohort 206. Final Birth Rate Expected by GAF Wives in 1960, Esti­ mated for Single Women Who Will Marry Later and for Other Ever-Married Women, and Adjusted for the Tend­ ency to Increase with Age, by Cohort 207. Projected Final Birth Rate of Ever-Married Women in Cohorts of 1916-1920 and Later, and Actual Rate for Prior Cohorts 208. Projected or Actual Final Birth Rate of Ever-Married Women, and Rate by Younger Ages as a Percentage of That Rate, by Cohort 209. Projected or Actual Cumulative Marriage Rate of All Women by Ages 45-49, and Rate by Younger Ages as a Percentage of That Rate, by Cohort 210. Projected or Actual Cumulative Birth Rates by Specified Ages, and Rates during Age Periods, for All Women, by Cohort 211. Actual or Projected Age-Specific Birth Rates of All Women, and Gross Reproduction Rate, by Five-Year Periods 212. Number of Births and Crude Birth Rates: Actual for 1910-1915 to 1955-1960, and Projected for 1960-1965 to 1980-1985 213. Projected Population of the U.S., by Age, July 1, 1960 to 1985, with Medium, Low, and High Fertility Trends (in millions)

373

375

376

379

380

381

387

389

392

396

397

398

APPENDIX TABLES

A-I. Response Rates for Women Eligible for the 1960 Study A-2. Per Cent Distribution by Selected Characteristics, for 1960 Sample of White and Nonwhite Wives 18-39 Years Old,

406

TABLES

A-3.

A-4. A-5. A-6.

B-1.

B-2.

B-3.

B-4.

and Independent Estimates for the Same Population Groups Sample Values and Sampling Errors for Per Cent Fecund, Per Cent Users, and Per Cent with Completely Planned Fertility, for White Wives, 18-39 Years Old, by Selected Characteristics Approximate Sampling Errors of Percentages Approximate Sampling Errors of Differences between Percentages Approximate Sampling Errors of Averages and of Differ­ ences between Two Averages, for Number of Births by Interview and Total Number of Births Expected Estimated Distribution of Couples in the 1955 Study, by 1955 Fecundity Classification, by 1960 Fecundity Classifi­ cation Per Cent Distribution by Characteristics on Which White and Nonwhite Couples Were Matched, for Nonwhites, Matched Whites, and All Whites Per Cent Distribution by Characteristics on Which White and Nonwhite Couples Not Living on Farm Were Matched, for Nonwhites, Matched Whites, and All Whites Per Cent Distribution by Cohort, Husband's Income, and Couple's Southern Farm Residence, for Nonwhites, Matched Whites, and All Whites

408

412 412 413

414

422

429

430

431

FIGURES 1. Per cent distribution by actual and expected number of births in 1955-1960, for wives eligible for the 1955 study 2. Actual and expected average number of births in 1955— 1960, for wives eligible for the 1955 study, by wife's education 3. Average most likely expected total number of births, for wives eligible for the 1955 study, by wife's education, 1955 and 1960 4. Births by 1960 and total number expected, for wives, by color 5. Number of children considered ideal for the average Ameri­ can family, wanted if life could be relived, wanted at time of interview, and expected, for wives, by color 6. Percentage of wives who expected fewer children than they wanted, the same number or more, by color 7. Births by 1960 and total number expected, for white wives, by age 8. Births by 1960 and total number expected, for wives, by religion 9. Births by 1960 and total number expected, for wives, by religion, by age 10. Births by 1960 and total number expected, for Catholic wives, by frequency with which they received the sacra­ ments 11. Births by 1960 and total number expected, for wives, by education, by religion 12. Births by 1960 and total number expected, for wives, by husband's income 13. Births by 1960 and total number expected, for wives, by years worked since first marriage and reason for working 14. Births by 1960 and total number expected, for wives, by husband's occupation 15. Births by 1960 and total number expected, for wives, by size of place of residence 16. Births by 1960 and total number expected, for wives, by region of residence 17. Per cent distribution by fecundity, white couples

15

18

26 49

51 55 57 70 75

82 94 104 108 112 117 122 133

FIGURES

18. Percentage of couples who had contraceptive or remedial operations, by wife's age 19. Percentage distribution by reason for operation, by kind of operation 20. Percentage of couples who had contraceptive or remedial operations, by wife's religion 21. Percentage of couples who had contraceptive or remedial operations, by region of residence 22. Percentage of couples who had contraceptive or remedial operations, by wife's education 23. Percentage of couples classified as Subfecund, by basis of classification, by duration of marriage 24. Average number of children expected and wanted by wife, by couple's fecundity 25. Per cent distribution by wife's attitude toward fertility control, for Protestant and Catholic wives, by education 26. Per cent distribution by use of contraception, intention to use, and fecundity ODFC status of those who will not use, by wife's education 27. Percentage of couples who ever used contraception before specified pregnancy, for couples with four pregnancies, by wife's education 28. Percentage of couples who have used or expect to use contra­ ception, by cohort 29. Percentage of couples who had used contraception by 1955 for the 1955 and 1960 studies, percentage who expected to use for the 1955 study, and percentage who began use in 1955-1960 for the 1960 study, for couples eligible for the 1955 study, by cohort 30. Percentage of couples who have used or expect to use con­ traception, by wife's education, 1955 and 1960 31. Per cent distribution by pregnancy planning status, for couples, by duration of marriage 32. Per cent distribution by pregnancy planning status, for couples, by wife's education 33. Per cent distribution by pregnancy planning status, for couples, by wife's education, by duration of marriage 34. Per cent distribution by pregnancy planning status, for couples, by wife's religion 35. Percentage of Users who have used specified method of con­ traception, 1955 and 1960 36. Per cent distribution by number of months between first marriage and first birth, white wives, 18-39 XXVlll

135 137 140 144 146 152 170 176

186

193 207

212 215 244 247 251 252 279 323

FIGURES

37. Births per 1,000 women 15-44 years old, for whites and nonwhites, United States 1920-1961 38. Average most likely expected total number of births for nonwhites, matched whites, and all whites, by couple's Southern farm residence 39. Average most likely expected total number of births for nonwhites, matched whites, and all whites, by wife's education 40. Per cent User for nonwhites, matched whites, and all whites, by couple's Southern farm residence 41. Per cent Excess Fertility for nonwhites, matched whites, and all whites, by couple's Southern farm residence 42. Per cent Excess Fertility for nonwhites, matched whites, and all whites, by wife's education 43. Actual and medium, low and high projected final birth rates, for all ever-married women living to middle age, by groups of cohorts 44. Cumulative medium birth rates of ever-married women by specified ages, in per cent of final birth rate, by groups of cohorts 45. Actual and medium projected cumulative birth rates from 15-19 to 45-49, for all women, by groups of cohorts

335

341 344 356 363 364

383

386 394

LIST OF ABBREVIATED REFERENCES ABBREVIATION

REFERENCE

Axelrod 1963

Morris Axelrod5 Ronald Freedman, David Gold­ berg, and Doris Slesinger, "Fertility Expectations of the United States Population: A Time Series," Population Index, Vol. 29, No. 1, pp. 25-31, January, 1963

Badenhorst 1963

L. T. Badenhorst, "Family Limitation and Methods of Contraception in an Urban Population," Popula­ tion Studies, Vol. XVI, No. 3, p. 290, March, 1963

The Catholic Hospital Association of the United Catholic Hospital Association 1959 States and Canada, Ethical and Religious Direc­ tives for Catholic Hospitals, Second edition, St. Louis, Missouri, 1959 Census 1940

Sixteenth Census of the United States, 1940, Popu­ lation: Differential Fertility 1940 and 1910, Women by Number of Children Ever Born, 1945

Census 1950

U.S. Bureau of the Census, 1950 Census of Popu­ lation, Special Reports, Vol. IV, Part 5, Chapter C, "Fertility"

Census 1958

U.S. Bureau of the Census, "Illustrative Projections of the Population of the United States, by Age and Sex: 1960 to 1980." Current Population Reports, Series P-25, No. 187, Nov. 10, 1958

Census 1960

U.S. Bureau of the Census, 1960 Census of Popu­ lation, "Fertility of the Population, 1960," Supple­ mentary Reports, PC (Si)-42, March 15, 1963

Census 1960a

U.S. Bureau of the Census, 1960 Census of Popu­ lation, PC(I)-ID

Census 1960b

U.S. Bureau of the Census, 1960 Census of the Population, PC(1)-46D

Census 1962

U.S. Bureau of the Census, Statistical Abstract of the United States, 1962, p. 58

Commonweal 1964 The Commonweal, Vol. 80, No. 11, June 5, 1964 (a special issue on "Responsible Parenthood") FPSPG

Ronald Freedman, Pascal K. Whelpton, and Arthur A. Campbell, Family Planning, Sterility, and Popu­ lation Growth, McGraw-Hill Book Company, Inc., New York, 1959 XXXt

LIST OF ABBREVIATED REFERENCES

Freedman 1954

Ronald Freedman and Harry Sharp, "Correlates of Values About Ideal Family Size in the Detroit Metropolitan Area," Population Studies, Vol. 8, No. 1, p. 37, July, 1954

Freedman 1961

Ronald Freedman, Pascal K. Whelpton, and John W. Smit, "Socio-economic Factors in Religious Differentials in Fertility," American Sociological Review, Vol. 26, No. 4, pp. 608-614, August, 1961

Freedman 1963

Ronald Freedman, David Goldberg, and Doris Slesinger, "Current Fertility Expectations of Married Couples in the United States," Population Index, Vol. 29, No. 4, pp. 366-391, October, 1963

Goldberg 1959

David Goldberg, Harry Sharp, and Ronald Freed­ man, "The Stability and Reliability of Expected Family Size Data," Milbank Memorial Fund Quar­ terly, 37 (October, 1959) No. 4, pp. 369-385

Grabill 1958

Wilson H. Grabill, Clyde V. Kiser, and Pascal K. Whelpton, The Fertility of American Women, John Wiley and Sons, Inc., New York, 1958

Grabill 1959

Wilson H. Grabill and Paul C. Glick, "Demo­ graphic and Social Aspects of Childlessness: Census Data," Milbank Memorial Fund Quarterly, Janu­ ary, 1959, Vol. 37, No. 1, pp. 60-86

Grabill 1961

Wilson H. Grabill and Robert Parke, Jr., "Mar­ riage, Fertility and Childspacing: August, 1959," U.S. Bureau of the Census, Current Population Re­ ports, Series P-20, No. 108, July 12, 1961

Greville 1957

Τ. Ν. E. Greville, "Illustrative United States Popu­ lation Projections," U.S. Social Security Adminis­ tration, Actuarial Study No. 46, May, 1957

Hill 1959

Reuben Hill, J. Mayone Stycos, and Kurt W. Back, The Family and Population Control, The University

of North Carolina Press, Chapel Hill, 1959 Indianapolis Study

Pascal K. Whelpton and Clyde V. Kiser, eds., Social and Psychological Factors Affecting Fertility,

Milbank Memorial Fund, New York; five volumes published in 1946, 1950, 1952, 1954, and 1958 Kelly 1955

Gerald A. Kelly, S. J., Medico-Moral Problems, Clonmore and Reynolds, Ltd., Dublin, Republic of Ireland, 1955 XXXll

LIST OF ABBREVIATED REFERENCES

Lenski 1961

Gerhard Lenski, The Religious Factor: A Socio­ logical Study of Religion's Impact on Politics, Economics, and Family Life, Doubleday and Com­ pany, Garden City, 1961, pp. 216-217

NCCC 1961

NOVS 1960

National Council of the Churches of Christ in the United States of America, "Responsible Parent­ hood," a policy statement adopted by the General Board, February 23, 1961 National Office of Vital Statistics, Vital Statistics of the United States, 1960

O'Brien 1963

POQ

Rev. John A. O'Brien, "Family Planning in an Ex­ ploding Population," Ave Maria National Catholic Weekly, Vol. 98, No. 8, pp. 5-7, August 24, 1963 Public Opinion Quarterly, Vol. IX, Fall, 1945,

p. 372 Potter 1959

Robert G. Potter, Jr., "Contraceptive Practice and Birth Intervals among Two-Child White Couples in Metropolitan America," in Thirty Years of Re­ search in Human Fertility: Retrospect and Prospect,

Milbank Memorial Fund, New York, 1959, pp. 74-92 Potter 1962

Robert G. Potter, Jr., Philip C. Sagi, and Charles F. Westoff, "Improvement of Contraception During the Course of Marriage," Population Studies, Vol. 16, No. 2, pp. 160-174, November, 1962

Rainwater 1960

Lee Rainwater, And the Poor Get Children, Quad­ rangle Books, Chicago, 111., 1960

Sagi 1962

Philip C. Sagi, Robert G. Potter, Jr., and Charles F. Westoff, "Contraceptive Effectiveness as a Func­ tion of Desired Family Size," Population Studies, Vol. 15, No. 3, pp. 291-296, March, 1962

Schachter 1958

Joseph Schachter, "Childspacing as Measured from Data Enumerated in the Current Population Sur­ vey: United States, April 1950 to April 1954," National Office of Vital Statistics, Vital Statistics— Special Reports, Vol. 47, No. 3, October 9, 1958

Siegel 1964

Jacob S. Siegel, Meyer Zitter, and Donald S. Akers, "Projections of the Population of the United States, By Age and Sex: 1964 to 1985," U.S. Bureau of the Census, Current Population Reports, Series P-25, No. 286, July, 1964 XXXlll

LIST OF ABBREVIATED REFERENCES

Tietze 1957

Christopher Tietze, "Reproductive Span and Rate of Reproduction among Hutterite Women," Fer­ tility and Sterility, Vol. 8, No. 1, January-February, 1957

Tietze 1962

Christopher Tietze, "The Use-Effectiveness of Con­ traceptive Methods," in Clyde V. Kiser (ed.). Re­ search in Family Planning, Princeton University Press, Princeton, N.J., 1962, pp. 357-369

WestofF 1957

Charles F. Westoff, Elliot G. Mishler, and E. Lowell Kelly, "Preferences in Family Size and Eventual Fertility Twenty Years After," American Journal of Sociology, 42, No. 5 (March, 1957), pp. 491-497

Westoff 1961

Charles F. Westoff, Robert G. Potter, Jr., Philip C. Sagi, and Elliot G. Mishler, Family Growth in Metropolitan America, Princeton University Press, Princeton, N.J., 1961

Westoff 1961a

Charles F. Westoff, Robert G. Potter, Jr., and Philip C. Sagi, "Some Estimates of the Reliability of Survey Data on Family Planning," Population Studies, Vol. 15, No. 1, pp. 52-69, July, 1961

Westoff 1963

Charles F. Westoff, Robert G. Potter, Jr., and Philip C. Sagi, The Third Child, Princeton Univer­ sity Press, Princeton, N.J., 1963

Whelpton 1947

Pascal K. Whelpton assisted by Hope Tisdale Eldridge and Jacob S. Siegel, Forecasts of the Popu­ lation of the United States, 1945—1975, U.S. Bureau of the Census, 1947

Whelpton 1954

Pascal K. Whelpton, Cohort Fertility, Princeton University Press, Princeton, N.J., 1954

Whelpton 1960

Pascal K. Whelpton and Arthur A. Campbell, "Fertility Tables for Birth Cohorts of American Women, Part 1," National Office of Vital Statistics, Vital Statistics—Special Reports, Vol. 51, No. 1, January 29, 1960

Whelpton 1963

Pascal K. Whelpton, "Why Did the United States' Crude Birth Rate Decline During 1957-1962?" Population Index, Vol. 29, No. 2, pp. 120-125, April, 1963

FERTILITY AND FAMILY PLANNING IN THE UNITED STATES

GLOSSARY OF SYMBOLS IN TABLES: .. The category described by the column and line heading is not possible. — The category described by the column and line heading is possible, but it contains fewer than 0.5 per cent of the cases (if the statistic de­ scribed is a proportion) or has an average that rounds to zero (if the statistic described is an average). * Average or proportion not computed because base contains fewer than 20 cases. Sampling variability would be very large for such small numbers.

CHAPTER 1

Background of the Study and the Reliability of Birth Expectations INTRODUCTION

This book is about the attitudes and behavior that determine the number of children born to married couples in the United States. The basic information comes from 3,322 wives who were interviewed in May, June, and July, 1960, by women on the staff of the University of Michigan's Survey Research Center. The interviews dealt with topics of great personal importance: the number of pregnancies and births the wife had had, the number of children she and her husband wanted, the number she expected altogether, whether she or her hus­ band had any physical defects that made future births unlikely or impossible, whether they had ever tried to prevent pregnancy, and, if they had, the methods they had used. In addition, the wife gave a wide variety of information about the way she and her husband lived: their religion, educational attainment, income, whether or not the wife worked, and so forth. This study is similar to one conducted in 1955, also by the Survey Research Center and the Scripps Foundation.1 Both are known as the Growth of American Families (or "GAF") studies. One of the reasons for the 1960 survey was to see how well wives in the earlier study had been able to predict the number of children that women like themselves would have during a five-year period, and to see whether the total number of children expected by such women changed significantly between 1955 and 1960. Also, the 1960 study was designed to explore more thoroughly some topics that received only brief attention in the 1955 survey. FERTILITY TRENDS THAT RAISED QUESTIONS

One of the main purposes of both the 1955 and 1960 studies is to gather information that will help to improve forecasts of numbers of births in the United States. Ever since the demobilization of the armed forces after World War II, birth rates have been well above the levels observed before the war. This is due partly to the fact that recently married women were having more babies while they were young than older women had had when they were as young. 1The findings of the 1955 study are reported in FPSPG . See page xxxi for the meaning of any abbreviated reference.

FERTILITY AND FAMILY PLANNING IN THE U.S.

This concentrated unusually many births in the years during which the transition to younger childbearing was taking place. Another cause of high postwar birth rates is the fact that younger couples will have more births altogether than older couples had. By 1955 it was not known which of these causes was more im­ portant. The popular explanation was that families were getting larger—that is, that young couples would have more children by the time they completed their childbearing than older couples had had. However, a few demographers knew that a reduction in the average age at childbirth can cause a short-run spurt in the birth rate even when the average size of completed families remains the same, and thought that this factor had not been given enough attention. One of the present authors showed, for example, that birth rates were higher in the postwar period than in the 1930s because (a) a higher proportion of young women had married, (b) a higher proportion of those who married had a first birth, and (c) a higher proportion of those who had a first birth went on to have more births. The least important of these factors was the third, which is closely related to average family size.2 It was obvious, however, that it would be impossible to disentangle the effects of changes in the age of childbearing and changes in average completed family size until we knew approximately how many children younger couples were going to have altogether. We had to know this in order to make reasonable forecasts about future trends in the birth rate. If we found, for example, that young women were having their children earlier but would have no more children alto­ gether than older women, then we could forecast a substantial decline in the birth rate after the tendency toward younger childbearing ceased. However, if we found that younger women were going to have more children altogether than older women had had, then we could foresee a more gradual decline in birth rates as the effects of the tendency toward younger childbearing diminished. Ronald Freedman and Pascal K. Whelpton, the designers of the first Growth of American Families Study, thought that a promising source of information about future births was the women who would be having those births. In the 1955 survey, information was obtained from interviews with white married women in the main reproductive years of life (taken here as 18-39 years of age) who were living with their husbands or whose husbands were temporarily absent in the armed forces. These women were questioned at length about their pregnancy histories and the number of children they expected to have in the future. Later in this chapter we shall see how well 2 GrabiU

1958, pp. 365-371.

BACKGROUND—RELIABILITY OF BIRTH EXPECTATIONS

their replies agreed with the number of births actually occurring in 1955-60. THE SAMPLE

Because the 1960 study was intended partly to check the reliability and stability of the birth expectations of the wives who were inter­ viewed in 1955, the survivors of such women had to be interviewed in 1960. We could not interview the same women because they were not asked to give their names and addresses in 1955. The authors of both studies felt that women would be more likely to give honest replies to questions about very personal matters if they were assured anonymity. Furthermore, from a methodological point of view it was considered undesirable to interview the same women. What we wanted to know was whether the average replies of a sample of women would give us a good indication of the future number of births that all women like themselves would have. Therefore, we thought that 1960 interviews with different women who had the same character­ istics as those interviewed in 1955 would provide a much more critical test of the 1955 data on expectations than would second interviews with the original sample.3 For this reason, one essential component of the 1960 sample had to consist of women who would have been eligible for the 1955 sample. They had to be white women who were 18-39 years old in 1955, married and living with their husbands at that time, or temporarily absent from the husband because of his service in the armed forces. This meant that they had to be 23-44 years old as of 1960 and married as of five years previously. They could have been divorced, separated, or widowed in the meantime. Other aims of the 1960 study required data for the same kind of women that were interviewed in 1955—that is, white wives 18-39 years old as of 1960. In addition, we wanted to get fertility data for nonwhite wives 18-39 years of age—the first such information to be collected for this group on a nationwide basis. All of these purposes could be served with a sample containing the following components: 1. White and nonwhite women 18-39 years old, married and living with husband or with husband temporarily absent in the armed forces. Many of the white wives in this group can be used to represent the 1960 survivors of wives who were 18-34 years old in the 1955 study. 1 In fertility studies with different purposes, it may be desirable to reinterview the same women. This is the case with the Princeton Study and with the Detroit Area Study of 1962. See Bibliography for sources.

FERTILITY AND FAMILY PLANNING IN THE U.S.

2. White wives 40-44 years old, married and living with husband or with husband temporarily absent in armed forces. Nearly all of these wives can be used to represent the 35-39-year-old wives in the 1955 study. 3. White women 23-44 years old who had been married and living with their husbands as of 1955 (or whose husbands were tempo­ rarily absent in the armed forces at that time), but were no longer married and living with husband as of 1960. These women represent those who were married and 18-39 years old in 1955, but who had become widowed, divorced, or separated by 1960. Table 1 shows the numbers of women in the various components of the 1960 sample, and the numbers who represent the survivors of the 1955 sample. TABLE l Number of Women Interviewed in 1960, by Characteristics Defining the 1960 Samp le, by Eligibility for the 1955 Sample Eligibility for 1955 sample Eligible Not eligible

Characteristics in 1960

Total

Total White, married, 18-44 years old Total 18-39 years old 40-44 years old White, previously married (as of 1955), 23-44 years old Nonwhite, married, 18-39 years old

3,322

2,406

916

2,986 2,414 572

2,341 1,787 554

645 627 18

66 270

65 ..

1 270

When we compare 1955 and 1960 data for women who were eligible for the 1955 sample, we use the total sample interviewed in 1955 (2,713 wives) and 2,341 of the 2,406 wives who represent their survivors in 1960. We have left the 65 previously married women out of such tabulations largely because we do not have certain kinds of information for them that we have for the rest of the 1960 sample. For example, when we compare the total number of children expected by the 1955 sample and by their representatives in the 1960 sample, we find that we do not have the total number of births expected for women who are no longer married because it was inappropriate to ask them how many children they expected to have in the future. In order to avoid the many problems associated with changing the bases of comparisons for different variables, we simply excluded the women who were no longer married. Fortunately, they are few in number, and their replies would not affect the comparisons

BACKGROUND—RELIABILITY OF BIRTH EXPECTATIONS

significantly. If there had been a substantially larger number of such women, we would have used their replies and made some accommoda­ tion for them in the tabulations. Most of this book deals with the 2,414 couples with white wives 18-39 years old in 1960. They represent about 18 million couples with similar characteristics in the United States. The 270 nonwhite couples in the sample represent about two million similar couples in the entire country. The sample was chosen under the direction of Leslie Kish of the Sampling Section, Survey Research Center, University of Michigan, by the method known as area probability sampling. The method of sampling and estimates of the sampling errors are described in Appendix A. After the sampling experts had chosen a number of dwelling units that would yield approximately the desired number of eligible re­ spondents, the interviewers visited these homes. In as many cases as possible, the interviewer's first visit was preceded by a letter from the Director of the Survey Research Center explaining the nature of the study and urging the women to cooperate. In the homes containing eligible respondents, the interviewers were able to complete 88 per cent of the interviews they sought. Only 6 per cent of the eligible respondents refused to be interviewed. Most of the other respondents were never found at home, even after repeated visits by the interviewer. These results compare very favor­ ably with those of national surveys on other subjects. THE INTERVIEW

Each interviewer had an interview schedule from which she read the questions to the respondent and on which she recorded her answers. This schedule is available on request.4 The interviews lasted about an hour and a quarter, on the average. In most cases, the respondent welcomed the interviewer and did not hesitate to give the information she requested, even about topics that might be considered sensitive—such as methods of contraception. Only 7 of the 2,414 white wives 18-39 years old refused to tell the interviewer whether or not they had ever used contraception. An additional 15 wives said that they had used contraception, but refused to identify the methods used. The interviewers noted that the respondent's cooperation was "very good" in 80 per cent of the interviews, "good" in 14 per cent, and "fair" or "poor" in only 6 per cent. Subjective judgments like these, however, do not tell us how ac4Write

to Scripps Foundation, Miami University, Oxford, Ohio.

FERTILITY AND FAMILY PLANNING IN THE U.S.

curately the respondents answered the questions. Some of them undoubtly found it difficult to recall specific details, even though they tried to report honestly. However, there are a few checks we can use to evaluate their replies. For example, we have several reasons for believing that the propor­ tion who have ever used contraception has been ascertained accurately. One is that the proportion reporting that they had begun contra­ ception by 1955 is virtually the same for couples in the 1955 study (70 per cent) as it is for couples who represent their survivors in the 1960 study (71 per cent). Another reason for believing that the use of contraception is re­ ported accurately is that a large majority of the couples who do not use contraception, according to the wife's report, have credible reasons for nonuse. For example, among the white wives 18-39 years old, 81 per cent said they had used contraception by 1960 and 19 per cent said they had not. Among this 19 per cent, 7 per cent said they intended to use contraception at some time in the future (most of them were young couples just starting their families) and 10 per cent gave evidence of reproductive impairments that made the use of contraception less urgent than for other couples or com­ pletely unnecessary. Thus, only 2 per cent of all the couples were able to bear children and intended never to use contraception. Nearly half of these couples were already limiting their fertility by a method the wife did not consider contraceptive (douching regularly soon after intercourse for cleanliness only). The remaining few couples had strong attitudes against any form of fertility control, but most of them had not yet had to face the reality of excessively large families. The only systematic inquiry about the reliability of survey data on family planning has come from the Princeton Study.5 In this study, the investigators compared replies by the same women to the same questions in two interviews three years apart. In general, they found that although some individual women gave different answers in the two studies, the distributions of the replies were nearly the same for both interviews. For example, some women gave different replies about methods of contraception used, but the distributions of replies were about the same. In other words, there were many compensating errors. One fact of great importance to the investigator emerges from the interviews for the present study as well as from those for other surveys of family planning: the vast majority of married women are willing to give information about their family planning practices and other highly personal topics in direct interviews, and they try to report 5

WestofiF 1961a.

BACKGROUND—RELIABILITY OF BIRTH EXPECTATIONS

as accurately as they can. Such excellent cooperation surprised many of the people involved in social research when it first became known, but it is now a familiar feature of surveys in this area. Much of the credit for such excellent results must go to the interviewers, whose efforts to gain the respondents' interest and confidence have been essential to the success of this study. RELATED STUDIES

The Growth of American Families Studies of 1955 and 1960 are the first nationwide studies of factors affecting the control of fertility in the United States, but there have been studies of similar or related topics based on more narrowly defined samples and with somewhat different objectives. The first such study was the Indianapolis Study of 1941.® The detailed analysis for this survey was based on 1,977 native white Protestant couples living in Indianapolis, married in 1927-29 when the wife was under 30, who lived in a large city for most of their lives and who had completed the eighth grade of school. The study was designed primarily to test specific hypotheses about factors affect­ ing fertility. Some of the factors were socioeconomic status and security, personality characteristics (such as feelings of inadequacy), fear of pregnancy, tendency to plan, interest in religion, and husbandwife dominance. The purpose of this survey was not so much to describe variations in fertility for different population groups as it was to try to investigate some of the underlying social and psychologi­ cal determinants of behavior affecting the control of fertility. The reason for sampling such a small segment of the United States popula­ tion was to eliminate many causes of variation in fertility (such as differences between urban and rural residents, Protestants and Catholics, whites and nonwhites, etc.) that were not under intensive study. The Princeton Study, the field work for which began in 1957, is a direct descendant of the Indianapolis Study. Its findings relate to a sample of white couples who were living in seven of the largest metropolitan areas of the country and who had had a second child five to seven months before the first interview in 1957. One of the purposes of this study was to investigate the factors determining whether or not the couple would have a third child. The reason for this focus is the fact that much of the higher fertility of the postwar period has resulted from the desire for more than two children. Among the wide variety of information obtained from both husband and wife in the first interview was the number of additional * See Bibliography for references to this and other related studies.

FERTILITY AND FAMILY PLANNING IN THE U.S.

children they wanted. The same sample was interviewed again in 1960 to see which couples had had a third child and to explore the factors influencing their control of fertility. As in the Indianapolis Study, many of the variables examined are psychological in nature, but there is also a strong emphasis on socioeconomic factors—par­ ticularly the major religious group with which the couple identifies. Again, a relatively narrow segment of the childbearing population was sampled in order to eliminate the influence of variables that were not under study. Several of the Detroit Area Surveys, sponsored by the University of Michigan, have dealt with topics related to fertility. The 1954 Detroit Area Survey pioneered questions on the number of children expected. Similar questions were asked in the 1955 and 1958 surveys. The aims of these surveys were to study socioeconomic differentials in past and expected childbearing in the Detroit area, and to get some information on the reliability and stability of birth expectations. A much more elaborate survey was conducted in early 1962 of 1,215 women in Detroit who had recently married or given birth to their first, second, or fourth child. These women were questioned again late in 1962 about their fertility since the first interview and their future birth expectations. The major aim of this study is to examine social and economic factors affecting fertility. Social Research, Incorporated, has conducted two surveys, under the sponsorship of the Planned Parenthood Federation of America, which are designed to examine psychological and interpersonal factors affecting the use of contraception in the United States. These studies are largely exploratory, so the sampling standards have not been rigid. The interviews have been "open-ended," because the aim of the in­ vestigators was not to test hypotheses, but to seek promising leads about the nature of less conscious attitudes affecting the use of contraception. The University of Michigan Population Studies Center has begun adding questions on past and expected childbearing to nationwide surveys of the Survey Research Center that deal principally with other topics. The investigators hope to use the answers to these ques­ tions to develop a time series of birth expectations for the United States that will extend the series begun in the Growth of American Families Studies. In contrast to the more analytical studies mentioned above, the Growth of American Families Studies seek simply to describe the distribution of certain fertility variables for the United States as a whole, to show how they differ for certain important subgroups of the population, and to trace their change over time.

BACKGROUND—RELIABILITY OF BIRTH EXPECTATIONS

PLAN OF THE BOOK

Although we show summary statistics for whites and nonwhites combined (for wives 18-39 years old and their husbands), we do not present detailed data for both groups combined. Instead, most of the detailed analysis is for white couples with the wife 18-39 years old. One important reason for adopting this practice is that it greatly facilitates comparisons with the 1955 survey, which was limited to white couples. Secondly, white and nonwhite couples have different levels of fertility and different socioeconomic differentials in fertility. It was thought best to treat these two population groups separately in order to bring out these important dissimilarities. Finally, the small size of the nonwhite sample (270) does not permit us to subdivide it to the extent that was possible with the white sample and still produce reasonably reliable estimates for subgroups. Therefore, most of the book is about white couples with the wife 18-39 years old. Unless otherwise specified, tables show data for this group. Tables presenting data for couples with wives in different age groups, for nonwhite couples only, or for both white and nonwhite couples will specify the relevant population groups in the titles. Although there are references to the total population and the nonwhite population in nearly every chapter in the book, the main comparisons between white and nonwhite couples are presented in Chapter 9. Inasmuch as the central focus of this book is the number of children wives expect to have, the early chapters deal with various aspects of this variable. The reliability and stability of wives' expecta­ tions about their additional births are discussed in the latter portion of this chapter. Chapters 2 and 3 deal with attitudes toward family size and the childbearing expectations of women in the present study. The next four chapters are concerned chiefly with factors affecting couples' control of fertility. Chapter 4 presents our findings about physiological limitations on the number of births couples can have. Chapter 5 discusses the extent to which couples use contraception to control their fertility. Chapter 6 describes how couples typically plan their families. Chapter 7 describes the different methods of con­ traception used and presents some data on their effectiveness. The next subject to be treated is the timing of births—that is, how soon after marriage births of various orders occur, the number of months between births, and the proportion of all births occurring by a given age or duration of marriage. This important topic, which was not included in the book about the 1955 study, is discussed in Chapter 8. As noted previously, Chapter 9 presents data for the nonwhite

FERTILITY AND FAMILY PLANNING IN THE U.S.

population, and compares this group with all white couples and with white couples who have social and economic characteristics that are similar to those of nonwhites. Finally, Chapter 10 discusses possible future trends in fertility, and presents projections of the population of the United States to 1985. Although this book covers many topics that were dealt with in the book reporting the findings of the 1955 study, it differs from the latter book in many respects. One important difference has already been mentioned: the fact that this book presents the first findings about the control of fertility among a nationwide sample of nonwhite couples. Another addition is the inclusion of a chapter on the timing of births (Chapter 8). An important difference between the two studies arises from the fact that we are now able to make time comparisons for certain variables with the use of two comparable nationwide samples. We can now see, for example, how birth expectations have changed be­ tween 1955 and 1960; whether operations preventing pregnancy have become more or less common; whether more couples are using contra­ ception. In other words, with this book our knowledge of topics per­ taining to the control of fertility in the United States begins to be cumulative. Finally, on the basis of the experience gained with the 1955 study, we were able to introduce more detailed questions about certain sub­ jects. For example, in the 1960 study, we asked not only whether the wife or husband had had an operation that prevented pregnancy (as was asked in 1955), but also what kind of operation and why it was performed. With respect to contraception, we added questions about methods used in each interpregnancy interval. Such additions have made possible somewhat more detailed analysis for some vari­ ables than was possible with data for the 1955 study. THE RELIABILITY OF BIRTH EXPECTATIONS

The question we are asking in this section is: Do groups of women actually have the number of births they expect to have? The question relates to groups of women because in the final analysis, the demographer wants to know about the fertility of aggregates of indi­ viduals, rather than particular individuals. We know that individuals are only mediocre predictors of their own fertility. This was brought out by a study of 145 engaged couples, mostly college-educated, who were originally interviewed in the middle 1930s. At that time they were asked how many children they wanted. Twenty years later they were contacted again to find out how many children they had had. The coefficient of correlation between the

BACKGROUND—RELIABILITY OF BIRTH EXPECTATIONS

number of children the engaged woman said she wanted and the number she eventually had was only .27, which indicates very poor agreement for individuals. However, the average numbers of children wanted just before marriage (2.8 for the women and 2.6 for the men) are very close to the average number born (2.6).7 This suggests very good agreement between prediction and performance for groups of persons. Many of these couples had fewer births than they originally wanted, and many others had more. These deviations nearly canceled each other. Approaching the problem from a somewhat different angle, the authors of the Princeton Study show that even though individuals are not perfect predictors of fertility, the number of pregnancies they have over a short period of time is correlated with the number of children they want eventually. The coefficient of correlation between the number of children the wife wanted in 1957 and her fertility during the three years between the 1957 and 1960 interviews is .48.8 Very similar results were obtained from 1955 and 1958 interviews with women included in the Detroit Area Survey.9 In the Growth of American Families Studies, as noted previously, we interviewed two groups of women in 1955 and 1960, who had the same characteristics as of 1955. There are several things we want to know about their expected and actual childbearing: 1. Are the two groups comparable with respect to the number of children they had borne by 1955? In other words, did we really sample the same kinds of women, as far as one highly relevant characteristic is concerned? 2. Did the wives interviewed in 1960 have the number of births in 1955-60 expected by wives interviewed in 1955? 3. Did the wives interviewed in 1960 expect the same final number of births as the wives interviewed in 1955? We can ask these questions about the entire sample and about subgroups, such as Protestants, high-school graduates, etc., in an effort to see where discrepancies are concentrated and what caused them. Data on all three questions are presented together in Tables 5 to 10 because they are highly interrelated. However, in describing these data, we will deal with one question at a time. Births by 1955—The average number of children born by 1955 is virtually the same for the 1955 sample (2.06) as for representatives of their survivors in the 1960 sample (2.01). Furthermore, the per­ centage distributions of the two groups by the numbers of children 'Westoff 1957. 8 Westoff 1963, p. 68. •Goldberg 1959, p. 378.

FERTILITY AND FAMILY PLANNING IN THE U.S.

ever born by 1955 are not significantly different (Table 2).10 Thus, we can be certain that, in the aggregate, the 1955 and 1960 samples represent similar women, as far as fertility up to 1955 is concerned. Subgroups of the two samples also show excellent agreement on fertility up to 1955. Protestant wives, for example, had 2.1 births by 1955, according to the 1955 sample, and 2.0 according to the 1960 sample. This small difference is not significant. Catholic wives had 2.1 births by 1955 according to both samples (Table 5). It is only when we subdivide the sample into small groups that we find large differences. For example, Jewish and "other" wives in the 1916-20 cohorts11 had 1.7 births by 1955, according to the 1955 TABLE 2 Per Cent Distribution by Births by 1955 for Wives 18-39 Years Old in 1955, According to Sample for 1955 and 1960

Number of births by 1955

1955 sample

1960 sample

Number of wives Per cent Total

2,713

2,341

100

100

15 22 31 17 7 4 3 2.06

17 22 30 18 7 4 3 2.01

0 1 2 3 4 5 6 or more Average

Difference (1960 minus 1955)»

—

2 -1 -2 1 — — —

-.05

* Difierences are computed from unrounded percentages; hence, they are not always equal to differences between rounded percentages.

sample, and 2.2, according to the 1960 sample. However, the two groups are very small (30 women in 1955 and 46 in 1960), and even a difference of 0.5 between groups of this size is not statistically "Throughout this book, we use the .95 confidence level in statements about statistical significance. In other words, the chances are 19 in 20 that a "significant" difference between two groups of the sample represents a real difference in the population from which the sample was drawn. See Appendix A for estimates of sampling errors. 11 In this boot, the term "cohort" refers to a group of women born in a 12month period between July 1 and June 30. The cohort of 1916, for example, refers to the women who were born between July 1, 1915 and June 30, 1916. Women in the five-year group of cohorts, 1916-20, were born between July 1, 1915 and June 30, 1920. We often use cohort rather than age to designate a group of women because we want to make it clear that we are referring to the same women (or samples of the same women) on different dates. Thus, "1916-20 cohorts" refers to the same population group both in 1955 and 1960.

BACKGROUND—RELIABILITY OF BIRTH EXPECTATIONS

significant. It could easily have occurred because of random differences between the two samples. One subgroup that shows significant difference between the samples of 1955 and 1960 in fertility by 1955 consists of wives in the 1916-20 cohorts with one to three years of high school. The 1955 sample of this group had 2.4 births by 1955; their representatives in the 1960 sample had 3.1 births by 1955, or nearly one-third more (Table 6). The difference of 0.7 between these averages is statistically significant and indicates that the two subsamples do not represent similar women, as far as fertility by 1955 is concerned. Another such group consists of Protestant wives with a gradeschool education. The 1955 group reported an average of 3.0 births by 1955, and their 1960 representatives reported only 2.6 (Table 7). The difference is significant. However, with the level of statistical significance we have chosen (.05), it will occasionally happen that subgroups show a statistically significant difference that is not due to a real difference in the popula­ tion sampled. Therefore, it is not at all surprising that we have found two such differences. In fact, even with the best sampling methods, it would be surprising if we did not find a few differences like this. Births in 1955-60—Having satisfied ourselves that the 1955 and 1960 samples and most subgroups of these samples are comparable with respect to fertility by 1955, we can proceed to the next question: how well did wives in 1955 predict the number of births that would occur to women like themselves in the next five years—1955 to I960?12 According to Table 3, the agreement between expected and actual numbers of births in 1955-60 is very close: the average expected was between .69 and .71, and the average actually born was .75.13 The difference is not significant. However, Table 3 and Figure 1 show some statistically significant differences between the percentage distributions of expected and actual births. The proportion of women in 1955 who expected no births in 1955-60 was 57 per cent, but the proportion of the 1960 sample who had no births in 1955-60 was 53 per cent. This indicates that those who underestimated their fertility outnumbered those who over­ estimated it. n Our operational definition of the period 1955-60 is the 5 years following interview for the 1955 sample, and the 5 years preceding interview for the 1960 sample. Because the 1955 interviews were conducted in February and March, and the 1960 interviews in May, June, and July, the period 1955—60 does not relate to exactly the same 60 months for both samples. However, they both overlap approximately the same 57 months, which is close enough for our purposes. 18 See Appendix B, Note 1, for method of estimating the number of births ex­ pected in 1955-60.

FERTILITY AND FAMILY PLANNING IN THE U.S.

When we subdivide the sample into the Fecund and Subfecund, we find that the discrepancy is concentrated among couples who were Subfecund in 1955, but not Definitely Sterile.14 (Obviously, the Definitely Sterile in 1955 could not have any children in 1955-60. They are not included in the figures shown for the Subfecund in TABLE 3 Per CentDistribution by Number of Births in 1955-1960, Expected by Wivesin the 1955 Sample and Reported by Wives in the 1960 Sample, for Wives 18-39 Years Old in 1955, by Fecundity in 1955

Number of births in 1955-60

Expected in 1955 (1955 sample) Minimum Maximum

Actual (1960 sample)

Differencea Actual Actual minus min. minus max. expected expected

Total — 100 100 57 53 -4 21 26 3 — 16 15 5 — 5 — 1 1 .71 .75 .06 Subfecund, but not Definitely Sterile in 1955 — 100 100 75 66 -9 12 25 12 9 -1 8 3 2 -1

Total 0 1 2 3 4 or more Average

100 57 23 15 5 1 .69

Total 0 1 2 3 4 or more Average

100 75 13 8 3 —

—

—

—

.40

.42

.05

Total 0 1 2 3 4 or more Average

100 44 29 19 6 1 .90

100 44 28 21 6 1 .93

.45 Fecund in 1955 100 45 29 18 6 2 .92

— —

-1 -1 1 1 .02

—

-4 5 -1 — —

.04 —.

-9 13 -2 -1 —•

.03 —

1 1 -3 — —

-.01

• Differences are computed from unrounded percentages; hence, they are not always equal to differences between rounded percentages.

Table 3 in order to limit the comparisons to Subfecund couples who might be able to have children.) Apparently, many of the wives in the Subfecund group in 1955 thought that they would not be "Fecund couples are normal in their ability to have children in the future; Subfecund couples are below normal; Definitely Sterile couples cannot have any more children. These terms are capitalized to remind the reader that they are used in a special sense in this study. See Chapter 4 for definitions.

BACKGROUND—RELIABILITY OF BIRTH EXPECTATIONS

able to have any more children, but similar wives in the 1960 sample reported that they had had some children. However, the most important finding in this section is that the Fecund majority had, as a group, a nearly perfect record of fertility prediction. Fecund wives in 1955 expected an average of almost one birth each (.90 to .93), and the 1960 wives who were Fecund in 1955 reported that they had had almost precisely that number (.92). NUMBER OP BIRTHS IN 1955-60 NONE

ONE

TWO

FOUR OR MORE

I

10

1 30 PERCENT

PECTED (1955 SAMPLE )

1— 40

—I— 50

-I 60

ACTUAL (I960 SAMPLE)

FIGURE 1 Per cent distribution by actual and expected number of births in 1955-1960, for wives eligible for the 1955 study

Furthermore, the expected and actual distributions by numbers of births are nearly identical. Such excellent agreement between expectations and reality was a complete surprise to the investigators. The authors of the 1955 study had thought that the information on the expected timing of future births was not worth very much. In the course of the interview, each woman who expected more births was asked when she expected the next, the one after that, and so forth, until she had accounted for all of her additional births. The replies seemed stereotyped and of little value, so comparable questions were not asked in the 1960 survey. We wish now that they had been.

F E R T I L I T Y AND FAMILY PLANNING IN T H E U . S .

As noted previously, good agreement between expected and actual childbearing for large aggregates of individuals does not necessarily mean good agreement for individuals. Nor does it necessarily mean good agreement for particular subgroups. We may find that some subgroups underestimated their fertility in 1955-60 while others over­ estimated it. In making comparisons between subgroups, we have averaged the minimum and maximum numbers of births expected by the 1955 sample so that we can show only one figure for births expected in 1955-60. We have also rounded the average numbers of expected TABLE 4 Per Cent Distribution by Most Likely Total Number of Births Expected, for Wives 18-39 Years Old in 1955, According to Samples for 1955 and 1960 Most likely total number of births expected Number of wives Per cent Total 0 1 2 3 4 5 6 or more Average

1955 sample

1960 sample

2,713

2,341

100

100

5 9 28 26 20 6 7 3.02

5 8 25 26 18 9 9 3.19

Difference (1960 minus 1955)"

—

-1 -2 —

-3 3 2 .17

•Differences are computed from unrounded percentages; hence, they are not always equal to differences between rounded percentages.

and actual births to one place past the decimal. This means, for example, that the expected and actual average numbers of births in 1955-60 are both 0.7 for the entire samples. We do this because the subgroups are so small that rounding to the second place past the decimal would give a false sense of accuracy. The one feature of the comparisons that stands out above all others is the fact that wives with the least education are the poorest predictors of the fertility of women like themselves (Figure 2). Within this group, the youngest women are the least accurate. The wives in the cohorts of 1931-37 with a grade-school education expected 1.1 births in 1955-60, according to the 1955 sample, and actually had 1.9 births, according to the 1960 sample (Table 6). Although this subgroup is small in both samples, the difference of 0.8 is sta-

BACKGROUND—RELIABILITY OF BIRTH EXPECTATIONS

tistically significant. Older wives with a grade-school education appear to be about as accurate as better-educated wives in predicting fertility. The poor record of the young less-educated wives is apparently related to inexperience with childbearing, for we find the greatest underestimates of future fertility among women with no children or TABLE 5 Consistency of 1955 and 1960 Statements about Numbers of Births, by Wife's Religion, by Cohort

Cohort

Births by 1955 1955 1960 sample sample"

1916-37 1931-37 1926-30 1921-25 1916-20

2 .1 1 .1 1 .9 2 .3 2 .6

1916-37 1931-37 1926-30 1921-25 1916-20

2 .1 1 .1 2 .0 2 .3 2 .6

1916—37 1931-37 1926-30 1921-25 1916-20

2 .1 1 .1 1 .9 2 .3 2..6

1916-37 1931-37 1926-30 1921-25 1916-20

1..6 *

1 .3 2..2 1..7

Births in 1955-60 Expected Actual (1955 (1960 sample)" sample)

Total .7 1.5 .9 .4 .2 Wife Protestant 2 .0 .6 1 .1 1.4 1 .9 .8 2 .3 .3 2 .7 .1 Wife Catholic 2 .1 .9 1 .1 1.8 2 .0 1.2 2 .4 .7 2.8 .3 Wife Jewish or other 1 .8 .6 * 1 .0 1 .5 1.0 2..0 .3 2..2 .1 2 .0 1 .1 1 .9 2 .3 2 .7

Most likely expected total no. of births 1955 1960»

.7 1 .5 .9 .5 .2

3..0 3.,2 3,.1 3..0 2 .9

32 3.4 3.3 3.0 3.0

.7 1 .4 .8 .4 .2

2 .9 2 .9 3..0 2 .9 2..8

3.1 3.2 3.1 2.9 3.0

.9 1 .8 1 .1 .7 .3

3..4 3. 8 3 .5 3.,3 3. 1

3.7 4.2 3.9 3.5 3.2

.6 1 .3 .8 .4 .1

2 .4

2.6 3.0 2.7 2.5 2.4

*

2,.5 2. 6 2. 0

a The 1960 data are for wives in the 1960 sample who were eligible for the 1955 sample.

only one. In 1955, the grade-school-educated wives with no children expected an average of 0.5 children in 1955-60, but comparable women in 1960 reported that they had had 1.1 births (Table 8 ) . The same averages were found for wives in the grade-school group with only one child. Apparently the less-educated wives who are just starting their families think that they will have better control of their fertility than

FERTILITY AND FAMILY PLANNING IN THE U.S.

they actually do. This is consistent with findings reported later in the book—for example, that less-educated couples wait the longest before beginning to use contraception, and that a higher proportion of them have families that they consider too large. We shall see in Chapter 3 that they do not want substantially more children than other couples, but they seem to be unable to control their fertility to the extent desired. WIFE'S EDUCATION TOTAL

COLLEGE

HIGH SCHOOL 1 I

HIGH SCHOOL,1-3

GRADE SCHOOL

Τ­ ο 0.5 I NUMBER OF BIRTHS IN 1955-60 \////\ EXPECTED II955SAMPLE) KSgXXXXS

ACTUAL

11960 SAMPLE)

FIGURE 2 Actual and expected average number of births in 1955-1960, for wives eligible for the 1955 study, by wife's education

The tendency for less-educated wives to have more births than they expect exists among both Protestants and Catholics, according to Table 7. The difference between expected and actual births is especially wide among Protestants with a grade-school education, but this may be related to the fact, noted in the previous section, that the 1955 sample of this group had higher fertility by 1955 than the 1960 sample. There appears to be at least a slight tendency to underestimate future fertility among wives who have bome four or more children.

BACKGROUND—RELIABILITY OF BIRTH EXPECTATIONS

For example, the college-educated wives with four or more births by 1955 said that they expected an average of 0.4 children in 1955-60, but similar women reported that they had had 0.7 children in this period (Table 8). The difference of 0.3 is not statistically significant because these subgroups are so small. Nevertheless, we are inclined TABLE 6 Consistency of 1955 and 1960 Statements about Numbers of Births, by Wife's Education, by Cohort

Births by 1955

Cohort

1955 sample

1960 sample"

1916-37 1931-37 1926-30 1921-25 1916-20

1. 8 8 1.,6 2..0 2,,4

1..7 .7 1. 5 2..0 2..2

1916-37 1931-37 1926-30 1921-25 1916-20

1..8 ,9 1..7 2..2 2 .3

1. 8 9 1..8 2,.1 2 .3

1916-37 1931-37 1926-30 1921-25 1916-20

2 .1 1 .4 2 .3 2 .3 2 .4

2 .3 1 .6 2 .1 2 .5 3 .1

1916-37 1931-37 1926-30 1921-25 1916-20

2 .9 1 .6 2 .5 3 .1 3 .4

2 .7 1 .5 2 .6 2 .8 3 .5

Births in 1955-60 Expected Actual

(1955 sample)

(1960 sample)'

Collegeb .8 1.7 1.2 .5 .2 High school, 4 .8 1.7 1.0 .5 .2 High school, 1-3 .6 1.2 .8 .3 .1 Grade school .4 1.1 .7 .3 .2

Most likely expected total no. of births

1955

1960®

.8 1 .6 .9 .6 .2

2.9 3.5 3.0 2.6 2.7

2.9 3.2 2.9 2.9 2.5

.8 1 .4 .9 .5 .2

2.9 3.1 2.9 3.0 2.6

3.0 3.2 3.3 2.8 2.6

.7 1 .3 .8 .4 .2

3.0 3.0 3.3 2.9 2.7

3.4 3.6 3.4 3.2 3.4

.7 1 .9 .8 .5 .3

3.6 3.3 3.5 3.6 3.7

3.9 4.3 3.7 3.6 4.0

"The 1960 data are for wives in the 1960 sample who were eligible for the 1955 sample. b Data for all wives by cohort are shown in Table 5.

to believe that there is a tendency in the direction stated because it occurs in all of the socioeconomic groups for which we have made the relevant comparisons. Since many of the couples with four or more births already had more children than they wanted in 1955, their greater-than-expected fertility in 1955-60 suggests that past fail­ ures to control family size often mean future failures. The tendency to have more births than expected in 1955-60 is

FERTILITY AND FAMILY PLANNING IN THE U.S.

stronger among couples who had not used contraception by 1955. The wives in this group expected an average of 0.8 births, but had 1.2 (Table 9); the difference is statistically significant. This again indicates that those who are least careful in controlling fertility are most likely to underestimate the number of births they will have. Such findings suggest that couples who are unwilling or unable to control their fertility either indulge in wishful thinking about the TABLE 7 Consistency of 1955 and 1960 Statements about Numbers of Births, by Wife's Religion, by Wife's Education Births in 1955-ω

Expected Actual 1960 (1955 (1960 sample" sample) sample)"

Births by 1955

Wife's education as of 1955

1955 sample

Total College High school, 4 High school, 1-3 Grade school

2..1 1 .8 1 .8 2. 1 3..0

Total College High school, 4 High school, 1-3 Grade school

2 .1 1 .9 1 .9 2 .1 2 .8

Total College High school, 4 High school, 1—3 \ Grade school J

1 .6 1..4 1. 6

Protestantb 2.0 .6 1.6 .7 1.8 .7 2.3 .6 2.6 .3 Catholic 2.1 .9 1.9 1.3 1.8 1.1 2.3 .7 3.0 .6 Other 1.8 .6 1.6 .8 1.7 .6

2. 2

2.2

.5

Most likely expected total no. of births

1955

I960»

.7 .8 .7 .7 .7

2..9 2..7 2..7 2..9 3..6

3.1 2.7 2.8 3.3 3.7

.9 1.0 1.0 .8 .9

3..4 3 .9 3 .3 3 .1 3 .7

3.7 3.7 3.5 3.5 4.3

.6 .6 .5

2 .4 2..5 2. 3

2.6 2.5 2.5

.6

2..7

3.1

•The 1960 data are for wives in the 1960 sample who were eligible for the 1955 sample. b Data for all wives by education are shown in Table 6.

number of children they will have, or do not realize how often they will conceive if they do not use contraception carefully and regularly. Couples who had begun using contraception by 1955 were ex­ cellent predictors of fertility in 1955-60. Their expected and actual childbearing were not significantly different (Table 9 ) . So far, we have described the kinds of wives who had more births than expected in 1955-60. Now we shall turn our attention to groups who had fewer than expected. First, we must note that errors of overestimating fertility are somewhat smaller and less con­ centrated in particular groups than the opposite errors. As a result,

BACKGROUND—RELIABILITY OF BIRTH EXPECTATIONS

most of the differences described below are not statistically significant. We regard the data as more suggestive than definitive. There is one small group that we would expect to overestimate fertility. This is composed of wives who do not deliberately try to prevent pregnancy, but who have adopted a practice that reduces somewhat the probability of conception: douching regularly for clean­ liness after intercourse. Wives in the 1960 sample who began this practice before 1955 reported that they had had fewer births than similar women had expected in 1955 (Table 9). Their failure to achieve expected fertility was probably due to the contraceptive effect of douching. Small deficits of actual below expected numbers of births also occur among the better-educated wives who are young (under 30 in 1955) or who had no births by 1955. College wives with no births in 1955, for example, expected 1.4 births in 1955-60 but had only 1.1 (Table 8). The difference is not significant, but the fact that young high-school graduates also show small deficits of actual below expected births suggests that there may be some relationship between overestimates of fertility and a good education among wives who are just beginning their families. It seems likely that this relationship is due largely to the fact that better-educated couples have better control over their fertility than other couples. Consequently, when fecundity impairments pre­ vent some of the better-educated from having the number of births they want, there are fewer couples who will compensate for this "lost" fertility by having an excessive number of births. Therefore, "lost" fertility tends to be greater than "excess" fertility, and the net result is fewer births than expected.15 We suspected that one particular group of wives interviewed in 1955 would not have as many births as they thought they would. It consists of young Catholics. They expected substantially more births than older Catholic wives and, in our opinion, had exaggerated their future fertility. In the report of the 1955 study, we said that it seemed likely "that the expectations of the younger Catholics will be reduced as they have more experience with bearing and raising children."16 Now we are able to check their expected and actual M TO support this explanation fully, we would have to know the number of births expected in 1955 by wives who later became Subfecund. We do not have this information, however, because we do not know which of the wives interviewed in 1955 became Subfecund by 1960. This is one of the disadvantages of using two samples of different women in 1955 and 1960 instead of interviewing the same women twice. However, the procedure we used has advantages that seem to outweigh the loss of this kind of information. " FPSPG , p. 280.

FERTILITY AND FAMILY PLANNING IN THE U.S. TABLE 8 Consistency of 1955 and 1960 Statements about Numbers of Births, by Wife's Education, by Number of Births in 1955 Number of births by 1955

Births in 1955-60 Expected Actual (1955 sample) (1960 sample)"

Most likely expected total no. of births 1955 I960·

Total Total 0 1 2 3 4 or more

.7 1 .2 1 .0 .6 .4 .4

Total 0 1 2 3 4 or more

.8 1 .4 1 .1 .6 .4 .4

Total 0 1 2 3 4 or more

.8 1 .4 1 .1 .6 .4 .4

Total 0 1 2 3 4 or more

.6 .9 .9 .5 .4 .4

Total 0 1 2 3 4 or more

.4 .5 .5 .4 .4 .4

.7 1.1 1.0 .5 .5 .6 College .8 1.1 1.2 .5 .5 .7 High school, 4 .8 1.2 .9 .5 .4 .7 High school, 1-3 .7 .9 1.1 .5 .5 .6 Grade school .7 1.1 1.1 .6 .6 .6

3.0 1.8 2.3 2.8 3.5 5.5

3.2 1.9 2.6 2.8 3.7 5.9

2.9 2.0 2.4 2.8 3.5 5.0

2.9 1.7 2.7 2.7 3.7 5.7

2.9 1.9 2.4 2.8 3.5 5.3

3.0 2.0 2.5 2.8 3.7 5.6

3.0 1.5 2.2 2.7 3.5 5.3

3.4 1.7 2.5 2.9 3.8 5.9

3.6 1.1 1.8 2.7 3.6 6.0

3.9 1.6 2.6 3.0 3.9 6.1

a The 1960 data are for wives in the 1960 sample who were eligible for the 1955 sample.

childbearing over a five-year period. When we do this, we find that these wives had just as many births as they thought they would. The 1955 group expected an average of 1.8 in 1955-60, and the 1960 group reported 1.8 (Table 5). In other words, it now appears that our original opinion was wrong and that they did not exaggerate their future fertility—at least, not over this five-year period. Further-

BACKGROUND—RELIABILITY OF BIRTH EXPECTATIONS

more, as we shall see in the next section, their total number of births expected went up instead of down. In general, religious preference does not seem to be related to the ability to predict fertility. Wives in each of the religious groups shown in Table 5 had approximately the number of births they ex­ pected to have in 1955-60. This simply means that the counter­ balancing tendencies for some to overstate and others to understate TABLE 9 Consistency of 1955 and 1960 Statements about Numbers of Births, by Use of Contraception, by Cohort

Cohort

Births by 1955 1955 1960 sample sample"

1916-37 1931-37 1926-30 1921-25 1916-20

2 ,2 1..2 2 .0 2..4 2. 7

2 .2 1 .3 2 .0 2 .5 2. 8

1916-37 1931-37 1926-30 1921-25 1916-20

1 .7 1 .0 1 .5 2 .1 2 .2

1 .7 1 .2 1 .5 1 .9 2 .0

1916-37 1931-37 1926-30 1921-25 1916-20

1 .8 1 .0 1 .8 2 .0 2. 3

1 .5 .7 1 .6 1 .9 2 .7

Births in 1955-60 Expected Actual (1955 (1960 sample) sample)" User by 1955b .7 1.4 .9 .4 .2 ODFC by 1955° .7 1.7 .9 .4 .2 Nonuser by 1955 .8 1.7 1.1 .5 .2

Most likely expected total no. of births 1955 I960"

.7 1.3 .8 .5 .2

3..0 3. 0 3. 1 3..0 3. 0

3. 2 3. 3 3. 3 3. 1 3. 1

.5 1.0 .7 .3 .1

2 .8 3,.2 2 .7 2 .7 2 .5

2. 6 3 .1 2 .9 2 .5 2 .3

1.2 2.0 1.3 .7 .4

3 .1 3 .7 3 .2 2..9 2. 7

3 .4 3 .7 3 .6 3..0 3. 2

" The 1960 data are for wives in the 1960 sample who were eligible for the 1955 b Data for all wives by cohort are shown in Table 5. sample. 0 Wives who used douche for cleanliness only and no other method of contraception.

their future fertility operate with the same results within each religious group. This can be seen to a limited extent in Table 7, which shows data for wives in the three major religious groups classified by their educational attainment. As noted previously, both Protestant and Catholic wives with a grade-school education had more births than they expected. There may be a slight tendency for Catholic wives with a college education to have fewer births than expected; they expected an average of 1.3 births in 1955-60, but had only 1.0. How­ ever, the difference is not statistically significant.

FERTILITY AND FAMILY PLANNING IN THE U.S.

Comparisons of 1955 and 1960 data for some of the independent variables we have not yet considered appear in Table 10. In general, they show excellent agreement between expected and actual numbers of births in 1955-60. Comparisons by duration of marriage suggest that the recently married wives are as able to predict fertility over TABLE 10 Consistency of 1955 and 1960 Statements about Numbers of Births, by Fecundity, Duration of Marriage, Husband's Occupation, Couple's Farm Experience, and Region of Residence Births in 7 955-60 Expected Actual 1955 1960 (1955 (1960 sample sample" sample) sample)® Births by 1955

Characteristic as of 1955 Total Fecundity: Definitely Sterile Other Subfecund Fecund Years married: Under 5 5-9 10-14 15 or more Husband's occupation: Upper white collar Lower white collar Upper blue collar Lower blue collar Farm Other Couple's farm experience: None Some, not on farm now On farm now Region of residence: Northeast North Central South West

2 .1

2..0

2. 4 1. 6 2 .2

2. 5 1. 7 2..1

.9 2 .1 2 6 2 .9

.7

.7

Most likely expected total no. of births

1955

I960»

3 .0

3.2

.4 .9

.5 .9

2. 4 2 .4 3 .4

2.5 2.4 3.5

.8 2 .1 2 .5 3 .0

1.5 .7 .3 .1

1.5 .7 .3 .2

3 .0 3 .0 3 .1 3 .1

3.2 3.2 3.0 3.3

1 .9 1 .9 2 .1 2 .2 2..7 1 .4

1 .8 1 .9 2 .1 2 .2 2 .4 1 .4

.7 .8 .6 .7 .7 1.2

.7 .7 .7 .7 .8 1.1

2 .8 2 .8 2 .9 3 .1 3 .7 3 .1

2.9 3.0 3.2 3.4 3.6 3.2

1. 8 2. 1 2. 8

1 .9 2..0 2..4

.8 .6 .6

.8 .7 .7

2 .9 3 .0 3 .7

3.1 3.2 3.5

1.9 2. 1 2..2 2. 0

2. 0 2. 1 1. 8 2. 2

.8 .7 .6 .7

.7 .8 .7 .7

3 .0 3. 1 3..0 3..0

3.1 3.4 3.0 3.3

—

—

a The 1960 data are for wives in the 1960 sample who were eligible for the 1955 sample.

a five-year span as those married longer. Also, there seems to be no substantial difference in this ability between wives classified by husband's occupation, couple's previous farm experience, or region of residence. Of all the socioeconomic factors considered in this sec­ tion, educational attainment is the only one closely related to the ability of groups of wives to predict their fertility.

BACKGROUND—RELIABILITY OF BIRTH EXPECTATIONS

In summary, we find that groups of married women can make good predictions of their fertility over a five-year period. There may be a general tendency to have more births than expected, but it is only slight. We do not know, of course, how far we can generalize these findings. It is possible that 1955-60 was unusual in some respects that brought about good agreement between expected and actual childbearing. Had a severe depression begun in mid-1955 and lasted until 1960, it is likely that actual births in the five-year period would have been below, instead of equal to, the 1955 expectations. Only later studies can tell us how the ability to predict fertility varies under different conditions. Meanwhile, we have good evidence that it is possible for groups of women to make accurate predictions of the number of births that women like themselves will have. Total Number of Births Expected—In this section we shall ask how the total number of births expected changes as wives grow older. An answer for the period 1955-60 appears in Table 4, which shows an increase over the five years. The average total number of births expected increased from 3.02 to 3.19, a statistically significant change of 6 per cent. The distributions of births expected for the two surveys show that the proportion expecting one, two, or four births declined, and the proportion expecting five or six increased.17 The increase in the total number of births expected was greatest in the younger cohorts (1926-30 and 1931-37). This has an im­ portant bearing on the use of birth expectations in preparing popula­ tion forecasts, which is discussed in Chapter 10. The increase was also greatest among the less-educated wives (Figure 3). In the youngest cohort group (1931-37), the average number of births expected by wives with a grade-school education rose from 3.3 in 1955 to 4.3 in 1960; for wives with one to three years of high school, the rise was from 3.0 to 3.6 (Table 6). Both changes are statistically significant. We also find that the greatest increases are for the least-educated wives with no births or only one in 1955 (Table 8). There are no similar changes for college-educated wives. Some of the cohorts within this group show declines in the total number of births expected and one cohort group shows a slight increase, but on balance the college wives expected the same number of births in 1960 as they did in 1955 (2.9). High-school graduates show a slight change upward, but the difference is not statistically significant. So, again we find that educational attainment has a major in" See Appendix B, Note 2 for method of determining the total number of births expected.

FERTILITY AND FAMILY PLANNING IN THE U.S.

fluence on the fertility predictions of younger women. Apparently, five years of experience with rapidly growing families lead less-edu­ cated wives to make higher, and probably more realistic, estimates of their total fertility. This finding has an important implication for the use of expecta­ tion data in estimating trends in fertility differences between socioeco­ nomic groups. In the report of the 1955 study we noted that among WIFE'S EDUCATION

TOTAL

COLLEGE

HIGH SCHOOL,4

HIGH SCH00L,l-3

GRADE SCHOOL

TOTAL NUMBER OF BIRTHS EXPECTED

X////A !955

1%¾¾%¾ I960

FIGURE 3 Average most likely expected total number of births, for wives eligible for the 1955 study, by wife's education, 1955 and 1960

wives with only a grade-school education, the youngest expected fewer children than the oldest. Among college-educated wives, however, the opposite was true. (This can be seen in Table 6.) These differences suggested the possibility of a downward trend in the fertility of the less-educated and an upward trend in that of the better-educated.18 Such tendencies eventually would have resulted in more children for the better-educated, which would have been an important reversal of a long-observed pattern. At the time these data were originally 18

FPSPG3 pp.

291-295.

BACKGROUND—RELIABILITY OF BIRTH EXPECTATIONS

reported, we noted that one of the most important areas of doubt was "whether the young grade-school wives will be able to avoid exceeding their relatively low expectations."19 Now we know that they were unable to keep from having more births than they expected, and it no longer appears likely that the customary tendency for lesseducated couples to have larger families will cease in the near future. Both Protestant and Catholic wives have raised their expectations since 1955. The tendency is especially marked for younger women. Catholics may have raised their expectations somewhat more than Protestants, but the difference is not great (Table 5). The feature of these comparisons that surprised us most, as noted earlier, is that the youngest Catholic wives (1931-37 cohorts) not only maintained their high expectations, but increased them from 3.8 to 4.2 births per couple. The relative change was about the same as for young Protestant wives. The rise in expectations among both Protestants and Catholics is concentrated among the less-educated. This tendency is especially marked among Catholics: those who had gone no further than gradeschool expected an average of 3.7 births in 1955 and 4.3 in 1960 (Table 7). This difference is statistically significant. The better-educated wives in each of the major religious groups did not change their expectations by a significant amount. There appears to be a slight reduction for the better-educated Catholics, but this could easily be due to sampling variability. Our experience with differentials in the childbearing expectations of religious and educational groups illustrates an important point about the use of such data in fertility research: that we can learn how best to use them only by comparing the results of consecutive surveys. One survey is not enough. In the case of educational differ­ ences, we were originally inclined to accept (with reservations) the low expectations of the less-educated younger wives in the 1955 sample. In the case of the religious differentials, however, we tended to reject the high expectations of the younger Catholic wives. In both instances, the 1960 data show that our first interpretations were wrong. Increases in the total number of births expected have been some­ what greater for couples who already had three or more children by 1955 than for couples with smaller families as shown in Table 8. (An important exception is the large increase for the less-educated wives with small families.) We found a similar pattern in the previous subsection, where expected and actual births in 1955-60 were com­ pared. In both cases, we think that the high past and expected fertility u

FPSPG, p. 293.

FERTILITY AND FAMILY PLANNING IN THE U.S.

of couples with several children results from the difficulty they en­ counter in preventing conception. Expectations have increased both for couples who have used con­ traception and for those who have not (Table 9). The only wives who showed reductions in the average number of births expected were those who regularly douched for cleanliness after intercourse. This may have resulted from their lower-than-expected fertility in 1955-60, which was noted in the previous section. Table 10 shows that birth expectations have increased for nearly all the independent variables we have not yet considered in this sec­ tion. It is true that many of the changes are not statistically significant, but the fact that they are generally in the same direction suggests that the upward tendency is real. The only major exceptions are farm occupation and residence. Wives living on farms in 1955, for example, reduced their expectations from 3.7 to 3.5. This difference, however, is not statistically significant. Furthermore, the 1955 and 1960 samples of farm wives are not quite comparable. By 1955, the earlier sample of these women had borne more children (2.8) than the later sample (2.4). This difference is almost large enough to be statistically significant. If it is true that the 1960 sample repre­ sents farm women with lower fertility than does the 1955 sample, the slight reduction in expectations between 1955 and 1960 is understandable. A major question we should try to answer here is whether the increase in birth expectations is primarily voluntary or involuntary. An increase could result from one or both of two tendencies: (1) the inability of some couples to control their fertility to the extent desired, and their subsequent realization that they will have more births than they originally expected, or (2) the desire of some couples for more children than they originally thought they would like to have. Both factors probably influenced the trend in birth expectations between 1955 and 1960, but the former cause appears to have been the more important. This opinion is based on the fact that birth expectations increased most among those couples who are least able to control their fertility—the least educated and those with the largest families. (Evidence that these groups have relatively poor control over their fertility is presented in Chapters 5 and 6.) However, the voluntary component of the rise in birth expectations should not be ignored. If excessive fertility had been the only cause of increase in the total number of births expected, we would probably have found the same kind of balancing of excess births and "lost" births (due to reproductive impairments) for the total number ex­ pected that we found for the number expected in 1955-60. Instead,

BACKGROUND—RELIABILITY OF BIRTH EXPECTATIONS

the total number of births expected has risen by 6 per cent for all groups combined. It has also risen for each of the major religious groups and for all users of contraception, even though these groups showed a balancing of actual and expected births in 1955-60. These facts suggest that there are some couples who had approxi­ mately the number of births they expected to have in 1955-60, but who plan to have more births altogether than the wives thought they would have when interviewed in 1955. In summary, between 1955 and 1960 there was a tendency for wives to increase their expected total number of births. It appears to be due primarily to the excessive fertility of some couples in this five-year period, but may also result from the desire of other couples for more children. Again, we do not know how far we can generalize the tendency for total birth expectations to increase over time. It may be peculiar to the 1955-60 period. Certainly, nothing comparable has been found in related studies. When the respondents for the 1955 Detroit Area Survey were reinterviewed in 1958, the investigators found that some had revised their expectations downward and some upward; the net result was virtually the same average number of births expected in both interviews—approximately 2.9.20 However, the Detroit Study also shows that changes in the number of births expected are related to the wife's age and the length of time she has been married. The younger and more recently married wives were most likely to revise their expectations, and the revisions tended to be upward.21 This is generally consistent with data from the present study. The Princeton Study shows a slight decline in the average number of children wanted between 1957 and 1960 interviews with the same women, from about 3.3 to 3.2.22 Although the Princeton Study uses children "desired," rather than "expected," the concepts are fairly close to one another as far as operational definitions are concerned. This is because the question asked about family size in the Princeton Study does not permit the wife to say that she wants fewer children than she already has. The wording used in the 1957 interview was, "How many children do you want to have altogether, counting the two you have now?"23 The main family-size question used in our 1960 study is, "How many children do you expect to have in all (counting those you now have)?" The only difference between the 20 Goldberg

1959, p. 376. 1959, pp. 378-380. 22 Westoff 1963, p. 71. 23Westoff 1961, p. 136.

21 Goldberg

FERTILITY AND FAMILY PLANNING IN THE U.S.

two questions is whether the additional children are "desired" or "expected." It is possible, of course, that a woman could expect more children in the future than she wants, because she thinks she may be unable to control fertility to the extent desired. This could explain why expectations increased slightly in the present study, but "desired family size" remained stable or declined in the Princeton Study. However, it seems doubtful that the difference in the wording of the two questions would have much effect on the averages. So, the evidence from related studies suggests that representative samples of women do not invariably raise their expectations over a short period of time. We must remember, however, that the data bearing on this topic are meager and not entirely comparable. Only more research can supply a definite answer. Although it is important to measure changes in birth expectations as accurately as possible, given the limitations of the sample, and to seek the reasons for them as carefully as we can, it must be empha­ sized that the differences we are concerned about are much smaller than the margin of doubt that would have surrounded the completed fertility of cohorts if we did not have data on birth expectations. For example, we note that most likely expectations for the 1931-37 cohort increased from 3.2 births per couple in 1955 to 3.4 in 1960, and we regard this as an important change. However, we would have a much less adequate notion of the eventual fertility of this group if we had no birth expectations to guide us. The book reporting the 1955 study shows, for example, that the use of quite reasonable methods of extrapolation could yield as few as 2.7 births per couple for the 1931-35 cohorts, or as many as 4.0.24 The 1955 expectations for this group ranged from 2.8 to 3.5, with 3.2 as the most likely. In 1960, they ranged from 3.2 to 3.7, with 3.4 as the most likely. These figures suggest that although the use of birth expectations have not eliminated uncertainty about the eventual fertility of cohorts still in the early portion of the childbearing period, they have reduced it substantially. This is one of the most important results of the com­ parisons between the 1955 and 1960 studies. SUMMARY

The broad aims of the 1955 and 1960 Growth of American Fami­ lies Studies are to see whether data on birth expectations can be used to help prepare forecasts of fertility rates and to describe many of the variables affecting family size in the United States. Some of these variables are the number of children couples want to have, the ability of couples to have children, whether or not couples use 24

FPSPGJ pp. 329-330.

BACKGROUND—RELIABILITY OF BIRTH EXPECTATIONS

contraception, and how they plan family growth. This book reports detailed findings for white married couples of childbearing age and broad findings for nonwhite couples. The interviews for both the 1955 and 1960 studies were highly successful. We found that wives were willing to give information about subjects that might be considered "sensitive" (such as their use of contraception). Consistency checks on the reported data suggest that they answered as accurately as they could. As a group, the wives in the 1955 study expected about the same number of births in 1955-60 as similar wives in the 1960 study reported that they had had. Some groups had fewer births than ex­ pected (probably because of unforeseen reproductive impairments) and some had more, but these deviations tended to cancel each other. Educational attainment is closely related to the ability to predict fertility. The younger wives with little education were least successful in forecasting fertility over a five-year period. They had substantially more births than they thought they would. The total number of births expected increased between 1955 and 1960 for the group of wives who were married and 18-39 years old at the time of the 1955 interview. The increases were greatest among the younger wives and among those with less education. These facts suggest that most of the rise results from the poor control some couples exercise over their fertility. However, it also appears likely that some couples wanted more children by 1960 than they had wanted by 1955. Similar data from related studies fail to show increases in childbearing expectations over short periods of time. Hence, we cannot be sure that there is a general tendency for expecta­ tions to increase with age.

CHAPTER 2

Ideal, Desired, and Expected Family Size Attitudes and desires are of prime importance in determining aver­ age family size in a country like the United States where family planning is such a widely accepted part of married life. For this reason, we questioned the wives not only about the number of children they expected, but also about their views on ideal family size and the number of children they actually wanted. The ideals of many wives call for more children than they feel it is prudent for them to have. This is not surprising, however. Similar discrepancies occur between ideals and realities in many areas of life. In deciding how many children they should try to have, most American wives are probably in­ fluenced by such conditions as their income, their health, their ability to care for more children, the size of their home, and their marital happiness. Of course, all wives do not expect to have the same number of children as they want even in their present circumstances because of a realistic consideration of other factors. Some expect fewer because of fecundity impairments, some expect more because of unwillingness or inability to prevent unwanted conceptions, and others expect fewer or more than they themselves want because of the family-size prefer­ ences of their husbands. On balance, the wives as a group think they will have a somewhat smaller number of children than they would prefer. In this chapter, we shall present our data on the number of children the wives considered ideal for the average American family, the number they said they themselves would have if they could live their lives over again under ideal conditions, the number they wanted under their actual and anticipated circumstances, and the number they expected. We shall then compare their expectations with their preferences. Finally, we shall examine the differences in family-size ideals, desires, and expectations that are associated with wife's age and other demographic variables such as wife's age at marriage, dura­ tion of marriage, and number of births by the time of the interview. Chapter 3 will deal with the religious, social, and economic variables. IDEAL FAMILY SIZE

All wives in our sample were asked: Ql. What do you think is the ideal number of children for the average American family?

IDEAL, DESIRED, AND EXPECTED FAMILY SIZE

Numerical replies were obtained from 2,642 respondents, or more than 98 per cent of the 2,684 wives who were 18—39 years of age.1 Since many stated a range, such as "three or four," it is convenient to present the results in terms of the minimum and maximum numbers considered ideal (Table 11). An overwhelming majority of these wives (about 9 out of 10) said they thought two, three, or four children are ideal for the average American family, with four clearly being the most popular number. Fewer than 0.4 per cent of the respondents stated an ideal of less than two children. At the other extreme, about 10 per cent said that the average American family should ideally include five or more children. The minimum and maximum replies average 3.4 and 3.5 children, respectively (Table 11). Averages are somewhat higher for nonwhites TABLE 11 Number of Children Considered Ideal for the Average American Family and Per Cent Distribution by Number Considered Ideal, for Wives, by Color"

Type of reply

Average number Numof ber children Per cent distribution by number considered ideal of considered 6 or wives ideal Total None 1 2 3 4 5 more

Minimum Maximum

2,642 2,642

3.4 3.5

100 100

Minimum Maximum

2,377 2,377

3.4 3.5

100 100

Minimum Maximum

265 265

3.6 3.8

100 100

Total — 22 — 18 White — — 22 — — 18 Nonwhite — 2 25 — 1 22 — —

31 28

39 43

4 5

5 6

32 29

39 43

3 4

4 5

18 19

38 39

7 7

11 13

" In this and subsequent tables in Chapter 2, all data are for wives who were 18-39 years old when interviewed unless otherwise specified.

than for whites, because of the larger proportion of nonwhite wives expressing an ideal in excess of four children (usually five or six). Nevertheless, the strong concensus on an ideal of two to four children was shared by both white and nonwhite wives. Results of the 1960 study are similar to those of the 1955 study of white wives, but two earlier surveys by the American Institute 1 Unless indicated otherwise, all data in this chapter are for the white and nonwhite wives who were 18—39 years old when interviewed in 1960. The white wives who were 40-44 will not be considered in this chapter. They were inter­ viewed in 1960 primarily to obtain data for the comparisons of the 1955 and 1960 results that are made in Chapter 1.

FERTILITY AND FAMILY PLANNING IN THE U.S.

of Public Opinion2 suggest that, while attitudes toward ideal family size may have remained rather stable in more recent years, they have changed significantly since the early 1940's. In 1941 and 1945, that organization asked a national sample of single and married women, 21 to 34 years old: "What do you consider is the ideal size of the family—a husband and wife and how many children?" The distribu­ tion of replies in Table 12 shows that, in both 1941 and 1945, approxi­ mately nine-tenths of the respondents reported an ideal of two, three, or four children—about the same proportions as in our 1955 and 1960 GAF studies. Within this range, however, the most popular TABLE 12 Per Cent Distribution by Number of Children Considered Ideal, 1941, 1945, 1955, and 1960

Number of children considered ideal All replies None 1 2 3 4

5 6 or more Average number

American Institute of Public Opinion® 1941

1945

100 — 1 40 32 21 3 3 3.0

100 — 1 25 33 31 7 3 3.3

Growth of American Familiesb

1960 1955 White wives White wives All wives 100

100

100

19 32 41 4 4 3.4

20 31 41 4 4 3.5

20 30 41 4 5 3.5

a Data

are for all single and married women, 21-34 years of age. are a mean distribution of the minimum and maximum replies of wives, 18-39 years of age. b Data

number shifted from two in 1941 to three in 1945 and then to four in 1955 and 1960, and the average increased from 3.0 to 3.5. It would thus appear that family-size ideals, as well as the actual size of the typical family, vary from time to time in the United States, perhaps in response to changes in economic conditions or the international situation, or perhaps merely in response to changes in the "fashion" of the times. We can only speculate about the relative roles of such factors, but it should be emphasized that attitudes toward family size have changed in the past and are likely to change in the future. After the wives in our 1960 sample had told how many children they thought were ideal for the average American family, they were asked: ' POQ. See page xxxi for the meaning of any abbreviated reference.

IDEAL, DESIRED, AND EXPECTED FAMILY SIZE

Q2. How many boys and how many girls would this be? Their responses reveal a slight preference for boys. About two-thirds of the wives said they thought the ideal family should include an equal number of boys and girls, but 64 per cent of the remaining third described an ideal in which boys outnumber girls. The average ideal of all wives includes 1.8 boys and 1.7 girls. Interestingly, enough, this ratio of 106 boys per 100 girls is quite similar to the actual sex ratio of newborn infants. In answering the question about the ideal number of boys and girls, many wives were probably influenced by the sex of the children they had already borne. The nonwhites stated a somewhat larger ideal than the whites, as mentioned earlier, but the sex ratio of their ideals is not significantly different. Two questions were asked to find out if American wives think family size should depend oil income: Q3. What is the ideal number (of children) for a family with a high income? Q4. . . . for a family with a low income? Apparently many wives do believe that income should influence family size, for the average number of children considered ideal is more than twice as large for a high-income family as for a low-income family (Table 13). Several things should be noted about these TABLE 13 Average Number of Children Considered Ideal for the Average American Family, a High Income Family, and a Low Income Family, for Wives, by Color Type of reply

Total

Minimum Maximum

3.4 3.5

Minimum Maximum

4.7 4.9

Minimum Maximum

2.1 2.2

White

Nonwhite

The average American family 3.4 3.6 3.5 3.8 A high income family 4.7 5.0 4.9 5.2 A low income family 2.1 2.1 2.2 2.3

answers, however. The interviewers did not define high and low in­ comes, and the terms undoubtedly meant different things to many wives. Furthermore, the difference between the numbers of children the wives said they considered ideal for families with high and low incomes does not resemble the differences that actually exist between

FERTILITY AND FAMILY PLANNING IN THE U.S.

the fertility of wives in the various income groups in the United States. As will be shown in Chapter 3, the wives in the lower income groups expect somewhat more children than those in the higher in­ come groups. Some wives may have given different replies to these two questions simply because they thought it was expected of them. Nevertheless, the results do suggest that most wives believe income ought to be an important consideration in determining ideal family size. In interpreting the data on family-size ideals, it should be re­ membered that, although many of the wives were probably projecting their own personal values to some extent, they were answering ques­ tions posed in terms of a general ideal for the total population—and not for themselves.3 As we shall see later, they often gave quite differ­ ent answers when asked how many children they personally wanted or expected. In order to determine the number of children the wives would want to have themselves under ideal conditions, we asked: Q92. If you could live your life over again, and have everything just the way you would like it to be, how many children would you have in all? Replies to this hypothetical question may be regarded as representing the personal ideals of the wives in contrast with the general ideals obtained from the questions about the average American family. Again there was a strong consensus on two, three, or four children. Almost 80 per cent of the 2,623 wives who gave a numerical reply stated an ideal within this range, and four was again the most frequent response by a substantial margin (Table 14). The average mini­ mum and maximum replies amount to 3.6 and 3.7 children, respectively—about two-tenths of a child more than the numbers considered ideal for the average American family (Table 11). Replies of nonwhite wives tend to be lower than those of white wives, even though the reverse is true in the case of the number of children considered ideal for the average American family. This apparent con­ tradiction can probably be attributed to the feeling of many nonwhite wives that their own economic and social position is so inferior that their personal family-size ideal should be smaller than that of the typical American wife. The average minimum and maximum responses of the white wives to the question about the number of children they would have if they could relive their lives were somewhat higher in 1960 than in 8

Freedman 1954.

IDEAL, DESIRED, AND EXPECTED FAMILY SIZE

1955 (3.6 and 3.7 as compared with 3.4 and 3.5). However, this difference is probably due largely to changes in the wording of the question. The wives who were interviewed in 1955 were more likely to have answered in terms of the realities of their married lives when they were asked: "Now, if you could start your married life over again and choose to have just the number of children you would want by the time you were 45, how many would that be?" Since the question was primarily intended to measure the wives' personal liking for children, it was revised in 1960 to include the phrase "and have everything just the way you would like it to be" in order to TABLE 14 Number of Children Wanted if Life Could be Relived and Per Cent Distribution by Number Wanted, for Wives, by Color

Type of reply

NumPer cent distribution by number wanted if life could be ber Average relived number of 6 or wives wanted Total None 1 2 3 5 more 4

Minimum Maximum

2,623 2,623

3.6 3.7

100 100

Minimum Maximum

2,364 2,364

3.6 3.7

100 100

Minimum Maximum

259 259

3.3 3.3

100 100

Total 2 23 1 21 White 1 1 22 1 1 20 Nonwhite 4 34 5 5 3 34

2 1

24 23

33 35

5 6

11 13

25 24

34 35

5 6

11 13

14 12

25 28

7 6

11 12

encourage the wives to report the number of children they would try to have if they did not have to consider the past, present, or future limitations of their circumstances. DESIRED FAMILY SIZE

The most realistic replies concerning attitudes toward family size were obtained when the wives were asked about the total number of children they themselves actually wanted to have at the time of the interview. As a group, they said they wanted an average minimum of 3.1 and maximum of 3.4 children (Table 15). These figures are significantly smaller than the numbers the wives said they considered ideal for the average American family or would have themselves if they could live their lives over again under ideal conditions. About three-fourths of the 2,684 wives wanted a total of from two to four children, with two being the most popular minimum and four the

FERTILITY AND FAMILY PLANNING IN THE U.S.

most popular maximum. Very few (1 per cent) said they wanted no children at all, and only 5 per cent wanted a maximum of one child. About 12 per cent indicated a preference for families that included at least five children. TABLE 15 Number of Children Wanted at Interview and Per Cent Distribution by Number Wanted, for Wives, by Color

Type of reply

Num­ ber Average Per cent distribution by number wanted at interview 6 or of number wives wanted Total None 1 2 3 4 5 more

Minimum Maximum

2,684 2,684

3.1 3.4

100 100

2 1

Minimum Maximum

2,414 2,414

3.1 3.5

100 100

2 1

Minimum Maximum

270 270

2.7 3.0

100 100

7 6

Total 9 32 5 25 White 8 31 4 24 Nonwhite 13 35 8 32

23 25

22 27

5 7

7 10

24 26

23 28

5 7

7 10

15 18

15 20

9 9

6 7

Here are several cases illustrating the very large majority of Ameri­ can wives who wanted two, three, or four children: CASE 0028. This respondent was 29, had been married nine years, and had two children, the youngest of whom was four years old. She did not want or expect to have any more, and remarked, "The way prices are now our income is not very high. If we had more children, we wouldn't have enough to go around." Her husband was a semi-skilled worker who earned between $3,000 and $4,000 a year. CASE 3019. A 23-year-old college graduate, who had been married for only a few months and had had no experience with motherhood, found it difficult to give a precise answer to the question about the number of children she wanted. She wanted to have at least two, however, and was quite certain that she would not want more than four. She explained that her husband, who was serving as a Second Lieutenant in the Army at the time of the interview, would become a school teacher when he completed his military service and that they would not be able to educate a larger number of children properly.

IDEAL, DESIRED, AND EXPECTED FAMILY SIZE

CASE 0257. An 18-year-old mother of one small baby said she definitely wanted to have two more children, because she herself had felt lonely and depressed as an only child. She would not want a larger family, however, for she and her husband, an aircraft mechanic, would not be able to provide more than three children with the material advantages she had enjoyed when she was growing up. CASE 0246. The wife of a young lawyer in private practice was 27 and had three children. She said she might want to have another child, but only one more, for if she and her husband had more than four, they could not afford to give them all a college education. The following cases are cited as examples of the extremely small number of wives who said they wanted no children at all or only one child: CASE 3050. This respondent was 20, had been married for two years, had had no children, and said that she did not want to have any in the future, adding, "I have never been fond of children. I think they are nerve-wracking. Why I can't even hold a baby without becoming nervous." She was a clerical employee at the same factory where her husband worked as a metal fitter. When asked about her reasons for working, she frankly replied, "Money to me is everything." CASE 3076. This 39-year-old wife was a free-lance writer and her husband was a university professor. Although they had been married for 15 years, they had intentionally avoided having any children, because as the wife said, "I do not feel that children would be happy in our home. We have many other interests; we travel a great deal. I am fond of children, but I feel there are many other things I must do." CASE 0627. A housewife, who was 39 and had been married for 16 years, had a nine-year-old daughter of whom she said she was very fond and proud. She and her husband, a business executive whose annual income was in the $15,000-and-over category, had not wanted to have any more children, because they wanted to take care of their one daughter "nicely" and be able to send her to private schools rather than divide their resources among more children. CASE 0836. This 35-year-old wife had been married a little more than two years and had a baby boy. She wanted to have only the one child, and said, "I don't think we should have more because we can't give them as much as we'd like. I want this one to have

FERTILITY AND FAMILY PLANNING IN THE U.S.

every advantage." The respondent had not worked since the birth of her child; her husband earned between $4,000 and $5,000 a year as a skilled worker. These cases illustrate the small group of wives who expressed a desire for unusually large families: CASE 3209. A 27-year-old wife, who had been married only a few months and had no children, said that she wanted and was very sure that she would have six or eight. She would not want to have fewer, because she thought children were better behaved, learned more, and were happier in large families. The wife was a Catholic and the husband was a Protestant, but the wife expected him to go along with her on family planning matters. CASE 3291. This Irish Catholic couple had been married almost 11 years and lived in a large eastern city where the husband was em­ ployed as a police investigator. The wife was 34 and had had seven children. She said she wanted to have at least 10 altogether but would be happy to accept "as large a family as fate gives me." She would not want to have a smaller family, because she loved children and believed that children in small families were likely to miss much of the "give and take" of learning to live together. CASE 0115. A very devout Protestant wife was 32 and had five children. She said that she wanted and definitely expected to have a total of 12, because as she remarked, "I just enjoy them. I feel you are blessed by the Lord when you have more children." Her husband was a heavy machinery operator in the logging and con­ struction industries. Husbands were not interviewed in the GAF survey, but the wives were asked how many children they thought their husbands wanted. Ninety per cent of the wives gave a numerical reply to this question, with most of the remaining 10 per cent simply indicating that they did not know how many children their spouses wanted. As shown in Table 16, the husbands were reported to want an average minimum and maximum of 3.1 and 3.2 children, respectively. Their preferences, as a group, are thus about the same as the average minimum number of children the wives wanted to have. Nineteen per cent of the wives who answered this question definitely wanted more children than their husbands, and 15 per cent definitely wanted fewer (Table 17). In 66 per cent of the cases, however, the wife either wanted the same number as the husband or the minimum and maximum numbers wanted by one spouse overlapped those of the other. Needless to

IDEAL, DESIRED, AND EXPECTED FAMILY SIZE

say, the validity of the figures depends upon the accuracy with which the wives perceived and reported the number of children their hus­ bands wanted. Nonwhite wives wanted a significantly smaller average number of children than white wives (Table 15). Forty-six per cent stated TABLE 16 Number of Children Wanted by Wife and by Husband, by Color Average number of children

Type of reply

Total

Number of wives Minimum Maximum

2,684 3.1 3.4

Number of husbands0 Minimum Maximum

2,421 3.1 3.2

White Wanted by wife 2,414 3.1 3.5 Wanted by husband 2,191 3.1 3.2

Nonwhite 270 2.7 3.0 230 3.1 3.1

s In 263 cases, the wife did not give a numerical answer to the question about the number of children her husband wanted.

TABLE 17 Percentage of Wives Who Wanted Fewer Children than Their Husbands, the Same Number or More, by Color Comparison

Total

White

Nonwhite

Number of wives Per Cent: Total Wife wanted: Fewera Same numberb More"

2(421

2,191

230

100

100

100

15 66 19

14 67 19

25 57 18

* The maximum number of children wanted by the wife is smaller than the minimum number wanted by the husband. b The minimum and/or maximum numbers of children wanted by the wife are equal to (or between) the minimum and maximum numbers wanted by the husband. 0 The minimum number of children wanted by the wife is larger than maximum number wanted by the husband.

a preference for families including no more than two children, as compared with only 29 per cent of the white wives. Another notable difference is the much larger proportion of nonwhite wives (25 per cent) who reported that their husbands wanted more children than they themselves did (Table 1 7 ) . The minimum and maximum numbers of children wanted by white wives at the time of the interview averaged 3.1 and 3.5 in

FERTILITY AND FAMILY PLANNING IN THE U.S.

the 1960 study and 3.0 and 3.1 in the 1955 study.1 These results are not comparable, however. In 1955, all of the wives were not asked the same question about the number of children they wanted to have; instead, the questions differed somewhat depending upon their replies to previous queries concerning the physiological ability of the couple to have offspring in the future. Because the different wording may have led some respondents to interpret these questions in different ways, an attempt was made to minimize this problem in the 1960 survey by asking all wives virtually the same question: Q56. I've asked about whether you expect to have [more] chil­ dren . . . now I'd like to ask you about how many children you want. If you could have just the number of children you want, and then stop, would you have less than you expect, the same number, or more than you expect?5 If they replied "less" or "more," they were asked: Q56a or b. How many children would you want to have in all? It was hoped that each wife would answer Question 56 by indi­ cating how many children she would have if she and her husband had no fecundity impairments and could have a child whenever she wanted to, if they could prevent conception at other times, and if all other conditions of their lives were the same as they had been thus far and were expected to be in the future. Such a concept is, of course, much more easily defined than obtained in an interview situation.6 Some wives were very uncertain about the number of children they wanted and perhaps answered in terms of the number 4We do not compare our data on the number of children wanted with those of the Princeton Study for two reasons. First, the Princeton Study sample represents white wives in Standard Metropolitan Areas with populations of 2,000,000 or more in 1950 (excluding Boston) who had their second birth in September 1956, five to seven months before the first interview. The GAF sample, in con­ trast, represents all American wives 18 to 39 years of age with husband present (or in the armed forces) in 1960, regardless of their place of residence, the number of children they had had, or the date at which their last birth (if any) had occurred. Second, because of the wording of the question, the wives who were interviewed in the Princeton Study could not say that they wanted fewer than the two children they already had, whereas the wives in the GAF study could say (as some did) that they wanted one child or no children at all. The results of the two phases of the Princeton Study are reported in: Westoff 1961 and Westoff 1963. 1 If a wife had reported that she did not expect to have any more children, the phrase "than you now have" was substituted for the phrase "than you ex­ pect" in the last sentence of this question. ° Several attempts were made to word Question 56 more precisely, but all were discarded as too long and involved for an interview of this type. However, the objectives of the question were explained to the interviewers in their in­ structions.

IDEAL, DESIRED, AND EXPECTED FAMILY SIZE

they would want to have under better or ideal circumstances. About one-tenth of the wives replied that they wanted at least as many children as they had, even though in answering Question 16 earlier in the interview they had said that, at the time of their most recent conception, they had not wanted another child either then or later.7 In such cases, the replies to Question 56 were revised to reflect the answers to Question 16.® In other cases where the wife may have misinterpreted Question 56 or answered unrealistically, there were no alternative questions that could be used as a basis for modifying the replies and they were accepted as given. Thus, it appears very likely that the data on the number of children wanted by the wives still overstate, to some extent, the size of families they would try to have in the absence of fecundity impairments and unwanted conceptions. Some of the wives who were thinking in terms of better circumstances than they were actually experiencing would not necessarily have given contra­ dictory answers to Questions 16 and 56. Other wives who already had more children than they really wanted probably said they wanted to have at least as many as they had in order to avoid the embarrass­ ment of admitting that any of their children were unwanted or that they had ever had a family planning failure. On the other hand, it is worth noting that although most wives who reported fecundity impairments will never have to consider the possibility of a large number of children, there was no indication that this encouraged them to overstate their family-size preferences unrealistically or to rationalize unusually small preferences that would conform more closely to their ability to bear children. The average number of children wanted by wives and the distribution of wives by number wanted were virtually the same for both Subfecund and Fecund couples.® ' Q16. Shortly before your last [present] pregnancy began did you really want another child sometime? ' This was done because we believe that in a large majority of these cases the conception that was reported as unwanted in Question 16 would not have been intentionally sought at a later date. In answering Question 56, many of these wives probably indicated that their latest child was wanted, because they had come to accept the child after it was born, or because they were thinking in terms of better or ideal circumstances. Of course, as the years pass, some wives change their minds about the number of children they want for various reasons, and some of the births in question would probably have been planned later if the unwanted conception had not occurred when it did. We believe, however, that this would be true in only a minority of the cases in which the wife gave contradictory answers to Questions 16 and 56. " The fecundity categories are defined and discussed in Chapter 4. They are capitalized to remind the reader that they are being used in a special way in this study.

FERTILITY AND FAMILY PLANNING IN THE U.S.

It should not be assumed, of course, that the number of children a wife wants remains constant throughout the childbearing years of her married life. Even if there were no basic economic and social changes in the society as a whole, the complex multitude of factors that affect the daily lives of individual couples would undoubtedly cause some wives to revise their personal family-size preferences up­ ward or downward from time to time. A very substantial proportion of the respondents in our sample appear to have changed their minds. In addition to Question 56 about the number of children wanted at the time of the interview, all wives were asked: Ql 1. Just before you were first married, how many children did you want to have in all? If the wife had had a live birth a year or more before the interview, she was also asked: Q15. A year after your first child was born, how many children did you want to have in all? The average replies to these questions suggest that the introduction to the realities of motherhood during the year following the birth of the first child is one experience that causes the number of children wanted by many wives to decline quite significantly: AVERAGE NUMBER OF CHILDREN WANTED BY WIFE

All wives White Nonwhite

Just before marriage

One year after birth of first child

At time of interview

3.1 3.2 2.4

2.6 2.7 2.1

3.3 3.3 2.9

Later, the number wanted apparently returns to a higher level than that reported for the period just before marriage. Replies to such retrospective questions should not be regarded as precise indicators of past attitudes, but it seems likely that they do reflect real fluctua­ tions. Certainly, they indicate that a large proportion of the wives thought that changes had occurred in their family-size preferences. In many cases where a wife had an unwanted conception, she may later have come to welcome the child, perhaps months before its birth. L. T. Badenhorst reported that more than one-fourth of the respondents in a sample survey of white wives in Johannesburg replied "yes" when asked: "Has it happened that a pregnancy which was unplanned and unwanted came to be welcomed in later years?"10 10

Badenhorst 1963.

IDEAL, DESIRED, AND EXPECTED FAMILY SIZE

EXPECTED FAMILY SIZE

All wives were also asked a series of questions concerning the number of children they had already borne by the time of the inter­ view and the total number they actually expected to have altogether. The methods by which the data on expectations were obtained are discussed in detail in Appendix B, Note 2. Very briefly, the number of children expected is the total number of children each wife thought she would have when her family was completed. This is believed to be a much more realistic indicator of size of completed family than the number of children wanted or any of the other measures of attitudes toward family size. The wording of the questions about expectations was intended to prevent wives from answering in terms of circumstances that were better or worse than those they were actu­ ally experiencing. Moreover, these questions were preceded by a series of questions concerning the wife's past pregnancy history and the physiological ability of the husband and wife to bear children in the future, in order to encourage the respondents to consider these important factors very carefully in reporting their expectations. By the time of the interview, the wives in our sample had been married for about a decade and had had 2.4 births on the average. Thus, a majority already had a very considerable amount of experi­ ence with both motherhood and family planning. This should have provided them with a realistic basis not only for deciding how many children they would try to have but also for predicting their ability to achieve such a goal. Of course, many of the wives will not have the number of children they expected. Some will have fewer because of unforeseen fecundity problems, some will have more because of accidental conceptions, and others will change their minds and revise their expectations ac­ cordingly. The comparison of the results of the 1955 and 1960 studies made in Chapter 1 indicates that, among all white wives who were 18-39 in 1955, these tendencies very nearly cancelled each other with respect to births during 1955-60. In other words, the five-year expectations of the 1955 wives as a group corresponded very closely with the number of births actually occurring to comparable 1960 wives during this period. However, the total number of births expected did increase slightly as the wives grew older between the two studies. Obviously, a definitive evaluation of the accuracy of the wives' total expectations cannot be made until they reach the end of their childbearing years, but thus far they seem to be quite reasonable for the sample as a whole. Table 18 shows the average number of births to wives by the

FERTILITY AND FAMILY PLANNING IN THE U.S.

time of the interview, the average number of children living then, the average number of additional births expected, and the average total number of births and children expected altogether. No allowance has been made for children who may die after 1960. The small difference between the total number of births expected and the total number of children expected is due solely to deaths that occurred prior to the interview. The 2,684 wives expected an average minimum and maximum of 2.9 and 3.5 births, with the "most likely" expected number being 3.2. The latter number is usually the same as the minimum and maximum numbers when they are alike or the average TABLE 18 Average Number of Births by Interview, Children Living at Interview, Additional Births Expected, Total Births Expected, and Total Children Expected, for All Wives (1960), White Wives (1955 and 1960), and Nonwhite Wives (1960)

Type of reply Births by interview Children living at interview Additional births expected Most likely Minimum Maximum Total births expected Most likely Minimum Maximum Total children expected Most likely Minimum Maximum

Total 1960

1955

White 1960

Nonwhite 1960

2.4 2.3

2.1 2.0

2.3 2.2

2.7 2.6

0.8 0.6 1.2

1.0 0.7 1.2

0.8 0.6 1.2

0.9 0.5 1.3

3.2 2.9 3.5

3.0 2.7 3.3

3.1 2.9 3.5

3.6 3.3 4.1

3.1 2.8 3.5

2.9 2.7 3.2

3.1 2.8 3.4

3.4 3.1 3.9

of the two when they are different. For example, in nearly all cases in which the wife said she expected to have "two or three" births, 2.5 is the most likely expected number. In effect, it is assumed that one-half of such wives will have two births, and the other half, three. Most of the exceptions to this rule involve couples with fecundity impairments which might prevent future births. In such cases, the wife is classified as expecting 0.3 births in addition to any she may already have had, if she said she would have a child sometime in the future if she could and if the fecundity impairment is not serious enough to indicate definite sterility. The most likely expected number of births is thus based on the assumption that 30 per cent of the wives in this category will be able to bear one child after the interview. If the fecundity impairment indicates definite sterility, however, the

IDEAL, DESIRED, AND EXPECTED FAMILY SIZE

wife is classified as expecting no births after the interview regardless of the answers she gave to the questions about her expectations.11 The distribution of wives by the most likely expected number of births in Table 19 indicates that a large majority—about 7 out of 10—thought they would have a total of two, three, or four births. A very large proportion of those who expected fewer than two or more than four apparently did so involuntarily. Subfecund couples12 account for 69 per cent of the wives who expected no births or only one birth, and 84 per cent of these wives said they would definitely prefer to have larger families. Among wives who expected five or more births, we find 48 per cent reporting that they wanted TABLE 19 Number of Births Expected and Per Cent Distribution by Number Expected, for Wives, by Color

Type of reply

Num­ ber Average Per cent distribution by number expected of number wives expected Total None 1 2 3 4 5

Most likely Minimum Maximum

2,684 2,684 2,684

3.2 2.9 3.5

100 100 100

4 7 3

Most likely Minimum Maximum

2,414 2,414 2,414

3.1 2.9 3.5

100 100 100

4 6 3

Most likely Minimum Maximum

270 270 270

3.6 3.3 4.1

100 100 100

6 9 4

Total 7 25 10 27 4 22 White 7 25 10 28 4 22 Nonwhite 11 22 13 25 5 23

6 or more

28 25 26

19 18 25

8 7 8

9 6 12

29 26 27

20 18 26

8 7 8

7 6 11

18 14 15

14 14 20

11 10 9

18 15 23

fewer than they expected. This group includes many of the couples who had been unsuccessful in preventing unwanted pregnancies (see Chapter 6). Distributions by the minimum and maximum numbers expected also show about 7 out of 10 wives anticipating two to four births (Table 19). However, two and three are the most frequent minimum expectations and three and four are the most frequent maximum expectations. From the standpoint of future population, it is very important which number of births wives have within the popular two-to-four range. The minimum and maximum expectations average 2.9 and 3.5, respectively. If American wives were to have the maxi11 The procedure used in obtaining the most likely expected number of births is described in detail in Appendix B, Note 2. 12 The fecundity classification is discussed in Chapter 4.

FERTILITY AND FAMILY PLANNING IN THE U.S.

mum number expected, the resulting rate of population growth would be about twice as high as it would be if they were to have the minimum number. Three-fourths of the most likely expected number of births—2.4 of 3.2—had already occurred by the time of the interview in 1960 (Table 18). More than 87 per cent of the wives had borne at least one child, and 68 per cent had borne two or more. Less than half (42 per cent) of the wives said they definitely expected to have at least one birth after the interview. Most of these were younger wives. Those under 30 account for about three-fourths of all the additional births expected. At the time of the interview, they already had six-tenths of the children they expected; wives who were 30-39 had almost nine-tenths of the number they expected. The expected size of completed family thus represents a considerable amount of prediction for the younger wives, but mainly past perform­ ance for the older ones. These findings dramatically reflect the tendencies of women to marry early and have their children soon after marriage; the findings will be discussed in Chapter 8. Eighty-seven per cent of the wives who expected at least one more birth were classified as Fecund, and another 8 per cent as Possibly Fecund. Their expectations appear to have a realistic basis, for most will be capable of bearing the number of children they anticipate. In fact, more than one-fourth of the wives who expected an additional birth were pregnant when interviewed. In 1962, the Population Studies Center of the University of Michigan began collecting data on the number of births to date and the total number of births expected that will be comparable to those from the 1955 and 1960 GAF studies. These fertility measures are being obtained along with information on the racial, religious, socioeconomic, and demographic characteristics of married couples by adding a small number of relevant questions to various nationwide surveys being conducted by the University's Survey Research Center. The study will provide valuable up-to-date information concerning the family-building expectations of American couples during each of the five years beginning in 1962.13 The first annual report, based on the replies of 1,383 respondents interviewed in 1962, generally confirms the results of the 1960 GAF study and indicates no major changes.14 The mean total number of births expected is the same in both studies (3.2), and the mean number of births that had occurred by the time of the interview 13 Axelrod

1963. 1963.

14 Freedman

IDEAL, DESIRED, AND EXPECTED FAMILY SIZE

is only slightly higher for 1962 than for 1960 (2.5 as compared with 2.4), possibly as a result of a continuation of the trend toward childbearing at younger ages. One methodological difference in the two studies should be mentioned, namely, the husband rather than the wife is the respondent for about one-half of the couples in the University of Michigan study. However, the average size of expected family is the same for all husbands and all wives as a group. Small differences emerge when the respondents are divided according to wife's age, but they do not appear to be statistically significant. Although nonwhite wives, as a group, want fewer children than white wives, they actually expect more (Figure 4, Table 18). The TOTAL

ιw//m/A

WHITE

mmv,i

NONWHITE T 1 1 Ι 2 3 AVERAGE NUMBER OF BIRTHS BIRTHS BY I960

TOTAL NUMBER EXPECTEO

Γ­ 4

MINIMUM MOST LIKELY \ w MAXIMUM

FIGURE 4 Births by 1960 and total number expected, for wives, by color

distribution of nonwhite wives by the most likely expected number of births in Table 19 shows a much larger proportion expecting five or more, as well as a significantly larger proportion expecting fewer than two. Only 54 per cent of the nonwhite wives thought they would have two to four births as compared with 74 per cent of the white wives. The higher percentage of nonwhite wives anticipating large families can apparently be attributed mainly to their more frequent failures to prevent unwanted conceptions. Among wives expecting five or more births, the most likely number ex­ ceeds the maximum number wanted in 79 per cent of the nonwhite cases and in only 42 per cent of the white cases. On the other hand, the larger proportion of nonwhites expecting no chil­ dren or only one child appears, for the most part, to reflect nonwhite preferences rather than involuntary factors. Recognized fecun-

FERTILITY AND FAMILY PLANNING IN THE U.S.

dity impairments are only slightly more prevalent among nonwhite than white couples (see Chapter 4). How reliable are the expectations of the nonwhite wives? By the time of the interview in 1960, they had already had an average of 2.7 births as compared with 2.3 for the white wives. Yet the average most likely expected number of additional births is only 0.1 larger for the nonwhites than for the whites (see Table 18). This suggests that the expectations of the nonwhite wives may be unrealistically low, perhaps as a result of wishful thinking. We know that the proportion of couples who have only a grade-school education, who have never used any method of contraception, and who have already had a large number of children by the time of the interview is substantially larger for nonwhites than for whites. In the 1955 survey, the expectations tended to be unrealistically low for white couples with such characteristics (see Chapter 1). For white wives, the average most likely expected total number of births amounted to 3.0 in 1955 and 3.1 in 1960 (see Table 18). Differences between the expectations of wives 25-39 years of age on the two dates account for all of this small increase; wives under 25 expected fewer births in 1960 than those of that age in 1955. Since the number of births by the time of the interview increased more than the total number expected for wives 18-39 as a group, the number of additional births expected declined somewhat between 1955 and 1960. In addition, the proportion of wives who definitely expected to have at least one birth after the interview fell from 49 to 42 per cent. These last two changes reflect the tendency for wives to marry and have their children at younger ages, and help to explain the decline in the birth rate since I960.15 IDEAL AND DESIRED FAMILY SIZE VS. EXPECTED FAMILY SIZE

The average number of children wives want to have when inter­ viewed is significantly smaller than the number they consider ideal for the average American family or the number they say they them­ selves would have if they could live their lives over again and have everything just the way they would like it to be (Figure 5, Table 20). Apparently, the family-size ideals of many wives call for more children than they feel it is prudent for them to have considering the limitations of their own circumstances. However, all of the data on ideal and desired family size reflect a strong consensus on moder­ ate-sized families of two, three, or four children. About 9 out of 10 wives considered two to four children to be ideal for the average American family, 8 out of 10 said they would have a number within 15 Whelpton

1963; Freedman 1963.

IDEAL, DESIRED, AND EXPECTED F A M I L Y SIZE NUMBER OF CHILDREN CONSIDERED IDEAL FOR AVERAGE

AMERICAN

FAMILY·'

NUMBER OF CHILDREN WANTED IFLIFE COULD BE RELIVED". WHITE

NONWHITE

NUMBER OF CHILDREN WANTED AT TIME OF INTERVIEW :

m

WHITE

M

NONWHITE

NUMBER OF CHILOREN EXPECTED: WHITE

NONWHITE

AVERAGE NUMBER OF CHILDREN MINIMUM

MAXIMUM

FIGURE 5 Number of children considered ideal for the average American family, wanted if life could be relived, wanted at time of interview, and expected, for wives, by color

this range if they could relive their lives under ideal conditions, and more than three-fourths wanted to have this many at the time of the interview. Although the proportion of wives actually expecting to have from two to four children is almost as large as the proportion wanting

FERTILITY AND FAMILY PLANNING IN THE U.S.

a number within this range, the average most likely expected number is only as large as the average minimum number wanted (Table 20). The most likely expected number is equal to (or between) the minimum or maximum number wanted in slightly less than half (48 per cent) of the cases. As shown in Table 21, 28 per cent of the wives expected to have fewer children than they wanted. This appears to have resulted chiefly from subfecundity and the smaller family-size preferences of some of the husbands. Fifty-four per cent of these wives indicated that either they or their husbands had some type of fecundity impairment, and another 23 per cent reported that their husbands wanted fewer children than they did themselves. Most of TABLE 20 Number of Children Considered Ideal for the Average American Family, Wanted If Life Could Be Relived, Wanted at Interview, and Expected, for Wives, by Color Type of reply

Minimum Maximum Minimum Maximum Minimum Maximum Most likely Minimum Maximum

Total

Average number of children White

Nonwhite

Considered ideal for average American family 3.6 3.4 3.4 3.5 3.5 3.8 Wanted if life could be relived 3.6 3.6 3.3 3.7 3.3 3.7 Wanted at time of interview 3.1 3.1 2.7 3.0 3.4 3.5 Expected 3.1 3.1 3 4 2.8 2.8 3.1 3.5 3.4 3.9

the remaining wives in this category probably misinterpreted the ques­ tion about the number of children wanted and overstated the number they would try to have under the conditions of life they were actually experiencing. Table 21 also shows that 24 per cent of the wives expected more children than they wanted; in fact, 16 per cent already had more than they wanted when interviewed. Most of the wives who expected more children than they wanted had had accidental conceptions, but the larger family-size preferences of some of their husbands was a contributing factor. More than one-third of these wives said their husbands wanted more children than they did, as compared with only one-twelfth of the other wives. After the wives were asked about the total number of children

IDEAL, DESIRED, AND EXPECTED FAMILY SIZE

they wanted and expected, they were asked what their reasons were for (1) not wanting a smaller family and (2) not expecting a larger family. It was believed that the answers to the second question would be more realistic if it were posed in terms of the number of children actually expected. In the first question, however, the wives were asked about their reasons for not wanting a smaller family. If they had been asked why they did not expect to have a smaller family, too many could only have replied that they already had the number of children they expected. Tables 22 and 23 present percentage dis­ tributions of the wives by the first reason they gave in response to these two questions. Half of the wives cited reasons for not wanting a smaller family that are related to the welfare or happiness of their TABLE 21 Percentage of Wives Who Expected Fewer Children than They Wanted, the Same Number or More, by Color Comparison

Total

White

Nonwhite

Number of wives Per Cent: Total Expected fewer than Wanteda Expected number wantedb Expected more than wanted0 Had number wanted or fewer Had more than wanted

2,684

2,414

270

100 28 48 24 8 16

100 29 50 22 7 15

100 23 37 41 11 30

" The most likely expected number is smaller than the minimum number wanted. b The most likely expected number is equal to (or between) the minimum or maxi­ mum number wanted. 0 The most likely expected number is larger than the maximum number wanted.

children (e.g., "Children need that many brothers and sisters for companionship"). Many of the other wives gave reasons that appear to be associated with the satisfaction or interests of the parents. Only 2 per cent said they did not want smaller families because of religious and moral reasons. Subfecundity was cited by 18 per cent of the wives as the first reason for not expecting a larger family (Table 23). If these wives are excluded, we find that 54 per cent of the remaining wives gave some type of economic reason, 10 per cent reported a reason related to their own health or the health of some other member of the family, and 9 per cent indicated that they would find it difficult to "care" or "have time" for a larger number of children. The replies of many wives probably reflect the views of their husbands, but fewer than 1 per cent mentioned their husbands' pref­ erences as the first reason for not wanting a smaller family or expecting

FERTILITY AND FAMILY PLANNING IN THE U.S.

a larger one. It should also be noted that a significant number of wives gave very vague answers (e.g., "I just think that is a nice number") or simply said they didn't know what their reasons were. Of course, most of the specific answers in Tables 22 and 23 could be used to explain a preference for families of almost any size. In deciding how many children they would try to have, many of the wives were probably influenced less by the reasons they gave than by the family-size norms or "fashions" of the religious, social, TABLE 22 Per Cent Distribution by First Reason for Not Wanting a Smaller Family, for Wives, by Color First Reason Number of wives Per Cent: Total Reasons related to welfare or happiness of children To provide companionship for children; better for children To avoid having an only child Reasons related to satisfaction or interests of parents Likes, loves children; this many makes for a happy, "full" life To avoid loneliness; to have children around the house now or when older To have desired proportion of boys and girls Religious and moral reasons (e.g., couples ought to have this many) Ambiguous reasons This number is right, nice, ideal; like parental family Don't know Other reasons Not ascertained

Total

White Nonwhite

2,684

2,414

270

100

100

100

50

51

40

36 14 20

37 14 20

29 11 25

14

14

15

3 3

3 3

2 8

2 12

2 12

1 12

10 2 6 9

10 2 5 8

9 3 9 13

and economic groups with which they and their families were closely associated. Most wives wanted to have about the same number of children as they said they thought their married friends of the same age would have. The family-size ideals, desires, and expectations of nonwhite wives follow a very different pattern from those of white wives. Nonwhite wives set the ideal number of children for the average American family higher than white wives, but want to have substantially fewer than white wives, and even say they would have fewer if they could relive their lives under ideal conditions (Figure 5, Table 20). How-

IDEAL, DESIRED, AND EXPECTED FAMILY SIZE TABLE 23 Per Cent Distribution by First Reason for Not Expecting a Larger Family, for Wives, by Color First reason Number of wives Per cent: Total

Total

White

Nonwhite

2,684

2,414

270

100

100

Subfecundity All other reasons Economic reasons Would cost too much; couldn't afford more, etc. Couldn't support more at desired stand­ ard of living Couldn't educate more Low income, insecure job, part-time work, unemployment Other economic reasons Poor health (parents and /or children); pregnancy unpleasant, or dangerous to wife or baby Hard to "care" or "have time" for more This is a "nice" number, number wife and/or husband wants; no more needed for happiness of family, etc. Other reasons, don't know Not ascertained

100

18 82 44

100 54

17 83 45

100 54

22 78 36

100 46

22

27

22

27

22

28

6 5

8 6

7 6

8 7

3 3

4 4

3 7

4 9

4 7

4 9

2 5

3 7

8 8

10 9

8 7

10 9

6 10

7 13

5 14 4

6 16 5

5 14 4

6 16 5

7 14 5

9 18 7

EXPECTED FEWER THAN WANTED

EXPECTED NUMBER WANTED

EXPECTED MORE THAN WANTED

PER CENT WHITE

NONWHITE

FIGURE 6 Percentage of wives who expected fewer children than they wanted, the same number or more, by color

FERTILITY AND FAMILY PLANNING IN THE U.S.

ever, as a group, they actually expect to have significantly more chil­ dren than white wives and considerably more than they themselves would prefer. Figure 6 and Table 21 show that 41 per cent expected more than they wanted, and 30 per cent thought they already had too many when interviewed. Although these percentages are almost twice as large as those for white wives, they probably understate the propor­ tion of nonwhite wives that will finally have more children than desired, for as previously mentioned, it appears that their expectations are often unrealistically low. Nonwhite wives generally gave reasons for not wanting a smaller family that were similar to those of the white wives (Table 22). However, a few differences should be noted. A smaller proportion of nonwhite wives cited reasons related to the welfare or happiness of their children, probably because they were more likely to have large families and adequate companionship for their children. On the other hand, a somewhat larger proportion gave reasons that appear to be associated with the satisfaction or interests of the parents, and a higher percentage of their reasons were not ascertained. One nonwhite wife out of twelve said she did not want a smaller family because this might keep her from having the desired proportion of boys and girls. Twenty-two per cent of the nonwhite wives reported subfecundity as the first reason for not expecting a larger family, as compared with 17 per cent of the white wives (Table 23). Among wives citing other reasons, we find that nonwhite wives were more Ukely to indicate that they would have trouble "caring" or "having time" for a larger number of children and less likely to mention economic and health reasons. Perhaps many of the nonwhite wives who expressed concern over the problem of "caring" or "having time" for a larger number of children did so because they were working or wanted to work to provide needed income for their families. In such cases, this concern should also be considered an economic reason. In the preceding sections of this chapter, we have described the family-size ideals, desires and expectations of the white and nonwhite wives as a group. It should not be assumed, of course, that this description applies to all the white and nonwhite subgroups in the sample. Some of our findings differ significantly with demographic, religious, and socioeconomic characteristics. Such differentials among the nonwhite wives will be discussed in Chapter 9. In the remainder of the present chapter, we shall examine the differences in the fertility values and expectations of white wives that are associated with such demographic factors as wife's age, wife's age at marriage, duration of marriage, and number of births by the time of the interview.

IDEAL, DESIRED, AND EXPECTED FAMILY SIZE

Chapter 3 will present data for white wives in the various religious, social and economic groups. EXPECTED FAMILY SIZE BY WIFE'S AGE AND OTHER DEMOGRAPHIC VARIABLES

Expected family size varies significantly with age of wife. As shown in Figure 7 and Table 24, the average most likely expected total number of births is lowest for the youngest and oldest white wives16 and highest for those who were 25-29 and 30-34 in 1960." The difference WIFE'S AGE

'///////////λ

18-24

V/////A

25-29 30-34

Ά

35-39

1

1

1

1 2 3 AVERAGE NUMBER OF BIRTHS

1

4

BIRTHS BY I960 TOTAL NUMBER EXPECTED (MOST LIKELY)

FIGURE 7 Births by 1960 and total number expected, for white wives, by age

between the average number of births expected by wives who were 18-24 and those who were 25-29 (3.0 vs. 3.4) is well above the minimum required for statistical significance. This is one of our most important findings. The lower expectations of the younger wives may 16 The remainder of this chapter will deal only with white wives. Nonwhite wives are discussed in detail in Chapter 9. 17 A similar pattern of expectations was found in the 1962 survey of the Univer­ sity of Michigan's Population Studies Center:

Freedman 1963, p. 369.

Wife's age

Mean total no. of children expected

18-24 25-29 30-34 35-39

3.1 3.4 3.2 2.9

FERTILITY AND FAMILY PLANNING IN THE U.S.

forecast a reversal of the postwar trend toward larger families, and the childbearing of these wives will have an important influence on the birth rate during most of the 1960's. Wives who were 18-24 account for 43 per cent of the additional births expected by all wives who were 18-39 years of age. The distributions of the wives by the most likely expected number of births in Table 24 show that a majority of the wives in all age groups expected two, three, or four births, with the proportion con­ sistently declining from 84 per cent at ages 18—24 to 66 per cent TABLE 24 Number of Births Expected and Per Cent Distribution by Number Expected, for White Wives, by Age

Wife's age

Num­ Average Per cent distribution by number expected ber of number wives expected Total None 1 2 3 4 5

Total 18-24 25-29 30-34 35-39

2,414 513 600 624 677

3. 1 3,.0 3. 4 3. 2 3. 0

100 100 100 100 100

Total 18-24 25-29 30-34 35-39

2,414 513 600 624 677

2 .9 2 .7 3 .1 3 .0 2,.8

100 100 100 100 100

Total 18-24 25-29 30-34 35-39

2,414 513 600 624 677

3 .5 3 .4 3 .7 3 .6 3..3

100 100 100 100 100

Most likely expected 7 25 4 2 5 27 2 4 22 7 24 4 11 27 8 Minimum expected 10 28 6 6 11 29 3 6 27 6 27 9 9 12 28 Maximum expected 3 4 22 1 2 21 1 17 2 3 3 22 7 27 6

6 or more

29 33 32 27 24

20 24 22 20 15

8 6 11 8 7

7 3 7 10 9

26 27 30 25 23

18 22 21 17 14

7 4 9 8 6

6 1 4 8 8

27 31 29 26 23

26 31 29 26 19

8 8 10 8 8

11 5 12 13 11

at ages 35-39. We may reasonably question whether such a large percentage of the younger wives will actually have two to four births, however. The proportion under 30 years of age expecting fewer than two and the proportion under 25 expecting more than four appear to be unrealistically low in view of our findings concerning the past experience of earlier cohorts. As the younger wives grow older, some will find that fecundity impairments will prevent their having as many children as they expected and others will have more than they ex­ pected because of family planning failures. The comparison of the results of the first and second GAF studies in Chapter 1 indicates that these two tendencies almost offset each

IDEAL, DESIRED, AND EXPECTED FAMILY SIZE

other between 1955 and 1960, with the younger wives, as a group, having about the same number of births as they expected during this period. However, the total number of births expected did increase somewhat as the wives grew older. This tendency will probably slightly diminish the difference between the future expectations of the wives who were 18-24 and the wives who were 25-29 in 1960, but unless it differs substantially from the pattern suggested by the 1955-60 comparisons, the 18-24-year-old group's expectations will remain sig­ nificantly lower.18 Expected family size also varies with wife's age at marriage. As pointed out in FPSPG,19 the age at which women marry is related to fertility in two ways. First, women who marry at younger ages are more likely to belong to socioeconomic groups with high fertil­ ity. Second, the earlier women marry, the longer the period they will have during which they can bear children before the onset of menopause or the development of premature fecundity impairments. The first of these two relationships is by far the more important, for the influence of age at marriage on family size depends chiefly on the number of children couples want and their willingness and ability to prevent unwanted births. Women do not necessarily have more children simply because they spend more childbearing years in marriage. Table 25 shows that the average most likely expected total number of births declines as wife's age at first marriage rises. Wives who married before they were 18 expected an average of 3.4 births, or about one-fourth more than the 2.7 expected by the 200 wives who married after they were 25. Even this difference between the youngest and oldest age-at-marriage groups is not extreme, however. Neither group expected very large or very small families. Two to four children are typical childbearing goals, and unless they marry at unusually late ages, most wives will have sufficient time to bear all the children they want. The average total number of births expected also declines as age at marriage increases within each of the different age groups as a general rule (Table 25). The exceptions are not statistically significant and are very likely due to sampling variability. As would be expected, the older wives generally had had more births by 1960 than the younger wives, regardless of age at marriage (Table 25). Wives who were 35-39 had already borne nine-tenths of all the babies they expected, but the 18-24-year-olds had borne "Chapter 10 discusses the effects of the tendency for expectations to rise as wives grow older. 19P. 231.

FERTILITY AND FAMILY PLANNING IN THE U.S.

TABLE 25 Number of Births by 1960 and Additional and Total Number Expected, for White Wives, by Age in 1960, by Age at First Marriage Wife's age in 1960

Total

Total 18-24 25-29 30-34 35-39

2.3 1.3 2.3 2.7 2.7

Total 18-24 25-29 30-34 35-39

0.8 1.7 1.0 0.6 0.3

Total 18-24 25-29 30-34 35-39

3.1 3.0 3.4 3.2 3.0

Wife's age at first marriage

Under 18

18-19

20-21

22-24

Births by 1960 2.4 2.7 2.2 2.1 1.3 1.9 0.7 0.5 3.0 2.7 2.3 1.6 3.0 3.1 2.7 2.3 3.1 3.2 2.6 2.5 Most likely expected additional births 0.9 0.7 0.9 0.9 1.7 2.7 1.3 2.1 0.8 1.0 1.5 0.5 0.5 0.3 0.5 0.8 0.2 0.3 0.2 0.3 Most likely expected total births 3.4 3.3 3.1 3.0 2.9 3.2 3.2 2.8 3.4 3.5 3.3 3.2 3.4 3.5 3.2 3.1 3.3 3.4 2.9 2.7

25-39

1.7 0.7 1.4 2.1 1.0 2.6 1.1 0.4 2.7 3.3 2.5 2.5

only a little more than four-tenths of their expected total. As a result, the average number of additional births expected after 1960 declines from 1.7 at ages 18-24 to 0.3 at ages 35-39. A very large majority of the wives over 30 did not expect any additional births: PER CENT EXPECTING'

Wife's age

1 or more additional births

2 or more additional births

3 or more additional births

Total 42 19 18-24 81 48 25-29 54 24 30-34 32 10 35-39 13 2 0 Based on the most likely expected number of births.

7 21 8 2 1

Wives under 30 account for 73 per cent of the additional births expected by all 18-39-year-old wives:

Wife's age

Per cent of additional births expected by all wives

Total 18-24 25-29 30-34 35-39

100 43 30 18 9

IDEAL, DESIRED, AND EXPECTED FAMILY SIZE

This emphasizes the important influence that the childbearing of the younger wives is likely to have on the birth rate during most of the 1960's. Since the number of births by 1960 varies with wife's age at interview and age at marriage, it also varies with duration of marriage. As the number of years since wife's first marriage rises, the average number of births by 1960 increases and the average number of addi­ tional births expected declines (Table 26). By 1960, wives who had TABLE 26 Number of Births by 1960 and Additional and Total Number Expected, for White Wives, by Age, by Years Since First Marriage Wife's age

Total

Total 18-24 25-29 30-34 35-39

2.3 1.3 2.3 2.7 2.7

Total 18-24 25-29 30-34 35-39

0.8 1.7 1.0 0.6 0.3

Total 18-24 25-29 30-34 35-39

3.1 3.0 3.4 3.2 3.0

Years since wife's first marriage

Under 5

5-9

10-14

Births by 1960 2.8 2.3 * 2.3 3.0 2.5 2.9 2.2 * 2.7 1.6 Most likely expected additional births 1.9 0.4 0.9 * 1.9 1.0 1.9 0.5 0.9 0.5 1.5 0.9 * 0.4 0.3 Most likely expected total births 3.3 2.9 3.2 * 2.9 3.3 3.5 3.2 3.4 3.3 2.3 3.0 * 3.0 2.1 1.0 1.0 1.3 0.8

15 or more

2.9 3.2 2.9 0.2

0.3 0.2 3.2 3.4 3.1

been married less than 5 years had one-third of the total number of births they expected, whereas wives married 15 or more years had nine-tenths of their expected births. The proportion of wives expecting at least one additional birth falls quite rapidly from 87 per cent for wives married under 5 years to 11 per cent for wives married 15 or more years: PER CENT EXPECTING

Years since wife's first marriage

1 or more additional births

2 or more additional births

3 or more additional births

Under 5 5-9 10-14 15 or more

87 51 23 11

56 18 4 2

24 6 2 0

FERTILITY AND FAMILY PLANNING IN THE U.S.

The number of additional births expected varies significantly with the number of births the wives had had when interviewed. As shown in Table 27, the average most likely expected number of additional births declines from 1.6 for wives with no births to 0.5 for those with three or four births, and then rises to 0.7 for those with five or more. The higher additional expectations of wives with five or more births can be attributed to the fact that this group primarily represents couples that wanted large families or had been unwilling TABLE 27 Additional and Total Number of Births Expected, for White Wives, by Age, by Number of Births by 1960 Wife's age

Total

0

Total 18-24 25-29 30-34 35-39

0 .8 1 .7 1 .0 0 .6 0 .3

1 .6 2 .5 2 .0 0 .9 0 .2

Total 18-24 25-29 30-34 35-39

3 .1 3.,0 3..4 3. 2 3..0

1 .6 2,,5 2.,0 0.9 0 .2

Number of births by I960 1 2 3

4

Most likely expected additional births 1.3 0.6 0.5 0.5 * 1.8 1.2 0.8 1.5 0.9 0.7 0.7 0.8 0.5 0.4 0.4 0.3 0.2 0.2 0.2 Most likely expected total births 2.3 2.6 3.5 4.5 * 2.8 3.2 3.8 2.5 3.7 2.9 4.7 3.4 1.8 2.5 4.4 1.3 2.2 3.2 4.2

5 or more

0.7 *

0.8 0.8 0.5 6.5 *

6.1 6. 6 6.7

or unable to prevent unwanted births. The percentages of wives ex­ pecting at least one, two, or three additional births are lowest for those who had three or four births by 1960: PER CENT EXPECTING

Number of births by I960 0 1 2 3 4 5 or more

1 or more additional births 61 65 37 27 28 34

2 or more additional births 50 33 13 6 7 12

3 or more additional births 25 12 3 2 2 5

When the number of births by 1960 is held constant, we find that younger wives generally expected a larger additional and total number of births than older wives (Table 27). Among wives with two previous births, for example, the most likely expected additional number of births averaged 1.2 for the 18-24-year-olds as compared with only 0.2 for those who were 35—39. There are two reasons

IDEAL, DESIRED, AND EXPECTED FAMILY SIZE

for this. First, younger wives obviously have more childbearing years ahead of them. Second, the younger the age at which a wife has a given number of births, the more likely she and her husband are to want more children or to be ineffective in their family planning. In the 1955 study, the total number of births expected was in­ versely related to wife's age, with the youngest wives expecting the largest families. In 1960, however, the youngest wives (18-24) ex­ pected smaller families than wives who were 25-29.20 Table 28 shows that the average most likely expected total number of births was lower in 1960 than in 1955 for wives who were 18-24 and higher for those who were 25 and over. The higher expectations of the latter group resulted partly from the fact that the younger wives with the higher expectations in 1955 had moved into the next higher age group by 1960 and partly from the tendency for the wives' expecTABLE 28 Number of Births by Interview and Additional and Total Number Expected for White Wives, by Age in 1955 and 1960

Wife's age Total 18-24 25-29 30-34 35-39

Births by interview 1955 1960 2.1 1.1 1.9 2.3 2.6

2.3 1.3 2.3 2.7 2.7

Most likely expected additional births 1955 1960

Most likely expected total births 1955 1960

1.0 2.0 1.2 0.6 0.3

3.0 3.2 3.1 3.0 2.9

0.8 1.7 1.0 0.6 0.3

3.1 3.0 3.4 3.2 3.0

tations to increase somewhat as they grew older between 1955 and 1960. On balance, the most likely expected total number of births for all wives 18-39 years of age increased slightly from 3.0 to 3.1 (this increase is slightly above the minimum required for statistical significance). The average number of births that had occurred by the time of the interview was higher in 1960 than in 1955 for all four age groups, reflecting the trend toward childbearing at earlier ages (Table 28). Since this increase was as large as, or larger than, the increase in total expectations for each age group, there was no increase in the expected number of additional births. In fact, the average number of additional births expected declined significantly for wives under 30. As previously mentioned, this may help to explain the lower birth rate in recent years. Wives under 30 account for almost threefourths of the additional births expected by all wives in our sample. " A s emphasized in cohort comparisons, the wives who were 20-24 in 1955 were 25—29 in 1960.

FERTILITY AND FAMILY PLANNING IN THE U.S.

A comparison of the 1955 and 1960 data on expected births by wife's age at marriage, duration of marriage, and number of births by interview reveal no significant differences that do not appear to be primarily related to wife's age. IDEAL AND DESIRED FAMILY SIZE vs. EXPECTED FAMILY SIZE, BY WIFE'S AGE AND OTHER DEMOGRAPHIC VARIABLES

Our principal measures of attitudes toward family size—the num­ ber of children the wives considered ideal for the average American family, the number they said they themselves would have if they could live their lives over again under ideal conditions, and the number they actually wanted to have at the time of the interview—all tend to be higher for older wives than for younger wives (Table 29). Averages of the first two of these measures—the wives' general and personal ideals—are higher than the average number of children the wives actually want which, in turn, is generally higher than the average number they expect. The difference between the average numbers of children wanted and expected is greatest for the oldest wives (3.4 vs. 2.9). As shown in Table 30, 40 per cent of the 35-39-year-old wives expected fewer children than they wanted and 23 per cent expected more. Older wives may have exaggerated the number they wanted to a somewhat greater extent than younger wives, but the higher proportion expecting fewer than wanted in the 35-39 group is due mainly to the greater prevalence of subfecundity among older couples (see Chapter 4). Wives who were 35-39 and reported no fecundity impairments expected more children than they wanted as a group, and there is no indication that Subfecund wives were any more likely to overstate their family-size preferences than those who were classified as Fecund. Both groups wanted virtually the same number of children. On the other hand, a smaller percentage of wives expected more children than they wanted in the 35-39 age group than in the 30-34 group, perhaps because older wives are less willing to admit that any of their children are unwanted. The proportion of wives expecting the number of children they want declines from 64 per cent at ages 18-24 to 38 per cent at ages 35-39. Of course, the younger wives are probably too optimistic about their ability to achieve their childbearing goals. As they grow older, many will find their plans upset by fecundity impairments and accidental conceptions. Duration of marriage is very closely associated with wife's age, and as would be expected, family-size ideals, desires, and expectations are related to these two variables in virtually the same ways (Tables 29 and 30).

IDEAL, DESIRED, AND EXPECTED FAMILY SIZE

There is little, if any, relation between the age at which wives marry and the number of children they consider ideal or want to have themselves. As pointed out earlier, however, there is an inverse relationship between wife's age at marriage and the number of chilTABLE 29 Number of Children Considered Ideal for the Average American Family, Wanted if Life Could be Relived, Wanted at Time of Interview, and Most Likely Expected, for White Wives, by Age, Years Since First Marriage, Age at First Marriage, and Number of Births by 1960 Average number of children

Considered Wanted ideal for averif life age American could be family·1 relived®

Wanted at time of interview

Most likely expectedb

3.7

3.3

3.1

3.2 3.5 3.4 3.6

3.4 3.7 3.7 3.8

3.1 3.3 3.3 3.4

3.0 3.3 3.2 2 .9

544 649 702 519

3.2 3.4 3.5 3.6

3.6 3.6 3.7 3.8

3.2 3.2 3.3 3.4

2.9 3.2 3.2 3.0

504 717 586 407 200

3.4 3.5 3.4 3.4 3.4

3.6 3.7 3.7 3.7 3.8

3.3 3.3 3.3 3.3 3.3

3.3 3.2 3.0 2.9 2.6

301 463 682 499 263 206

3.2 3.2 3.1 3.5 3.9 4.5

3.4 3.3 3.2 3.7 4.3 5.6

3.0 2.9 2.8 3.3 3.9 5.6

1.6 2.3 2.6 3.4 4.4 6.2

Demographic characteristic

Number of wives

Total Wife's age: 18-24 25-29 30-34 35-39 Years since wife's first marriage: Under 5 5-9 10-14 15 and over Wife's age at first marriage: Under 18 18-19 20-21 22-24 25-39 Number of births by 1960: None 1 2 3 4 5 or more

2,414

3.4

513 600 624 677

a

A small number of wives did not give numerical replies to these questions. The expected number of children is slightly smaller than the expected number of births for most groups, because it excludes children who died prior to the interview. b

dren expected. Consequently, wives who marry at later ages are more likely to expect fewer children than they want and less likely to expect more than they want (Tables 29 and 30). Among wives with two or more births by 1960, we find a direct relation between the number of children they had already borne and the numbers they considered ideal and actually wanted to have (Table

FERTILITY AND FAMILY PLANNING IN THE U.S.

29). Wives with no births or only one tended to report somewhat higher family size ideals and desires than those with two births, but fecundity impairments had prevented many of them from having as many births as they would have liked. The average number of children expected is smaller than the average number wanted for TABLE 30 Percentage Who Expected Fewer Children than They Wanted, the Same Number or More, for White Wives, by Age, Years Since First Marriage, Age at First Marriage, and Number of Births by 1960

Demographic characteristic

Number of wives

Total 2,414 Wife's age: 18-24 513 25-29 600 624 30-34 35-39 677 Years since wife's first marriage: Under 5 544 5-9 649 10-14 702 15 and over 519 Wife's age at first marriage: Under 18 504 18-19 717 20-21 586 22-24 407 25-39 200 Number of births by 1960: None 301 463 1 2 682 499 3 4 263 5 or more 206

Total

Ex­ pected Ex­ fewer pected than number wanted" Wantedb

Expected more than wanted" Had Had number more than wanted Total or fewer wanted

100

29

50

22

7

15

100 100 100 100

22 22 29 40

64 56 44 38

14 22 27 23

7 9 7 5

7 14 20 17

100 100 100 100

23 24 30 38

68 53 41 37

8 24 29 25

6 9 8 5

2 15 21 20

100 100 100 100 100

26 25 29 32 37

45 50 51 51 52

29 25 20 17 11

8 9 7 6 3

21 16 13 11 8

100 100 100 100 100 100

52 38 25 24 12 15

45 57 56 44 43 37

3 4 19 31 46 48

3 4 10 7 9 9

— —

9 24 37 38

* The most likely expected number is smaller than the minimum number wanted. b The most likely expected number is equal to (or between) the minimum or maxi­ mum number wanted. 0 The most likely expected number is larger than the maximum number wanted.

wives with less than three births and larger for those with three or more. Table 30 shows that the proportion of wives expecting fewer births than they wanted generally declines, and the proportion expect­ ing more than they wanted increases, as the number of births by 1960 rises. Fifty-two per cent of the wives with no births and 38

IDEAL, DESIRED, AND EXPECTED FAMILY SIZE

per cent of those with only one birth expected fewer children than they wanted. On the other hand, almost half of the wives with four or more births expected more children than they wanted. In fact, many of these wives already had more than they wanted. Tables 31 and 32 show percentage distributions of wives in the different age groups according to the first reason they gave for not wanting a smaller family and for not expecting a larger one. There are no large differences in the reasons given for not wanting a smaller TABLE 31 Per Cent Distribution by First Reason for Not Wanting a Smaller Family, for White Wives, by Age

First Reason Number of Wives Per cent: Total Reasons related to welfare or happiness of children To provide companionship for children; better for children To avoid having an only child Reasons related to satisfaction or interests of parents Likes, loves children; this many makes for a happy, "full" life To avoid loneliness; to have children around the house now or when older To have desired proportion of boys and girls Religious and moral reasons (e.g., couples ought to have this many) Ambiguous reasons This number is right, nice, ideal; like parental family Don't know Other reasons Not ascertained

18-24

Wife'sι Age 25-29 30-34

35-39

513

600

624

677

100

100

100

100

55

53

50

50

39 16 18

39 14 19

34 16 23

37 13 20

12

14

17

14

3 3

2 3

3 3

4 2

3 13

3 13

1 14

3 11

11 2 3 9

11 2 6 7

11 3 5 8

9 2 7 9

family that are associated with wife's age, but a somewhat larger proportion of the wives under 30 cited reasons related to the welfare or happiness of their children. This is probably because more younger wives had only one or two children by 1960 and were therefore concerned with the problem of providing them with adequate companionship. The proportion of wives reporting subfecundity as the first reason for not expecting a larger family rises from only 8 per cent at ages 18-24 to 28 per cent at ages 35-39 (Table 32). Among the remaining wives, we find that the younger ones generally had smaller family incomes, and were much more likely to cite economic reasons for

FERTILITY AND FAMILY PLANNING IN THE U.S.

not expecting a larger family, and less likely to cite health reasons or other reasons that are not classified in Table 32. In addition, a somewhat larger proportion of the younger wives indicated that they would find it difficult to "care" or "have time" for a larger family, perhaps because they were more likely to have infants and small children at the time of the interview. TABLE 32 Per Cent Distribution by First Reason for Not Expecting a Larger Family, for White Wives, by Age

First reason Number of wives Per cent: Total Subfecundity All other reasons Economic reasons Would cost too much; couldn't afford more, etc. Couldn't support more at desired standard of living Couldn't educate more Low income, insecure job, parttime work, unemployed Other economic reasons Poor health (parents and/or children); preg. unpleasant, or dangerous to wife or baby Hard to "care" or "have time" for more This is a "nice" no., no. wife and/or husband wants; no more needed for happiness, etc. Other reasons, didn't know Not ascertained

Wife'sι age 25--29 30--34

18--24 513

600

100 100 64

32

677

624

100

8 92 59

35--39

100

100

11 89 100 53 59

19 81 100 42 52

28 72 31

100 42

35

25

28

20

24

14

19

10 5

11 5

7 8

8 9

6 6

8 7

4 4

6 6

5 7

5 8

4 8

4 9

3 7

4 9

3 6

4 8

5

5

8

9

8

11

11

16

9

10

9

10

7

9

4

6

6 7 6

7 8 7

5 11 4

5 13 4

5 15 3

6 19 4

4 18 4

6 25 5

Differences in the reasons given for not wanting a smaller family and for not expecting a larger family that are associated with duration of marriage are practically the same as those associated with wife's age. Wives who married at younger ages tended to have more births by 1960 and to report lower family incomes than wives who married at older ages. As a result, they were less likely to cite reasons for not wanting a smaller family that are related to the adequacy of their children's companionship, and more likely to cite economic rea­ sons for not expecting a larger family.

CHAPTER 3

Family Size by Religion and Socioeconomic Status Differences in the fertility of various religious, social, and economic groups arise mainly from two sources: differences in the number of children couples want and differences in their willingness and ability to control fertility. Other causes of variation, such as differences in fecundity and age at marriage, are of much less importance. Our results indicate that some groups are similar with respect to the average number of children wanted, but differ significantly in their ability to prevent unwanted pregnancies. This appears to be true in the case of groups with varying amounts of education, for example. On the other hand, there are some groups that differ primarily with respect to the number of children wanted. This is true of the two major religious categories: Protestants and Catholics. In an effort to show not only how widely groups differ in fertility, but also whether the differences are due to family-size preferences or other factors, we shall present data in this chapter on the number of children wanted and whether or not the wife has had, or expects to have, fewer or more children than she wants, as well as data on past and expected childbearing. In considering some of these differ­ entials, the reader may find it helpful to refer to discussions in Chap­ ters 5 and 6 of group differences in the use of contraception and in the ability to prevent unwanted conceptions. The characteristics that are considered in the present chapter are religious preference, degree of religiousness, educational attainment, income, wife's work experience, husband's occupation, and place of residence. This chapter will deal only with the white wives who were 18-39 years of age at the time of the interview in 1960. Group differences among the nonwhites will be discussed in Chapter 9. RELIGION

Differences between the attitudes and values of the two major religious groups—the Protestants and the Roman Catholics—are very important considerations in studies of American fertility. The Catholic Church teaches that the primary purpose of marriage is the procrea­ tion and education of children and that all other aims must be re­ garded as secondary. It does not urge Catholic couples to have the maximum possible number of children, however, nor does it say how

FERTILITY AND FAMILY PLANNING IN THE U.S.

many they should have. This is a decision that is to be made by the husbands and wives themselves after giving careful thought to the conditions affecting their own situation. For serious, unselfish rea­ sons such as medical, eugenic, economic, and social considerations, it is permissible for them to avoid childbearing "for a long time, perhaps even the whole duration of the marriage." Indeed, in recent years, many Catholic authorities have increasingly stressed that the responsibilities of parenthood require couples to have no more children than they can properly educate and adequately provide with material and spiritual needs. Nevertheless, should they decide to regulate the number and spacing of births, the only methods that are morally acceptable in the eyes of the Church are periodic continence (rhythm) and prolonged abstinence form sexual intercourse.1 WIFE'S RELIGION PROTESTANT

V/////A

CATHOLIC

JEWISH

1

1

1

I

2

3

AVERAGE NUMBER OF BIRTHS BIRTHS BY I960

TOTAL NUMBER EXPECTED (MOST LIKELY)

FIGURE 8 Births by 1960 and total number expected, for wives, by religion

Most Protestant denominations strongly support the principle of responsible parenthood, but differ from the Catholic Church by plac­ ing less emphasis upon the importance of reproduction as a purpose of marriage and by approving the use of all methods of contraception that are "mutually acceptable, non-injurious to health, and appro­ priate to the degree of effectiveness required in the specific situation."2 Religious Differentials in Fertility—These differences in attitudes are apparently reflected in actual reproductive behavior. Figure 8 and Table 33 show that Catholic wives had already bome more 1 See FPSPG, Appendix A, Note 10, for a more comprehensive statement of the attitude of the Roman Catholic Church regarding the control of family size. For a more recent discussion of the Catholic position on this subject, see, for example, O'Brien 1963 and Commonweal 1964. See page xxxi for the meaning of any abbreviated reference. 2NCCC 1961.

RELIGION AND SOCIOECONOMIC STATUS

children on the average than Protestant wives by the time of the interview in 1960. Furthermore, there is a substantially larger differ­ ence between the numbers of children they thought they would have when their families were completed. The average most likely expected total number of births is 3.7 for Catholics and 2.9 for Protestants. The range between the minimum and maximum numbers of expected TABLE 33 Births by 1960 and Total Number Expected, for Wives, by Religion of Wife, Husband, and Couple, and by Who Performed the Marriage

Religion Total Wife's religion Protestant Catholic Jewish Other Husband's religion Protestant Catholic Jewish Other Couple's religion Both Protestant Married by minister Othera Both Catholic Married by priest Other" Wife Protestant, husband Catholic Married by priest Other® Wife Catholic, husband Protestant Married by priest Other" Both Jewish All others a

Number of wives

Average number of births Total expected by wives Most By 1960 likely Minimum Maximum

2,414

2 .3

3 .1

2 .9

3.5

1,596 668 106 44

2 .2 2 .5 2 .1 2 .2

2 .9 3 .7 2 .5 3 .0

2 .7 3 .4 2 .4 2 .7

3.2 4.2 2.6 3.3

1,581 640 111 82

2 .2 2 .6 2 .0 2 .5

2 .9 3 .8 2 .5 3 .3

2 .7 3 .4 2 .4 3 .1

3.2 4.2 2.7 3.5

1,454 1,178 276 525 462 63

2 .2 2 .1 2 .5 2 .6 2 .6 2 .6

2 .9 2 .9 3 .2 3 .9 3 .9 3 .7

2 .7 2 .6 2 .9 3 .5 3 .5 3 .4

3.2 3.2 3.5 4.4 4.4 4.1

106 40 66

2..4 2. 4 2,.3

3..2 3.,2 3.,1

2 .9 3..0 2. 9

3.5 3.4 3.6

114 74 40 104 111

2..2 2. 3 2. 1 2. 1 2. 4

3. 2 3. 3 2. 9 2. 5 3. 2

2. 8 2. 9 2..7 2. 4 3. 0

3.6 3.8 3.4 2.6 3.5

Includes common law marriages.

births is considerably wider in the case of Catholic wives, suggesting that they are less certain about the number of children they will try to have, or about their ability to regulate family size. Nevertheless, there is no doubt that Catholics anticipate larger families. Even their minimum expectations are greater than the maximum expectations of Protestant respondents.

FERTILITY AND FAMILY PLANNING IN THE U.S.

The proportion of Catholic wives expecting five or more births (usually five to eight) is considerably more than twice as large as that of Protestant wives (see Table 34). However, about three-fourths of the Protestants and three-fifths of the Catholics thought that they would have a total of from two to four. Within this popular range, we find that Protestants were more likely to expect two or three and Catholics were more likely to expect four. Jewish wives as a group expected fewer births than either the Catholics or the Protestants, as shown in Table 33. The very small residual category of "other" wives in this table consists chiefly of respondents of the Eastern Orthodox faith and those who stated that they had no religious preference. TABLE 34 Per Cent Distribution by Total Number of Births Expected, for Wives, by Religion

Type of reply

Number of wives

Total

None

Most likely Minimum Maximum

2,414 2,414 2,414

100 100 100

4 6 3

Most likely Minimum Maximum

1,596 1,596 1,596

100 100 100

5 7 3

Most likely Minimum Maximum

668 668 668

100 100 100

4 5 3

a

Number of births expected 12 Totala 25 28 22 Protestant 7 27 10 30 4 24 Catholic 6 16 9 19 3 14 7 10 4

3

4

5

6 or more

29 26 27

20 18 26

8 7 8

7 6 11

31 28 29

19 16 26

6 6 7

5 4 7

23 21 19

23 25 27

13 10 13

15 11 21

Includes 150 Jewish and "other" wives.

The differences between the average numbers of births expected by Protestant, Catholic, and Jewish wives are all statistically significant at the .95 confidence level. It should also be emphasized that the higher expectations of Catholics cannot be attributed to differences in socioeconomic characteristics. Within every major social and eco­ nomic group, we find Catholic wives expecting a substantially larger number of births than Protestant and Jewish wives. The ProtestantJewish differential, on the other hand, is considerably smaller, and the lower expectations of the Jews appear in large measure to reflect their greater concentration in large urban areas and their higher edu­ cational, occupational, and income status. Comparisons of the average numbers of births expected by Protestants and Jews with similar socio-

RELIGION AND SOCIOECONOMIC STATUS

economic characteristics reveal no differences that are large enough to be statistically significant.3 Table 33 also presents the average numbers of births by 1960 and the average total numbers of births expected for wives classified on the basis of the religion of the husband and the couple. Grouping the wives according to their husband's religious preference rather than their own has very little effect upon the Protestant, Catholic, and Jewish averages, since the vast majority of respondents were married to husbands of the same major religious faith. Only 9 per cent of the wives were in mixed marriages in which one spouse was Protestant and the other Catholic. Such wives thought they would have more births on the average than Protestant wives married to Protestant husbands, but their expectations were consider­ ably closer to those of wives in exclusively Protestant marriages than to those of wives in exclusively Catholic marriages (Table 33). Whether the wife or the husband was Cathohc is apparently of little importance. Protestant wives with Catholic husbands expected vir­ tually the same number of births as Catholic wives with Protestant husbands.4 Of more importance is whether or not the marriage was performed under the auspices of the Catholic Church. The wives who were married by a priest anticipated more births than other wives in mixed marriages, regardless of which spouse was Catholic.5 However, even these wives thought they would have smaller families than Cathohc wives married to Cathohc husbands. The relatively small number of Catholic wives with Catholic hus­ bands who were not married within the Church expected somewhat fewer births than the large majority who were married by a priest (Table 33). In the case of marriages in which both spouses were Protestant, there was a different relationship: the wives' expectations were generally higher when the marriage was not performed by a Protestant minister. This can apparently be attributed more to social and economic differences than to differences of a strictly religious nature. Couples not married by a minister included a larger proportion 3 These findings are consistent with those of a special analysis of the 1955 GAF data in which the family-size expectations for 66 Jewish couples were com­ pared with those for 66 Protestant and 66 Catholic couples selected to match the Jewish couples as closely as possible on the basis of duration of marriage and five socioeconomic characteristics. See Freedman 1961. 4 I n the 1955 study, it was found that the expectations of wives in mixed Protestant-Catholic marriages tended to be somewhat higher in cases where the wife rather than the husband was the Catholic partner, but the difference was not statistically significant. See FPSPG, pp. 284-285. sA similar relationship was found with respect to the total number of children desired in the first phase of the Princeton Study and the number of pregnancies actually occurring between the first and second phase-s of this study. See Westoff 1961, pp. 180-181 and Westoff 1963, pp. 91-92.

FERTILITY AND FAMILY PLANNING IN THE U.S.

of wives of lower socioeconomic status, and as we shall see later, they tended to expect larger families than Protestants of higher status. It is not feasible to compare our data for the various Protestant denominations because of the limited number of cases and the small differences in the fertility of most Protestant subgroups. However, Table 35 shows the average number of births by 1960 and the average total number of expected births for wives stating a preference for three types of Protestant denominations: those classified as liberal, intermediate, and fundamentalist in nature.6 The figures are quite similar for the liberal denominations and the very large intermediate group, but are significantly higher in the case of the fundamentalists. This cannot be explained by differences in educational attainment, income, occupation, or size of place of residence. Even when these socioeconomic factors are held constant, we find that the fundamenTABLE 35 Births by 1960 and Total Number Expected, for Protestant Wives, by Type of Denomination

Type of denomination

Number of wives

By 1960

Most likely

Minimum

Maximum

All Protestants4 Liberal Intermediate Fundamentalist

1,596 274 1,077 199

2.2 2.1 2.2 2.5

2.9 2.8 2.9 3.3

2.7 2.6 2.7 3.0

3.2 3.1 3.2 3.6

a

Average number of births Expected total

Includes 46 wives whose denomination was not specified.

talist wives had bome the largest number of children by 1960 and expected to have the largest completed families. Catholic wives already had more births than Protestant wives by the time of the interview in all age groups with the exception of the youngest (see Figure 9 and Table 36). In the 18-24 group, the average number of births by 1960 is the same for Catholics as for Protestants (1.3), but the Catholic wives had been about seven months older on the average when they were married and had had less time in which to bear children. Table 36 shows that even among the 18-24-year-olds Catholics generally had more births than Protes­ tants when they had married at approximately the same ages. Catholic wives expected significantly larger completed families in all four age groups. However, Catholic as well as Protestant expectations conform to the general cohort pattern described in the preceding chapter. 'For a list of the denominations included in each of these three categories, see Appendix B, Note 4.

RELIGION AND SOCIOECONOMIC STATUS

The smallest families were anticipated by the youngest (18-24) and oldest (35-39) wives in both religious groups. Data for the Jewish wives are not shown separately in detailed comparisons like this be­ cause of the small number of such respondents in our sample. The higher fertility of Catholics is also reflected in Table 37 in which the wives are classified according to the number of years WIFE'S AGE ANO RELIGION 18-24·.

W/////M

PROTESTANT

'//////////////7Ά

CATHOLIC

25-29: PROTESTANT

V//////////A

CATHOLIC

30-34: PROTESTANT

CATHOLIC

35-39 PROTESTANT

CATHOLIC

-I

1

1

1 2 AVERAGE NUMBER B I R T H S BY I 9 6 0

r

3 4 OF BIRTHS

TOTAL NUMBER EXPECTED (MOST LlKELV)

FIGURE 9 Births by 1960 and total number expected, for wives, by religion, by age.

since their first marriage. In each duration-of-marriage group, the Catholic wives not only had the largest number of births by 1960, but also thought they would have the largest number of additional births after the interview. Until recent years, it was often assumed that the differences be­ tween the fertility of Catholics and non-Catholics would gradually disappear as the Catholics became more similar to the remainder

FERTILITY AND FAMILY PLANNING IN T H E U.S. TABLE 36 Births by 1960 and Total Number Expected, for Wives, by Religion, by Age in 1960, by Age at First Marriage Wife's age in 1960 and religion

Total

Wife's age at first marriage Under 18 18--19 20 -21 22-24

25-39

Births by 1960 Total All religions" Protestant Catholic 18-24 All religions" Protestant Catholic 25-29 All religions" Protestant Catholic 30-34 All religions" Protestant Catholic 35-39 All religions" Protestant Catholic Total All religions" Protestant Catholic 18-24 All religions" Protestant Catholic 25-29 All religions" Protestant Catholic 30-34 All religions" Protestant Catholic 35-39 All religions" Protestant Catholic

2.3 2.2 2.5

2.7 2.6 3.0

2 .4 2 .4 2 .7

2 .2 2 .1 2 .5

2.1 1.9 2.3

1.3 1.3 1.3

1.9 1.8 1.9

1 .3 1 .2 1 .4

0 .7 0 .7 0 .8

0.5 0.5

2.3 2.3 2.5

3.0 2.8

2 .3 2 .0 2..8

1.6 1.7 l.o

0.7 0.7

*

2 .7 2 .5 3 .0

2.7 2.6 3.0

3.1 3.0 3.7

3 .0 3 .1 3 .2

2 .7 2 .5 3 .1

2.3 2.1 2.9

1.4 1.2

2.7 2.7 3.0

3.1 3..2 2. 6 2.5 3..1 2. 6 2.3 3.1 3. 7 3.2 2. 7 2.8 Most likely expected total births

2.1 1.9 2.7

3.1 2.9 3.7

3.4 3.2 4.1

3 .3 3 .1 3 .8

3 .1 2 .8 3..7

3.0 2.6 3.6

3.0 2.8 3.6

3.2 3.0 3.9

2. 9 2. 8 3. 4

2. 8 2. 6 3. 2

3.2 2.8

3.4 3.1 4.1

3.5 3.3

3. 3 2. 8 4. 3

3.2 3.0 3.4

3.3 2.6

*

3. 4 3. 2 4. 2

3.2 3.0 3.8

3.4 3.2 4.3

3..5 3..5 3. 7

3..2 3. 0 3. 7

3.1 2.5 4.3

2 5 2.3

3.0 2.8 3.4

3.3 3.2 3.8

3..4 3. 2 4. 0

2. 9 2. 8 3. 3

2.7 2.5 3.2

2.5 2.1 3.4

" Includes Jewish and "other" wives.

1.7 1.4 2.1

#

*

*

2.7 2.3 3.5

*

*

*

RELIGION AND SOCIOECONOMIC STATUS

of the population in social and economic status. These differences have persisted, however, and they now appear to be widening. More­ over, there is no indication that they are reduced by the upward socioeconomic mobility of Catholics. In fact, the largest CatholicProtestant differentials in family-size expectations are usually found in the higher social and economic groups, as we shall show in later sections of this chapter. Trends between 1955 and 1960—In the 1955 study, Catholic wives expected substantially larger completed families than Protestant TABLE 37 Births by 1960 and Additional and Total Number Expected, for Wives, by Religion, by Years Since First Marriage Years since wife's first marriage

Total Under 5 5-9 10-14 15 or more Total Under 5 5-9 10-14 15 or more Total Under 5 5-9 10-14 15 or more

Total

Wife's religion Protestant Catholic

Other

Births by 1960 2.2 2 .1 2.5 1.0 1.1 0 .8 2.2 2 .1 2.7 2.7 3.2 2 .4 2 .4 2.9 3.1 Most likely expected additional births 0 .8 0.7 1.2 0,.5 1..9 1.7 1. 8 2.4 0..9 0.7 1.3 0,.7 0..3 0..4 0.4 0.6 0..2 0.2 0,.1 0.4 Most likely expected total births 2.9 2..6 3. 1 3.7 2. 9 2.7 3.5 2. 5 3. 2 2.9 2. 8 3.9 3. 3 3.1 3.9 2..7 3. 2 3.1 3.6 2. 5 2 .3 1 .0 2 .3 2 .8 2 .9

wives, but had the same average number of births (2.1) by the time of the interview (see Table 38). This apparent discrepancy can be explained in part by the fact that the Catholics had not been married as long as the Protestants, as a group, because of their tendency to marry at later ages. It was found that Catholic wives had already borne a somewhat larger number of children than Protes­ tant wives who had been married for the same length of time.7 The results of the Current Population Survey of the Bureau of the Census in March, 1957, were quite similar; the number of children 7 FPSPG,

pp. 275-277.

FERTILITY AND FAMILY PLANNING IN THE U.S.

ever born per 1,000 ever-married women, 15 to 44 years old, was only slightly higher for Catholics than for Protestants.8 However, a comparison of the data from the 1955 and 1960 GAF studies in Table 38 indicates that there was an increase during this five-year period in the average number of births by the time of the interview and that this increase was much greater for Catholics than for Protestants. As a result, Catholic wives had already had significantly more births than Protestant wives when the 1960 survey was conducted (2.5 as compared with 2.2) in spite of the somewhat shorter average duration of their marriages.9 The number of births by 1960 was higher for Catholics in all age groups with the exception TABLE 38 Births by Interview and Additional and Total Number Expected, for Wives, by Religion, by Age, in 1955 and 1960

Wife's religion and age Totala AU Protestants 18-24 25-29 30-34 35-39 All Catholics 18-24 25-29 30-34 35-39

Number of wives 1955 1960 2,713 1,817 394 461 505 457 787 146 220 212 209

2,414 1,596 370 396 405 425 668 126 176 174 192

Births by Most likely expected interview Additional births Total births 1955 1960 1955 1960 1955 1960 2 .1 2 .1 1 .1 2 .0 2..4 2 .6 2 .1 1 .2 1 .9 2 .4 2 .6

2 .3 2 .2 1 .3 2 .3 2 .6 2 .7 2 .5 1 .3 2 .5 3 .0 3 .0

1 .0 0 .8 1 .8 1 .0 0..5 0..2 1 .3 2 .6 1 .5 0 .9 0 .5

0. 8 0,.7 1..5 0 .8 0..5 0..2 1 .2 2,,2 1 .6 0 .8 0 .5

3 .0 2 .9 2 .9 3 .0 2 .9 2..8 3..4 3 .8 3 .4 3 .3 3 .1

3..1 2..9 2..8 3 .1 3..0 2..8 3 .7 3..6 4 .1 3 .8 3 .4

"Includes Jewish and "other" wives. None of the 1955-1960 changes for these small groups is large enough to be statistically significant at the .95 confidence level.

of the youngest (18-24), and as previously mentioned Catholics had borne more children than Protestants in this age group when they had married at the same ages. Catholic wives who were 30-39 in 1960 already had as many children when they were interviewed as Protestant wives of the same age thought they would have altogether when their families were completed. Table 38 also shows that the average total number of births ex­ pected by the 18-39-year-old Catholic group was significantly larger in 1960 than in 1955 (3.7 vs. 3.4). Expectations of Protestants were 8Census 1962, p. 58. * A substantial widening of the difference between the average numbers of children ever born to white Protestants and Catholics is also recorded in the results of the Detroit Area Study surveys conducted in that metropolitan area in 1952 and each year from 1954 to 1959. See Lenski 1961, pp. 216-217.

RELIGION AND SOCIOECONOMIC STATUS

the same in both years, however, and the difference between the averages for the Catholics and Protestants rose from 0.5 in 1955 to 0.8 in 1960. The larger expected size of completed family for all 18-39-year-old wives in the 1960 study can thus be attributed almost entirely to the Catholics.10 Two factors are involved in the widening of the Catholic-Protestant differential. First, the wives who were 18-34 in 1955 and thus 23-39 in 1960—and eligible for both studies—tended to revise their expectations upward as they grew older during 1955-1960, and this tendency may have been somewhat more pronounced for Catholics than for Protestants (see Chapter 1). Second, the Catholics who were included in the 1960 sample of 18-39-year-old wives, but were not eligible in 1955, expected to have larger families than those who were included in 1955, but were not eligible in 1960. In other words, the Catholics who entered the sample between the 1955 and 1960 studies thought they would have more children than those who left it. The Protestants who moved into the sample, on the other hand, expected fewer children than those who moved out of it. Both of these factors thus contributed to the increase between 1955 and 1960 in the average number of births expected by the 18-39-year-old Catholic group. In the case of the Protestant group, however, the number expected remained unchanged, because the in­ crease resulting from the tendency for wives to revise their expectations upward as they grew older during 1955-1960 was offset by the fact that the Protestants who entered the sample between the two studies expected fewer births than those who left it. Let us consider the groups entering and leaving the sample in more detail. The group entering the sample between 1955 and 1960 consisted of: 1. Wives who were 18-22 and hence too young to have been in the 1955 study, and 2. Wives who were 23 and over in 1960 but were married for the first time after 1955 (or were married earlier but were not living with their husbands when the 1955 survey was conducted). The vast majority (93 per cent) of these wives were 23-29 in 1960. The group leaving the sample included: 1. Wives who were 35-39 in 1955 and hence too old for the 18-39 group in 1960, and 2. A small number of wives who were 18-34 in 1955, but were widowed, divorced, or separated in 1960. 10 The small group of Jewish and "other" wives also expected more births in 1960 than in 1955, but this increase was not large enough to be statistically significant. The average numbers of births expected by all non-Catholic wives combined were the same in both studies when rounded to one decimal place.

FERTILITY AND FAMILY PLANNING IN THE U.S.

A comparison of respondents who moved into the sample with those who moved out of the sample thus amounts almost entirely to a comparison of wives who were 18—29 and married less than five years in 1960 with all wives who were 35-39 in 1955. The average numbers of expected births for Protestants and Catholics in these two categories are as follows: Expected in 1960 by wives who were 18-29 and married less than five years Expected in 1955 by all 35-39-year-old wives

Protestants

Catholics

2.6

3.6

2.8

3.1

The fact that Catholic wives who entered the sample expected more births than those who left it—while the reverse was true for Protes­ tants—largely reflects the much higher expectations of younger than older wives among the Catholics(see Table 38). The widening of the Catholic-Protestant fertility differential was forecast by the expectations of the 18-24-year-old wives in 1955, but when the results of the 1955 study were reported, it was believed that the Catholic wives in this age group were probably being un­ realistic in expecting such large families.11 There were several reasons for this skepticism. It was found that Catholics who were 18—24 had not borne a substantially larger number of children by the time of the interview than Protestants of this age, even when they had been married for approximately the same length of time. Furthermore, the difference between the expectations of Catholics and Protestants was much smaller in all of the older age groups. The older Catholic wives expected considerably smaller families than the younger ones, and indicated that they wanted significantly fewer children when they were interviewed than they recalled having wanted just before they were first married. It was, therefore, believed that the younger Catholic wives would also be likely to revise their expectations down­ ward as they grew older and had more experience with marriage and the bearing and rearing of children, and that although they would probably have more children than the younger Protestant wives, the difference between the sizes of their completed families would not be as large as their expectations would indicate. As we have seen in Chapter 1, the results of the 1960 study indicate that this skepticism concerning the higher expectations of the younger Catholics in 1955 was not justified. They had substantially more births than the younger Protestants between 1955 and 1960 and did not revise their expectations downward during this period. 11FPSPG,

pp. 275-280.

RELIGION AND SOCIOECONOMIC STATUS

On the contrary, they revised them upward, and perhaps to a some­ what greater extent than Protestants. Both Protestants and Catholics expected a smaller average number of additional births to occur after the interview in 1960 than in 1955 (Table 38). This helps to explain the decline in the birth rate since 1960 and can be attributed primarily to the trend toward childbearing at younger ages that is discussed in Chapter 8.12 The 18-39-year-old Protestants already had 76 per cent of the total number of births they expected in 1960 in contrast with 71 per cent in 1955. For Catholics, the comparable proportions are 69 per cent in 1960 and 62 per cent in 1955. Table 38 shows that the decline in the number of additional births expected is most pronounced for the 18-24 age group where there was not only an increase in the number of births by the time of the interview, but also a decrease in the total number expected. This is particularly significant from the stand­ point of current fertility trends, for the wives who were 18—24 in 1960 will account for an important proportion of the childbearing during most of the 1960's. The substantial religious differences in the number of births by the time of interview and in the expected size of completed family in the 1960 GAF study are confirmed by the following results of the 1962 surveys of the Population Studies Center of the University of Michigan: Mean number o f b i r t h s b y i n t e r v i e w Mean total number expected

Catholics

Non-Catholics

2.8 3.9

2.2 2.9

In fact, these differentials are larger than those of the 1960 study, and suggest that the diverging trend noted during 1955-1960 may have continued between 1960 and 1962. We cannot be certain that such a change actually occurred during this very short period, how­ ever, for two reasons. First, comparisons of data for the religious groups from the 1960 and 1962 surveys are subject to a considerably larger margin of sampling error than the 1955-1960 comparisons, because the 1962 sample is only about one-half as large as the GAF samples. Second, there are differences in the procedures of the 1960 and 1962 studies, and although they are believed to have little effect upon most of the results, the 1960 and 1962 data are not exactly comparable.13 Religiousness—We asked more questions about religion in 1960 12See

also Whelpton 1963 and Freedman 1963. the results of the 1962 survey and a discussion of the differences in the procedures used in this study and the GAF studies, see Freedman 1963. 13For

FERTILITY AND FAMILY PLANNING IN THE U.S.

than in 1955. Most of the new questions were designed to measure the strength of adherence to the values and doctrines of the Protestant and Catholic faiths as indicated by attendance at religious services, education in church-related schools, religious activities in the daily life of the family, and expressed interest in religious matters. If re­ ligious affiliation is significantly related to fertility, it would seem likely that the degree of devoutness or involvement would also be an important consideration. The only questions that shed any light on this in the 1955 study were those that asked how often the wives FREQUENCY WITH WHICH SACRAMENTS WERE RECEIVED NEVER

ONCEAYEAR OR LESS

V/////A '///////Λ

A FEW TIMES AYEAR ONCE A MONTH

'/////////λ

TWO OR THREETIMES A MONTH ONCE A WEEK OR MORE

V//////////A ,

1

1

2 3 4 AVERAGE NUMBER OF BIRTHS BIRTHS BY I960

1

5

TOTAL NUMBER EXPECTED (MOST LIKELY)

FIGURE 10 Births by 1960 and total number expected, for Catholic wives, by frequency with which they received the sacraments

and husbands went to church. Replies to these questions and similar ones in the 1960 study reveal no significant relation between the frequency of church attendance and the number of births expected by Protestants. For Catholics, however, there is a direct relationship, with the more frequent churchgoers clearly expecting larger families (especially in I960).14 M This finding is confirmed by the results of the 1962 surveys of the Population Studies Center of the University of Michigan. See Freedman 1963, p. 379. In the Princeton Study, a similar relationship is found with respect to the frequency of church attendance and the number of children desired. See Westoff 1961, pp. 195196.

RELIGION AND SOCIOECONOMIC STATUS

In the 1960 survey, we also asked how often Catholic wives and husbands received the sacraments of the church, since it was believed that this would probably provide a better indication of their religious­ ness than church attendance. In any event, our results indicate that fertility is more closely related to the frequency with which the sacra­ ments are received. The average most likely expected total number of births rises from 3.2 for Catholic wives who never received the sacraments to 4.4 for those who received them once a week or more, as shown in Figure 10 and Table 39. This is one of the larger differen­ tials in expectations to be found in our analyses, but it should be TABLE 39 Births by 1960 and Total Number Expected by Wife, for Catholic Wives and Husbands, by Frequency with Which They Received the Sacraments

Frequency with which sacraments were received

Number of wives

All Catholic wives Never Once a year or less A few times a year Once a month Two or three times a month Once a week or more All Catholic husbands Never Once a year or less A few times a year Once a month Two or three times a month Once a week or more

668 84 61 160 191 78 94 640 120 95 158 143 49 75

Average number of births Total expected by wives Most By 1960 likely Minimum Maximum 2..5 2..4 2. 4 2. 6 2. 5 2. 6 2..6 2. 6 2..5 3. 0 2..5 2 .6 2 .7 2 .5

3.7 3.2 3.3 3.7 3.7 4.1 4.4 3.8 3.2 3.8 3.8 3.7 4.1 4.4

3.4 3.0 3.0 3.4 3.3 3.7 3.8 3.4 2.9 3.4 3.5 3.4 3.8 3.9

4.2 3.5 3.6 4.0 4.2 4.7 5.2 4.2 3.6 4.1 4.1 4.4 4.6 5.0

noted that even the Catholic wives who never received the sacraments were expecting to have more children than the average Protestant wife. Table 39 also shows that family-size expectations follow the same general pattern when the respondents are grouped according to the frequency with which their Catholic husbands received the sacraments. Although the Catholics who received the sacraments more fre­ quently thought they would have substantially larger completed fami­ lies than other Catholics, they had not had a substantially larger number of births by 1960 (Figure 10 and Table 39). This apparent discrepancy can be explained mainly by the fact that they included a larger proportion of the better-educated from higher socioeconomic

FERTILITY AND FAMILY PLANNING IN THE U.S.

groups who had married at later ages and had not had as much time in which to have children. For example, the Catholic wives who received the sacraments at least once a week had been married 2.6 years less on the average than those who never received them: Frequency with which sacraments were received

Average number of years since first marriage

Never Once a year or less A few times a year Once a month Two or three times a month Once a week or more

11 2 9.5 10.0 10.1 8.9 8.6

CathoKcs with more education, higher incomes, and higher occu­ pational status tend to go to church and receive the sacraments more often than other Catholics, but socioeconomic factors such as these do not account for the differences in expected family size that are TABLE 40 Births by 1960 and Total Number Expected by Wife, for Catholic Wives and Husbands, by Amount of Elementary and Secondary Education in Catholic Schools Amount of elementary Number and secondary education in of Catholic schools wives All Catholic wives None Some All» All Catholic husbands None Some All»

668 373 191 104 640 355 194 91

Average number of births Total expected by ι wives

By 1960

Most likely

Minimum

Maximum

2.5 2.6 2.5 2.5 2.6 2.7 2.4 2.7

3.7 3.6 3.7 4.2 3.8 3.6 3.7 4.3

3.4 3.3 3.4 3.7 3.4 3.3 3.4 3.9

4.2 4.1 4.2 4.8 4.2 4.1 4.1 4.8

a Includes only those wives and husbands who went beyond the eighth grade. Those who did not go beyond the eighth grade are included in the two preceding categories.

associated with the frequency of church attendance and the frequency with which the sacraments are received. We find the Catholics who go to church and receive the sacraments more frequently expecting substantially larger families than those who do so less frequently within each major socioeconomic group differentiated in this study. The amount of education Catholics receive in Catholic schools is generally regarded as one of the more important measures of their exposure to the teachings of the Catholic Church during their forma­ tive years. Forty-four per cent of the Catholic wives and 45 per cent of the Catholic husbands in our sample had received at least

RELIGION AND SOCIOECONOMIC STATUS

a part of their education in church-related institutions. Table 40 compares the number of births by 1960 and the total number of births expected for Catholic wives who reported that none, some, or all of their elementary and secondary education had been obtained in Catholic schools. Higher education is excluded from the comparison, because only 14 per cent of the 668 Catholic wives had attended Catholic colleges or universities. There is a direct relationship between the amount of Catholic education and the expected size of completed family, with the average most likely expected total number of births rising from 3.6 for Catholic wives who had never gone to Catholic schools to 4.2 for those who had attended Catholic schools exclusively. This difference is well above the minimum required for statistical significance. However, the expec­ tations of the respondents who had received only a part of their elementary and secondary education in Catholic schools are not sig­ nificantly higher than those of Catholic wives whose education was entirely secular. This suggests that Catholic education below the high school level has very little influence on fertility. Most of the 191 Catholic wives who had received only a part of their education in Catholic schools had never attended Catholic schools beyond the eighth grade. Table 40 also shows that the differentials in family-size expectations are of a similar order of magnitude and follow the same pattern when the wives are classified on the basis of the amount of education their Catholic husbands had received in Catholic schools. Even though the Catholics who had received all of their ele­ mentary and secondary education in church-related schools expected to have substantially larger completed families than those with an entirely secular education, they had not had a larger number of births by the time of the interview (Table 40). Again, the discrepancy between the number of births by 1960 and the total number expected can be explained by differences in duration of marriage. The wives who had attended Catholic schools exclusively had been married for 2.6 years less on the average than other Catholic wives. As a group, they had not only married at a somewhat later age, but were also younger at the time of the interview. More than 19 per cent of the 18-29-year-old Catholic wives had received all of their elementary and secondary education in Catholic schools as compared with less than 13 per cent of the 30-39-year-olds. This probably reflects the enrollment of a greater proportion of children in Catholic school systems in more recent years. Wives under 30 account for virtually all of the differences in the expected size of completed family that are associated with the amount of education in Catholic schools (Table 41). No sig-

FERTILITY AND FAMILY PLANNING IN THE U.S.

nificant differences are found among respondents who were 30-39, suggesting that less emphasis may have been placed upon the teaching of Catholic values relating to marriage and the family when the older wives were attending school. This contrasts sharply with our findings concerning other indicators of involvement in the Catholic faith, such as the frequency of church attendance and the frequency with which the sacraments are received. Catholics who went to church and received the sacraments more frequently expected larger families than other Catholics in all age groups. It is also interesting to note that although there is a direct relation­ ship between the number of births expected and the amount of Catholic elementary and secondary education for all wives with more TABLE 41 Most Likely Expected Total Number of Births, for Catholic Wives, by Amount of Elementary and Secondary Education in Catholic Schools, by Age Amount of elementary and secondary education in Catholic schools

Total

18-24

Wife's age 25-29

30--34

35-39

All Catholic wives None Some Alla

3.7 3.6 3.7 4.2

3.6 3.1 3.9 4.0

4.1 4.0 3.9 4.6

3..8 3,,9 3. 9 3. 8

3.4 3.4 3.4 *

8 Includes only those wives who went beyond the eighth grade. Those who did not go beyond the eighth grade are included in the two preceding categories.

than a grade school education, the relationship becomes stronger as the total amount of education increases (Table 42). By far the largest differences are found among the respondents who had attended college. In the case of wives with only a grade school education, however, there is an inverse relationship between expectations and the amount of Catholic education, with the wives who had never gone to Catholic schools anticipating larger families. Catholics with a background of Catholic education demonstrate greater current evidence of religiousness than other Catholics by at­ tending church and receiving the sacraments more frequently. More than 43 per cent of the Catholic wives who had gone to Catholic elementary and secondary schools reported receiving the sacraments twice a month or more as compared with only 17 per cent of those with a secular education. This raises a question concerning the association between family-size expectations and Catholic education. Does the association reflect current involvement in the Catholic faith

RELIGION AND SOCIOECONOMIC STATUS

rather than past experience in Catholic schools per se? Apparently not, for the association continues in our analyses when we hold con­ stant current involvement as measured by the frequency of church attendance and the frequency with which the sacraments are received. In interpreting the association between fertility and Catholic edu­ cation, it is also necessary to consider two selective factors that may determine which children go to Catholic schools. Socioeconomic status operates as one selective factor, because it generally costs more money to send children to parochial schools than to public schools. The second selective factor arises from the fact that the more religious Catholic parents are more likely to make the effort to give their children a Catholic education. However, there is no indication that the higher expectations of Catholics with a Catholic education can TABLE 42 Most Likely Expected Total Number of Births, for Catholic Wives, by Amount of Elementary and Secondary Education in Catholic Schools, by Education Amount of elementary and secondary education in Catholic schools

Total

College

All Catholic wives None Some All»

3.7 3.6 3.7 4.2

4.1 3.6 4.2 4.8

Wife's education High High school, 4 school, 1-3 3.6 3.4 3.6 4.0

3.6 3.6 3.5 4.0

Grade school 4.3 4.4 4.1

" Includes only those wives who went beyond the eighth grade. Those who did not go beyond the eighth grade are included in the two preceding categories.

be attributed merely to their higher socioeconomic status or more religious family background. Catholics with an exclusively Catholic elementary and secondary education expected substantially more children than other Catholics within virtually all socioeconomic groups. Furthermore, our results suggest that Catholic education exerts a significant influence on fertility in addition to that of a religious childhood home environment.15 We still find Catholic wives with a Catholic education expecting considerably larger families than those with a secular education when the comparison is limited to wives who said that religious activities were very important in the daily life of their families when they were children. Respondents of all religious faiths were asked if there were any religious activities in the daily life of their families when they Were children, such as having family prayers, reading the Bible, saying 15 Similar results were reported in the second phase of the Princeton Study. See Westoff 1963, pp. 100-101.

FERTILITY AND FAMILY PLANNING IN THE U.S.

grace at meals, and listening to religious programs on radio or tele­ vision. If they answered "yes," they were asked whether these activities were "very important, important, or not so important" in their family life. Replies to such retrospective questions are, of course, apt to be subject to considerable bias. However, they provide our only indica­ tion of the religiousness of the respondents' childhood home environ­ ment, and there is some evidence that the replies of many wives were probably quite accurate. We find, for example, that the Catholic wives who said religious activities were "very important" or "im­ portant" in their childhood family life were much more likely than TABLE 43 Total Number of Births Expected, for Wives, by Religion, by Importance of Religious Activities in Daily Life of Family When Wife Was a Child Total"

Importance of religious activities All wives Some activities Very important Important Not so important No activities Not ascertained

No. of wives

Protestant

Catholic

Average Average Average most likely most likely most likely expected expected expected No. of no. of No. of no. of no. of births wives births wives births

2,414

3.1

1,596

2.9

668

3.7

777 568 100 962 7

3.3 2.9 3.0 3.2

474 402 74 641 5

3.0 2.8 2.9 3.0

266 140 16 245 1

3.9 3.4

*

*

*

3.7 *

• Includes 150 Jewish and "other" wives.

other Catholics to have attended Catholic elementary and secondary schools. Table 43 shows the average numbers of births expected by Protestant and Catholic wives grouped according to their replies to the questions about religious activities in the daily life of their families when they were children. None of the differences are large enough to be statistically significant for Protestants, however. In fact, the Protestant wives who said religious activities were "very important" in their childhood family life expected to have the same number of births as those who said there were no religious activities at all in their family life when they were growing up. The expectations of the Catholics who replied that religious activi­ ties were "very important" in the daily life of their families during their childhood are higher than those of other Catholics, but there is no clear evidence of a direct relationship between the expected

RELIGION AND SOCIOECONOMIC STATUS

size of family and the religiousness of the childhood home environment even for Catholics. The pattern of association is irregular, and the difference between the expectations of wives reporting "very important" religious activities and those reporting none at all is not statistically significant. Similar questions were asked about the current importance of religious activities in the daily life of the wives' own families. As shown in Table 44, there is a much clearer association between the number of births expected and the answers the respondents gave to these questions. Among both Protestants and Catholics, we find that expectations are substantially higher for the wives who said that

TABLE 44 Total Number of Births Expected, for Wives, by Religion, by Importance of Religious Activities in Daily Life of Family at Present Total a

Importance of religious activities All wives Some activities Very important Important Not so important No activities Not ascertained a

Protestant

Catholic

Average Average Average most likely most likely most likely expected expected expected No. of no. of No. of no. of No. of no. of wives births wives births births wives 2,414

3.1

1,596

2.9

668

3.7

918 647 110 733 6

3.5 3.1 3.2 2.8

588 461 86 455 6

3.2 2.9 3 1 2.7

297 152 18 201

4.1 3.6

*

—

*

*

3.3 *

Includes 150 Jewish and "other" wives.

religious activities were "very important" in the daily life of their present families than they are for those who reported no religious activities in their family life, and the differences are large enough to be statistically significant by a considerable margin. The wives were also asked if their interest in religion had been "decreasing, staying about the same, or increasing" during recent years. Almost 44 per cent replied that their interest had been in­ creasing (Table 45). They thought they would have more children than the wives who said their interest in religion had been staying about the same or decreasing in both the Protestant and the Catholic groups, but the differences are statistically significant only in the case of the Catholics.

FERTILITY AND FAMILY PLANNING IN THE U.S.

Ideal and Desired Family Size vs. Expected Family Size—Our measures of ideal and desired family size—the number of children the wives considered ideal for the average American family, the num­ ber they said they themselves would have if they could live their lives over again under ideal conditions, and the number they actually TABLE 45 Total Number of Births Expected, for Wives, by Religion, by Change in Wife's Interest in Religion in Recent Years

Change in wife's interest in religion All wives Increasing Same Decreasing Not ascertained β

Total" Protestant Catholic Average Average Average most likely most likely most likely expected expected expected No. of No. of no. of no. of no. of No. of wives births wives births wives births 2,414 1,059 1,155 194 6

3.1 3.3 3.1 2.9 *

1,596 706 753 133 4

2.9 3.0 2.9 2.8

668 291 328 48 1

*

3.7 4.0 3.6 3.3 *

Includes 150 Jewish and "other" wives.

TABLE 46 Number of Children Considered Ideal for the Average American Family, Wanted if Life Could Be Relived, Wanted at Time of Interview, and Most Likely Expected, for Wives, by Religion

Wife's religion

Number of wives

Total Protestant Catholic Jewish Other

2,414 1,596 668 106 44

Average number of children Considered Wanted ideal for aver- if life Wanted at age American could be time of family" relived" interview 3.4 3.3 3.8 2.9 3.2

3.7 3.4 4.4 3.1 3.3

3.3 3.1 4.0 2.7 3.1

Most likely expectedb 3.1 2.9 3.7 2.5 2.9

* A small number of wives did not give numerical replies to these questions. b The expected number of children is slightly smaller than the expected number of births for some groups, because it excludes children who had died prior to the interview.

wanted to have at the time of the interview—all tend to be sub­ stantially higher for Catholics than for Protestants and, in turn, higher for Protestants than for Jews (Table 46). The religious differentials in these ideals and desires are thus quite similar to those found in the wives' expectations, indicating that the differences in the numbers of children expected by the Protestant, Catholic, and Jewish groups

RELIGION AND SOCIOECONOMIC STATUS

can be attributed more to differences in family-size preferences than to differences in the ability to have the number of children preferred. Table 47 shows the percentages of Protestant, Catholic, and Jewish wives who expected fewer children than they wanted when inter­ viewed, the same number as they wanted, or more than they wanted. There are no statistically significant differences between the percent­ ages for Protestants and Catholics. The proportion of wives expecting fewer children than they wanted is about the same for Jews as for Protestants and Catholics, but a considerably smaller proportion of Jewish wives expected to have more children than they wanted (only 15 per cent as compared with 23 per cent of the Protestants and 22 per cent of the Catholics). Although the differences between these TABLE 47 Percentage Who Expected Fewer Children than They Wanted, the Same Number or More, for Wives, by Religion

Wife's religion

Number of wives

Total

Total Protestant Catholic Jewish Other

2,414 1,596 668 106 44

100 100 100 100 100

Ex­ pected Ex­ fewer pected than number wanted" Wantedb 29 28 31 29 25

50 50 47 56 57

Expected more than wanted° Had Had number more wanted than Total or fewer wanted 22 23 22 15 18

7 6 9 5 7

15 16 13 10 11

s

The most likely expected number is smaller than the minimum number wanted. most likely expected number is equal to (or between) the minimum or maximum number wanted. 0 The most likely expected number is larger than the maximum number wanted. b The

percentages are not statistically significant because of the small number of Jewish cases, they are probably real and not due to sampling error, for Jewish couples tend to be more successful than Protestant or Catholic couples in regulating family size, as we shall see in Chapter 6. Protestant and Catholic wives generally gave similar answers to the questions about their reasons for not wanting a smaller family (Table 48). However, a few differences should be noted. Catholics were less likely to give reasons related to the welfare or happiness of their children, probably because more of them already had several children by 1960 and were therefore less concerned with the problem of providing their offspring with adequate companionship. On the other hand, a larger proportion of Catholic wives gave reasons that appear to be related to the satisfaction or interests of the parents

FERTILITY AND FAMILY PLANNING IN THE U.S.

or reasons that were very ambiguous. Only 2 per cent of the Protestant wives and 3 per cent of the Catholic wives said they did not want smaller families because of religious and moral reasons. Nineteen per cent of the Protestant wives reported fecundity im­ pairments as their first reason for not expecting a larger family, as compared with only 15 per cent of the Catholic wives (Table 49). This probably reflects the fact that fewer Catholic couples had had operations that prevent pregnancy (see Chapter 4). Among wives not TABLE 48 Per Cent Distribution by First Reason for Not Wanting a Smaller Family, for Wives, by Religion First reason Number of wives Per cent: Total Reasons related to welfare or happiness of children To provide companionship for children; better for children To avoid having an only child Reasons related to satisfaction or interests of parents Likes, loves children; this many makes for a happy "full" life To avoid loneliness; to have children around the house now or when older To have desired proportion of boys and girls Religious and moral reasons (e.g., couples ought to have this many) Ambiguous reasons This number is right, nice, ideal; like parental family Don't know Other reasons Not ascertained

Totala

Protestant Catholic

2,414

1,596

668

100

100

100

51

55

43

37 14 20

39 16 18

33 9 25

14

12

20

3 3

2 3

4 1

2 13 10 2 5 8

2 12 10 2 5 8

3 15 13 2 6 9

" Includes 150 Jewish and "other" wives.

citing subfecundity as their first reasons for not expecting a larger family, we find that Catholics were less likely to mention economic reasons and more likely to give other unclassified reasons or to indicate that they could not say what their reasons were (Table 49). In the following sections of this chapter, we shall examine the family-size expectations, ideals, and desires of wives in the various social and economic groups. We shall, in many cases, show the socio­ economic differentials separately for Protestants and Catholics, be­ cause the association between fertility and socioeconomic factors is often quite different within these two religious groups.

RELIGION AND SOCIOECONOMIC STATUS TABLE 49 Per Cent Distribution by First Reason for Not Expecting a Larger Family, for Wives, by Religion First reason Numberofwives Per cent: Total

Total»

Protestant

Catholic

2,414

1,596

668

100

Subfecundity All other reasons Economic reasons Would cost too much; couldn't afford more, etc. Couldn't support more at desired standard of living Couldn't educate more Low income, insecure job, part-time work, unemployment Other economic reasons Poor health (parents and/or children); pregnancy unpleasant, or dangerous to wife or baby Hard to "care" or "have time" for more This is a "nice" number, number wife and/or husband wants; no more needed for happiness of family, etc. Other reasons, don't know Not ascertained

100

100

—

17 83 45

100 54

19 81 47

100 58

15 85 42

100 49

22

27

23

29

22

25

7 6

8 7

7 6

9 7

6 4

7 5

4 7

4 9

3 8

4 9

4 6

5 8

8 7

10 9

8 7

9 8

10 7

11 9

5 14 4

6 16 5

5 11 4

6 13 4

4 16 6

5 19 7

" Includes 150 Jewish and "other" wives.

EDUCATION

The relationship between the number of expected births and the educational attainment of the wife tends to be negative, with sig­ nificantly larger families being anticipated by the less-educated (Figure 11 and Table 50). Wives with only a grade school education thought they would have more children than those who had gone to high school for one to three years and substantially more than those whose education had extended beyond the third year of high school. There is almost no difference, however, between the expecta­ tions of the wives who had completed four years of high school and the wives who had gone on to attend college, the two groups which together account for almost two-thirds of the sample. When the comparison is made on the basis of the educational attainment of the husband, rather than of the wife, the differences in expectations are somewhat smaller, suggesting that the husband's educational background may have less influence on fertility than the wife's (Table 50). There is also one minor difference in the pattern

FERTILITY AND FAMILY PLANNING IN THE U.S. WIFE'S RELIGION AND EDUCATION ALL RELIGIONS: COLLEGE HIGH SCHOOL,4

HIGH SCHOOL, 1-3 GRADE SCHOOL

PROTESTANT: COLLEGE

HIGH SCHOOL, 4 HIGH SCHOOL,1-3 GRADE SCHOOL CATHOLIC: COLLEGE

y///////////A

HIGH SCHOOL,4

''///////A

HIGH SCHOOL,I-3

GRADE SCHOOL

O AVERAGE NUMBER OF BIRTHS BIRTHS BY I960

S

V7\ TOTAL NUM BER EXPECTED Δ (MOST LIKELY)

FIGURE 11 Births by 1960 and total number expected, for wives, by education, by religion

of association. Wives whose husbands had attended college thought they would have slightly more births than those whose husbands had completed only four years of high school, but this difference is not statistically significant. Although the relationship is more irregu­ lar, the largest differences in expectations are found when the educa­ tional categories are based on the education of both the husband and the wife.

RELIGION AND SOCIOECONOMIC STATUS

The educational differences in the number of births by 1960 are larger than those in the expected size of completed family (Table 50), because the less-educated already had a comparatively large proportion of the total number of children that they anticipated. TABLE SO Births by 1960 and Total Number Expected, for Wives, by Education of Wife, Husband, and Couple

Number of wives

Education Total Wife's education College High school, 4 High school, 1-3 Grade school Husband's education College High school, 4 High school, 1-3 Grade school Couple's education Both college One spouse college, other high school, 1-4 Both high school, 4 One spouse high school, 4, other high school, 1-3 Both high school, 1-3 One spouse high school, 1-4, or college, other grade school Both grade school

Average number of births Total expected by wives Most Minimum Maximum By 1960 likely

2,414

2.3

3.1

2.9

3.5

427 1,153 579 255

2.0 2.1 2.6 3.1

3.0 3.0 3.3 3.7

2.8 2.7 3.0 3.4

3.3 3.4 3.6 4.1

632 856 554 372

2.1 2.1 2.5 2.8

3.1 3.0 3.3 3.5

2.8 2.7 3.0 3.2

3.4 3.3 3.6 3.9

307

2.0

3.1

2.8

3.4

428 568

2.2 2.0

3.0 2.9

2.8 2.6

3.4 3.2

399 234

2.4 2.6

3.2 3.3

3.0 3.0

3.5 3.6

329 149

2.7 3.2

3.3 3.9

3.1 3.6

3.6 4.3

In fact, the less-educated wives thought they would have fewer births after the interview than the better-educated wives:

Wife's education College High school, 4 years High school, 1-3 years Grade School

Anerage most likely expected number of additional births 1.0

0.9 0.7 0.6

This can be explained in part by the fact that the wives with less education had been married longer on the average than those with

FERTILITY AND FAMILY PLANNING IN THE U.S.

more education and that recognized fecundity impairments were more frequent among the less-educated (see Chapter 4). However, it is also quite probable that the expectations of many of the less-educated wives are unrealistically low. As pointed out in Chapter 1, the younger wives with only a grade school education significantly underestimated their future childbearing when interviewed in the 1955 survey. Among Protestants, the relationship between the wife's educational attainment and the expected size of completed family tends to be inverse and is very similar to that described for the total sample of white wives (Figure 11 and Table 51). The Protestant wives who had attended college or completed four years of high school thought they would have fewer births than those who had completed TABLE 51 Births by 1960 and Total Number Expected, for Wives, by Religion and Education Average number of births Wife's religion and education

Number of wives

By 1960

Most likely

All Protestants College High school, 4 High school, 1-3 Grade school All Catholics College High school, 4 High school, 1-3 Grade school

1,596 284 752 392 168 668 79 341 168 80

2.2 2.0 2.0 2.5 2.9 2.5 2.3 2.3 2.7 3.5

2.9 2.8 2.8 3.1 3.5 3.7 4.1 3.6 3.6 4.3

Expected total

Minimum 2.7 2.6 2.6 2.9 3.2 3.4 3.7 3.2 3.3 4.0

Maximum 3.2 3.1 3.1 3.4 3.8 4.2

4.8 4.1 4.0 4.7

from one to three years of high school and considerably fewer than those who had not gone beyond grade school. In the case of the Catholics, however, the relationship between the wife's educational attainment and the expected size of family is almost U-shaped, rather than inverse, with the college-educated wives expecting nearly as many births as those with only a grade school education and substantially more than the wives with varying amounts of high school education. Catholics thought they would have significantly more children than Protestants in all educational groups, but the largest CatholicProtestant differentials in expectations are found among the college educated. Although the Catholic wives who had attended college expected to have significantly larger completed families than those with only a high school education, they had not had more births by 1960 (Table 51). This is probably primarily because they had married

RELIGION AND SOCIOECONOMIC STATUS

at later ages and had thus had less time in which to bear children. Perhaps a substantial proportion of the college-educated Catholics have overestimated their future fertility and will tend to revise their expectations downward as they have more experience with the bearing and rearing of children. There is some indication that this was true in the 1955 survey, but the evidence is not conclusive because of the small number of college-educated Catholics in the sample (see Chapter 1). The relatively high expectations of the Catholic college group appear to be a recent development, for they can be attributed almost TABLE 52 Total Number of Births Exjjected, for Wives, by Age, by Education, by Religion Total'·

Wife's age and education 18-29 College High school, 4 High school, 1-3 Grade school 30-39 College High school, 4 High school, 1-3 Grade school

Protestant

Catholic

Average Average Average most likely most likely most likely Number expected Number expected Number expected of no. of of no. of of no. of wives births wives births wives births 180 564 282 87

3.2 3.0 3.3 4.0

124 381 204 57

2.8 2.8 3.1 3.7

37 169 66 30

4.7 3.5 3.9 4.6

247 589 297 168

2.9 3.0 3.3 3.5

160 371 188 111

2.8 2.8 3.2 3.3

42 172 102 50

3.7 3.6 3.4 4.2

* Includes 150 Jewish and "other" wives.

entirely to respondents who were under 25 years of age in the 1955 study and under 30 in the present study. Table 52 shows that among the Catholic wives who were 18-29 in 1960 the college-educated thought they would have 34 per cent more births than the high school graduates and slightly more than even the wives who had not gone beyond grade school. Among the Catholic wives who were 30-39, however, the college-educated did not expect significantly more births than the high school graduates and expected 12 per cent fewer than the wives in the grade school group. The pattern of association between the expected size of completed family and the educational attainment of the wife in 1960 is quite similar to that found in the 1955 study in the case of the Protestants (Table 53). Although expectations were higher in 1960 than in 1955 for Protestant wives who had gone beyond grade school and lower

FERTILITY AND FAMILY PLANNING IN THE U.S.

for those who had not, the differences are not large enough to be statistically significant. Among the Catholics, on the other hand, ex­ pectations were higher in 1960 than in 1955 in all educational groups, with the 1955-1960 differences being considerably larger than those for Protestants, particularly in the lower educational groups. The Catholic wives who had gone to college anticipated more births than those with only a grade school education in 1955 while the opposite was true in 1960, but the difference between the expectations of TABLE 53 Total Number of Births Expected, for Wives, by Religion, by Education, 1955 and 1960

Wife's religion and education All religions" Total College High school, 4 High school, 1-3 Grade school Protestant Total College High school, 4 High school, 1-3 Grade school Catholic Total College High school, 4 High school, 1-3 Grade school

Number of wives 1955 1960

Average most likely expected number of births 1955 1960

2,713» 417 1,236 681 377

2,414 427 1,153 579 255

3.0 2.9 2.9 3.0 3.6

3.1 3.0 3.0 3.3 3.7

1,817 306 794 457 260

1,596 284 752 392 168

2.9 2.7 2.7 2.9 3.6

2.9 2.8 2.8 3.1 3.5

787 73 396 208 110

668 79 341 168 80

3.4 3.9 3.3 3.1 3.7

3.7 4.1 3.6 3.6 4.3

" Includes Jewish and "other" wives. b Includes 2 wives whose education was not ascertained.

the wives in these two educational groups is not statistically significant in either the 1955 or the 1960 results. In the 1960 study, the educational differences in attitudes toward family size generally follow about the same pattern as those in ex­ pected family size for wives of all religious faiths combined. As shown in Table 54, the number of children the wives considered ideal for the average American family, the number they said they themselves would have if they could live their lives over again under ideal con­ ditions, and the number they actually wanted to have at the time of the interview all tend to be higher for the wives who had not gone beyond grade school than for those with more education. This

RELIGION AND SOCIOECONOMIC STATUS

may, however, be due in large measure to the fact that the respondents with only a grade school education already had substantially more children than the better-educated by 1960 and were thus more likely to have exaggerated their replies to the questions about their family-size ideals and desires in order to avoid admitting that any of their children were unwanted or that they had ever had a family TABLE 54 Number of Children Considered Ideal for the Average American Family, Wanted if Life Could Be Relived, Wanted at Time of Interview, and Most Likely Expected, for Wives, by Religion, by Education Average number of children

Wife's religion and education All religions0 Total College High school, 4 High school, 1-3 Grade school Protestant Total College High school, 4 High school, 1—3 Grade school Catholic Total College High school, 4 High school, 1-3 Grade school

Number of wives

Considered ideal for average Wanted if Wanted at American life could time of family8 be relived" interview

Most likely expectedb

2,414 427 1,153 579 255

3.4 3.3 3.4 3.5 3.7

3.7 3.7 3.6 3.7 4.0

3.3 3.3 3.2 3.3 3.5

3.1 3.0 3.0 3.2 3.6

1,596 284 752 392 168

3.3 3.2 3.2 3.5 3.6

3.4 3.4 3.4 3.5 3.6

3.1 3.1 3.0 3.2 3.1

2.9 2.8 2.7 3.1 3.3

668 79 341 168 80

3.8 4.1 3.7 3.7 4.1

4.4 5.2 4.2 4.2 4.8

4.0 4.8 3.9 3.6 4.3

3.7 4.1 3.5 3.5 4.2

" A small number of wives did not give numerical replies to these questions. b The expected number of children is slightly smaller than the expected number of births for some groups, because it excludes children who had died prior to the interview. 0 Includes 150 Jewish and "other" wives.

planning failure. When we compare wives with the same number of births by 1960, we find that those in the grade school group did not want more children than those in the higher educational groups. Although the educational differences in family-size ideals and de­ sires follow about the same pattern as those in expectations, they tend to be smaller (Table 54). This suggests that educational attain­ ment distinguishes the wives less in terms of family-size values than

FERTILITY AND FAMILY PLANNING IN THE U.S.

in terms of other factors that determine the number of children they actually expect to have, such as the ability to control fertility. The important association between educational attainment and the ability to control fertility is clearly indicated by the data in Table 55 showing that the proportion of wives expecting more children than they wanted rises from only 15 per cent for the college group to 38 per cent TABLE 55 Percentage Who Expected Fewer Children than They Wanted, the Same Number or More, for Wives, by Religion and Education

Wife's religion and education All religions'1 Total College High school, 4 High school, 1-3 Grade school Protestant Total College High school, 4 High school, 1-3 Grade school Catholic Total College High school, 4 High school, 1-3 Grade school

Expected Ex­ fewer pected than number wanted" Wantedb

Expected more than wanted Had number Had wanted more or than Total fewer wanted

Num­ ber of wives

Total

2,414 427 1,153 579 255

100 100 100 100 100

29 30 29 27 27

50 55 52 46 35

22 15 19 27 38

7 6 6 8 10

15 9 12 19 28

1,596 284 752 392 168

100 100 100 100 100

28 28 29 26 24

50 57 52 46 35

23 15 19 28 41

6 5 6 7 11

16 10 13 21 30

668 79 341 168 80

100 100 100 100 100

31 32 33 28 30

47 51 50 46 36

22 18 18 26 34

9 11 8 11 9

13 6 10 15 25

• The most likely expected number is smaller than the minimum number wanted. b The most likely expected number is equal to (or between) the minimum or maxi­ mum number wanted. 0 The most likely expected number is larger than the maximum number wanted. d Includes 150 Jewish and "other" wives.

for the grade school group. Even the latter figure very likely under­ states the proportion of wives with only a grade school education who will actually have more children than they want by the time their families are completed, for as previously mentioned, the expecta­ tions of many of the less-educated respondents are probably unrealistically low. Most of the wives who expected more children than they wanted already had more than they wanted when interviewed.

RELIGION AND SOCIOECONOMIC STATUS

The Catholics are an exception to the generalization that educa­ tion has less influence upon family-size values than upon the number of children actually expected. When the Catholic wives are grouped according to educational attainment, we find that their personal family-size ideals and desires vary more than their expectations. This can be attributed to the college-educated Catholics. They expected to have slightly fewer children than Catholics with only a grade TABLE 56 Per Cent Distribution by First Reason for Not Wanting a Smaller Family, for Wives, by Education

College

High school, 4

High school, 1-3

Grade school

Number of wives

427

1,153

579

255

Per cent: Total

100

100

100

100

53

57

46

36

38 15

40 17

35 11

28 8

22

18

19

24

17

13

13

19

4 1

2 3

3 3

2 3

3 11

1 12

3 15

3 17

8 3 6 4

10 2 5 7

13 2 6 11

13 4 7 14

First reason

Reasons related to welfare or happiness of children To provide companionship for children; better for children To avoid having an only child Reasons related to satisfaction or interests oi parents Likes, loves children; this many makes for a happy "full" life To avoid loneliness; to have children around the house now or when older To have desired proportion of boys and girls Religions and moral reasons (e.g., couples ought to have this many) Ambiguous reasons This number is right, nice, ideal; like parental family Don't know Other reasons Not ascertained

school education, but said they would have more if they could live their lives over again under ideal conditions and actually wanted to have more at the time of the interview (Table 54). The number of children the Catholic wives wanted if they could relive their lives and the number they wanted when interviewed are thus considerably higher for the college group than for any of the other educational groups. This would seem to be consistent with the previously men­ tioned finding that the better educated demonstrated more evidence of associational involvement with the Catholic faith than the less

FERTILITY AND FAMILY PLANNING IN THE U.S.

educated by attending church and receiving the sacraments more frequently and by being more likely to have attended church-related schools. Almost 28 per cent of the college-educated Catholic wives had received all of their elementary and secondary education in Catholic schools as compared with only 18 per cent of the Catholic high school graduates. Tables 56 and 57 show percentage distributions of the wives in TABLE 57 Per Cent Distribution by First Reason for Not Expecting a Larger Family, for Wives, by Education

First reason Number of wives Per cent: Total Subfecundity All Other Reasons Economic reasons Would cost too much; couldn't afford more, etc. Couldn't support more at desired standard of living Couldn't educate more Low income, insecure job, parttime work, unemployment Other economic reasons Poor health (parents and/or chil­ dren); pregnancy unpleasant, or dangerous to wife or baby Hard to "care" or "have time" for more This is a nice number, number wife and/or husband wants; no more needed for happiness of family, etc. Other reasons, don't know Not ascertained

College

High school, 4

High school, 1-3

Grade school

427

1,153

579

255

100

100

100

100

11 89 100 44 48

16 84 100 55 47

21 79 45

100 59

22 78 40

100 52

14

15

23

27

26

33

23

29

7 11

8 12

7 6

8 7

8 2

10 3

4 2

5 3

3 9

3 10

3 8

4 9

4 5

6 7

5 6

7 8

8

9

8

9

9

11

10

13

9

8

7

8

6

8

8

11

3 22 4

4 26 4

6 13 3

7 16 4

5 9 4

6 12 5

2 10 7

3 13 9

the various educational groups according to the first reasons they gave for not wanting a smaller family and for not expecting a larger one. Only a few of the educational differences are large enough to be significant. The less-educated wives were not as likely as the better-educated wives to give reasons related to the welfare or happi­ ness of their children, probably because they were more likely to have families that were already large enough to provide adequate companionship for their offspring. On the other hand, a larger propor-

RELIGION AND SOCIOECONOMIC STATUS

tion of the less-educated wives' reasons were ambiguous or were not ascertained by the interviewers. The proportion of wives reporting fecundity impairments as the first reason for not expecting a larger family is inversely related to their educational attainment, rising from 11 per cent for the wives who had attended college to 22 per cent for those who had not gone beyond grade school. Among wives not citing subfecundity as their first reason, we find that those with less education were somewhat more likely to report economic reasons in general, but were less likely to mention specifically the cost of educating children and less likely to give reasons other than those that are classified in Table 57. ECONOMIC FACTORS

Income—Couples in the lower-income groups have probably been having more children than couples in the higher-income groups ever since the process of industrialization and urbanization began in the United States. However, the negative association between income and fertility has diminished considerably since the depression years of the 1930's. We find that the wives whose husbands were earning less than $5,000 per year expected a somewhat larger number of births than the wives of husbands with higher incomes, but the difference is very small (Figure 12 and Table 58). In fact, most of the differ­ ences between the expectations of the wives in the various income groups are not large enough to be statistically significant, and the pattern of the differences is irregular. The wives of husbands in the higher-income classes had actually borne more children by the time of the interview in 1960 than the wives of husbands in the lower-income classes, but this can be at­ tributed primarily to differences in age. The wives whose husbands had higher incomes were older and had been married longer on the average, and had thus had more time in which to have children. When we compare only those wives who are in the same age groups, we generally find that the wives of husbands with higher incomes not only expected to have a smaller number of children when their families were completed but also had a smaller number when inter­ viewed. The income differentials are quite small within most age groups, however, and indicate no significant intercohort trend in the relationship between fertility and income. The Protestant wives whose husbands were earning less than $5,000 annually thought they would have a slightly larger average number of births than those whose husbands were in the higherincome groups (Table 59). Among the Catholics, on the other hand, there are larger income differentials in expectations, and the rela-

FERTILITY

AND FAMILY

PLANNING

IN THE

U.S.

FIGURE 12 Births by 1960 and total number expected, for wives, by husband's income

tionship tends to be U-shaped, rather than inverse, with the wives of husbands in the lowest and highest income groups expecting more children than the wives of husbands in the middle income group. Within most of the socioeconomic classes differentiated in this study, we find a slight inverse relationship between the wives' family-size TABLE 58 Births bv 1960 and Total Number ExDected. for Wives, bv Husband's Income

Husband's Income

Number of wives

By 1960

Total $10,000 or more $ 7,000-$9,999 $ 6,000-86,999 $ 5,000-55,999 S 4,000-54,999 $ 3,000-$3,999 Under $3,000

2,414 261 405 312 423 380 306 327

2.3 2.5 2.3 2.4 2.3 2.3 2.2 2.1

104

Average number of births Expected total Most likely Minimum Maximum 3.1 3.1 3.0 3.1 3.1 3.2 3.3 3.2

2.9 2.8 2.8 2.9 2.9 2.9 3.0 2.9

3.5 3.3 3.4 3.4 3.5 3.6 3.7 3.6

RELIGION AND SOCIOECONOMIC STATUS

expectations and their husbands' income. Although a large proportion of the farm wives are in the lower-income groups, the association between expectations and income is not appreciably changed when they are excluded from our analysis. Farm wives whose husbands had annual money incomes of less than $5,000 did not anticipate significantly larger families than the wives of husbands who made less than $5,000 in nonfarm occupations. TABLE 59 Total Number of Births Expected, for Wives, by Religion, by Husband's Income

Husband's Income Total $7,000 or more $5,000-$6,999 Under $5,000

Totala Average most likely Number expected no. of of births wives 2,414 666

735 1,013

Protestant Catholic Average Average most likely most likely Number expected Number expected no. of no. of of of wives births births wives

3.1 3.0 3.1 3.2

1,596 428 462 706

2.9 2.9 2.9 3.0

668 148 247 273

3.7 3.8 3.5 3.9

» Includes 150 Jewish and "other" wives. TABLE 60 Number of Children Considered Ideal for the Average American Family, Wanted if Life Could Be Relived, Wanted at Time of Interview and Most Likely Expected, for Wives, by Husband's Income Average number of children

Husband's income

Number of wives

Considered ideal for average American family8

Total $7,000 or more $5,000-$6,999 Under $5,000

2,414 666 735 1,013

3.4 3.3 3.4 3.5

Wanted if Wanted at life could time of be relived" interview 3.7 3.7 3.6 3.7

3.3 3.3 3.3 3.3

Most likely expected 3.1 3.0 3.1 3.2

• A small number of wives did not give numerical replies to these questions.

Replies of the wives in the various income groups to the questions on attitudes toward family size reveal no important differences (Table 60). The wives of husbands with incomes of less than $5,000 set the ideal size of the average American family a little higher than the other wives, but there are almost no income differentials in the numbers of children the wives said they themselves would have if they could live their lives over again under ideal conditions, or in

FERTILITY AND FAMILY PLANNING IN THE U.S.

the numbers they said they actually wanted to have at the time of the interview. In all income classes, the wives indicated that they expected fewer children than they wanted on the average. The propor­ tion expecting fewer or more than they wanted varies very little with income (Table 61). A somewhat larger proportion of the wives whose husbands earned less than $7,000 a year anticipated more children than they wanted, but the difference is not statistically significant. Thus far we have considered only the husband's income as a measure of economic status. When the wives are classified on the basis of the family's income,1® instead of the husband's income alone, TABLE 61 Percentage Who Expected Fewer Children than They Wanted, the Same Number or More, for Wives, by Husband's Income Expected more than wanted 0

Husband's income

Number of wives

Total

Total 37,000 or more Ϊ5,000-J6,999 Under $5,000

2,414 666 735 1,013

100 100 100 100

Expected Exfewer pected than number wanted® Wantedb 29 28 29 28

50 52 48 49

Total 22 20 23 23

Had number Had wanted more or than fewer wanted 7 5 8 8

15 15 15 15

The most likely expected number is smaller than the minimum number wanted. The most likely expected number is equal to (or between) the minimum or maxi­ mum number wanted. 0 The most likely expected number is larger than the maximum number wanted. a

b

we find that the relationship between fertility expectations and income is stronger, although the pattern of association is still irregular (Table 62). The difference between the average most likely expected numbers of births for the highest and lowest income groups is 0.4 when family income is used as compared with only 0.1 when husband's income is used. The stronger relationship between family income and fertility is, however, probably due much less to the influence of family income per se than to the influence of the wife's employment status. In many cases, a higher family income simply reflects the fact that the wife as well as the husband was gainfully employed, and wives who work tend to have fewer children than wives who do not work. M In the vast majority of cases in this study, the family's income consists either of the husband's income alone or of the incomes of both the husband and the wife. There are very few instances in which other family members were reported to be contributing to the family's support.

RELIGION AND SOCIOECONOMIC STATUS

Wife's work history—The wife's work history is of considerable interest in the analysis of American fertility because of the great extent to which today's women combine jobs with marriage and childbearing. More than 23 per cent of the white wives in our sample were gainfully employed at the time of the interview, and another 48 per cent reported that they had been employed at some time during their married lives. Only 22 per cent said they had never worked since they were married and definitely did not expect to do so in the future. The wives who had worked after their marriage had not only had fewer births by 1960 than those who had not worked, but also anticipated a significantly smaller total number of births (Figure 13 and Table 63). Furthermore, we find an inverse relationship between fertility and the duration of the wife's work experience. The wives who had worked for longer periods of time TABLE 62 Births by 1960 and Total Number Expected, for Wives, by Family Income

Family income

Number of wives

By 1960

Total $10,000 or more $ 7,000-59,999 5 6,000-56,999 5 5,000-55,999 S 4,000-54,999 S 3,000-S3,999 Under 53,000

2,414 339 526 321 426 316 259 227

2.3 2.1 2.2 2.5 2.4 2.5 2.3 2.2

Average number of births Expected total Most likely Minimum Maximum 3.1 2.8 2.9 3.2 3.3 3.4 3.3 3.2

2.9 2.6 2.7 3.0 3.0 3.1 3.0 2.9

3.5 3.1 3.3 3.5 3.6 3.7 3.7 3.7

had had fewer births by 1960 and expected to have fewer after the interview than those who had worked for shorter periods. The wives with past work experience were asked if they had worked mainly because they liked to, because they wanted to get extra things for themselves and the family, or because the family really needed the income. In our analysis, we have combined the wives who said they worked to get extra things for themselves and the family with those who said they "liked to work," and as shown in Figure 13 and Table 63, this group tended to have fewer children and to expect smaller completed families than the larger group of "other" wives who said they worked mainly because their families really needed the income. The causal relationship between the wife's work experience and fertility is quite complex. Many wives are probably motivated to limit the size of their families because of their desire to work, and many wives who cannot have as many children as they want as a

FERTILITY AND FAMILY PLANNING IN THE U.S.

result of fecundity impairments probably work primarily because they have comparatively few responsibilities at home. These hypotheses are supported by the finding of a negative association between wife's work experience and family-size expectations among wives in both the Fecund and the Subfecund groups17 (Table 64). On the other hand, the wives who said they worked mainly because their families YEARS WIFE WORKED SINCE FIRST MARRIAGE AND REASONS FOR WORKING

Υ/////Λ

HAS NOT WORKED

HAS WORKED 1.5 YEARS OR LESS

Υ//////Λ

LIKES TO WORK

OTHER REASONS 1.5-4.5 YEARS LIKES TO WORK

VZ/7/Ά

OTHER REASONS

4.5 YEARS OR MORE LIKES TO WORK OTHER REASONS

T~ AVERAGE NUMBER OF BIRTHS BIRTHS BY I960 I TOTAL NUMBER EXPECTED (MOST LIKELY)

FIGURE 13 Births by 1960 and total number expected, for wives, by years worked since first marriage and reason for working

needed the income had and expected significantly more children than those who worked because they "liked to," and this suggests that some wives may hold jobs because of economic needs arising from the fact that they have or expect to have a relatively large number of children. Wives who had worked during their marriage expected fewer " The fecundity classification is discussed in Chapter 4.

RELIGION AND SOCIOECONOMIC STATUS TABLE 63 Births by 1960 and Total Number Expected, for Wives, by Years Wife Worked Since First Marriage, by Reason for Working

Years wife worked since first marriage by reason Total" Has not worked Has worked Under 1.5 years: likes to work All others 1.5-4.5 years: likes to work All others 4.5 years or more: like to work All others

Number of wives

Average number of births Expected total Most By 1960 likely Minimum Maximum

2,414 683

2 .3 2.,6

3 .1 3. 6

2..9 3..3

3,.5 3. 8

235 462

2,.1 2. 4

3..2 3..4

3. 0 3. 2

3. 6 3..7

166 441

1 .9 2 .3

2. 8 3 .1

2..5 2 .8

3..2 3..4

116 293

1 .4 2 .0

1 .9 2 .4

1 .6 2 .2

2 .6 2 .8

' Includes 18 wives whose work experience was not ascertained. TABLE 64 Total Number of Births Expected, for Wives, by Years Wife Worked Since First Marriage and Reason for Working, by Fecundity of Couple

Years wife worked and reason Total" Has not worked Has worked Under 1.5 years: likes to work All others 1.5-4.5 years: likes to work All others 4.5 years and over: likes to work All others

Fecund Subfecund Average Average most likely most likely Number expected Number expected of no. of of no. of wives births wives births 1,674 503

3.4 3.8

740 180

2.6 3.0

182 368 114 314 51 135

3.4 3.5 3.0 3.3 2.4 2.7

53 94 52 127 65 158

2.6 3.3 2.2 2.6 1.5 2.2

* Includes 7 Fecund wives and 11 Subfecund wives whose work experience was not ascertained.

births than those who had not, and wives who had worked for longer periods expected fewer than those who had worked for shorter periods among both Protestants and Catholics (Table 65) and within most major social and economic groups. Moreover, there is no evidence that the importance of this association is diminishing, for there is a strong inverse relationship between wife's work experience and family-size expectations among the younger as well as the older wives.

FERTILITY AND FAMILY PLANNING IN THE U.S. TABLE 65 Total Number of Births Expected, for Wives, by Years Wife Worked Since First Marriage and Reasons for Working, by Wife's Religion

Years wife worked and reason Totalb Has not worked Has worked Under 1.5 years: Likes to work All others 1.5 years and over: Likes to work All others β b

Total" Protestant Catholic Average Average Average most likely most likely most likely Number expected Number expected Number expected of no. of of no. of of no. of wives births wives births wives births 2,414 683

3 .1 3 .6

1,596 439

2..9 3..4

668 189

3..7 4 .3

235 462

3 .2 3 .4

157 294

3 .0 3 .2

58 149

4 .1 4 .1

282 734

2 .4 2 .8

189 503

2 .2 2 .7

76 193

2 .9 3..2

Includes 150 Jewish and "other" wives. Includes 18 wives whose work experience was not ascertained.

TABLE 66 Number of Children Considered Ideal for the Average American Family, Wanted if Life Could Be Relived, Wanted at Time of Interview, and Most Likely Expected, for Wives, by Years Wife Worked Since First Marriage and Reason for Working

Years wife worked and reason Total0 Has not worked Has worked Under 1.5 years: Likes to work All others 1.5-4.5 years: Likes to work All others 4.5 years and over: Likes to work All others

Number of wives

Average number of children Considered ideal for Wanted the average if life Wanted at Most American could be time of likely family" relived" interview expectedb

2,414 683

3 .4 3 .5

3..7 3 .8

3.3 3.5

3.1 3.5

235 462

3 .5 3..4

3 .7 3..7

3.4 3.4

3.2 3.4

166 441

3 3 3..3

3..5 3. 6

3.1 3.2

2.7 3.0

116 293

3.,3 3..4

3,.3 3..5

3.0 3.0

1.9 2.4

A small number of wives did not give numerical replies to these questions. The expected number of children is slightly smaller than the expected number of births for some groups, because it excludes children who had died prior to the interview. 0 Includes 18 wives whose work experience was not ascertained. a

b

RELIGION AND SOCIOECONOMIC STATUS

There also tends to be an inverse relationship between work expe­ rience and the number of children the wives considered ideal for the average American family, the number they said they themselves would have if they could live their lives over again under ideal condi­ tions, and the number they actually wanted to have at the time of the interview (Table 66). The differences in these ideals and desires are considerably smaller than the differences in expectations, TABLE 67 Percentage Who Expected Fewer Children than They Wanted, the Same Number or More, for Wives, by Yezirs Wife Worked Since First Marriage and Reason for Working Expected more than wanted 0

Years wife worked and reason Totald Has not worked Has worked Under 1.5 years: Likes to work All others 1.5-4.5 years: Likes to work All others 4.5 years and over: Likes to work All others

Ex­ pected Ex­ fewer pected than number wanted· Wantedb

Had num­ Had ber wanted more or than Total fewer wanted

Number of wives

Total

2,414 683

100 100

29 23

50 52

22 25

7 8

15 17

235 462

100 100

26 20

55 55

19 25

8 8

11 17

166 441

100 100

33 29

50 48

18 23

8 7

10 16

116 293

100 100

56 41

32 40

12 19

3 5

9 14

* The most likely expected number is smaller than the minimum number wanted. b The most likely expected number is equal to (or between) the minimum or maxi­ mum number wanted. 0 The most likely expected number is larger than the maximum number wanted. d Includes 18 wives whose work experience was not ascertained.

however, since work experience during marriage is associated not only with the wives' family-size values but also with their ability to have the number of children they want. The wives who had not worked since they were married expected the same average num­ ber of children as they said they wanted to have at the time of the interview, but those who had worked expected fewer than they wanted as a group, with the difference between the average numbers they expected and wanted increasing with the duration of their work experience. As shown in Table 67, only 23 per cent of the wives who had

FERTILITY AND FAMILY PLANNING IN THE U.S.

not worked at all expected fewer children than they wanted as com­ pared with almost half of those who had worked 4.5 years or more. This reflects the fact that a very large proportion of the wives with longer periods of work experience could not have as many children as they wanted because of fecundity impairments; 55 per cent of those who had worked 4.5 years or more are classified as Subfeeund. The proportion expecting more children than they wanted declines with the duration of work experience and is smaller for the wives who had worked because they "liked to" than for those who had worked mainly because their families needed the income. HUSBAND'S OCCUPATION UPPER WHITE COLLAR LOWER WHITE COLLAR

UPPER BLUE COLLAR

V777Z\

LOWER BLUE COLLAR

FARM

1

1

1 2

1

1

3

4

AVERAGE NUMBER OF BIRTHS BIRTHS BY I960 TOTAL NUMBER EXPECTED (MOST LIKELY)

FIGURE 14 Births by 1960 and total number expected, for wives, by husband's occupation

Husband's occupation—Although the fertility differentials associ­ ated with the husband's occupation are still somewhat sharper than those associated with the husband's income, they are substantially smaller than they were before World War II. In the present study, we find that the wives of husbands in farm occupations (farmers and farm laborers) expected about 13 per cent more births than the wives of husbands in other occupations (Figure 14 and Table 68). The differences in the expectations for the various nonfarm occupational groups are considerably smaller, however. The wives whose husbands were in blue-collar occupations anticipated only 7 per cent more births than the wives of husbands who were in whitecollar occupations, and there is virtually no difference between the

RELIGION AND SOCIOECONOMIC STATUS

expectations for the upper and lower blue-collar groups or for the upper and lower white-collar groups. Among Protestants, the pattern of association between the hus­ band's occupational status and the expected size of completed family is similar to that described for the total sample of white wives, but the difference between the expectations of the wives in the white-collar and blue-collar groups is a little larger (Table 69). In the case of TABLE 68 Births by 1960 and Total Number Expected, for Wives, by Husband's Occupation

Husband's occupation

Number of wives

Total Upper white collar Lower white collar Upper blue collar Lower blue collar Farm Other"

2,414 725 312 465 670 154 88

Average number of births Expected total By 1960 Most likely Minimum Maximum 2.3 2.2 2.1 2.4 2.5 2.6 1.7

3.1 3.0 3.0 3.2 3.2 3.5 3.0

2.9 2.8 2.7 3.0 3.0 3.2 2.8

3.5 3.3 3.3 3.6 3.6 3.8 3.3

• Includes members of the armed forces, students, and cases in which the husband's occupation was not ascertained. TABLE 69 Total Number of Births Expected, for Wives, by Husband's Occupation, by Wife's Religion

Husband's occupation Total Upper white collar Lower white collar Upper blue collar Lower blue collar Farm Other s

Totala Protestant Catholic Average Average Average most likely most likely most likely Number expected Number expected Number expected of no. of of no. of of no. of wives births wives births wives births 2,414 725 312 465 670 154 88

3.1 3.0 3.0 3.2 3.2 3.5 3.0

1,596 457 200 310 435 128 66

2.9 2.8 2.7 3.1 3.1 3.3 2.8

668 165 95 143 219 26 20

3.7 3.9 3.6 3.5 3.7 4.5 3.9

Includes 150 Jewish and "other" wives.

the Catholics, however, the association tends to be U-shaped, with larger families being expected by the wives of husbands in upper white-collar, lower blue-collar, and farm occupations, and smaller families being expected by the wives of husbands in lower white-collar and upper blue-collar occupations. In the nonfarm group, the largest differences between the expectations of Protestants and Catholics are found among the wives in the upper and lower white-collar classes.

FERTILITY AND FAMILY PLANNING IN THE U.S.

The differences in the numbers of births expected by the wives of husbands in the various occupational groups are considerably smaller in the present study than in the 1955 study because of an increase in the expectations of the wives in the nonfarm groups and a decrease in the expectations of the farm wives (Table 70). These changes occurred among both the Protestants and the Catholics, but the increase in the expectations of the wives of husbands in nonfarm occupations was more pronounced for the Catholics. The 1960 results TABLE 70 Total Number of Births Expected, for Wives, by Husband's Occupation, 1955 and 1960

Husband's occupation Total Upper white collar Lower white collar Upper blue collar Lower blue collar Farm Other

Number of wives 1955 1960

Average most likely expected number of births 1955 1960

2,713 620 286 644 765 242 156

3.0 2.8 2.8 2.9 3.1 3.7 3.1

2,414 725 312 465 670 154 88

3.1 3.0 3.0 3.2 3.2 3.5 3.0

suggest that the contraction of the occupational differentials may be continuing, for the differences in the most likely expected numbers of births are of much less importance among the younger wives than among the older wives: WIFE'S AGE

Husband's occupation All wives Upper white-collar Lower white-collar Upper blue-collar Lower blue-collar Farm

18-29 3.2 3.2 3.2 3.3 3.2 3.4

30-39 3.1 2.9 2.8 3.2 3.3 3.5

As shown in Table 71, the wives of farmers and farm laborers tended to want slightly more children than the other wives at the time of the interview, but there are no differences in the average numbers of children wanted by the wives of husbands in the four major nonfarm occupational groups. The wives in all groups actually expected fewer children than they wanted on the average. However, the proportion of wives expecting fewer children than they wanted is smaller for the farm group than for the large nonfarm groups (Table 72). On the other hand, the proportion expecting more

RELIGION AND SOCIOECONOMIC STATUS

TABLE 71 Number of Children Considered Ideal for the Average American Family, Wanted if Life Could Be Relived, Wanted at Time of Interview, and Most Likely Expected, for Wives, by Husband's Occupation Average number of children

Husband's occupation

Number of wives

Considered ideal for average American family"

Wanted if life could be relived"

Wanted at time of interview

Most likely expectedb

Total Upper white collar Lower white collar] Upper blue collar Lower blue collarJ Farm Other]|

2,414 725 312 465 670 154 88

3.4 3.3 3.5 3.4 3.5 3.7 3.4

3.7 3.7 3.6 3.6 3.7 3.9 3.4

3.3 3.3 3.3 3.3 3.3 3.5 3.0

3.1 3.0 2.9 3.2 3.2 3.4 3.0

a

A small number of wives did not give numerical replies to these questions. The expected number of children is slightly smaller than the expected number of births for some groups, because it excludes children who died prior to the interview. b

TABLE 72 Percentage Who Expected Fewer Children than They Wanted, the Same Number or More, for Wives, by Husband's Occupation Expected more than wanted 0

Husband's occupation

Number of wives

Total

Total Upper white collar Lower white collar Upper blue collar Lower blue collar Farm Other

2,414 725 312 465 670 154 88

100 100 100 100 100 100 100

Expected Exfewer pected than number wanted" wantedb 29 29 32 28 30 23 18

50 54 47 51 43 53 58

Had numHad ber more wanted than Total or fewer wanted 22 17 21 22 27 24 24

7 6 6 6 9 8 6

15 12 14 15 18 16 18

" The most likely expected number is smaller than the minimum number wanted. b The most likely expected number is equal to (or between) the minimum or maxi­ mum number wanted. 0 The most likely expected number is larger than the maximum number wanted.

children than they wanted is lowest in the case of the wives of hus­ bands in upper white-collar occupations and highest in the case of wives whose husbands were in lower blue-collar and farm occupations. PLACE OF RESIDENCE

The differences in fertility that are associated with place of resi­ dence reflect, in varying degree, many of the religious, social, and

FERTILITY AND FAMILY PLANNING IN THE U.S.

economic factors that have already been discussed in this chapter. For example, Catholics are more heavily concentrated in the larger metropolitan areas than Protestants, and persons with a college educa­ tion are more likely to live in cities and towns than persons with less education. But selective influences such as these are not the only causes of residential differentials in fertility, for differences in the ways of life in various types of communities also affect family size. There is evidence that rural couples have been having more children than urban couples in the United States ever since the Colonial period. Although many other factors are involved, the higher rural fertility has probably been due in large measure to differences in the roles children play in rural and urban communities. In the rural agricul­ tural environment, the family has usually been the principal unit of production, and children have often been regarded as economic assets since they are a source of labor. In the urban environment, on the other hand, there have been comparatively few family eco­ nomic enterprises in which children could work, and children have usually contributed more to the family's expenses than to its income. Rural-urban fertility differentials are no longer as significant as they once were, however. Urbanization has proceeded to the point where less than 9 per cent of the American people were still living on farms in 1960. Furthermore, the gap between the fertility of rural and urban couples has narrowed considerably since the depression years of the 1930's, because the subsequent rise in fertility has been greater in the cities and towns than in the rural communities. Greater communication between all types of communities and a high level of mobility are blurring many of the distinctive features of rural and urban life, and economic and technological changes in agriculture are eliminating the advantages of large farm families. The rural farm wives who lived outside the Standard Metropolitan Statistical Areas expected to have about 16 per cent more children than the other wives in the present study, but substantially smaller differences are found between the expectations of the large numbers of wives in the various nonfarm residential categories (Figure 15 and Table 73). The wives who lived in rural areas but not on farms did not anticipate significantly larger families than the wives who lived in cities and towns. Within the Standard Metropolitan Statistical Areas, we find that the wives in the central cities with populations of 50,000 to 149,999 thought they would have somewhat more births than the wives in the larger central cities. However, there is no differ­ ence between the expectations of the wives in the 12 large central cities and those of the wives in the smaller central cities with 150,000 or more inhabitants, and the expectations of the wives in the central

RELIGION AND SOCIOECONOMIC STATUS

cities of all sizes differ very little from those of the wives in the suburban rings. These residential differences follow about the same pattern among both Protestants and Catholics (Table 74). The rural-urban difference in expected fertility was smaller in the 1960 study than in the 1955 study, partly because of a decline SIZE OF PLACE OF RESIDENCE IN STANDARD METROPOLITAN STATISTICAL AREAS' 12 l a r g e c i * ι e s

Othercities of 150,000 + Other cities of 50,000-149,999

mm

R i n g s o f 12 l a r g e c i t i e s

Ringsofother cities of 150,000 +

NOT IN STANDARD METROPOLITAN STATISTICAL AREAS'Cities of 2,500-49,999

ΉΖΆ

Rural nonform Rural farm

AVERAGE NUMBER OF BIRTHS BIRTHS BV I960 TOTAL NUMBER EXPECTED (MOST LIKELY)

FIGURE 15 Births by 1960 and total number expected, for wives, by size of place of residence

in the expectations of rural wives, but primarily because of an increase in the expectations of urban wives, which was more pronounced among the Catholics than among the Protestants. This change can be attributed largely to the younger wives. Rural-urban differences in expected family size are considerably smaller among the wives

FERTILITY AND FAMILY PLANNING IN THE U.S.

TABLE 73 Births by 1960 and Total Number Expected, for Wives, by Size of Place of Residence Average number of births Size of place of residence

Number of wives

By 1960

Most likely

Mini­ mum

Maxi­ mum

2,414

2.3

3.1

2.9

3.5

187 243 184 443

2.1 2.2 2.3 2.2

3.0 3.0 3.3 3.1

2.8 2.7 3.0 2.8

3.3 3.3 3.7 3.4

537

2.4

3.1

2.9

3.5

282 335 203

2.2 2.4 2.7

3.0 3.1 3.6

2.8 2.8 3.3

3.3 3.5 3.9

Total

Expected total

In Standard Metropolitan Statistical Areas:

12 large cities· Other cities of 150,000+ Cities of 50,000-149,999 Rings of 12 large cities" Rings of other cities of 50,000 + Not in Standard Metropolitan Statistical Areas:

Cities of 2,500-49,999 Rural nonfarm Rural farm

• The cities are Baltimore, Boston, Chicago, Cleveland, Detroit, Los Angeles, New York, Philadelphia, Pittsburgh, St. Louis, San Francisco, and Washington.

TABLE 74 Total Number of Births Expected, for Wives, by Size of Place of Residence, by Religion Total"·

Size of place of residence Total

Protestant

Catholic

Number of wives

Average most likely expected no. of births

Average Average most likely Num- most likely expected ber expected no. of of no. of births wives births

Number of wives

2,414

3. 1

1,596

2 .9

668

3.7

187 243 184 443

3. 0 3..0 3. 3 3.1

62 143 99 236

2..8 2..8 2..9 2. 8

81 83 72 160

3.5 3.5 4.0 3.6

537

3..1

357

2..9

161

3.8

282 335 203

3 .0 3 .1 3 .6

227 291 181

2 .9 3 .0 3 .4

49 40 22

3.9 3.8 5.2

In Standard Metropolitan Statistical Areas:

12 large Citiesb Other cities of 150,000+ Cities of 50,000-149,999 Rings of 12 large Citiesb Rings of other cities of 50,000 + Not in Standard Metropolitan Statistical Areas:

Cities of 2,500-49,999 Rural nonfarm Rural farm

• Includes 150 Jewish and "other" wives. b The cities are Baltimore, Boston, Chicago, Cleveland, Detroit, Los Angeles, New York, Philadelphia, Pittsburgh, St. Louis, San Francisco, and Washington.

RELIGION AND SOCIOECONOMIC STATUS

who were under 30 years of age in 1960 than among those who were 30-39. Our measures of attitudes toward family size indicate that the rural farm wives wanted as well as expected more children than the other wives (Table 75). However, the difference between the number of children the rural farm wives wanted and the number the other wives wanted is not as great as the difference between TABLE 75 Number of Children Considered Ideal for the Average American Family, Wanted if Life Could be Relived, Wanted at Time of Interview, and Most Likely Expected, for Wives, by Size of Place of Residence Average number of children

Size of place of residence Total

Number of wives

Considered ideal for Wanted average if life Wanted at Most American could be time of likely family" relived" interview expectedb

2,414

3..4

3.7

3.,3

3..1

187 243 184 443

3..2 3..3 3..5 3. 4

3.5 3.6 3.9 3.7

3..2 3 .2 3,.4 3.,4

3..0 2,.9 3 .2 3..0

537

3. 4

3.8

3..4

3..1

282 335 203

3 .4 3 .5 3 .8

3.5 3.5 3.8

3 .0 3 .3 3 .5

3 .0 3 .1 3 .5

In Standard Metropolitan Statistical Areas:

12 large cities" Other cities of 150,000 + Cities of 50,000-149,999 Rings of 12 large cities0 Rings of other cities of 50,000+ Not in Standard Metropolitan Statistical Areas:

Cities of 2,500-49,999 Rural nonfarm Rural farm

" A small number of wives did not give numerical replies to these questions. b The expected number of children is slightly smaller than the expected number of births for some groups, because it excludes children who died prior to the interview. 0 The cities are Baltimore, Boston, Chicago, Cleveland, Detroit, Los Angeles, New York, Philadelphia, Pittsburgh, St. Louis, San Francisco, and Washington.

the numbers they expected. Rural farm wives thought they would have the same average number of children as they said they wanted to have at the time of the interview, while the wives in most of the other residential groups expected fewer children than they wanted on the average. This reflects the fact that a smaller proportion of the rural farm wives expected fewer children than they wanted, and a larger proportion expected m o r e t h a n they w a n t e d ( T a b l e 7 6 ) . The wives who lived in the 12 large cities and their suburban rings were the least likely to expect more than they wanted. We also determined where the wives had lived for the longest

FERTILITY AND FAMILY PLANNING IN THE U.S.

TABLE 76 Percentage Who Expected Fewer Children Than They Wanted, the Same Number or More, for Wives, by Size of Place of Residence Expected more than wanted0

Size of place of residence Total

ExHad pected ExnumHad Number fewer pected ber more of than number wanted than wives Total wanted0 Wantedb Total or fewer wanted 2,414

100

29

50

22

7

15

187 243 184 443

100 100 100 100

29 27 26 31

52 50 54 51

19 23 20 18

5 7 7 5

14 16 13 13

537

100

29

48

23

9

14

282 335 203

100 100 100

27 30 23

47 46 52

26 24 26

7 8 9

19 16 17

In Standard Metropolitan Statistical Areas:

12 large cities4 Other cities of 150,000 4Cities of 50,000-149,999 Rings of 12 large cities'1 Rings of other large cities of 50,000 + Not in Standard Metropolitan Statistical Areas:

Cities of 2,500-49,999 Rural nonfarm Rural farm

* The most likely expected number is smaller than the minimum number wanted. b The most likely expected number is equal to (or between) the minimum or maxi­ mum number wanted. 0 The most likely expected number is larger than the maximum number wanted. d The cities are Baltimore, Boston, Chicago, Cleveland, Detroit, Los Angeles, New York, Philadelphia, Pittsburgh, St. Louis, San Francisco, and Washington.

period of time since they were first married. Although we cannot classify them on this basis in as much detail as we can according to place of residence at the time of the interview, the longest place of residence since marriage appears to be related to expected fertility in much the same way: Longest Place of Residence Since First Marriage

Average Most Likely Expected Number of Births

Cities of 100,000 or more Other urban areas and rural nonfarm RuralFarm

3.0 3.2 3.4

The expectations for the rural farm group are somewhat lower when the longest place of residence since marriage is used, but the difference is not statistically significant. The influence of past farm residence on fertility is of great interest because of the large numbers of Americans who have moved from farms to cities and towns. Do persons who have ever lived on farms tend to have higher fertility than persons with no farm background?

RELIGION AND SOCIOECONOMIC STATUS

Our data show that the wives who had lived on farms at some time since they were first married but were not living on farms at the time of the interview, expected almost as many births as the wives who were still living on farms (Table 77). However, the earlier farm experience of neither the husband nor the wife appears to have much effect upon the expected fertility of couples that had not lived on farms after marriage. In fact, the expected number of births is somewhat lower for such couples than for couples with no farm background at all. Although there are no longer any great regional differences in the fertility of white wives, we do find that the white wives who lived in the South had borne fewer children by 1960 and expected TABLE 77 Births by 1960 and Total Number Expected, for Wives, by Fiirm Residence of Wife and/or Husband

Farm residence of wife and/or husband Total None at any time None since first marriage Wife some before, husband none Wife none before, husband some Both some before Some since first marriage, but not on farm at interview On farm at interview

Number of wives

Average number of births Expected total By Most Mini­ Maxi­ 1960 likely mum mum

2,414 1,053

2.3 2.3

3.1 3.2

2.9 2.9

3.5 3.5

269 378 316

2.3 2.1 2.3

3.1 3.0 3.0

2.8 2.7 2.7

3.5 3.3 3.4

148 250

2.7 2.6

3.4 3.5

3.0 3.2

3.7 3.8

to have smaller completed families than those who lived in the re­ mainder of the country (Figure 16 and Table 78). The differences between the actual and expected fertility of the Southern wives and the wives in the other three major regions are all statistically significant at the .95 confidence level. Nevertheless, this pattern of differences is probably due partly to sampling variability, since it is not entirely consistent with the results of the 1960 Census with respect to the number of children ever born. According to the 1960 Census, the number of children ever born per 1,000 ever-married white women who were 18-39 years old was a little lower for the Northeast than for the South. In any event, the results of the present study and the Census both indicate that the Southern white wives had borne somewhat fewer children than the white wives in the North Central and Western regions, in spite of the more rural character of the South. This can

FERTILITY AND FAMILY PLANNING IN THE U.S.

REGION OF RESIDENCE NORTHEAST NORTH CENTRAL

0

1 1 12

1 3

1 4

AVERAGE NUMBER OF BIRTHS BIRTHS BY I960 §§§§//] TOTAL NUMBER EXPECTED (MOST LIKELY) FIGURE 16 Births by 1960 and total number expected, for wives, by region of residence

be explained, to a great extent, by the fact that the rise in fertility since the 1930's has been sharper in the cities and towns than in the rural communities of the United States. In addition, Southern farm residence no longer appears to be associated with large families. The wives who lived on farms in the South thought they would have fewer children than the wives who lived on farms outside the TABLE 78 Births by 1960 and Totzd Number Expected, for Wives, by Region of Residence

Region of residence

Number of wives

Average number of births By 1960

Most likely

Minimum

Maximum

Total Northeast North Central South West

2,414 581 715 744 374

2.3 2.4 2.4 2.1 2.4

3.1 3.2 3.3 2.9 3.2

2.9 3.0 3.0 2.6 3.0

3.5 3.6 3.7 3.2 3.5

Expected total

South, and the expectations for couples with past farm experience in the South are lower than the expectations for couples with past farm experience in other regions. SUMMARY

Religious preference is very closely related to the fertility of white wives. Catholics expected about 28 per cent more children than Prot-

FERTILITY AND FAMILY PLANNING IN THE U.S.

estants and 48 per cent more than Jews. The majority of wives in each of the three major religious groups thought they would have a total of from two to four children, but within this popular range, Catholics were more likely to expect four. Moreover Catholics were almost three times as likely to expect five or more children. Although many of the traditional social and economic differentials in fertility have narrowed, the religious differential appears to have widened in recent years. The average number of births by the time of the inter­ view and the average total number of births expected for Catholics and non-Catholics differed more in the 1960 study than in the 1955 study, as a result of a rise in the Catholic averages during this five-year period. There is no significant association between most of our measures of religiousness and the expected fertility of Protestants. However, it is clear that the Catholics who reported greater involvement and inter­ est in their faith were anticipating larger families than the other Cath­ olics. Among Catholics, the expected size of completed family is di­ rectly related to the frequency of church attendance, the frequency with which the sacraments are received, the amount of education ob­ tained in church-related elementary and secondary schools, the im­ portance of religious activities in the daily life of the family, and in­ creasing interest in religious matters. Educational attainment no longer has as much influence on fer­ tility as it once did, but it is still of considerable importance. The wives with only a grade-school education expected about 12 per cent more births than those who had gone to high school for one to three years, and 23 per cent more than those whose education had extended beyond the third year of high school. There is virtually no difference, however, between the expectations of the wives who had completed four years of high school and the wives who had gone on to attend college, the two groups which together account for almost two-thirds of the sample. Our data on expected family size indicate that many of the tra­ ditional group differences in fertility may almost disappear. We find very small differences in the expectations of nonfarm wives when they are classified on the basis of husband's income and occupation. The farm wives expected about 16 per cent more children than the non­ farm wives, but the expectations of the many wives in the various nonfarm residential categories differ very little. Furthermore, there are no longer any large differences in the fertility of white wives living in the four major regions of the country. Many of the socioeconomic differentials in fertility would be some­ what larger if they did not follow different patterns within the two

FERTILITY AND FAMILY PLANNING IN THE U.S.

large religious groups. Among Protestants, for example, we find that expected fertility is inversely related to such measures of socioeconomic status as wife's education, husband's income, and husband's occupa­ tion. Among Catholics, on the other hand, the relationship tends to be almost U-shaped, with the expected size of completed family being larger for couples in the lower and higher status groups and smaller for those in the middle groups. Catholics thought they would have more children than Protestants in all socioeconomic groups, but the largest Catholic-Protestant differentials are found in the higher socio­ economic groups. Group differences in fertility arise mainly from differences in the number of children couples want and differences in their willingness and ability to control fertility. The religious differences in expected family size can apparently be attributed primarily to differences in family-size preferences. We find that the differences between the aver­ age numbers of children actually expected by the Protestant, Catholic, and Jewish wives correspond very closely to the differences between the average numbers of children they said they wanted, and there are no large religious differences in the proportion of wives expecting fewer or more children than they wanted. Most of the other group differences, in contrast, appear to be due in large measure to differ­ ences in the willingness and ability to control fertility. When we group the wives on the basis of their education, husband's income and occu­ pation, and size of place of residence, we find that they differ consider­ ably less with respect to the number of children wanted than with respect to the number actually expected. The association between edu­ cational attainment and the ability to control family size is especially strong. Only 15 per cent of the wives with a college education ex­ pected more children than they wanted as compared with 38 per cent of those with a grade-school education. We also find a strong association between the wife's work experi­ ence and fertility. The wives who had worked after their marriage had not only had fewer births by I960 than those who had not worked, but also expected significantly smaller completed families. Furthermore, the number of births by 1960 and the total number of births expected are both inversely related to the duration of the wife's work experience. The reasons for this important association are quite complex. Many wives limit the size of their families because of their desire to work; others work primarily because fecundity im­ pairments prevent their having as many children as they would like to have; and some work chiefly because of economic needs arising from the fact that they have or expected to have children.

CHAPTER 4

Fecundity Impairments Impairments of the reproductive system are fairly common among married couples in the main reproductive ages. Among white couples in our sample, with the wife 18 to 39 years old, 31 per cent have some evidence of less-than-normal ability to have children in the future. Within this group, the ability to have children varies widely. For 11 per cent, future births are impossible; for 12 per cent, there is enough evidence of impaired fecundity to make childbearing in the future seem improbable; and for 8 per cent, there is some evidence of impaired fecundity, but not enough to make future childbearing seem improbable. These findings are based on two kinds of information collected in the survey: (1) the wife's opinion of her and her husband's ability to have children in the future, frequently supplemented by the medical basis for her opinion, and (2) the wife's report of her pregnancy history and of her and her husband's use of contraception. This information was used to estimate each couple's fecundity—that is, their ability to have children in the future. The method we used to classify couples by fecundity will be described later. First, we shall describe the kind of information that the fecundity classification is based on. REPORTS OF FECUNDITY IMPAIRMENTS

All the wives were asked whether they thought that they and their husbands could have children in the future if they wanted them. Most of the wives thought they could, but a significant minority (21 per cent) thought they could not, or were uncertain whether or not they could. About half of these couples had had an operation or injury that made further childbearing impossible. Of the remainder, two-thirds (or 7 per cent of the white wives, 18-39 years old) had seen doctors to find out whether or not they could have more children. The couples who had talked with a doctor about their fecundity problems reported a wide variety of conditions responsible for their subnormal ability to have children. Among the more frequently men­ tioned are one ovary and one tube removed, repeated fetal deaths, low sperm count, blocked Fallopian tubes, tipped uterus, infantile uterus, and the onset of menopause. We accepted such reports as evidence of impaired ability to reproduce only where the medical basis for the opinion seemed sound or where the couple had not used contraception for many years and the wife had not conceived.

FERTILITY AND FAMILY PLANNING IN THE U.S.

Inasmuch as the fecundity classification relies heavily on informa­ tion provided by the wife and not by her doctor, we may well question its accuracy. Judicious suspicion is necessary in a study that attempts to represent a cross section of the population. The sample includes some wives with relatively low intelligence and little education. It also includes some women who are average in these respects, but who are overly anxious about their ability to reproduce or who exag­ gerate their infirmities. Here are some examples of inadequate replies to questions about impairments to the reproductive system: CASE 6093. A 35-year-old mother of two children reports, "I have a skin disease, 'isotherius.' They don't know what causes it, but preg­ nancy might cause death. I got it two years ago and it can recur if I go out in the sun." We asked a gynecologist whether there is such a disease, or any like it. He said that he had never heard of the disease or any like it that would make pregnancy dangerous, and checked with a derma­ tologist to verify this impression. CASE 5640. An 18-year-old woman, married for seven months, thinks that it may be impossible for her to get pregnant because she and her husband have not used contraception for two months and she has not yet conceived. Reports like the two cited above are unusual. The large majority of wives who expressed some doubt about their or their husband's abil­ ity to have children gave good reasons for their opinions. Examples of more adequate reports are the following: CASE 5661. A 39-year-old mother of one child says, "My tubes are partially deflated. I had them blown open and then I got pregnant, but have never done that since my daughter was born (13 years previously) and the doctor says that's what I need. I'm not actually trying to get pregnant. Unless I have my tubes opened, I won't get pregnant, and I don't want to go through having them blown open; it's too painful." Neither she nor her husband had ever used contraception. CASE 5653. A 27-year-old wife, married for eight years, had never been pregnant. She reported that the doctor was uncertain whether she and her husband could have children. She says, "My husband has a low sperm count. The doctor was giving him shots for a couple of weeks. The doctor said to take shots for seven or eight months, but he didn't. Others told him it would do no good."

FECUNDITY IMPAIRMENTS

Even reports like these do not form an ideal basis for classifying couples, but they help us make an approximate judgment about a couple's capacity for future childbearing, particularly when they are viewed in conjunction with pregnancy and contraceptive histories. Another source of inaccuracy in our classification of couples by fecundity is the inability of couples to report impairments of which they are not aware. We have no basis for estimating how many such cases there are in our sample. Most of them are probably among couples who have all the children they want, have used contraception for many years, and have not discovered any fecundity impairment because they have not tried to have a baby. There are also a few women who seem fairly certain that they will have more chidren and who mention the existence of a fecundity impairment in connec­ tion with another question. For example: CASE 3352. A 25-year-old wife who had two live births and a stillbirth

(which occurred in the seventh month of gestation) and who expects one or two additional children was asked why she did not expect any more than that. She said that her reasons were "mosdy financial" and also, "My husband and I have Rh factor about blood. I never understood all that, but it has something to do with our blood." It is possible, of course, that this couple's one stillbirth was not related to their Rh incompatibility, but we assumed that Rh incompat­ ibility was discovered in connection with the stillbirth. If Rh in­ compatibility already caused one stillbirth, it is quite possible that it will lead to another, if the wife becomes pregnant again. Therefore, we classified the couple as below normal in their ability to have more children. Such instances are rare. We found only four women who thought they could have more children in spite of serious fecundity impair­ ments. There were 94 other women (or 4 per cent of the white married wives, 18 to 39 years old) who thought they could have children in spite of less serious fecundity impairments, but their esti­ mate of their ability to reproduce had some justification. Fecundity impairments are not necessarily permanent. A number of wives (14 per cent of the 18-39-year-old white married women) thought at the time of interview that they could have more children but said that there was a time when they thought that they might not be able to do so. In most cases, these doubts arose before the first child was bom, and often there was little substantial basis for them. Of the wives who consulted doctors about their fears, over one-third were examined and found to have no apparent defects.

FERTILITY AND FAMILY PLANNING IN THE U.S.

In many such instances, the doctor told his patient to relax and not worry about conceiving. In other cases, however, the doctor did find something wrong, and corrected it with an operation or treatments. THE FECUNDITY CLASSIFICATION

The classification of couples by fecundity represents an attempt to estimate their ability to bear children in the future. Their past childbeaiing record is taken into account only insofar as it gives us some indication of what their future ability is likely to be. We did not try to summarize a couple's ability to have children during all of their married life. Such a classification might be useful for some purposes, but inasmuch as this study is oriented toward future childbearing, our greatest need is for some measure of future ability to have children. Thus, we have classified some couples who have had many live births as Definitely Sterile, indicating that they are unable to have any more children. At the other end of the continuum, we have classified many couples who have not yet had any children as Fecund, indicating that they are probably able to have one or more live births in the future, if they wish to do so. It must be emphasized that we are classifying couples, not wives only, by fecundity status. There are a number of cases, for example, in which the wife has no reproductive impairment, but the husband cannot produce enough normal sperm to fertilize the wife's ova. The fecundity classification system used in the present study differs from that used in the 1955 study in several respects. Although it is recognized that a change in the classification system reduces the comparability of the two studies, it seemed desirable to make certain improvements in the classification system that became feasible with the somewhat more detailed information collected in 1960. The most important improvements were to try to rank the fecundity groups along a continuum from the most impaired to the least impaired, and to eliminate the category for "indeterminate" cases that was used in the 1955 system. These and other changes are described later in the present chapter and in Appendix B, Note 3. The fecundity classification consists of five major groups defined as follows: 1. The Definitely Sterile:1 Couples in this category cannot have children in the future. A large majority of them (93 per cent) have had operations that make conception impossible. The prevalence of operations will be discussed in detail in a later section. 1The titles of the fecundity categories are capitalized in order to remind the reader that these designations are being used in a special way in this study.

FECUNDITY IMPAIRMENTS

2. The Probably Sterile: These couples are probably unable to have any children in the future. A couple is included in this category under the following conditions: (a) the wife reported an impairment that may make conception impossible, and (1) a doctor told her that she will probably not be able to have a live birth, or (2) she had not become pregnant after the onset of the con­ dition in spite of not using contraception; or (b) the couple had not used contraception for ten or more years and the wife had not become pregnant. Examples of couples classified as Probably Sterile are as follows: CASE 5315. A 29-year-old woman, married for ten years, had an ovary removed three months before she was interviewed. The doctor told her there was only a "slight chance" that she would become pregnant. She and her husband had not used contraception for seven years, and she had never been pregnant. CASE 5501. This 39-year-old woman reported that she and her husband had never used contraception, and that she had been preg­ nant only once—14 years ago. 3. The Possibly Sterile: These couples may be unable to have any children in the future, but there is less evidence that future children are unlikely than there is for couples classified as Probably Sterile. Couples are classified as Possibly Sterile under the following conditions: (a) the wife reported an impairment that may make pregnancy impossible, and (1) the doctor was uncertain whether birth can occur, or (2) the couple had not tested their fecundity sufficiently (by not using contraception) to make conception seem highly improbable; or (b) the couple had not used contraception for two to nine years if the wife had never been pregnant (or three to nine years if she had been pregnant), and she did not conceive during or after this long period.2 2 The minimum criteria of two years for the never-pregnant and three years for the ever-pregnant are used here because they were used in the 1955 study and the Indianapolis Study. In addition to these criteria, the wife must have had fewer than one birth for every three years of married life. If she had more births, she is classified as Fecund.

FERTILITY AND FAMILY PLANNING IN THE U.S.

Examples of couples classified as Possibly Sterile are: CASE 4886. This 35-year-old woman reported two births and five miscarriages. She said, "There is something that causes me trouble around the third month and causes me to lose them. He (the doctor) is giving me hormone shots to help me keep them. He is not sure yet if I am pregnant or not." CASE 4834. This 34-year-old woman had a miscarriage nine years ago. She and her husband had not used contraception since then, but she had not become pregnant. She had not seen a doctor to find out whether she could have a baby. 4. The Possibly Fecund: These couples probably can have children in the future, but there is some reason to believe that their ability to reproduce is less than normal. Couples are included in this group under the following conditions: (a) the wife reported an impairment of the reproductive system that is not serious enough to prevent childbearing, or (b) she reported a fairly serious condition, but had one or more births after the onset of the condition, or (c) the wife did not become pregnant during a period of three years or more when contraception was not used, but con­ ceived later. Examples of couples classified as Possibly Fecund are as follows: CASE 5651. This couple never used contraception during 13 years of married life, and had four children—the latest two years before the interview. The wife, who was 36 years old, reported, "My doctor thinks I've had about all I'm going to have. I'm not sure of the exact reason, but he said I'm so small on the inside he is surprised I've had any children." CASE 1390. This 23-year-old wife reported that she and her husband did not use contraception for three and a half years, between 1954 and 1958, and that she did not become pregnant. Later she had a child, and then began contraception. 5. The Fecund: These are couples for whom there is no evidence of impaired ability to have children. There is a small group of wives (10 per cent of the white wives 18-39 years old) who reported that they and their husbands had not used contraception, but that they douched for purposes of cleanli­ ness soon after intercourse. Apparently, many of these wives did not recognize the contraceptive effectiveness of douching. Some of them reported that they thought they couldn't have children because

FECUNDITY I M P A I R M E N T S

they had not been pregnant in spite of not using contraception for years. For example: CASE 4908. This 35-year-old mother had her only child 11 years ago. She would like to have two more children, and she and her husband have not used contraception for the last 10 years. Because she has not conceived, she thinks that she is not able to, although her family doctor examined her a year after her delivery and said, as she reported it, "everything was normal." She douches immediately after intercourse with plain water. Concerning this practice she says, "It might be good in some cases (to prevent pregnancy), but I never heard of it working."

In the 1955 study, such wives were classified as Indeterminate, because it was impossible to tell, from the available information, whether douching or some fecundity impairment was responsible for their not becoming pregnant. In the present study, however, we asked additional questions to find out how soon after intercourse the douch­ ing ordinarily occurs. If it occurs within one-half hour, we treat the couple the same way we would treat a couple using contraception; where no specific impairment is mentioned, we assume that the wife has not become pregnant because douche was being used, and we classify the couple as Fecund. If the wife usually douched more than half an hour after intercourse, we assume that douching did not keep her from becoming pregnant, and classify the couple in one of the other fecundity categories, depending on how much time had passed since the last pregnancy. Although the method of classifying couples by fecundity in the present study differs from that used in the 1955 study, both classi­ fication schemes show approximately the same broad picture of fecundity impairments in the United States. The main differences are that the present system tends to give a slightly lower proportion Definitely Sterile because of a stricter application of the definition for this category, and a slightly lower proportion Subfecund, because some of the cases that would have been classified as Indeterminate in the 1955 study are classified as Fecund in the 1960 study. A more detailed comparison between the 1955 and 1960 methods of classification is shown in Appendix B3 Note 3. We may think of the fecundity classification presented here as related to the probability that a couple will have a live birth following a period when contraception is not used. This may be defined as the probability of conceiving within a given period of noncontraceptive exposure (or "fecundability") multiplied by the probability of having a live birth after conception occurs. These measures were used in the Princeton Study to express the reproductive ability of

FERTILITY AND FAMILY PLANNING IN THE U.S.

all the couples, as a group, but not of individual couples.3 For the Definitely Sterile in the present study, the probability of having another child is zero. For the Probably Sterile, the probability is above zero, but very slight; a few of the couples in this group may be able to have a child in the future. The Possibly Sterile have a greater probability of having children in the future, and the Possibly Fecund are almost as able as the Fecund to do so. We would like TABLE 79 Number and Per Cent Distribution by Fecundity, for White and Nonwhite Couples

Fecundity

Number of couples

Total Subfecund: Definitely Sterile Probably Sterile Possibly Sterile Possibly Fecund Fecund

2,684 830 274 105 235 216 1,854

Total Subfecund: Definitely Sterile Probably Sterile Possibly Sterile Possibly Fecund Fecund

2,414 740 257 85 199 199 1,674

Total Subfecund: Definitely Sterile Probably Sterile Possibly Sterile Possibly Fecund Fecund

270 90 17 20 36 17 180

Per cent of all couples Total 100 31 10 4 9 8 69 White 100 31 11 4 8 8 69 Nonwhite 100 33 6 7 13 6 67

Per cent of Subfecund couples

100 33 13 28 26

100 35 11 27 27

100 19 22 40 19

to assign probability values to the intermediate fecundity categories, but have no basis for estimating them. Therefore, we shall have to remain content with verbal statements of the probabilities. The number and proportion of couples in each of the five fecundity categories are shown in Table 79. Proportions are shown 8Westoff 1961, pp. 45-69. (See page xxxi for the meaning of any abbreviated reference.) It should be noted that there was little need for estimating the fecundity of each couple in the Princeton Study, because it was reasonably certain that all of the couples interviewed in this study were able to have more children. AU of them had had two births, the last occurred about six months before interview, and only six per cent were uncertain whether they could have more children (Ibid., p. 140).

FECUNDITY I M P A I R M E N T S

in Figure 17. Altogether, 31 per cent of the white couples with the wife 18-39 years old have been classified as below normal in fecundity. Throughout the book, we shall refer to such couples as the Subfecund, but it should be remembered that this group includes couples who differ greatly in their ability to have children. One-third of them, the Definitely Sterile, cannot have children in the future. At the other end of the scale, the Possibly Fecund, who constitute one-quarter of the Subfecund, probably can have children in the future, but at a lower rate than the Fecund. Some consideration was given to including the Possibly Fecund with the Fecund, but this idea was

FECUND

POSSIBLY

FIGURE 17 Per cent distribution by fecundity, white couples

rejected for two reasons. First, by leaving the Possibly Fecund with the Subfecund, we keep the Fecund group "pure"—that is, we are as certain as we can be, given the limitations of our data, that a very large majority of the Fecund are able to have children in the future. Second, by leaving the Possibly Fecund with the Subfecund, we compensate (to an unknown extent) for the presence of hidden subfecundity among couples classified as Fecund. As the system of classification now stands, there are some couples who should be taken out of the Fecund group and some who should be added to it. Those who should be taken out are couples who are below normal in fecundity but who do not know this because they have not tried to have children during the past few years; since we cannot find

FERTILITY AND FAMILY PLANNING IN THE U.S.

them, we cannot classify them as Subfecund. Those who should be added to the Fecund group are the couples now classified as Sub­ fecund who actually are normal or nearly normal in their ability to have children. If we added the latter group, but failed to subtract the former group, the resulting proportion of couples classified as Fecund would surely be an overestimate. In order to avoid estimates that we know would overstate the proportion Fecund, we decided not to include the Possibly Fecund with the Fecund. Nonwhites have about the same proportion Subfecund as white couples (Table 79). The small difference between the two groups is not significant. However, nonwhites have a smaller proportion Definitely Sterile (reflecting fewer operations) and greater proportions Probably and Possibly Sterile. The latter differentials may be due to the fact that nonwhites use contraception less than whites, and, as will be shown later, lesser use of contraception makes the discovery of fecundity impairments more likely. In later sections of this chapter, we shall see how the proportions of couples in the fecundity categories differ for various groups in the population, and how fecundity impairments affect fertility. In the next section, however, we will deal with one unique and important aspect of fecundity—the prevalence of operations that make pregnancy impossible. OPERATIONS THAT PREVENT PREGNANCY

Operations that make pregnancy impossible are a common cause of sterility among couples in the reproductive ages. For 10 per cent of the white couples, with the wife 18 to 39 years old, either the husband or wife had undergone such an operation. This is one-third of all Subfecund couples. In over half of these cases (57 per cent) the operations were performed in order to make conception impossi­ ble; because such operations serve as a substitute for contraception, we shall refer to them as "contraceptive operations." In the remaining cases, the operations were performed in order to correct diseased conditions of the reproductive system, such as cancer, and not for the purpose of preventing conception; we shall refer to these as "remedial operations." Although there are more couples with contraceptive than remedial operations in the sample, we cannot say with certainty that contra­ ceptive operations are more numerous in the population represented by the sample. The difference between the percentage of couples with contraceptive operations (5.6 per cent) and remedial operations (4.2 per cent) is not statistically significant. Contraceptive operations are apparently more common than remedial operations if the wife

FECUNDITY IMPAIRMENTS

is under 35 or has been married less than 15 years (Table 80 and Figure 18). But if she is 35-39 or married over 15 years, the two types of operations are almost equally common. A large majority of the wives reporting operations gave enough information to classify them as contraceptive or remedial in intent. TABLE 80 Percentage of Couples with Contraceptive and Remedial Operations, by Wife's Age, Duration of Marriage, and Number of Births

Characteristic Total Wife's age: 18-24 25-29 30-34 35-39 Years married: Under 5 5-9 10-14 15 or more Number of births: None 1 2

Number of couples

All operations that prevent Contraceptive pregnancy operations

2,414

10

6

513 600 624 677

2

2

1

5 12 18

4 7 9

1 4 9

544 649 702 519

1 5 11 23

1 4 7 12

1 2 4 11

301 463

9 4 8 10 17 19

1 1 4 7 14 14

8 3 4 3 4 6

682

3 4 5 or more

Remedial operations

499 263 206

4

WIFE'S AGE

18-24 YEARS

25-29 YEARS

30-34 YEARS

35-33 YEARS

1 O PERCENT

PERCENT ALL TYPES OF OPERATIONS

CONTRACEPTIVE OPERATIONS

1

5 IO PERCENT

REMEDIAL OPERATIONS

FIGURE 18 Percentage of couples who had contraceptive and remedial operations, by wife's age

FERTILITY AND FAMILY PLANNING IN THE U.S.

Ambiguous information was given in only a few cases (about 3 per cent of all the operations reported). Some of the wives may have reported the purpose of the operation incorrectly. For example, they may have said that it was performed to treat a pathological condition of the reproductive system, although it was actually performed in order to prevent pregnancy. Such misreporting was probably rare, in view of the fact that the kind of operation performed is generally consistent with the purpose of the TABLE 81 Per Cent Distribution by Reason for Operation, for Couples with Operations that Prevent Conception, by Type of Operation

Total"

Hysterectomy

Fallopian tubes cut or tied

Vasectomyb

Number of couples

238

84

85

53

Per cent: Total

100

100

100

100

43 5 38 57

94 12 82 6

Reason for operation

Remedial Cancer Other pathology Contraceptive Wife should not have any more cesareans Pregnancy would en­ danger wife's health in other ways Eugenic reasons Economic reasons Child-care reasons Other

8

28 1 8 4 8

—

5 — '

1 — —

14 —

— •—

14 86

100

20

2

52 1 5 2 6

28 4 25 15 26

—

• The toal includes 13 wives and 1 husband with operations other than those speci­ fied, and two couples for whom it is not known whether wife or husband had the operation. b The number of reported vasectomies is 45. In addition, 8 husbands were reported to have had an operation for the purpose of preventing conception, and the kind of operation was not specified. It is assumed that all of these operations were vasectomies.

operation (Table 81 and Figure 19). Most of the hysterectomies (94 per cent) were remedial in intent. A woman is not likely to have a hysterectomy simply to prevent pregnancy. Most of the tubal ligations and all of the vasectomies were contraceptive in intent. Tubal ligations are occasionally performed to correct a pathological condition of the Fallopian tubes, but this use of the operation is rare. Vasec­ tomies are never performed for the purpose of relieving a diseased condition, although they are sometimes performed on older men dur­ ing an operation for the removal of the prostate gland in order to prevent infections of the epididymes and testicles.

FECUNDITY IMPAIRMENTS

As we shall see, different factors affect the incidence of contra­ ceptive and remedial operations. For that reason the two kinds of operations are treated separately in the two subsections following. Contraceptive Operations —From one point of view, contraceptive operations should not be classified as fecundity impairments. Because they are performed voluntarily to prevent further childbearing, they are a substitute for contraception and could be included in a list of contraceptive methods. However, it seems somewhat more reason­ able to adhere strictly to the concept of a fecundity impairment as KINO OF OPERATION

ALL OPERATIONS

HYSTERECTOMY

'//////////////////////////////ZZZA msum

TUBES CUT OR TIED

VASECTOMY

—ι— 20

Γ

40,

T-

.60

PERCENT CONTRACEPTIVE OPERATIONS:

η IOO

I 80

PREGNANCY WOULD ENDANGER WIFE'S HEALTH DON'TWANTMORE CHILDREN FOR ECONOMIC REASONS DON'T WANT MORE CHILDREN FOR OTHER REASONS REMEDIAL OPERATIONS

ΙΟΟΦΟΗ [////[

FIGURE 19 Percentage distribution by reason for operation, by kind of operation

any physiological limitation on the ability of a couple to have children in the future. This view is supported by the fact that contraceptive operations are unlike contraception insofar as their effects are irreversi­ ble. In other words, if a couple who uses contraception wants to have a child, they can stop using the method and, in the absence of a serious impairment, the wife can conceive. However, a couple with a contraceptive operation cannot usually do anything to have more children. In only a minority of cases can the effect of the operation be reversed. The predominant reason given for having a contraceptive opera­ tion is to protect the wife's health, often with a doctor's recommenda­ tion. This is the reason given for 63 per cent of such operations. For example:

FERTILITY AND FAMILY PLANNING IN THE U.S.

CASE 4711. This woman had her tubes cut and tied when she was

25 years old after her first pregnancy ended in a therapeutic abortion. Her doctor recommended the operation because she had a rheumatic heart condition and he thought that pregnancy would endanger her health. CASE 6901. This woman had her tubes tied when she was 31 years old, soon after the delivery of her fifth child. She said, "I have had pneumonia nine times in nine years, also an asthmatic condition. I have extremely difficult labor for 36 to 48 hours, and the lung condition during these labors could kill me."

One of the health reasons frequently cited for preventing preg­ nancy by means of a contraceptive operation is that the wife has had one or more cesarean births, and her doctor does not think that she should have any more. Fourteen per cent of all contraceptive operations and nearly one-quarter of those that were tubal ligations were performed for this reason. One of these cases is cited below: CASE 4725. A 25-year-old mother of three children had her tubes

cut and tied at the delivery of her last child. She said this was done "because it was dangerous for me to have more children. My three came by cesarean section, and my doctor said that in my case three was enough." About one-fifth of the wives reporting operations for contraceptive purposes said that the operation was performed because she and her husband felt they could not take care of any more children, or because they did not want any more. For example: CASE 4704. This wife reported a vasectomy on her husband which

was performed because, "We had all we wanted when we got our boy." The couple has two girls and a boy. CASE 4719. A mother of three boys and a girl said that her husband

had a vasectomy because, "We already had three boys to educate and put through college—all we felt financially able to educate." CASE 5554. This 32-year-old woman had no children. She said her husband had a vasectomy because they didn't want any children. The interviewer noted, "She was very emphatic in telling me she did not want any children at all. She said her career meant more to her than children could. She also said she and her husband had a very active social life and were seldom home."

FECUNDITY IMPAIRMENTS

A very similar distribution of reasons for having contraceptive operations is reported in a survey of women in Puerto Rico. According to this study, 60 per cent of the women gave health as the reason for being sterilized and 23 per cent cited economic reasons.4 The corresponding proportions in our study are 63 and 21 per cent, respectively. The reason for having a contraceptive operation often influences whether the husband or the wife has the operation. In cases where the motive is to protect the wife's health, the operation is usually performed on the wife. Where there are other reasons for wanting to avoid conception, the husband is more likely to have the operation. Contraceptive operations are very rare among couples with no births or only one. Three of the couples with contraceptive operations had no births; two because they had not wanted any births (one is Case No. 5554, cited above) and one because pregnancy would have endangered the wife's health. Four of the couples with contra­ ceptive operations had only one birth. Among all couples with no birth or one, only 1 per cent had a contraceptive operation. The proportion with contraceptive operations increases as the number of births becomes greater until it reaches 14 per cent among couples who had five or more births. As would be expected, the proportion with contraceptive operations also rises with age and dura­ tion of marriage (Table 80). Because the prevalence of contraceptive operations increases with parity, couples with such operations tend to have had more births than other couples. On the average, the couples with contraceptive operations have had 3.6 births, compared to 2.3 births for all couples. However, the average number of births for couples with contraceptive operations is only moderately above the total of 3.1 births expected by all couples. This comparison shows that contraceptive operations are by no means limited to couples with unusually large families. In fact, three-quarters of the couples with contraceptive operations had two to four births, which is the range desired and expected by most couples. It might be supposed that couples who have contraceptive opera­ tions for economic, child-care, or other nonhealth reasons have had more births than couples with operations intended to protect the wife from pregnancies that would endanger her health. However, this is not the case. In fact, couples with nonhealth reasons for contra­ ceptive operations had fewer births (3.4) than those with health reasons (3.6). The difference, however, is not statistically significant. It is evident, then, that the couples who seek contraceptive operations 4 HiU

1959, p. 181.

FERTILITY AND FAMILY PLANNING IN THE U.S.

because they do not want to take care of more children do not have extremely large families. Only five out of fifty such couples had six or more children. Even though their families are moderate in size, couples with contraceptive operations are more commonly found among those who have had accidental conceptions and among couples who have had more pregnancies than they want. The prevalence of contraceptive operations is 11 per cent among couples who had accidental con­ ceptions, and 4 per cent among couples who used contraception with-

WlFE S RELIGION

PROTESTANT

CATHOLIC

JEWISH

OTHER

T 0 PERCENT ALL TYPES OFOPERATIONS

5 IO PERCENT

CONTRACEPTIVE OPERATIONS

Ο

"Ι­ 5 IO PERCENT

REMEDIAL OPERATIONS

FIGURE 20 Percentage of couples who had contraceptive and remedial operations, by wife's religion

out an accident (Table 82). It is 15 per cent among couples who had more pregnancies than they wanted, and 4 per cent among other couples. Altogether, half of the couples with contraceptive oper­ ations reported accidental pregnancies, unwanted pregnancies, or both. Thus, the prevalence of contraceptive operations is clearly related to the ability to limit fertility with the use of contraception. Those who have demonstrated their inability to control conception are ap­ proximately three times as likely as other couples to have contraceptive operations. Few couples rely on contraceptive operations as the only means of limiting family size. Most of those with such operations (82 per cent) had used contraception before sterilization made further use unnecessary.

FECUNDITY IMPAIRMENTS TABLE 82 Percentage of Couples with Contraceptive and Remedial Operations, by Use of Contraception, whether Couple Had an Accidental Conception, and Pregnancy Planning Status

Characteristic Total Use of contraception Users Had accidents· Had no accidents' Nonusers Pregnancy planning statusb Completely planned Partially planned Excess pregnancies

Number of couples

All operations that prevent Contraceptive pregnancy operations

Remedial operations

2,414

10

6

4

1,948 423 1,525 466

8 13 7 17

6 11 4 5

2 1 3 12

515 1,486 413

5 9 19

3 4 15

3 5 4

8 An accidental conception is defined here as one that occurred while some method of contraception was being used regularly. b Detailed definitions of the planning status categories are given in Chapter 6. Briefly, couples with "completely planned" fertility planned each pregnancy by stop­ ping contraception in order to conceive, or used contraception regularly and never conceived; those with "partially planned" fertility had not had more pregnancies than they wanted; those with "excess pregnancies" did not want their last conception when it occurred, or later.

Among the socioeconomic variables, religious preference shows one of the strongest relationships to the prevalence of contraceptive operations. The proportion of wives reporting contraceptive operations is much lower among Catholics (2 per cent) than among Protestants (8 per cent) (Table 83 and Figure 20). This is consistent with the explicit prohibition by the Catholic Church of operations intended solely to prevent pregnancy. As stated in a semiofficial guide for use in Catholic hospitals: "Procedures that induce sterility, whether permanent or temporary, are permitted when: (a) they are immediately directed to the cure, diminution, or prevention of a serious pathological condition; (b) a simpler remedy is not reasonably available; and (c) the sterility itself is an unintended and, in the circumstances, an unavoidable effect."5 Our data show that there are very few Catholics who do not follow this principle. Because contraceptive operations are so rare among Catholics, we reviewed the 12 interviews with Catholic wives in which such 8 Catholic

Hospital Association 1959, p. 6.

FERTILITY AND FAMILY PLANNING IN THE U.S.

operations were reported in order to see whether these couples had any distinctive characteristic in common that might account for their nonconformist behavior. For example, were they nominal, nonpracticing Catholics? Was the spouse who had the operation non-Catholic at the time the operation was performed? Did these couples favor the use of methods of contraception that the Church disapproved of? TABLE 83 Percentage of Couples with Contraceptive and Remedial Operations, by Wife's Religion, Education, Region of Residence, and Size of Place of Residence

Characteristic Total Wife's religion: Protestant Catholic Jewish Other Wife's education: College High school, 4 High school, 1-3 Grade school Region of residence: Northeast North Central South West Size of place of residence: Twelve central cities" Other cities 50,000 + Rings of 12 central cities Rings of other cities Cities of 2,500-49,999 Rural nonfarm Rural farm

Number of couples

AH operations that prevent Contraceptive Remedial pregnancy operations operations

2,414

10

6

4

1,596 668 106 44

12 5 6 9

8 2 3 2

5 3 3 7

427 1,153 579 255

4 9 14 15

3 5 8 8

2 4 5 7

581 715 744 374

5 8 13 15

2 4 7 10

3 4 5 5

187 427 443 537 282 335 203

5 7 9 10 12 14 13

2 4 5 5 6 10 8

3 4 3 4 6 5 5

* The cities are Baltimore, Boston, Chicago, Cleveland, Detroit, Los Angeles, New York, Philadelphia, Pittsburgh, St. Louis, San Francisco, and Washington.

The 12 Catholic wives reporting contraceptive operations seemed to be at least as devout as other Catholic wives. Seven attended church once a week, four attended two to three times a month, and one never went to church. Eleven had always been Catholic; only one was a convert. Nine of the twelve wives favored family limitation, and seven had used methods of contraception that were not approved by the Church. Thus, they may be somewhat less orthodox than other

FECUNDITY IMPAIRMENTS

Catholics with respect to family planning methods. Their number of births was not unusual—3.1 on the average. The reasons given for the operations were similar to the reasons mentioned by the Protestant wives. The predominant reason was the mother's health (mentioned in nine cases). Nine of the operations were tubal ligations; three were vasectomies. In one case of vasectomy, the husband was Protestant. Among the characteristics available for study the only one that most of these couples have in common is a lack of education in Catholic schools. Ten of the twelve wives had never attended a Catholic school. This is unusual in view of the fact that among all Catholic wives 44 per cent had some of their education in Catholic schools. Only one wife showed any awareness of a conflict between a contraceptive operation and the teachings of the Church: CASE 5512. A 33-year-old mother of one child said she had her tubes tied, "because the doctor said pregnancy did not agree with me. I was in the hospital almost the whole nine months. I had blood transfusions and intravenous feedings." She told the interviewer that she and her husband were leaving the Catholic Church because of trouble over her operation.

Among the few (106) Jewish wives in our sample, only three (or 2.8 per cent) reported contraceptive operations. The proportion is well below that for Protestants, but the number of Jewish wives interviewed is so small that we cannot say with certainty that they are less likely than Protestant wives to report such operations. Yet, a low incidence of contraceptive operations among Jewish couples is consistent with their higher educational attainment. Approximately half of the Jewish wives interviewed had gone to college, as compared with only one-sixth of the Protestant wives. As we shall see, the better educated are less likely than the less educated to have contra­ ceptive operations. This relationship may account for the relatively low proportion of Jewish wives reporting them. The proportion of couples with contraceptive operations varies considerably between the different regions of the United States (Figure 21). Very few (2 per cent) of the wives living in the North­ east (the New England States, New York, New Jersey, and Pennsyl­ vania) reported contraceptive operations on themselves or their husbands. None of the wives living in New York City or Boston did so. (This does not mean that there are no couples in these cities who have had a contraceptive operation, but only that the sample

FERTILITY AND FAMILY PLANNING IN THE U.S.

was so small that there was a very low probability of finding one or more such couples among the small proportion that exists.) The North Central States have a higher proportion with contraceptive operations (4 per cent) and again, none were found in some of the large cities: Cleveland, Detroit, and St. Louis. The proportion is still higher in the South (7 per cent), and highest in the West (10 per cent). The regional differences in the prevalence of contraceptive opera­ tions are even more impressive when we compare couples among REGION OF RESIDENCE NORTHEAST

NORTH CENTRAL

SOUTH

WEST

O

1 IO

PERCENT

PERCENT

CONTRACEPTIVE OPERATIONS

REMEDIAL OPERATIONS

PERCENT ALL TYPES OF OPERATIONS

"I 5

FIGURE 21 Percentage of couples who had contraceptive or remedial operations, by region of residence

whom such operations are most heavily concentrated—that is, couples with Protestant wives 30-39 years old. The percentages of such couples with contraceptive operations are shown below. Region of residence

United States Northeast North Central South West

Per cent of Protestant wives 30 to 39 reporting contraceptive operations

11 5 8 12

19

Contraceptive operations appear to be a fairly common means of preventing the birth of additional children among Protestants Kving in the West and the South.

FECUNDITY IMPAIRMENTS

In the South, the high proportion of contraceptive operations reflects a larger proportion of wives who undergo tubal ligations (and, less commonly, other operations) to prevent pregnancy for health reasons. Altogether, 5 per cent of the wives in the South re­ ported that they had contraceptive operations because pregnancy would endanger their health, and in many cases they reported that a doctor had recommended the operation. The proportion is sig­ nificantly higher for the South than for the remainder of the country (2 per cent) (Table 84). TABLE 84 Percentage of Couples with Contraceptive Operations, by Region of Residence, by Spouse on Whom Operation was Performed, by Reason for Operation Spouse on whom operation was performed and reason for operation Number of couples Percentage with contraceptive operations: On wife' On husbandb To prevent pregnancy for health reasons: On wife On husband To prevent pregnancy for non-health reasons: On wife On husband

Total

North­ east

North Central

South

West

2,414

581

715

744

374

6 3 2

2 2

4 3 1

7 6 2

10 3 7

3 2

6 5 1

4 2 2

1 1 1

6 1 5

4 3 1 2 1 2

—

1 1 —

1 — —

—

2 1 1

» Most of these operations (88 per cent) are tubal ligations. b Probably all of these operations are vasectomies; 85 per cent are reported to be vasectomies, and 15 per cent are unspecified. However, it is virtually certain that all of the latter operations are also vasectomies.

The greater use made of tubal ligations in the South than in other regions may be due in part to the lesser influence of the Catholic Church on medical practice in the hospitals of the South. As noted previously, the Catholic Church does not permit operations for contra­ ceptive purposes to be performed in Catholic hospitals, and there are relatively fewer Catholic hospitals in the South than in other regions. This explanation suggests that the low proportion of couples with contraceptive operations found in the Northeast is due largely to the influence of the Catholic Church. This is plausible, but it must be recognized that it is a speculative explanation. In the West, the greater prevalence of contraceptive operations reflects primarily a higher proportion of husbands who undergo

FERTILITY AND FAMILY PLANNING IN THE U.S.

vasectomy to prevent pregnancy for economic or other nonhealth reasons. The proportion is 5 per cent, compared to 1 per cent in the remainder of the country. The difference is statistically significant (Table 84). We do not know why vasectomies appear to be more popular in the West than elsewhere. Inasmuch as vasectomies are simple opera­ tions that can be performed in a doctor's office, regional differences in hospital policies toward sterilizing operations would be less likely to affect the incidence of vasectomies than tubal ligations. It is possible WIFE S EDUCATION COLLEGE

HIGH SCHOOL,4

HIGH SCHOOL, 1-3

GRADE SCHOOL

-ι O

Γ

5 IO PERCENT

Ο

"Τ­ 5 IO PERCENT

ALL TYPES OF

CONTRACEPTIVE

OPERATIONS

OPERATIONS

1 O

1

5 IO PERCENT

REMEDIAL OPERATIONS

FIGURE 22 Percentage of couples who had contraceptive and remedial operations, by wife's education

that one or more of the western areas chosen for the sample contains an unusually high proportion of husbands with vasectomies. Most of the contraceptive operations in the West are concentrated in three of the seven areas sampled, and it may be that these three areas overrepresent the true incidence of contraceptive operations in the West. This is not a likely explanation, however, for even after sampling errors are adjusted to allow for the effect of "clustering,"8 the prev­ alence of vasectomies in the West is significantly higher than it is in the remainder of the country. Apparently many couples in the West want a very effective method of preventing pregnancy after " This is an allowance for the slightly increased probability of finding similar respondents that occurs when the places of residence included in the sample are chosen so that they will be close together rather than randomly distributed over a large area.

FECUNDITY IMPAIRMENTS

they have had all the children they want, and their doctors are willing to help them by performing vasectomies. The proportion of couples with contraceptive operations varies inversely with educational attainment (Figure 22). Of the wives whose education proceeded no further than grade school or who had only 1 to 3 years of high school, 8 per cent reported contraceptive operations; in contrast, only 3 per cent of the wives who had been to college reported such operations (Table 85). Only one college graduate reported a contraceptive operation. TABLE 85 Percentage of Couples with Contraceptive Operations, by Region of Residence, by Wife's Religion and Education, and Size of Place of Residence

Characteristic

Total

North­ east

North Central

South

West

All couples Wife's religion: Protestant Catholic Other Wife's education: College High school, 4 High school, 1-3 Grade school Size of place: Cities 50,000 and over Smaller cities and rings Rural farm and nonfarm

6

2

4

7

10

8 2 3

3 2

6 1 3

8 3

14 4

*

*

2 2 4

3 3 6 10

2 7 12 8

7 10 15 10

1 2 4

1 5 6

5 6 11

8 11 12

3 5 8 8 3 6 9

—

—

The relationship between the prevalence of contraceptive oper­ ations and educational attainment cannot be explained by variations in the effectiveness of contraception or in the proportion of couples with excess pregnancies. It is true that fewer of the less-educated couples use contraception, that those who use it do so less effectively, and that the less educated are more likely than others to have unwanted pregnancies. But none of these factors accounts for the fact that they are much more likely than other couples to have contracep­ tive operations. None of the other demographic or socioeconomic factors used in the study explains a significant part of this relationship. We can only speculate about the reasons for this relationship, but it seems reasonable to suspect that doctors are more likely to recommend contraceptive operations to couples with less education than to other couples. Presumably, doctors would be likely to trust

FERTILITY AND FAMILY PLANNING IN THE U.S.

a well-educated couple to use contraception effectively, if they were told that pregnancy would endanger the wife's health. However, they would not be as confident of a less-educated couple's ability to prevent pregnancy. Although this explanation is speculative, it is consistent with the fact that less-educated wives are more likely than the better-educated to cite health reasons for contraceptive operations. Very few of the wives with little education said that the operation was performed for financial or child-care reasons. Only two women with a gradeschool education gave such a reason. Although we do not know in how many cases the doctor recommended an operation and in how many cases the wife or husband asked for the operation, various statements in the interviews suggest that in most of the cases where pregnancy would endanger the wife's health, the operation was recom­ mended by a doctor. Hill, Stycos, and Back in their study of fertility in Puerto Rico, also emphasized the role of the physician in recommending contra­ ceptive operations to less-educated women. They report that many of the physicians, ". . . feel discouraged about the effectiveness of mechanical and chemical contraceptives for a relatively uneducated population and see sterilization as the only effective solution in the near future."7 Although there are many important social and economic differences between the United States and Puerto Rico, the attitudes of doctors toward their less-educated patients may be similar. Wives living in cities of 50,000 or more are least likely to report contraceptive operations, and those living in rural areas are most likely to do so. This is consistent with the relationship of contraceptive operations to religion (since rural areas have a liigher proportion of Protestants), to region of residence (since over half of the couples living in rural areas are in the South), and to education (since rural areas have a higher proportion with little education). However, none of these characteristics accounts completely for the difference between city and countryside in the proportion of couples with contraceptive operations. We do not know whether it is due to differences in medical practice or to different tendencies of rural and city couples to ask to be sterilized. Wives of farmers are more Ukely to report contraceptive operations than are other wives. This, of course, is consistent with the higher proportion reporting such operations in rural areas. Wives of skilled workers ("upper blue collar") are almost as likely as wives of farmers to report contraceptive operations. The differences are small between wives whose husbands have other occupations. 'Hill 1959, p. 181.

FECUNDITY IMPAIRMENTS

In summary, the contraceptive operation is found most commonly among the less educated and in rural areas. We are not sure why this is so, but we think that it may be due to a tendency of doctors to recommend such operations to couples who may not use contra­ ception effectively when they believe that pregnancy would endanger the wife's health. Apparently there is an upward trend in the proportion of couples with all types of operations that prevent pregnancy. In 1955, 9 per cent of the wives reported such operations (contraceptive and remedial combined); in 1960, 10 per cent did so. Although the increase be­ tween 1955 and 1960 is not statistically significant, it arises entirely as a result of a significant increase in the proportion of husbands who had operations. The data are as follows: Spouse on whom operation was performed

Percentage of couples 1955 I960

Differenceb (1960 minus 1955)

8.7» 9.9 Husband and/or wife 1.2 7.6 Wife 7.8 -.2 2.2 Husband .9 1.4 • The number of couples with operations shown in the report of the 1955 study is 245, or 9.0 per cent of the sample. In order to insure consistency with the definitions used in the 1960 study, 10 of these cases were reclassified because there was some evidence that the operation did not absolutely prevent pregnancy. The remaining 235 cases represent 8.7 per cent of the 1955 sample. b Rounded independently.

Data drawn solely from the present study confirm the fact that the proportion of husbands with operations was very low in 1955. Among women in the 1960 sample who were also eligible for the 1955 sample (i.e., they were 18-39 years old, married, and living with husband in 1955), the proportion reporting that their husbands had had an operation by 1955 was 0.9 per cent—exactly the same proportion as found for the 1955 sample. Although the proportion of husbands with operations is low in both years, it is more than twice as high in 1960 as it was in 1955. This rise reflects an increase in the popularity of the vasectomy as a method of birth control. (All but one of the male operations reported in the present study are vasectomies intended to prevent pregnancy.) We do not know whether the rise reflects a greater tendency on the part of doctors to recommend vasectomies, or a greater tendency on the part of patients to ask for them. The proportion of couples with all types of contraceptive opera­ tions may have increased. Again referring to women in the 1960 study who were 18-39, married, and living with husband in 1955, we find that 4 per cent of them reported a contraceptive operation

F E R T I L I T Y AND F A M I L Y PLANNING IN T H E U.S.

by 1955. Among the wives 18-39 years old in 1960, the proportion with contraceptive operations is 6 per cent. Comparisons for wives classified by age (Table 86) suggest that this increase has been con­ fined largely to older couples. We are not entirely sure that an increase in all types of contra­ ceptive operations has occurred, because the 1955 and 1960 propor­ tions of 4 and 6 per cent are not based on independent samples (if they were, the rise would be statistically significant), but are based, in part, on reports of the same women—i.e., wives who were married and 18-34 years old as of 1955 and five years older, or 23-39 years old in 1960. The main reason for believing that contraceptive opera­ tions have become more common is the significant rise in the proporTABLE 86 Percentage of Couples with Contraceptive and Remedial Operations, by Wife's Age, 1955 and 1960 Wife's age

1955»

All operations 1955 1960

Total

9

8

10

4

6

4

4

20-24 25-29 30-34 35-39

2 6 10 17

2 6 10 14

3 5 12 18

1 4 5 5

2 4 7 9

2 5 8

1 1 4 9

Contraceptive 1955 1960

Remedial 1955 1960

• This is for the 1955 sample. All other figures are for the 1960 sample.

tion of husbands with operations between the independently drawn 1955 and 1960 samples. If there is an increase in the proportion of couples with contra­ ceptive operations, it is probably due to an improved understanding of the nature and effects of such operations. People are probably more aware than previously that tubal ligations and vasectomies are relatively simple and very safe surgical procedures, that they cause neither impotence nor a decline in sexual desire, and that they afford virtually perfect protection against unwanted pregnancies. Remedial Operations—Among white wives 18—39 years old, 4 per cent reported operations that prevented pregnancy but were per­ formed for the purpose of correcting a pathological condition. The prevention of pregnancy was a side effect of these operations, rather than the primary purpose. These "remedial" operations were some­ what less numerous than contraceptive operations, discussed in the preceding section. The following case is an example of a remedial operation:

FECUNDITY IMPAIRMENTS

CASE 5550. When she was 22 years old and had been married one year, this wife had a hysterectomy. She said, "I thought I was preg­ nant, but it was a tumor." Later she and her husband adopted a child, and at the time of interview they were about to adopt another. A large majority (77 per cent) of the remedial operations are hysterectomies. Most of the remainder involved the removal of diseased Fallopian tubes or ovaries. Nearly all of them were on the wife; only one remedial operation was reported for a husband. Few of the operations were performed to remove cancerous tissue—only 11 per cent of all remedial operations, which represents only one-half of 1 per cent of the white wives 18 to 39 years old. In contrast with couples having contraceptive operations, those with remedial operations had relatively few births. About a quarter had no births, as in Case 5550, cited above. About half had the two to four births expected and desired by most wives. On the average, those with remedial operations had 2.1 births, which is well below the final number of 3.1 expected for all couples. The proportion of wives with remedial operations is highest (8 per cent) among wives with no births, but the prevalence of such operations does not appear to be closely related to number of births for couples who have had one or more children. On the other hand, the proportion with remedial operations rises sharply with age and duration of marriage, reaching 9 per cent for wives 35-39 and 11 per cent for those married at least fifteen years (Table 80). The prevalence of remedial operations is not related to success in limiting fertility. Wives who have had accidental pregnancies or excess pregnancies are no more likely than other wives to report remedial operations (Table 82). This contrasts sharply with the situ­ ation for contraceptive operations, which are closely related to success in family planning. The fact that remedial operations are no more likely to be reported by unsuccessful planners than successful planners reinforces the opinion that the purpose of the operation has been accurately reported in a large majority of cases. If there were any tendency to say that the purpose of an operation was remedial even though it was actually contraceptive, we would probably find a higher proportion of remedial operations among couples with accidental and excess pregnancies. Remedial operations are more common among couples who have not used contraception (12 per cent) than among those who have (2 per cent). The higher incidence among Nonusers8 is due to the 8 Throughout this book, couples who have ever used contraception are referred to as Users, and those who have never used contraception are referred to as Nonusers. See Chapter 5 for definitions.

FERTILITY AND FAMILY PLANNING IN THE U.S.

fact that for many of the couples in this group, the operation occurred before it became necessary to limit family size by using contraception. Catholics may be less likely than Protestants to have operations that sterilize but are purely remedial in intent (Figure 20). We are not sure that this is true because the difference between the two religious groups could be due to sampling variability. The data are as follows: Wife's religion

Number of couples

ProtestEint Catholic

1,596 668

Percentage Number of with remedial remedial operations operations

77 19

4.8 2.8

If the difference between the percentages were 2.3 or greater, we could be reasonably certain (with a probability of 95 in 100 of being YEARS MARRIED

UNDER 5

5-9

10-14

15 OR MORE

I

10

ι

20

r~

PERCENT

30

—I—

"Ί—

40

50

BASIS OF CLASSIFICATION: OPERATION RECOGNIZED PATHOLOGY LONG PERIODS OF NONCONTRACEPTIVE EXPOSURE

FIGURE 23 Percentage of couples classified as Subfecund, by basis of classification, by duration of marriage

right) that in the population represented by our sample, Catholic wives are less likely to report remedial operations on themselves or their husbands than are Protestant wives.9 However, the actual differ­ ence of 2.0 percentage points leaves us uncertain. It is possible that * The estimate of sampling error used here allows for the fact that the sample is not strictly random, but "clustered." See Appendix A.

FECUNDITY IMPAIRMENTS

there is a smaller likelihood of remedial operations among Catholics in view of the care taken in Catholic hospitals to prevent an operation that sterilizes the patient unless it is the only reasonably available form of treatment that will remove the pathology. That such care is taken is suggested by the following discussion of the morality of sterilizing operations for the treatment of excessive uterine bleeding, written by a Catholic interpreter of moral law: [The moralist] ". . . must be very careful when giving general answers or when solving particular cases that involve unknown physicians; for there are some doctors who have what I might term a 'sterilising mentality.' They believe that certain classes of patients should be sterilised. Yet they realise that a conscientious Catholic woman will not permit this, and they also know that they will not be permitted to perform a patently sterilising operation in a Catholic hospital. Consequently, under the guise of attacking pathology, they recommend treatments or operations which produce the desired result of sterilisation though under another name."10 Reverend Kelly summarizes the Catholic position with respect to such operations as follows: "Hysterectomy, oophorectomy, or suppression of ovarian function by irradiation may be allowed to remedy uterine bleeding when such bleeding is a source of serious detriment to health and when less extreme remedies are not reasonably available. The patient's consent should be had; and both patient and doctor should sincerely wish to remove the pathology and not merely seek an excuse for a contra­ ceptive measure."11 Like Catholics, Jewish wives also report a low proportion of remedial operations, but the number of Jewish wives is so small that there is no basis for believing that they are less likely than Protestant wives to have such operations (Table 83). UnKke contraceptive operations, remedial operations do not show a strong relationship to region of residence (Figure 21). A higher proportion of wives report remedial operations in the South and the West than in the Northeast and North Central States, but the differ­ ences are not statistically significant (Table 83). Similarly, there are no close relationships between the prevalence of remedial operations and size of place of residence or husband's occupation. The socioeconomic characteristic that is most strongly related to 10KeIly 11 Kelly

1955, pp. 48-49. 1955, pp. 54-55.

FERTILITY AND FAMILY PLANNING IN THE U.S.

the prevalence of remedial operations is educational attainment (Table 83 and Figure 22). This finding is difficult to account for. It does not seem probable that less-educated couples are more likely to have impairments of the reproductive system severe enough to make surgery advisable. Perhaps the less-educated are somewhat more exposed to the risk of infections of the reproductive system, but such infections are not ordinarily treated surgically. In the 1955 study, the prevalence of contraceptive and remedial operations combined varied inversely with the wife's educational at­ tainment. No attempt was made in the earlier study to discover the reasons for the operations, so we cannot make the contraceptiveremedial distinction. Consequently, we don't know whether the inverse relationship in 1955 was due to variations in the prevalence of contra­ ceptive or remedial operations or of both kinds. We know only that the 1955 data do not contradict the 1960 data. A possible explanation for the higher proportion reporting re­ medial operations among the less-educated is that these wives may tend to say that an operation was performed entirely for noncontraceptive reasons, even though the operation was contraceptive in intent. To test this hypothesis, at least superficially, we reviewed all of the interviews with women with an eighth-grade education or less who reported having had remedial operations. However, there are no cases in which it seems likely that the intent of the operation was contra­ ceptive rather than remedial. Even though less-educated wives are no more likely than others to have diseases of the reproductive system, perhaps they do not have these conditions treated early enough to avoid surgery. Or per­ haps doctors tend to recommend operations to the less-educated wives because they think that they may not follow through on longer-term treatments that would require frequent visits to the doctor's office or to a clinic. Whatever the explanation is, the present study does not supply the kind of information needed to bring it out. There appears to be no definite trend in the proportion of couples with remedial operations. Among women in the 1960 sample who were 18-39 years old, married, and living with a husband in 1955, the proportion reporting a remedial operation by 1955 is 4.1 per cent. Among wives who were 18-39 in 1960, the proportion with such operations by 1960 is almost identical: 4.2 per cent. VARIATIONS IN SUBFECUNDITY

In order to understand the important factors underlying variations in the proportion of couples who have fecundity impairments, it is useful to divide the Subfecund into three broad groups:

FECUNDITY IMPAIRMENTS

(1) Couples who have had operations that prevent pregnancy; (2) Couples with recognized pathologies that may prevent con­ ception or make it difficult; (3) Couples who have no recognized pathologies, but who have had long periods of noncontraceptive exposure (i.e., they did not use contraception for a long time and the wife did not conceive). Each of these groups accounts for about one-third of the Subfecund. The factors associated with variations in the first group (couples with operations) have been discussed in the preceding section. In this section we shall be concerned primarily with variations in the over-all proportion Subfecund and in the proportions in the two latter components of the Subfecund. Variations in the proportion in the two latter categories emphasize the fact that it takes time and exposure to the risk of conception to discover fecundity impairments. This is especially true of the third category, which contains couples with long periods of noncontraceptive exposure. None of these couples would have been able to meet the requirements for inclusion among the Subfecund if they had not had a period of noncontraceptive exposure of at least three years (if ever pregnant) or two years (if never pregnant). The longer a couple has been married, the greater is the possibility that they have had such a period of noncontraceptive exposure. Consequently, the propor­ tion of couples in this group varies more closely with duration of marriage than with age (Table 87). The proportion rises from 2 per cent for those married under 5 years to 17 per cent for those married 15 or more years (Figure 23). There is a comparable rise with age (from 4 per cent for couples with the wife 18-24 years old to 17 per cent at ages 35-39), but within each age group below age 35 there is a distinct rise with duration of marriage. (The most obvious irregularity in the table, the 22 per cent at ages 35-39 for couples married 5 to 9 years, is probably a random variation due to sampling variability; the proportion is based on only 37 couples.) Duration of marriage also affects the proportion of couples who find that they have recognizable pathological conditions (blocked tubes, low sperm count, etc.) that make it difficult or impossible to have children in the future. However, duration of marriage is not as closely associated with this category of fecundity impairment as it is with the category based solely on long periods of noncontra­ ceptive exposure. The fact that the discovery of impairments often requires not only time, but also exposure to the risk of conception, leads us to

F E R T I L I T Y AND FAMILY PLANNING IN T H E U.S.

the conclusion that Nonusers should have a higher proportion Subfecund than Users. This is true, but there is a second factor that affects the relationship between fecundity and use of contraception—simply that the Subfecund have less need for contraception than the Fecund. In other words, Nonusers are more likely than Users to discover that they are Subfecund, and after they have discovered their condiTABLE 87 Percentage of Couples Classified as Subfecund because of Operation Preventing Pregnancy, Recognized Pathology, or Long Period of Noncontraceptive Exposure, by Wife's Age, by Duration of Marriage Years married Wife's age

Total

Total 18-24 25-29 30-34 35-39

31 13 21 36 47

Total 18-24 25-29 30-34 35-39

10 2 5 12 18

Total 18-24 25-29 30-34 35-39

10 7 9 12 12

Total 18-24 25-29 30-34 35-39

11 4 7 13 17

Under 5

5-9

10-14

Subfecund, total 24 38 * 24 20 38 28 36 * 39 49 Operation preventing pregnancy 1 11 5 * 1 6 1 12 5 6 11 6 * 11 5

10 10 7 25

Recognized pathology 12 10 * 7 12 9 12 12 * 22 11 Long period of noncontraceptive exposure 2 15 8 * 2 9 2 14 6 9 13 10 * 22 17 6 7 4 9

15 or more

52 * ; #

51 52 23 * *

25 23 12 * *

11 12 17 * *

15 17

tion they have less need to use contraception than do Fecund couples. Because both of these factors operate, Nonusers have a much higher proportion Subfecund than Users: 60 per cent, as compared to 24 per cent (Table 88). Among couples who have had many years of noncontraceptive exposure (Nonusers married 15 or more years), the proportion Subfecund is very high: 85 per cent. Not only do Nonusers have a higher prevalence of subfecundity than Users, but even among Users we find that those who began

FECUNDITY IMPAIRMENTS

the use of contraception later have greater proportions Subfecund. For example, among older couples (wife 30-39) who used contra­ ception since the beginning of marriage, 28 per cent are Subfecund. The Subfecund gradually rises as resort to contraception is postponed until it reaches 44 per cent for those who began use following the fourth pregnancy (Table 89). The relationship between the use of contraception, fecundity, and socioeconomic variables is illustrated by an analysis of the differences TABLE 88 Percentage of Couples Classified as Subfecund, by Duration of Marriage, by Use of Contraception Years married

Total

Users

Nonusers

Total Under 5 5-9 10-14 15 or more

31 10 24 38 52

24 5 17 30 42

60 25 64 74 85

TABLE 89 Percentage of Couples Classified as Subfecund, by Wife's Age, by Number of Conceptions before Contraception Was Begun Number of conceptions before contraception was begun Total Total 0 1 2 3 4 5 or more

24 20 22 29 34 41 44

Wife's age 18-29 13 10 11 17 23 36 *

30-39 33 28 33 38 41 44 42

in fecundity between Protestants and Catholics. The proportion Sub­ fecund is nearly the same for both religious groups (approximately 31 per cent), but the proportions in the particular Subfecund cate­ gories are quite different (Table 90). The proportion Definitely Sterile is over twice as high for Protestants as for Catholics, and the propor­ tions in other Subfecund categories are lower for Protestants. We have already discussed the factors associated with differences in the proportions Definitely Sterile in the preceding section: Protestants are more likely than Catholics to have operations that prevent con­ ception. But why are Catholics more likely than Protestants to be in the other Subfecund categories? The answer lies in the differential

FERTILITY AND FAMILY PLANNING IN THE U.S.

between the two religious groups in the use of contraception. If we eliminate couples classified as Definitely Sterile and subdivide the re­ maining (nonsterile) couples by their use of contraception, we see that the proportion Subfecund is nearly the same for Protestants and TABLE 90 Per Cent Distribution by Fecundity, for Couples by Wife's Religion and Education

Wife's religion and education

Num­ ber of couples

Total 2,414 Wife's religion: Protestant 1,596 Catholic 668 Other 150 Wife's education: 427 College High school, 4 1,153 High school, 1--3 579 Grade school 255

Subfecund Defi­ Proba­ Pos­ Pos­ sibly nitely bly sibly Total Sterile Sterile Sterile Fecund

Total

Fecund

100

69

31

11

4

8

8

100 100 100

69 70 75

31 30 25

13 5 7

3 4 4

7 12 9

8 9 5

100 100 100 100

78 72 62 60

22 28 38 40

5 10 14 15

2 3 4 7

8 8 8 11

6 7 12 7

TABLE 91 Per Cent Distribution by Fecundity, for Non-Sterile Couples, by Wife's Religion, by Use of Contraception

Wife's religion

Number of couples

Total

Total Protestant Catholic Other

2,157 1,385 632 140

100 100 100 100

Total Protestant Catholic Other

1,786 1,205 451 130

100 100 100 100

Total Protestant Catholic Other

371 180 181 10

100 100 100 100

Fecund

Total

Subfecund Probably Possibly Possibly Sterile Sterile Fecund

All nonsterile couples 78 4 22 79 21 4 74 26 4 80 20 4 Users 83 2 17 84 2 16 82 2 18 83 17 3 Nonusers 50 13 50 49 51 14 51 10 49 *

*

*

9 8 12 10

9 9 10 6

6 5 8 8

9 9 8 5

25 24 25

12 12 13

*

*

Catholics in each use category (Table 91). More Catholics are classi­ fied as Subfecund simply because more of them are Nonusers. The same kinds of relationships account for different proportions with fecundity impairments between different education groups.

FECUNDITY IMPAIRMENTS

Couples in which the wife has only a grade-school education have a very high proportion Subfecund (40 per cent) because such couples are more likely than others to have operations, as we saw in the preceding section, and less likely to use contraception, as we shall see in the next Chapter. There is some tendency for the less-educated to have higher proportions Subfecund among Users (Table 92), but the Users in these groups have made less use of contraception (in that they began use later in marriage and used contraception less TABLE 92 Per Cent Distribution by Fecundity, for Non-Sterile Couples, by Wife's Education, by Use of Contraception

Wife's education

Number of couples

Total

Total College High school, 4 High school, 1-3 Grade school

2,157 404 1,038 498 217

100 100 100 100 100

Total College High school, 4 High school, 1-3 Grade school

1,786 360 879 397 150

100 100 100 100 100

Total College High school, 4 High school, 1-3 Grade school

371 44 159 101 67

100 100 100 100 100

Subfecund

Probably Possibly Possibly Fecund Total Sterile Sterile Fecund All nonsterile couples 22 78 4 83 17 2 79 21 4 72 28 4 71 29 8 Users 83 17 2 14 86 2 15 85 2 78 22 2 19 3 81 Nonusers 50 50 13 61 39 7 53 47 12 51 49 12 52 48 19

9 9 9 9 12

9 6 8 14 8

6 6 5 7 7

9 7 7 13 8

25 30 28 18 24

12 2 13 19 9

regularly) than have Users among the better educated. The low pro­ portion Subfecund among the Nonusers who had gone to college is due largely to the fact that most of these wives had been married only a short time. Nonuse of contraception is very rare in the college group and is limited largely to the recently married. In general, the differences in fecundity between social and economic groups are due largely to differences in contraceptive prac­ tices, which make the discovery of impairments more likely for some groups than for others. There is no evidence that socioeconomic groups differ in important degree in the inherent ability of their members to reproduce.

FERTILITY AND FAMILY PLANNING IN THE U.S.

In socioeconomic groups where it is not customary to begin using contraception early in marriage, we find many couples who rely solely on fecundity impairments to limit their fertility. For example, among wives 35-39 years old, 28 per cent of those with a grade-school educa­ tion had a serious impairment (Definitely, Probably, or Possibly Sterile) and had never used contraception; the comparable proportion for college-educated wives is 10 per cent. The proportions for Catholic and Protestant wives in this cohort group are 24 and 15 per cent, respectively. Similar differences are found between all socioeconomic groups that differ in the extent to which they use contraception. Such differences emphasize an important fact about the control of fertility in the United States: nearly all couples have or intend to have some limitation on their fertility before the end of the childbearing period. The limitation may be voluntarily imposed by the use of contraception or a contraceptive operation, or it may be in­ voluntarily acquired through some pathological condition; but regard­ less of its nature, it is nearly always present. TRENDS IN FECUNDITY, 1955 ΤΟ 1960 Has the prevalence of fecundity impairments changed in the years between the 1955 and 1960 studies, or has it remained relatively constant? If we were referring to the inherent ability of couples to reproduce, we would expect only a small change over a five-year period. Perhaps improved health standards and better treatment of fecundity impairments would result in a slight increase in the ability to have children, but the rise would probably not be noticeable in a small sample. However, we are not referring to the inherent ability of couples to reproduce, but to fecundity impairments as operationally defined in this study. As we have seen, the prevalence of such impair­ ments is subject to various influences, the most important being the use of contraception and the tendency to have operations that prevent pregnancy. Have such influences produced any significant changes over a five-year period? An approximate answer to this question can be obtained from Tables 93 and 94. These tables compare distributions for two over­ lapping components of the 1960 sample. The 1960 distributions by fecundity are for white couples with wives 18-39 years old; the 1955 distributions relate to the 1955 fecundity status of white couples with wives 18-39 years old in 1955 who were interviewed in 1960, when they were 23-44. Data from the present study are used for both 1955 and 1960 in order to insure that the two distributions are comparable. First, the comparisons show the slight increase in the proportion

FECUNDITY IMPAIRMENTS

with contraceptive operations noted previously. Again, the rise is not great enough to be statistically significant, but evidence cited in an earlier section suggests that it is a real increase due to a greater proportion of husbands with vasectomies. Second, Table 94 shows a small increase in the proportion of couples classified as Subfecund on the basis of the wife's report of some pathological condition (from 8 per cent to 10 per cent). It is TABLE 93 Per Cent Distribution by Fecundity, for Couples by Wife's Age, 1955 and 1960

Wife's age

Number of couples

Total

Total 18-19 20-24 25-29 30-34 35-39

2,414 73 440 600 624 677

100 100 100 100 100 100

69 93 86 79 64 53

Total 18-19 20-24 25-29 30-34 35-39

2,411 97 471 599 670 574

100 100 100 100 100 100

72 95 89 78 65 55

Fecund

Subfecund Definitely Probably Possibly Possibly Total Sterile Sterile Sterile Fecund 1960

Total 18-19 20-24 25-29 30-34 35-39

-2 -2 -4 1 -1 -2

31 7 14 21 36 47

11 —

3 6 13 19 1955"

28 8 — 5 11 2 22 6 35 11 45 15 1960 minus 1955b 2 2 4 -1 1 2

4 —

1 2 3 7 3 — —

2 4 7

2

—

—

—

1 —

2 4

1 —

-1 —

8 1 5 5 9 13

8 5 5 8 11 8

9 3 5 7 10 16]

7 2 3 7 10 8

-1 -2 —

-2 -1 -3

1 3 1 1 1 1

• The data for 1955 related to the 1955 age and fecundity of women who were inter­ viewed in 1960 and were married and living with husband in 1955. b Computed from unrounded percentages.

impossible to tell whether this difference is real or not. It may be due to a tendency for wives who reported such a condition as of 1960 to say that they did not have it five years ago, even though some of them did, but there is no evidence with which this hypothesis can be tested. It may also be due to a greater tendency for couples who suspect that they have impairments to go to doctors to see what is wrong. Whatever the cause, the rise is found in each age group. Finally, Table 94 shows a decrease in the proportion of couples classified as Subfecund on the basis of long periods of noncontraceptive

FERTILITY AND FAMILY PLANNING IN THE U.S.

exposure. However, the drop is slight and not statistically significant. A decline in this proportion is consistent with the greater use of contraception found in 1960 than in 1955, inasmuch as Nonusers are more likely than Users to have such long periods, as was demon­ strated in the preceding section. TABLE 94 Per Cent Distribution by Evidence of Subfecundity, for Couples by Wife's Age, 1955 and 1960

Wife's age

Total

Total 18-19 20-24 25-29 30-34 35-39

100 100 100 100 100 100

69 93 86 79 64 53

Total 18-19 20-24 25-29 30-34 35-39

100 100 100 100 100 100

72 95 89 78 65 55

Total 18-19 20-24 25-29 30-34 35-39

Fecund

-2 -2 -4 1 -1 -2

Subfecund

31 7 14 21 36 47

Basis for classifying couple as Subfecund Contra­ Reme­ Patho­ Nonceptive dial logical contra­ oper­ operacon­ ceptive ation ation dition exposure 1960 6 —

2 4 7 9 1955» 4

4 —

1 1 4 9

28 4 — — 5 — 11 1 22 4 2 35 5 5 5 45 8 1960 minus 1955b — 2 2 — •— 2 — 1 4 •— -1 -1 1 3 -1 3 2 1

10 5 7 9 12 12

11 1 4 7 13 17

8 2 4 8 10 10

13 3 5 9 15 22

2 3 3 1 2 3

-2 -2 -1 -2 -2 -5

a The data for 1955 relate to the 1955 age and fecundity of women who were inter­ viewed in 1960 and were married and living with husband in 1955. b Computed from unrounded percentages.

The changes described above, which are partially compensating, increase the proportions Subfecund by two percentage points between 1955 and 1960. This amount is not statistically significant. FECUNDITY IMPAIRMENTS AND CHILDLESSNESS

Nearly all American couples now in the childbearing ages have a strong aversion to childlessness. Only 1 per cent of the 18-39 year-old wives said that they would rather not have any children,

FECUNDITY IMPAIRMENTS

and some of these women already felt overburdened with families they considered excessively large. The proportion of young couples who will remain childless throughout marriage, however, will probably be in the neighborhood of 6 to 8 per cent. (It is impossible to state an exact figure inasmuch as an unknown proportion of couples with no children will have one or more.) The major cause of this childless­ ness is the existence of fecundity impairments. Voluntary childlessness is nearly extinct. Such a low prevalence of childlessness represents a return to a condition that probably existed in the United States over a century ago. The earliest cohorts of native white women for which we have any data (women born in 1836—45, approximately) had 9 per cent childless among the ever-married, according to reports for surviving women in the 1910 Census.12 The proportion childless increased until it reached 20 per cent in the cohorts of 1901-05.13 Since then, the proportion has declined rapidly. According to the 1960 Census, the percentage childless among ever-married white women had already dropped to 10 per cent for women who were 30-34 in 1959 (the cohorts of 1926-30, approximately). This is close to data from the present study, which show slightly over 8 per cent childless among the cohorts of 1926-30. Younger cohorts may have even lower propor­ tions childless by the end of the childbearing period. Although fecundity impairments are now the chief cause of child­ lessness, this was not always true in the United States. Apparendy the rise from 9 to 20 per cent was brought about chiefly by the deliberate prevention of pregnancy and not by any increase in the proportion of couples with fecundity impairments. We have some evidence for this statement from the 1941 study of a sample of Indianapolis Protestant white couples married 12 to 14 years, with most of the wives born in 1901-10. Nineteen per cent of these couples had never had a child (approximately the same proportion as that for comparable wives in the entire country) and between one-third and one-half of these couples were childless by design rather than by accident.14 If fecundity impairments had been the sole cause of childlessness, the proportion of couples with no children would have been between 9 and 13 per cent. It is difficult to say with any assurance how low the proportion childless will become in the United States. We can get some idea from the data in Table 95 which shows the proportion of couples who are childless and whether or not they have any fecundity impair13 Census

1940, Table 4, p. 9. proportion is for ever-married white women. Census 1960a. Table 190. " Indianapolis Study, Vol. 2, p. 336. 13This

FERTILITY AND FAMILY PLANNING IN THE U.S.

ments, classified by the wife's age. The proportion who are childless and have fecundity impairments is highest for ages 35-39, where it is 8 per cent. In the next younger age group (30-34) the proportion childless is already down to 8 per cent for the Fecund and Subfecund combined. We may consider 8 per cent as an upper limit on the prevalence of childlessness that will be achieved by white couples in the United States. The actual proportion will probably be lower but we can only speculate about how much lower. Extremely low proportions with no children have been reported for certain population groups. For example, Tietze found only 2.4 per cent childless among 209 TABLE 95 Percentage of Couples Who Are Childless and in Specified Fecundity Categories, by Wife's Age Subfecund

Wife's age

Number of couples

Total

Fecund

Total

Defi­ nitely Sterile

Total 18-24 25-29 30-34 35-39

2,414 513 600 624 677

12 24 10 8 10

7 20 6 3 1

6 4 4 5 8

2 1 1 2 3

Prob­ ably Sterile

Possi­ bly Sterile

2

2 2 2 2 2

—

1 1 3

Possi­ bly Fecund 1 —

— •—

Hutterite women who had married before age 25, had spent 20 years or more with the same husband, and had not used contraception.15 Among ever-married women 30 to 44 years old living in rural-farm areas of Utah, the proportion childless is only 3.1 per cent, according to the 1960 Census.16 Proportions between 3 and 5 per cent are shown by the 1910 Census for ever-married women over 40 who were born in Russia or Poland.17 Although a proportion of 3 per cent childless may never be reached by the population of the United States as a whole, it is apparent that the proportion is falling, and may become lower than any previously recorded for the country,. The low prevalence of childlessness in the United States is due in part to the fact that a majority of couples marry while the wife is under 21 and the husband under 23, and do not begin using contraception until at least one child is born. Thus, the couple is exposed to the possibility of conception at ages that are relatively uTietze

1957, p. 90. Census 1960b, pp. 46-187. " Grabill 1959, p. 63. 16

FECUNDITY IMPAIRMENTS

free from fecundity problems. Among couples with the wife 18 or 19 years old in the present study, only 7 per cent had any fecundity impairments, and these were so minor that most of them were classified as Possibly Fecund. The influence of age at marriage on the proportion of couples with no children is striking. Among wives in the 1921-25 cohorts, the proportion childless for those who married at ages 21 or less (61 per cent of all wives) is 6 per cent, for those marrying at ages 22 to 25 it is 11 per cent, and for those marrying at ages 25 and over it is 20 per cent. The fecundity impairments that are most closely associated with childlessness are those of couples classified as Probably Sterile. Nearly TABLE 96 Percentage of Couples Who Are Childless, by Fecundity, by Wife's Age

Fecundity

Total

18-24

Total Fecund Subfecund Definitely Sterile Contraceptive operation Remedial operation Other Probably Sterile Possibly Sterile Possibly Fecund

12 10 18 16 2 25 74» 45 23 4

24 23 33

a Based

Wife's age 25-29 30-34

*

10 7 18 14

8 5 14 15

*

—

—

*

*

25

35-39 10 3 18 16 2 24

*

*

*

*

*

*

50 20

30 4

38 20 1

47 15 —

on 19 couples.

half of such wives are childless (Table 96). Operations on the repro­ ductive system lead to some childlessness, but only if the purpose of the operation is to correct a pathological condition. Almost none of the couples who had operations for contraceptive purposes have no children—only three out of 136, or 2 per cent. Among couples with less severe fecundity impairments (the Possibly Sterile), nearly a quarter are childless, and among those with minor impairments (the Possibly Fecund), only 4 per cent are childless. (The proportion is so low for the latter group partly because the definition of a large component requires the couple to become pregnant after the discovery of an impairment or after a long period of noncontraceptive ex­ posure.) Among Fecund couples in the latter part of the childbearing period (wife aged 30-44), only 4 per cent are childless, and some of these wives expect to have children.

F E R T I L I T Y AND F A M I L Y PLANNING IN T H E U . S .

If the tendencies to marry young and to conceive soon after mar­ riage continue, very few couples will have serious fecundity impair­ ments before the first child is born. FECUNDITY IMPAIRMENTS AND FAILURE TO REACH DESIRED FAMILY SIZE

The most common effect of fecundity impairments is not to keep couples childless, but to prevent them from having as many children as they want. The most likely number of living children expected for couples who expect fewer children than wanted is below the minimum number of children that the wife says she wants.18 In spite of the rise in average family size in the United States, over a quarter (29 per cent) of the white wives 18 to 39 years old expect to have fewer children than they would like to have. Fecundity impairments are a major factor in keeping fertility lower than desired for these couples, but they are not the only factor. Four­ teen per cent of the wives expect to have fewer children than they want in spite of the fact that they are Fecund. Most of these wives want one more child than other wives (four children rather than three, on the average). The proportion of wives who expect fewer children than they want and are Fecund is approximately constant by age, varying only between 12 and 16 per cent (Table 97). However, the proportion of wives who expect fewer children than they want and are Subfecund increases rapidly with age: 6 per cent at ages 18-24 and 26 per cent at ages 35-39. Thus, near the end of the childbearing period, about 40 per cent of the wives have fewer children than they want, and the majority of them are prevented by fecundity impairments from having additional births. Different kinds and degrees of fecundity impairments result in different proportions expecting fewer children than they want. Among the wives reporting contraceptive operations on themselves or their husbands, only a quarter expect fewer children than they want. This "We used the number of children the wife wants rather than the number the husband wants or a combination of the numbers for wife and husband, because we regard the wife's report of her own desires as more reliable than her report of her husband's desires. In addition, fewer wives failed to report their own desires than those of their husbands. There is some evidence that number of children wanted overstates the number of children that the couple would have if husband and wife agreed on the number wanted and if they had perfect control of their fertility (Chapter 2). This leads to an overstatement of the proportion of couples expecting fewer children than wanted. However, there is no evidence that this proportion is more exaggerated for Subfecund than for Fecund couples; both groups have the same average number wanted, 3.3 (Table 100).

FECUNDITY IMPAIRMENTS

is, of course, consistent with the purpose of the operation. The wives who reported such operations and said that they expected fewer chil­ dren than they wanted were predominantly those whose doctors ad­ vised them not to have any more children for health reasons. Among the wives reporting remedial operations that resulted in sterility, two-thirds said that they wanted more children (Table 98). TABLE 97 Percentage of Wives Who Expect Fewer Children than Wanted and Are in Specified Fecundity Categories, by Wife's Age Subfecund

Wife's age

Number of couples

Total

Fecund

Total 18-24 25-29 30-34 35-39

2,414 513 600 624 677

29 22 22 29 40

14 16 13 12 14

Total

Defi­ nitely Sterile

Prob­ ably Sterile

Possi­ bly Sterile

Possi­ bly Fecund

15 6 9 16 26

5 1 2 6 10

3 1 2 2 5

4 3 3 4 7

3 2 3 3 3

TABLE 98 Percentage of Wives Who Expect Fewer Children than Wanted, by Fecundity, by Wife's Age

Fecundity

Total

18-24

Total Fecund Subfecund Definitely Sterile Contraceptive operation Remedial operation Other Probably Sterile Possibly Sterile Possibly Fecund

29 20 49 47 26 67 89» 80 54 32

22 18 46

a Based

* *

Wife's age 25-29 30-34 22 16 42 29 17

29 19 45 51 27 79

35-39 40 27 54 52 33 65

*

*

*

*

*

*

*

*

59 36

57 32

71 50 28

82 54 36

on 19 couples.

Thus, most of these wives were compelled, for reasons of health, to stop childbearing before they wanted to. Aside from operations that prevent pregnancy, the more severe the fecundity impairments, the higher the proportion of wives who expect fewer children than they want. The most limiting effects of fecundity impairments are felt by couples classified as Definitely Sterile who did not have operations. There are only 19 wives in this group; 17 of them (or 89 per cent) did not have as many children as they wanted, and 14 of them had never borne a child. Most of

FERTILITY AND FAMILY PLANNING IN THE U.S.

the couples in this group were Subfecund at the time of marriage. Since only three of them had ever used contraception, there is no evidence that contraception was responsible for the extremely low fertility of this group. Also, these couples were slightly younger at the time of marriage than the average white couple, so we know that late age at marriage did not contribute to their low fertility. Almost as severe in their limitation on fertility are the impairments of couples classified as Probably Sterile; 80 per cent of the wives in this group had fewer children than they wanted. Among the Possibly Sterile, the proportion is considerably lower (54 per cent), and among the Possibly Fecund only 32 per cent of the wives expected to have fewer children than they wanted. Since the proportion of such wives among the Fecund is 20 per cent, fecundity impairments may be said to contribute heavily to the proportion expecting fewer children than desired when the impairment is relatively severe and when it does not result from an operation performed solely to prevent preg­ nancy. FECUNDITY IMPAIRMENTS AND AVERAGE NUMBER OF BIRTHS

It is generally agreed that in the United States and other countries with moderate or low birth rates, fecundity impairments have con­ siderably less effect on fertility than does the use of contraception. On the baas of the 1955 study, we estimated that if all the couples who had been married 15 or more years had never had any fecundity impairments and had never used contraception, they would have had an average of 9.16 births. Their actual average number of births was only 2.91, however. This reduction of 68 per cent was due to the combined effect of contraception and fecundity impairments. Since Fecund couples married 15 or more years had an average of 3.52 births, which was a reduction of 62 per cent below the potential of 9.16, we inferred that the use of contraception was much more important than fecundity impairments in keeping fertility below a hypothetical maximum value.19 A repetition of the analysis on data from the present survey shows essentially the same results. A very similar evaluation of the relative effects of contraception and subfecundity has been made on the basis of data from the Indianapolis Study.20 Although, as these studies show, fecundity impairments are not as important as contraception in keeping the average number of births well below a theoretical maximum, they are important in relation to the average number of children born and the average number 19 FPSPG, pp. 238-242 and 424-427. " Indianapolis Study, vol. 2, pp. 303-357.

FECUNDITY IMPAIRMENTS

wanted. As Table 99 shows, the average number of children expected by the Subfecund (2.5) is about one-quarter below the average num­ ber wanted (3.3). The deviation from the number wanted is very small for those with contraceptive operations, much larger (almost two children) for couples with remedial operations, and largest (almost three children) for the few couples who are Definitely Sterile because of a pathological condition (Figure 24). From that point, the deviation of expected from wanted children becomes smaller as im­ pairments become less severe. The Possibly Fecund expect to have TABLE 99 Average Number of Births by Interview, Average Number of Births and Children Expected, Average Number of Children Wanted by Wife, and Per Cent that Number of Children Expected Is Above or Below Number Wanted, for Couples by Fecundity

Fecundity Total Fecund Subfecund Definitely Sterile Contraceptive operation Remedial oper­ ation Other Probably Sterile Possibly Sterile Possibly Fecund

Number of couples

Average number of children Average number Per of births cent Most Most Exexpected By likely likely Wanted pected above or ex­ inter­ ex­ minus below by view pected pected wife wanted wanted

2,414 1,674 740 257

2.3 2.4 2.2 2.8

3.1 3.4 2.6 2.8

3.1 3.3 2.5 2.7

3.3 3.3 3.3 3.6

-.8 -.9

-24 -25

136

3.6

3.6

3.4

3.5

-.1

-3

102 19 85 199 199

2.1 .8 1.1 1.7 2.4

2.1 .8 1.3 2.3 3.1

2.0 .7 1.2 2.2 3.1

3.7 3.6 3.1 3.2 3.3

-1.7 -2.9 -1.9 -1.0 -.2

-46 -81 -61 -31 -6

-.2 —

-6 —

almost as many children as they want. These variations are in the same direction as those described previously for the proportion of couples who are childless and the proportion who expect fewer children than they want. In contrast to the Subfecund, the Fecund expect to have about as many children as they want on the average. Some Fecund couples expect more than they want and some expect fewer, but these groups tend to compensate for each other. The effect of fecundity impairments on the fertility of all white couples with wives 18-39 is to keep the number of children expected about 6 per cent below the number wanted. However, the effect of fecundity impairments on the comparison between numbers of

FERTILITY AND FAMILY PLANNING IN THE U.S.

children expected and wanted varies with age, as is shown in Table 100. These comparisons indicate that the effect of impairments is greatest in the oldest age group, where the discrepancy between the number of children expected (2.9) and wanted (3.4) for all couples is 15 per cent. These results give us an approximate indication of the effects of fecundity impairments on fertility. A more careful analysis yields FECUNDITY ALL COUPLES

CONTRACEPTIVE OPERATION REMEDIAL OPERATION

Υ7777777777ΖΛ

OTHER DEFINITELY STERILE

γζζζζζζζζζζζζζζζζζζζζλ

PROBABLY STERILE

Υ////////////Λ

POSSI BLY STERILE

ν//////////7λ

POSSIBLY FECUND

FECUND

I

T

Γ

2

3

AVERAGE NUMBER OF CHILDREN EXPECTED (MOST LIKELY) V//A WANTED BY WIFE

FIGURE 24 Average number of children expected and wanted by wife, by couple's fecundity

nearly the same results. The objective of this analysis is to estimate the number of births Subfecund couples would have had by interview and the number they would expect if they had had no impairments. These estimates were then added to the number of births by interview and the number expected by the Fecund to obtain a hypothetical number for all couples. The latter totals were compared with observed births and expectations to indicate the extent to which fecundity impairments reduced fertility. The following assumptions were used to derive estimates of the

FECUNDITY IMPAIRMENTS

number of births the Subfecund would have had and expected if they had had no impairments. 1. The ratio of children by interview to children wanted is the same for Subfecund as for Fecund wives in each age group. In other words, both groups are assumed to have reached the same stage of family growth by a given age. TABLE 100 Average Number of Children Expected and Wanted by Wife, for Couples by Fecundity, by Wife's Age All couples

Wife's age Total 18-24 25-29 30-34 35-39

Fecund couples

Subfecund couples

ChilExChilExChilExdren Chil- pected dren Chil- pected dren Chil- pected dren minus ex­ dren minus dren minus ex­ ex­ pected wanted wanted pected wanted wanted pected wanted wanted 3.1 3.0 3.3 3.2 2.9

-.2 -.1

3.3 3.1 3.3 3.3 3.4

—

-.1 -.5

3.3 3.1 3.5 3.5 3.3

.—.

3.3 3.2 3.4 3.3 3.4

-.1 .1 .2 -.1

2.5 2.1 2.7 2.6 2.4

3.3 3.1 3.3 3.3 3.4

-.8 -1.0 -.6 -.7 -1.0

TABLE 101 Average Number of Births by Interview and Average Total Number Expected, for Fecund and Subfecund Couples, and Estimates of These Averages for Subfecund Couples and All Couples Assuming No Fecundity Impairments, by Wife's Age

Wife's age

Total

Total 18-24 25-29 30-34 35-39

2.3 1.3 2.3 2.7 2.7

2.4 1.3 2.4 2.9 3.0

Total 18-24 25-29 30-34 35-39

3.1 3.0 3.4 3.2 3.0

3.4 3.2 3.5 3.6 3.4

Observed averages

Fecund Subfecund

Hypothetical averages, assuming no fecundity impairments

Total

Subfecund

Births by interview 2.2 2.5 1.2 1.3 2.2 2.4 2.9 2.3 3.1 2.3 Total births expected 2.6 3.4 2.1 3.1 2.8 3.5 2.7 3.6 2.5 3.5

Per cent reduction due to fecundity impairments*

Total

Subfecund

2.8 1.3 2.4 2.9 3.2

7 1 1 7 13

21 11 7 20 26

3.5 3.1 3.5 3.6 3.5

8 4 4 9 14

26 32 19 25 29

" Computed from unrounded figures.

2. The ratio of total children expected to children wanted is the same for Subfecund as for Fecund wives in each age group. In other words, both groups are assumed to be equally capable of bearing children. The results of these assumptions are shown in Table 101. They indicate that in the absence of impairments, the Subfecund would

FERTILITY AND FAMILY PLANNING IN THE U.S.

have and expect as many births as the Fecund, within each age group. This is due to the fact that both the Fecund and Subfecund want approximately the same number of children, and the number of children wanted was used as the basis for estimating the number that would occur to the Subfecund if they had no impairments. The estimates show that by ages 35—39 fecundity impairments have reduced the number of births to Subfecund couples by 26 per cent, and births to all couples by 13 per cent. Total numbers of births expected are reduced 29 per cent for the Subfecund and 14 per cent for all couples. These figures should, of course, be regarded as approximate rather than as precise estimates, but they do indicate that fecundity impair­ ments have a fairly important effect on births in the United States. SUMMARY

About one-third of the white couples with wife 18-39 years old are below normal in their ability to have births in the future. Eleven per cent (the Definitely Sterile) cannot have any more births, 12 per cent may not be able to (the Probably and Possibly Sterile), and 8 per cent (the Possibly Fecund) probably can have more, but at a lower rate than Fecund couples. Most of the Definitely Sterile couples have had an operation on the husband or the wife that makes pregnancy impossible. In slightly over half of these cases, the operation was performed for the expressed purpose of preventing pregnancy; these are referred to as contraceptive operations because they serve as a substitute for contraception. The remaining operations were performed to treat a pathological condition, and are referred to as remedial operations. The reason for nearly two-thirds of the contraceptive operations is to protect the wife against pregnancies that would endanger her health. The remaining operations were performed because the couple did not want additional children for financial or other reasons. The proportion of couples with contraceptive operations is highest in the South and West, among Protestants, in rural areas, and among the less-educated. These differentials may be accounted for partly by differences in policies toward such operations (particularly by the Catholic prohibition of contraceptive operations) and partly by an apparent tendency for doctors to recommend contraceptive operations to less-educated couples when they believe that additional pregnancies would be harmful to the wife's health. Remedial operations show much less variation with socioeconomic variables. However, the proportion of couples with remedial operations is greater for the less-educated. The reasons for this differential are not known.

FECUNDITY IMPAIRMENTS

There appears to be an upward trend in the proportion of couples with contraceptive operations, reflecting an increase in vasectomies. The proportion with remedial operations, however, has not changed since 1955. The proportion of couples with fecundity impairments other than those induced by operations is influenced greatly by the length of time married and the use of contraception. Couples married longer and couples who do not use contraception have greater opportunities to discover fecundity impairments than the recently married and the Users. Socioeconomic groups that make less use of contraception have higher proportions Subfecund than other groups. Fecundity impairments reduce the number of children born below ever, the influence of fecundity impairments on childlessness appears to be declining as more couples have their first birth at relatively young ages, when the incidence of subfecundity is low. Fecundity impairments reduce the number of children born below the number wanted. Toward the end of the childbearing period, about a quarter of the wives have fewer births than they want because of fecundity impairments. Among wives 35-39 years old the number of births is about 13 per cent lower than it would have been if there were no fecundity impairments. Subfecund couples in this age group had about 26 per cent fewer births than they would have had without impairments.

CHAPTER 5

The Control of Fertility The central topic of this chapter is the proportion of couples who try to control their fertility by using contraception. It should be under­ stood at the outset that attempts to use contraception are not in­ variably successful. Many couples fail to control their fertility satis­ factorily—by having more children than wanted, by having children sooner than wanted, or both. The effectiveness of contraception is a separate subject that will be considered in Chapters 6 and 7. The term contraception as used in this book, refers to any method used with the intention of preventing conception, other than a sterilizing operation. This definition is consistent with that used in other studies (the Indianapolis Study, 1955 Growth of American Families Study, and the Princeton Study). The methods included in the term are all of the mechanical and chemical methods used to prevent intercourse from resulting in conception; in addition, they include abstinence,1 rhythm, and withdrawal. In other words, contra­ ception is any deliberate pattern of sexual behavior intended to pre­ vent conception. In classifying couples as Users or Nonusers of contra­ ception, no account was taken of the effectiveness of the method used. For example, if a wife says that she and her husband use the rhythm method, they are classified as Users even if they calculate the "fertile" days incorrectly and regularly have intercourse when conception is most likely. This chapter refers only to the white population. The use of con­ traception among nonwhites will be discussed in Chapter 9. Inasmuch as attitudes are important determinants of practice, we shall first describe the attitudes of the wives toward contraception. Following this, we shall investigate the proportion of couples who have used or expect to use contraception, and show how the prevalence of use varies with socioeconomic status. These variations are found to result primarily from differences in when couples customarily begin using contraception, so the timing of first use and differentials in timing become central topics of this chapter. Finally, trends in the use of contraception are described and related to trends in the timing of first use. ATTITUDES TOWARD CONTRACEPTION

The widespread practice of contraception in the United States is supported by attitudes sanctioning and even encouraging its use. 1 Abstinence was defined for the respondent as "no intercourse for many weeks or months." Only 3 per cent of the wives reported ever having used this method.

THE CONTROL OF FERTILITY

That practice is consistent with values in this area of married life is shown by the high proportion of wives (80 per cent) who said that they approved of couples' efforts to limit family size and space their births. The remaining wives said they were against such efforts when first asked, but a majority of them responded favorably when asked whether they thought it was all right for couples to use the rhythm method. These findings are based on the replies to several attitude questions which preceded the questions about the couple's practice of contra­ ception. The first attitude question is: Q61. Many married couples do something to limit the number of pregnancies they have or to control the time when they get pregnant. In general, would you say you are against this, for it, or what? Eighty per cent of the wives replied that they were "for" such efforts, and a majority of this group said they were strongly in favor of them. The 20 per cent who replied "against" were asked: Q61c. Some married couples use only a natural method—rhythm or safe period—to keep from getting pregnant too fre­ quently. Would you say you are against this, for it, or what? This question was designed primarily for Catholic wives who might think that the first question referred to the "artificial birth control" prohibited by their Church, even though we tried to word the question to avoid such misinterpretation. About one-third of the Catholic wives, as well as a few other wives, did misinterpret the first question. They said they were against efforts to control fertility, but then said, in response to Q61c, that they thought it was all right for couples to use the rhythm method. They account for 13 per cent of all white wives 18-39 years old. Only a small minority (5 per cent of the wives) said they were against all methods of fertility control, including rhythm. These wives were asked a series of questions to discover whether there were any conditions under which they thought it would be all right for couples to try to prevent pregnancy (Questions 6Id, e, f, and g). Slightly more than half of them thought that it would be all right if the couple had a low income, if the wife were in poor health, or if the couple could not have normal children. However, the remaining wives (2 per cent of all white wives 18-39) thought that all married

FERTILITY AND FAMILY PLANNING IN THE U.S. WIFE'S EDUCATION

PROTESTANT

COLLEGE

HIGH SCHOOL,4

HIGH SCHOOL.I-3

GRAOE SCHOOL

40

PERCENT

60

CATHOLIC COLLEGE

mmmmmsiss

HIGH SCHOOL,4

HIGH SCHOOL,1-3

GRADE SCHOOL

I

20

I

40

60 PERCENT

STRONGLY FOR •/Λ MODERATELY FOR OR "OTHER" AGAINST,BUT FOR RHYTHM

m

BO

~I00

AGAINST BUT ALL RIGHT FOR SOME NOT ALL RIGHT FOR ANY

NOT ASCERTAINED

FIGURE 25 Per cent distribution by wife's attitude toward fertility control, for Protestant and Catholic wives, by education

couples should have as many children as they could. Obviously, wives with such extreme feelings are rare. Two socioeconomic characteristics are strongly related to attitudes toward fertility control: religion and education (Figure 25). The chief difference between the two major religious groups is that Catholics are less likely than Protestants to say they favor fertility control in general (in response to Q61), and more likely to say that they are for rhythm. However, a large majority of wives in both religious groups favor some kind of control, including rhythm: 96

THE CONTROL OF FERTILITY

per cent for Protestants, 85 per cent for Catholics (Table 102). Jewish wives were most favorable toward fertility control (Table 103). The proportion of wives who oppose all forms of control and think that couples should have as many children as they can is sig­ nificantly larger for Catholics (5 per cent) than for Protestants (1 per cent), but the important fact is that such wives form a very small minority of each religious group. TABLE 102 Per Cent Distribution by Wife's Attitude toward Fertility Control, for Wives, by Religion and Education

Wife's education and religion All religions Total College High school 4 High school 1-3 Grade school Protestant Total College High school 4 High school 1-3 Grade school Catholic Total College High school 4 High school 1-3 Grade school

Wife's attitude toward fertility control Against, qualified Not All all For No. right right Moder­ For for for of wives Total Strongly ately Other rhythm some any 2,414 427 1,153 579 255

100 100 100 100 100

49 58 49 48 41

21 20 23 21 18

10 8 10 11 9

13 11 13 13 14

3 1 2 3 7

1,596 284 752 392 168

100 100 100 100 100

57 65 57 55 49

23 22 25 21 19

11 9 11 12 11

5 3 5 7 8

2 1 2 2 7

668 79 341 168 80

100 100 100 100 100

25 14 26 30 25

19 16 20 21 16

8 9 9 7 5

33 49 33 29 28

4 1 4 5 6

2 1 1 3 8 1 —

1 2 4 5 6 3 5 14

NA 2 1 2 2 4 1 — —

1 2 5 4 6 4 6

Extreme feelings against contraception are more closely associated with poor education than with religion. The proportion of wives who were against any form of fertility control is 1 per cent for those with a high-school or college education, and 8 per cent for those with only a grade-school education. The education differences are found among both Protestants and Catholics (Table 102). Among all wives who said that they thought couples ought to have all the children they can, nearly two-thirds had three years of high school or less.

FERTILITY AND FAMILY PLANNING IN THE U.S.

Protestants have about the same attitude toward fertility control regardless of how frequently they attend church, or what type of denomination they belong to (Tables 103 and 104). Even among those who go to "fundamentalist" churches (see Appendix B, Note 4 for definition), there is wide acceptance of fertility control as a part of married life. Only 2 per cent of the wives who belonged to such groups said that they thought fertility control was wrong and that couples should have all the children they could. TABLE 103 Per Cent Distribution by Wife's Attitude toward Fertility Control, for Wives by Religion, by Frequency of Church Attendiince

Wife's religion No. and frequency of of church attendance wives Total 2,414 Protestant Total 1,596 Once a yr. or less 186 2-12 times a year 446 More than once a month 964 Catholic Total 668 Once a yr. or less 29 2-12 times a year 62 More than once a month 577 Jewish 106 Other 44

Wife's attitude toward fertility control Against Not qualified All all For right right For for Moder­ for Total Strongly ately Other rhythm some any

NA

100

49

21

10

13

3

2

2

100 100 100

57 60 57

23 20 23

11 8 11

5 6 5

2 4 2

1 1 2

1 1 1

100

56

24

11

5

2

1

1

100 100 100

25 52 42

19 17 29

8 —.

33 10 13

4 10 2

5 7 3

5 3 2

100 100 100

22 83 52

19 12 11

8 3 16

5 1 7

5 1 5

10

36

4

—

—

7

2

Among Catholics, devoutness in religious practice strongly in­ fluences attitudes toward fertility control. The association between these variables is most clearly revealed in Table 104, which shows that Catholic wives who take the sacraments least frequently are most favorable toward fertility control, and least likely to say they are against fertility control in general but for rhythm. There appears to be a slight association between frequency of taking the sacraments and the proportion of Catholic wives who are against all forms of fertility control. This proportion is highest (9 per cent) among wives who take the sacraments most frequently, and lowest (1 per cent) among those who never take them. However, the proportion varies irregularly for the intermediate categories, and none of the differences is large enough to be statistically significant.

THE CONTROL OF FERTILITY

The amount of education a Catholic wife has had in Catholic schools apparently influences her attitude toward fertility control to a considerable extent, except in one respect: regardless of how long they attended church schools, about 85 per cent of the Catholic wives favor some form of fertility control. The education groups differ in the proportion of wives who at first say they are against the control of fertility, and then say that it is all right to use rhythm. Wives with all of their education in Catholic schools were more likely to TABLE 104 Per Cent Distribution by Wife's Attitude toward Fertility Control, for Protestant Wives by Type of Denomination, and for Catholic Wives by Frequency of Sacraments Wife's attitude toward fertility control Not All all For right right Moder­ For for for Total Strongly ately Other rhythm some any NA qualified

Type of denomination or frequency of sacraments

No. of wives

Total" 2,414 Protestant 1,596 Total Liberal 274 Intermediate 1,077 Fundamentalist 199 Other and NA 46 Catholic Total 668 84 Never Once a year or less 61 A few times a year 160 Once a month 191 2-3 times a month 78 Once a week or 94 more

100

49

21

10

13

3

2

2

100 100 100 100 100

57 62 56 53 76

23 23 23 23 9

11 11 10 13 9

5 3 6 6 2

2 1 2 3 2

1 1 1 2 2

1 1 1 1

100 100 100 100 100 100

25 39 48 30 19 17

19 23 25 22 21 12

8 10 5 7 8 6

33 19 15 29 35 49

4 4 3 2 4 8

5 1 5 4 8 3

100

11

13

9

47

6

9

—

5 5 —

6 5 6 6

» Includes Jewish and other wives.

respond in this way than those with some or none of their education in such schools. In spite of the small number of cases, the differences are large enough to be statistically significant (Table 105). Is marriage outside the wife's own religion associated with her attitude toward fertility control? Apparently not. Wives in a mixed Protestant-Catholic marriage have attitudes more like the wives of their own religion than of their husband's religion (Table 106). Within marriages between Protestants and Catholics, there is a slight tendency for those who were married by a priest to be less favorable to fertility control in general and more favorable to rhythm

FERTILITY AND FAMILY PLANNING IN THE U.S. TABLE 105 Per Cent Distribution by Wife's Attitude toward Fertility Control, for Catholic Wives by Amount of Education in Catholic Schools

Amount of edu­ cation in Catholic schools

No. of wives

Total

Total None Some All»

668 373 191 104

100 100 100 100

Wife's attitude toward fertility control Against, Not qualified all All right right For for for For ModerOther rhythm some any Strongly ately 25 28 28 13

19 22 18 14

8 7 7 12

33 29 32 47

4 4 7 1

NA

5 6 3 6

5 4 5 7

a Includes only those wives who went beyond the eighth grade. Those who did not go beyond the eighth grade are included in the two preceding categories.

TABLE 106 Per Cent Distribution by Wife's Attitude toward Fertility Control, for Wives, by Religion of Wife EUid Husband

Wife's religion

Hus­ band's religion

No. of wives

Total Prot. Prot. Catholic Catholic

Total Prot. Catholic Prot. Catholic

2,414 1,454 106 114 525

Wife's attitude toward fertility control Against, qualified Not All all For right right Moder­ For for for Total Strongly ately Other rhythm some any 100 100 100 100 100

49 57 53 24 26

21 23 14 26 18

10 11 11 9 8

13 5 10 28 34

3 2 6 4 4

2 1 2 4 5

than for those who were not married in the Church. However, the differences are not statistically significant. Have wives' attitudes toward fertility control changed since the 1955 study? We cannot be entirely certain because different questions were used to elicit attitudes in the 1955 and 1960 studies, and different coding procedures were used to classify the replies. However, if there has been a change, it has probably been toward more favorable atti­ tudes, as the following comparisons suggest: Wife's attitude toward fertility control Total Favorable (including pro-con) Unfavorable N.A.

1955 100 78 21 1

1960 100 80 18 2

NA 2 —

4 6 5

THE CONTROL OF FERTILITY

Another indication of trends in attitudes toward fertility control is provided by the replies of wives in the different cohorts in the 1960 study (Table 107). All cohorts have nearly the same distribution of attitudes toward fertility control. The fact that the younger wives were just as favorable as the older wives may mean that fertility control is becoming more popular. This interpretation is based on the assumption that attitudes toward family limitation become more favorable with age, as experience with children makes the need for fertility control more apparent. If such a tendency exists, comparisons of different cohorts at the same age would show a rising trend in the proportion favoring fertility control. TABLE 107 Per Cent Distribution by Wife's Attitude toward Fertility Control, for Wives by Cohort, from 1916 to 1942

Cohort

No. of wives

1941-42 1936-40 1931-35 1926-30 1921-25 1916-20

73 440 600 624 677 572

Total 100 100 100 100 100 100

Wife's attitude toward fertility control Against, qualified Not All all For right right ModerFor for for Strongly ately Other rhythm some any NA 47 47 49 51 49 48

25 23 22 20 20 20

10 8 10 9 11 10

11 13 13 14 12 12

3 3 2 2 3 5

3 3 2 2 3 3

3 2 2 1 2 3

After expressing her own opinions about the control of fertility, each wife was asked about her husband's attitudes, as follows: Q63. How does your husband feel about married couples trying to limit or control pregnancy? Is he against this, for it, or what? Q64. Does he feel strongly about this, somewhat strongly, or not strongly? The distribution of the replies to these questions is very similar to the comparable distribution for wives—in fact, there are no significant differences. Thus, the typical wife thinks that her husband's attitude toward fertility control is about the same as her own. To discover which spouse the wife felt should take responsibility for preventing pregnancy, the following question was asked: Q89. If a couple wants to prevent pregnancy, who do you think should be the main person to make sure that it's done . . . the husband or the wife?

FERTILITY AND FAMILY PLANNING IN THE U.S.

Over half of the wives (57 per cent) thought that both husband and wife should share the responsibility. However, 26 per cent thought the wife should make sure that contraceptive measures are taken, and 18 per cent thought the husband should do so. These attitudes are related to socioeconomic status. The opinion that the wife should have the main responsibility was more popular among the collegeeducated wives; the husband was mentioned more frequently by the wives with only a grade-school education. However, a majority of wives in all groups saw the control of fertility as a joint responsibility. QUESTIONS ASKED ABOUT THE USE OF CONTRACEPTION

In the interview, the questions about contraception followed those about attitudes toward fertility control. The first was: Q65. Here is a card with the names of methods2 some married couples use to keep from getting pregnant. Have you or your husband ever used any of them? If the wife said "yes," she was asked which methods had been used. If she said "no," she was asked: Q65b. Have you ever used any methods not shown on this card? Wives who reported douching were asked: Q72. Did you douche partly because you thought it would help you to keep from getting pregnant, or entirely for other reasons? Wives who douched to prevent pregnancy were considered Users of contraception, but those who douched entirely for other reasons were considered Nonusers, if douche was the only method they reported in Q65 or Q65b. In some previous studies of family planning (e.g., FPSPG and Indianapolis Study), a distinction was made between "Motive" Users and "Action" Users. Motive Users were defined as couples who made some attempt to prevent pregnancy, regardless of the effectiveness of the method they used. For example, wives who douched several hours after intercourse, when their action would no longer prevent preg­ nancy, were considered Motive Users if their purpose in douching was to prevent conception. This was consistent with the classification as Motive Users of couples who use other methods incorrectly. Action Users, on the other hand, included couples who unwittingly prevented pregnancy by a practice they did not consider to be contra' Rhythm, abstinence, condom, withdrawal, diaphragm, douche, jelly or creme, suppositories, foam tablets, and tampon.

THE CONTROL OF FERTILITY

ceptive, i.e., by douching. In previous studies, couples were classified as Action Users, but not as Motive Users, if the wife douched regularly for cleanliness only and did not try to prevent pregnancy in any other way. In the present study, such couples are not called Action Users, but are referred to as ODFC, indicating that they "only douched for cleanliness" and did so regularly within one-half hour after intercourse.3 In the present study, the term User has the same meaning as Motive User. In other words, we classify as Users couples who tried to prevent pregnancy, regardless of which methods they used or how effective they were. Whether a method was used in such a way that it would be effective is a separate subject. The questions referred to above were asked to determine whether or not couples should be classified as Users. In addition to these, the wives were asked many other questions about contraception— when it was used, whether two (or more) methods were used together, how effective it was, and so forth. These questions will be described where they are relevant to the findings being presented. THE PROPORTION OF COUPLES WHO USE CONTRACEPTION

A large majority of couples in the United States try to control the number of children they have by using some form of contraception. Most of the couples in our sample who had not used contraception either had serious fecundity impairments or had married relatively recendy and planned to have one or more children before be­ ginning use. There were very few couples who were able to have children, had not used contraception, and did not intend to do so. Among white couples with the wife 18 to 39 years old, 81 per cent had used contraception by 1960, and an additional 7 per cent expected to begin use later. Of the remaining couples who expected not to use contraception, the majority (10 per cent of all couples) were Subfecund. Most of these couples did not need to use contra­ ception to limit their fertility in the future. An additional 1 per cent of all couples were unintentionally limiting their fertility because the wife douched for cleanliness within one-half hour after intercourse. Altogether, then, 98 per cent of the white couples had or expected some form of limitation on their fertility. For 10 per cent it was ' In the other studies referred to previously, the initials DFCO were used to designate wives who douched for cleanliness only; we are using a different arrange­ ment of the initial letters to indicate that the definition used here differs from that used before. We limit the category ODFG to wives who douched for cleanli­ ness within one-half hour after intercourse and used no other method at any time. The category DFCO used in the other studies designates this practice, but without the time limit.

FERTILITY AND FAMILY PLANNING IN THE U.S.

subfecundity; for 1 per cent it was douching for cleanliness. But most couples (87 per cent) had used contraception or expected to do so. Although nearly nine out of ten couples had used or expected to use contraception, the proportion varies somewhat between groups, as can be seen in Tables 108 and 109. For example, it is 93 per cent for the college-educated wives and 72 per cent for those with TABLE 108 Percentage of Couples Who Have Used or Expect to Use Contraception, by Fecundity, Wife's Age, Duration of Marriage, and Number of Births

Characteristic Total Fecund Subfecund Possibly Fecund Possibly Sterile Probably Sterile Definitely Sterile Wife's age 18-24 25-29 30-34 35-39 Years married Under 5 5 to 9 10 to 14 15 or more Number of births 0 1 2 3 4 5 6 or more

Have used or expect to use

Do not expect to use

—

87 96 68 84 67 49 63

13 4 32 16 33 51 37

78 84 83 77

14 7 4 2

92 91 88 80

8 9 12 20

544 649 702 519

75 86 82 78

17 5 4 2

91 91 86 80

9 9 14 20

301 463 682 499 263 119 87

55 74 89 89 87 80 76

28 15 7 8 10 16 22

No. of couples

Have used

Expect to use

2,414 1,674 740 199 199 85 257

81 89 62 77 54 45 63

7 7 6 7 14 5

513 600 624 677

17

72

12

85

3 3 3 4 2

93 92 90 84 78

a grade-school education (Figure 26); it is 90 per cent for Protestant wives and 80 per cent for Catholic wives. What accounts for these differences? In order to explore the factors responsible, we have subdivided the couples who expect never to use contraception into three groups: 1. The Subfecund, many of whom will never need to use contra­ ception; 2. The Fecund who use ODFC and thus have some limitation on their fertility; 3. The Fecund who do not use ODFC.

THE CONTROL OF FERTILITY TABLE 109 Percentage of Couples Who Have Used or Expect to Use Contraception, by Wife's Religion and Education, Husband's Income and Occupation, Region of Residence, and Size of Place of Residence

Characteristic Total Wife's religion Protestant Catholic Jewish Other Wife's education College High school 4 High school 1-3 Grade school Husband's income $10,000 or more $7,000-89,999 $6,000-$6,999 $5,000-$5,999 $4,000-$4,999 $3,000-$3,999 Under $3,000 Husband's occupation Upper white collar Lower white collar Upper blue collar Lower blue collar Farm Region of residence Northeast North Central South West Size of place of residence 12 central cities' Other cities 50,000 + Rings of 12 central cities Rings of other cities Cities 2,500-49,999 Rural nonfarm Rural farm

Have used Do not or expect expect to use to use

No. of couples

Have used

Expect to use

2,414

81

7

87

13

84 70 95 77

5 10 7

90 80 95 84

10 20 5 16

427 1,153 579 255

88 83 78 66

5 7 7 6

93 90 85 72

7 10 15 28

261 405 312 423 380 306 327

89 84 85 80 81 77 70

2 5 4 8 7 8 11

91 89 89 88 88 85 82

9 11 11 12 12 15 18

725 312 465 670 154

86 84 79 76 81

5 5 8 8 4

90 89 86 84 84

10 11 13 16 16

581 715 744 374

77 82 83 80

7 6 5 10

85 88 88 89

15 12 12 11

187 427 443 537 282 335 203

76 77 79 82 82 84 84

9 8 7 6 7 5 4

84 85 86 88 89 89 88

16 15 14 12 11 11 12

1,596 668 106 44

•

—

a The cities are Baltimore, Boston, Chicago, Cleveland, Detroit, Los Angeles, New York, Philadelphia, Pittsburgh, St. Louis, San Francisco, and Washington.

According to Tables 110 and 111, where the proportions in these groups are shown, the widest differences between groups are found for the proportion of couples who are Subfecund and do not intend to use contraception. There are smaller variations in the proportions who are Fecund but refuse to use contraception even though they

FERTILITY AND FAMILY PLANNING IN THE U.S. WIFE'S EDUCATION COLLEGE

HIGH SCHOOL,4

HIGH SCHOOL,1-2

GRADE SCHOOL

20

40

PERCENT

60

Υ//Λ

HAS USED

'/

EXPECT TO USE

A

D| Λ

*

80

IOO

WON'T USE l FECUND 1 ObFC WON'T USE FECUND,NON-ODFC

WON'T USE,SUBFECUND

FIGURE 26 Per cent distribution by use of contraception, intention to use, and fecundity ODFC status of those who will not use, by wife's education

will very probably have more children. The smallest differences are found for the proportion classified as Fecund, ODFC. Thus, group differences in the past or prospective use of contra­ ception are due largely to different proportions of couples who refuse to use contraception and different proportions who have little need to use contraception. Because these two kinds of couples have essen­ tially different reasons for not using contraception, they are discussed separately below. FECUND COUPLES WHO REFUSE TO USE CONTRACEPTION

As noted previously, only a small minority (2 per cent) of the couples were Fecund Nonusers who, according to the wife, would not limit their fertility by using contraception. First of all, can we trust the wives' replies? All wives who said that contraception had not been used were asked: Q88. Do you expect to use a method sometime later on to keep from getting pregnant? If the wife replied "no" or was uncertain whether or not she or her husband would use a method, the interviewer asked the following

THE CONTROL OF FERTILITY

probing questions to discover whether the wife really meant what she said: Q88c. If you never do anything later on to keep from getting pregnant, aren't you liable to have a child every two or three years until you are forty-five? (If yes) Q88d. Is this all right with you, or do you think you might do something later on to prevent pregnancy? Wives were classified as expecting never to use contraception if they said in reply to Q88d that it would be all right with them if they had a child every two or three years until they were 45, or if they said that they did not really expect to have children that frequently, but still would not use contraception. The following cases are examples of wives who thought they would remain Nonusers: CASE 3210. This 31-year-old mother of five children wanted and expected seven, but said she would not try to keep from having more. She replied that it would be all right with her if she had a child every two or three years until she was 45. (Such frequent childbearing would bring her about six more children, or eleven alto­ gether.) She is Catholic and reported that her priest gave her per­ mission to use the rhythm method, but that she would not use it because she disapproved of any limitation on her fertility. She felt she should accept all of the children Godsent her. CASE 3211. This 22-year-old Catholic wife had not yet had any children, and expected no more than four altogether. She strongly disapproved of all methods of contraception, including rhythm, and intended never to do anything to prevent pregnancy. She denied that she was liable to have as many as one child every two or three years until she was 45.

In nearly every case of refusal to use contraception, we find that the couple would have more children than the wife wanted or ex­ pected if she and her husband remained Fecund until she was 45 years old. Only, two of the 52 wives in this group wanted as many children as they would probably have eventually, if they never used contraception. It seems likely, then, that some of these couples will use contraception in spite of the wife's current intention not to do so. Others will probably become Subfecund before family limitation becomes a pressing necessity. Perhaps fewer than half will persist in their determination not to use contraception in spite of continued childbearing.

FERTILITY AND FAMILY PLANNING IN THE U.S.

Tables 110 and 111 show the differences between various groups in the proportions who are Fecund and intend never to use contra­ ception. In general, these differences are not very important because so few couples are involved. The main point is that they account for a relatively minor portion of the difference between groups in the proportion who have used or expect to use contraception. In other words, the primary reason that groups differ in the propensity to use contraception is not that certain kinds of couples refuse to TABLE 110 Percentage of Couples Who Expect Not to Use Contraception, by Fecundity by Use of ODFC for Fecund Couples, by Wife's Age, Duration of Marriage, and Number of Births

Chiiracteristic Total Wife's age 18-24 25-29 30-34 35-39 Years married Under 5 5-9 10-14 15 or more Number of births 0 1 2 3 4 5 6 or more

No. of couples

Expect not to use contraception Fecund ODFC non-ODFC Total Subfecund

2,414

13

10

1

2

513 600 624 677

8 9 12 20

4 5 10 18

1 1 1

4 2 2 1

544 649 702 519

9 9 14 20

3 6 12 17

1 1 1 1

4 2 2 1

301 463 682 499 263 119 87

28 15 7 8 10 16 22

25 10 6 8 7 8 7

1 1 1

2 3 1

—

—

—

1 1 1

2 7 14

use contraception even when some form of fertility limitation is clearly needed or soon will be. This is true even for the differences in contra­ ceptive practices between religious groups. Catholics have a higher proportion of Fecund "never-users" than Protestants (5 per cent, as compared to 1 per cent), but this accounts for less than half of the 10 percentage point spread between the religious groups in the proportion of couples who expect never to use contraception (20 per cent for Catholics, 10 per cent for Protestants). Obviously, group differences in the use of contraception are brought about by some­ thing more important than the intention of a few Fecund couples never to use contraception.

THE CONTROL OF FERTILITY TABLE 111 Percentage of Couples Who Expect Not to Use Contraception, by Fecundity, by Use of ODFC for Fecund Couples, by Wife's Religion and Education, Husband's Income and Occupation, Region of Residence, and Size of Place of Residence Expect not to use contraception Characteristic Total Wife's religion Protestant Catholic Jewish Other Wife's education College High school 4 High school 1-3 Grade school Husband's income $10,000 or more $7,000-$9,999 $6,000-®6,999 $5,000-$5,999 $4,000-$4,999 $3,000-$3,999 Under $3,000 Husband's occupation Upper white collar Lower white collar Upper blue collar Lower blue collar Farm Region of residence Northeast North Central South West Size of place of residence 12 central cities® Other cities of 50,000 + Rings of 12 central cities Rings of other cities Cities of 2,500-49,999 Rural nonfarm Rural farm

Fecund

No. of couples

Total

Subfecund

2,414

13

10

1

2

1,596 668 106 44

10 20 5 16

8 14 5 11

1 1

1 5

—

—

2

2

427 1,153 579 255

7 10 15 28

4 8 12 21

261 405 312 423 380 306 327

9 11 11 12 12 15 18

7 9 10 10 9 9 13

725 312 465 670 154

10 11 13 16 16

8 7 11 12 12

581 715 744 374

15 12 12 11

10 10 10 9

187 427 443 537 282 335 203

16 15 14 12 11 11 12

10 13 10 9 8 9 8

ODFC non-OI

— —

2 2 — — —

2 1 1 1 —

2 1 1 —

2 —

1 1 3 —

1 1 1 1 —

2 2 2 5 2 2 1 —

2 4 4 1 2 2 3 3 4 2 1 1 3 2 2 2 2 1 3

• See footnote, Table 109.

SUBFECUND COUPLES WHO DO NOT NEED TO USE CONTRACEPTION

Ten per cent of white couples with the wife 18-39 years old were Subfecund Nonusers who do not expect to use contraception. They had had 1.7 births, on the average (Table 112); but most of them could not have any more children. Forty per cent were

FERTILITY AND FAMILY PLANNING IN THE U.S.

Definitely Sterile (in the majority of these cases, the wife had had a remedial operation) and 18 per cent were Probably Sterile. Why didn't these couples use contraception before becoming Subfecund? There are several possible answers to this question: 1. Some couples knew when they were married that childbearing would be difficult or impossible and never had any pressing need to try to prevent conception. 2. Some opposed contraception, intended never to use it, and become Subfecund before frequent childbearing put pressure on them to change their attitude. 3. Some delayed using contraception because they wanted children as soon as possible or for other reasons, and became Subfecund before they felt the need to begin contraception. TABLE 112 Per Cent Distribution by Degree of Subfecundity, and Average Number of Births, for Subfecund Nonusers Who Expect Not to Use Contraception

Fecundity

No. of couples

Per cent

Average number of births

Subfecund Definitely Sterile Probably Sterile Possibly Sterile Possibly Fecund

234 94 43 65 32

100 40 18 28 14

1.7 1.9 0.8 1.8 2.3

It seems probable that most of the Subfecund couples who will not use contraception are in the third category, because they are relatively few in the first two. First of all, we know that very few couples are Subfecund at the time of marriage, and most of these are not aware of their subfecundity. Among white couples married less than two years, for example, only 7 per cent are classified as Subfecund, and most of them did not have serious impairments. Therefore, we can eliminate the first explanation as a major reason for not using contraception. Undoubtedly it applies to only a minority of the Subfecund Nonusers who will not use contraception. Secondly, few Subfecund Nonusers failed to use contraception because they disapproved of it. On the contrary, most of these couples were favorably disposed toward the use of contraception even though they had little or no need for it themselves.4 4 It is true that Subfecund Nonusers who expected never to use contraception were less favorable toward family limitation than other couples, but a majority of them (69 per cent) approved of fertility control through rhythm or other methods. The comparable proportion for all white wives 18-39 is 93 per cent.

THE CONTROL OF FERTILITY

Therefore, it seems likely that most of the Subfecund Nonusers are couples who delayed using contraception, became Subfecund, and found that they had little or no need for any additional limitation on their fertility. The following case is an example of a couple who delayed using contraception and then found that they didn't need to begin: CASE 5666. This couple was married 12 years ago, when the wife was 26 years old. She says that when she was married she wanted six children. She and her husband did not use contraception, and waited over three years for their first child. They did not begin using contraception after their first child because they wanted more children. So far (eight years later), they have not been able to have any more. The wife reports that both she and her husband are strongly in favor of using contraception; the only reason that they don't use it is because they don't have to. Since socioeconomic groups differ greatly with respect to when couples customarily begin using contraception, they also differ in the proportion of couples who eventually use contraception. For example, 86 per cent of the couples in the grade-school group did not use contraception before the first pregnancy, but only 39 per cent of the couples in the college group delayed use to this extent. As a result of the timing differentials that these proportions represent, 21 per cent of the couples in the grade-school group found that sub­ fecundity probably would make it unnecessary for them to begin using contraception, as compared with only 4 per cent of the couples in the college group. (These are couples classified as Subfecund Nonusers who do not expect to use contraception in Table 111.) The 17-point spread between these percentages accounts for most of the 21-point difference between the grade-school and college groups in the proportion who have used or expect to use contraception—72 per cent and 93 per cent, respectively (Table 109). The relationship between delayed use of contraception and the onset of subfecundity also contributes to differences between the religious groups in the past or prospective use of contraception, al­ though here the differentials are smaller than those for educational attainment. Catholics are more likely than Protestants to delay con­ traception, and as a result they have relatively more Subfecund Nonusers who expect not to use contraception. A majority of the Subfecund Nonusers who expect not to use contraception have had no more children than they wanted. That is, they became Subfecund before they needed to use contraception to prevent unwanted births. It may seem surprising that the proportion

FERTILITY AND FAMILY PLANNING IN THE U.S.

of such couples can be high enough to account for most of the differ­ ences between groups in the proportion using or expecting to use contraception. Can we make estimates of this proportion that are independent of our data on differential use and timing? For example, if all couples delayed using contraception until they had all the children they wanted or until they became Subfecund but did not have more children than wanted, and if only the Fecund began using contraception to prevent unwanted births, how many couples would not begin? First of all, we know that at least 4 or 5 per cent would never need contraception because they cannot have any children (see Chapter 4). In addition, other couples become Subfecund after having one or more children. We estimate that in the white population of the United States, the proportion who can have no more than three children—the average number wanted—is between 10 and 20 per cent, and probably in the upper part of this range. (See Appendix B Note 5 for more detail.) The proportion of couples who would not have more children than they want, even if they never used contraception, is probably somewhat higher, inasmuch as many couples want four children, and a few want five or more. If 15 to 25 per cent could not have more children than wanted, we could conclude that most of the differences between groups in the inci­ dence of contraception result from variations in the proportion discovering that they don't need to use contraception to prevent un­ wanted births. For example, a proportion in this range is consistent with the 21 per cent of grade-school couples classified as Subfecund Nonusers who expect not to use contraception. In addition, we must remember that some of the Subfecund couples who will never use contraception had unwanted births. This is particularly true of couples with lower socioeconomic status. This means that we should increase our first estimate of 15 to 25 per cent (which represents couples who would become Subfecund before having too many children) by an allowance for couples who would have more children than wanted before using contraception or be­ coming Subfecund. This allowance would vary with socioeconomic status. Such an estimate would also help us to connect differentials in when use begins with differentials in the eventual use of contra­ ception, but only to a minor degree. The essential point is that the timing of first use and the prevalence of eventual use of contraception are causally related. WHEN USE OF CONTRACEPTION BEGINS

As we have seen, the proportion who have used or expect to use contraception does not vary greatly between groups. Where large

THE CONTROL OF FERTILITY

variations exist (such as those between education groups), they can be accounted for mainly by the proportion who are Subfecund Nonusers who will not use; this proportion, in turn, can be accounted for by different practices regarding the timing of first use. Couples classified by wife's educational attainment show the greatest differences in the timing of first use of contraception (Figure 27). As noted in the previous section, the college group begins use ORDINAL NUMBER OF PREGNANCY

'////////////////Χ

vyyyyyyyyyyyyyyyyyyA

PERCENT m m COLLEGE

E S 3 HIGH SCHOOL, 1-3

I'/////a HIGH SCHOOL,4

Κ//Λ GRADE SCHOOL

FIGURE 27 Percentage of couples who ever used contraception before specified pregnancy, for couples with four pregnancies, by wife's education

earliest and the grade-school group begins latest (Table 113). The widest differences in the timing of first use are for the proportion who began before the first pregnancy. Among couples with three pregnancies, for example, the proportion who used contraception be­ fore the first was 55 per cent for the college group and 18 per cent for the grade-school group. By the time these couples had had three pregnancies, the two education groups were closer together:

FERTILITY AND FAMILY PLANNING IN THE U.S.

TABLE 113 Cumulative Percentage of Couples Who Used Contraception before Specified Pregnancy, and Percentage Who Are Fecund or Subfecund Nonusers, by Number of Pregnancies by Wife's Education No. of pregnancies and wife's education No pregnancies Total College High school 4 High school 1-3 Grade school One pregnancy Total College High school 4 High school 1-3 Grade school Two pregnancies Total College High school 4 High school 1-3 Grade school Three pregnancies Total College High school 4 High school 1-3 Grade school Four pregnancies Total College High school 4 High school 1-3 Grade school Five pregnancies Total College High school 4 High school 1-3 Grade school Six or more preg­ nancies Total College High school 4 High school 1-3 Grade school

No. of couples Total

When contraception was 1st used: before

Users

Nonusers

SubFirst Second Third Fourth Fifth Total Fecund fecun

216 53 120 32 11

100 100 100 100 100

52 70 47 56

52 70 47 56

48 30 53 44

12 9 13 6

37 21 40 38

*

*

*

*

*

373 72 204 71 26

100 100 100 100 100

73 79 74 75 46

49 69 49 38 31

73 79 74 75 46

27 21 26 25 54

14 14 14 11 27

13 7 13 14 27

628 102 324 153 49

100 100 100 100 100

86 94 90 76 78

42 68 45 25 18

70 83 75 55 57

86 94 90 76 78

14 6 10 24 22

7 5 5 13 2

7 1 6 10 20

498 103 226 120 49

100 100 100 100 100

87 92 92 79 76

38 55 42 23 18

63 80 64 54 45

74 88 75 66 61

87 92 92 79 76

13 8 8 21 24

3 3 1 8 6

9 5 8 13 18

334 47 146 95 46

100 100 100 100 100

88 94 90 86 76

29 53 36 17 9

55 66 64 48 26

71 81 79 66 46

78 83 84 75 57

88 94 90 86 76

12 6 10 14 24

4 4 3 2 9

9 2 7 12 15

180 32 70 47 31

100 100 100 100 100

85 97 87 81 74

22 41 20 17 13

42 69 46 30 23

59 81 66 51 32

64 91 71 55 35

72 97 79 66 42

15 3 13 19 26

7 3 9 6 10

4 13 16

185 18 63 61 43

100 100 100 100 100

77

13

31

46

52

61

23

14

10

*

*

*

*

*

*

*

*

*

89 77 51

19 7

38 25 9

54 44 21

65 51 23

73 61 33

11 23 49

6 13 28

5 10 21

—

8 —

THE CONTROL OF FERTILITY

92 per cent of the college group had begun, and 76 per cent of the grade-school group. This tendency for socioeconomic differences in the use of contra­ ception to narrow as more children are born is due to the increasing awareness of Nonusers that they must limit their fertility]. As more children are born, the implications of unrestricted childbearing become clearer. However, the proportion of Usens among the less-educated never quite reaches the proportion attained by the better-educated. As shown in Table 113, for example, the proportion of the gradeschool group using contraception never rises above 78 per cent. Most of the remaining couples encountered fecundity impairments before they felt impelled to begin contraception. Similar differences appear when couples are classified by husband's income (Table 114), although the differences are not as great as they are between education groups and there are more irregularities. In general, however, the lower-income groups have lower proportions beginning before the first conception, and the differences between groups narrow as more pregnancies occur. The three major religious groups also show significant differences in the timing of first use of contraception (Table 115). The highest percentages beginning use before the first conception are for Jews and others, the next highest are for Protestants, and the lowest are for Catholics. The differences narrow as more pregnancies occur, but never disappear. Among Protestants the maximum is 93 per cent using contraception for those with four pregnancies, but the proportion for Catholics never rises above the 79 per cent attained by couples with two pregnancies. Why do couples differ so much with respect to when they begin using contraception? It doesn't seem to be due to substantial differ­ ences in the number of children wanted altogether. As we have seen in Chapter 3, the socioeconomic groups do not differ greatly in this respect. According to the wives interviewed, differences in the timing of first use of contraception are due mainly to differences in the proportion who want children as soon as possible. This finding is based on replies to the following questions, which were asked of all wives who began using contraception after having had one or more pregnancies: Q80b or f. What are the reasons why you and your husband didn't begin using a method earlier? Q80c or g. Before you began using a method, did you want to have children as soon as possible?

FERTILITY AND FAMILY PLANNING IN THE U.S. TABLE 114 Cumulative Percentage of Couples Who Used Contraception before Specified Pregnancy, and Percentage Who Are Fecund or Subfecund Nonusers, by Number of Pregnancies, by Husband's Income Number of pregnancies & husband's income No pregnancies Total $7,000 + S5,000-36,999 Under 55,000 One pregnancy Total $7,000 + $5,000-$6,999 Under $5,000 Two pregnancies Total $7,000 + 5,000-$6,999 Under $5,000 Three pregnancies Total $7,000 + $5,000-$6,999 Under $5,000 Four pregnancies Total $7,000 + $5,000-$6,999 Under $5,000 Five pregnancies Total $7,000 + $5,000-86,999 Under $5,000 Six or more preg­ nancies Total $7,000+ 85,000-$6,999 Under $5,000

No. of couples Total

Users

When contraception was 1st used: Nonusers before SubFirst Second Third Fourth Fifth Total Fecund fecun


«-J

gj

23.9

22.,

23.,

23..

b

2"¾ 9

25 5 ^59 97 » ο 27

27 4

25,7

25 5

25,3

Secondbirth Thirdbirth

b

b

b

oc 4 25.4

;j

21.1 26 5 26 7 20 4 29.4

* Source: The medians for first marriages, and for births to the cohorts of 1880-89 to 1910-19 are from GrabiIl 1961, Tables 5 and 7. Those for later cohorts are low and high estimates made by the Scripps Foundation from the cumulative birth rates in Table 7 of Grabill 1961 and the following low and high estimates of final birth rates based on the GAF data: First births

Cohorts Low 1935-39 800 1930-34 880 1925-29 870 1920-24 870 bNot estimated because too Census study was made.

Second births

Third births

High High Low High Low b b b b 950 720 750 940 880 470 640 920 780 520 840 520 890 740 760 490 few births had occurred by August 1959 when the

were fourth rose from about 28 in the cohorts of 1880-89 to 38 in the cohorts of 1910-19, while the percentage that were sixth and higher fell from about 51 to 39. Accompanying the above mentioned changes in median age at first marriage and at birth of children have been corresponding changes in the length of the intervals between marriage and first, second, and third births. Among the white women first marrying from 1900-09 to 1930-34 and having one, two, or three or more children, there was a strong tendency to delay family growth more and more. The number of months elapsing before half the babies were born rose from 17.8 to 21.6 for first births, from 48.5 to 62.1

FERTILITY AND FAMILY PLANNING IN THE U.S.

for second, and from 79.2 to 100.6 for third (Table 166). Little change took place among brides of the late 1930's or early 1940's, but postwar brides have greatly accelerated the childbearing pattern. Approximately half of all the first babies that will be borne by the brides of 1950-54 arrived by 16 or 17 months after marriage—about four months or 20 per cent sooner than for the brides of 1925 to 1944. Corresponding reductions for second and third births are in the neighborhood of 15 months and 20 months (respectively) or around 25 per cent. TABLE 166 Median Length of Interval between Marriage and First Birth, Second Birth, and Third Birth, for White Women Who First Married in 1900-1909 to 1955-1959" Year of first marriage Birth order First Second Third

195559 12.3 14.3 b b b b

195054 16.2 17.5 44.8 50.0 67.7 85.5

1945- 1940- 1935- 1930- 1925- 192049 44 39 34 29 24 17.0 21.2 21.1 21.6 19.6 17.8 17.9 21.8 51.0 62.9 62.2 62.1 56.7 49.6 54.1 65.0 79.3 96.5 96.9 100.6 90.1 80.0 92.0 104.9

1910- 190009 19 16.7

17.8

48.2

48.5

79.6

79.2

" Source: The medians for marriages of 1900-09 to 1935-39 are from Grabill 1961 Table 16. Those for later marriages are low and high estimates made by the Scripps Foundation from the cumulative birth rates in Table 16 of Grabill 1961 and the follow­ ing low and high estimates of final birth rates based on GAF data: Year of First births Second births Third births High Low Low High marriage Low High b b b b 1955-59 820 970 700 900 450 1950-54 900 960 760 820 480 600 1945-50 760 890 930 430 490 720 750 1940-44 870 890 b Not estimated because too few of these births had occurred by August 1959 when the Census study was made.

Nor was the turning point reached by the brides of 1950-54. It now appears highly probable that those of 1955-59 will set a new record for the speedy starting of families, with half of the first babies coming in 12 to 14 months—two to four months quicker than for the brides of five years earlier. It is too early to make reliable estimates of the speeding up of second and third births by the brides of 1955-59, but here too new records may be set. RECENT TRENDS ACCORDING TO TWO OTHER MEASURES OF TIMING

In discussing trends and differentials in the timing of births among various socioeconomic groups later in this chapter, it will be desirable

TIMING AND SPACING OF BIRTHS

to use other yardsticks than median age at childbirth and median length of intervals from marriage to birth. One objection to median age at birth of nth child or length of nth interval is that their computation requires actual or estimated figures regarding the pro­ portion of women that will eventually have an nth child or inter­ val. The high and low estimates shown in Tables 165 and 166 are almost certain to bracket the final values but there is less basis for making corresponding estimates for wives classified by religion, education, and other characteristics. The other measures that will be used are mean length of intervals, and cumulative birth rates by various ages expressed as percentages of the most likely final rate expected by wives. Before considering the differentials shown by these rates it is desirable to consider their trends, and their ad­ vantages and disadvantages in measuring changes in the timing of births to cohorts still in the reproductive ages. Mean Length of Intervals—A glance at the top deck of Table 167 shows important decreases from the cohorts of 1916-20 to those of 1936-403 in the mean length of intervals for the births occurring before the time of the study. The smallest decrease is that for the first birth interval—from 30 months for the two earliest groups of cohorts to 20 or less for wives born in the 1930's. Larger reductions in terms of months have occurred in the length of the intervals between the first and second births, the second and third, and the third and fourth. Each of these intervals lasted about 41 to 46 months for wives in the cohorts of 1916-20 but so far is only about two-thirds as long for the 1931-35 cohorts, and still shorter for the next group. This rate of decline is about the same as that for the first birth interval. Part of the decrease is due, of course, to the fact that the wives of earlier cohorts were near the end of the childbeaiing period in 1960 while those in recent cohorts were in the early part of it. The younger wives had not been married as long as those in their 40's and could not have had the intervals of ten years or more that had been experienced by some of the latter. It is helpful, there­ fore, to look at the lower decks of Table 167, where wives are classified by years since first marriage. Comparisons of cohorts within each of these decks shows that controlling for duration of marriage reduces the trend toward closer birth spacing, but does not eliminate it. For example, five of the six changes from the cohorts of 1931-35 to those of 1936-40 are decreases, and two of them exceed three months or 10 per cent (second and third decks). Downward trends 1 In this table and those that follow the cohorts of years y to y + 4 consist of the women born from July 1 of year y — 1 through June 30 of year y -J- 4.

FERTILITY AND FAMILY PLANNING IN THE U.S.

from the cohorts of 1926-30 to 1931-35 also appear for first, second, and third births to wives married 5 to 9 years, and for first births to those married 10 to 14 years. Comparisons involving 200 or more members of earlier cohorts show little change for some birth orders and shorter intervals for others, but no consistent pattern. TABLE 167 Age at First Marriage and Length of Birth Intervals, for White Wives, by Cohort, by Years Since First Marriage

Years since first marriage, a n dc o h o r t Total 1936-40 1931-35 1926-30 1921-25 1916-20 Under 5 1936-40 1931-35 1926-30 5 to 9 1936-40 1931-35 1926-30 1921-25 10 to 14 1931-35 1926-30 1921-25 1916-20 15 and over 1926-30 1921-25 1916-20

Average number of months* Mar. 1st 2nd 3rd to to to to 1st 2nd 3rd 4th G D E F

1st G

19.,0 20..1 20,.7 21..5 22..2

(17) 20 27 30 30

(22) 29 35 38 43

77 90 92 90 92

41 73 79 78 81

14 43 50 49 51

5 19 25 26 28

310 121 32

19 .8 23..7 29.,4

14 15 16

19 18

70 74 62

26 40 19

4 12 3

—

127 351 134 37

17,.2 20..0 23 .7 29 .5

95 93 90 73

76 80 77 51

39 47 37 22

17 19 10 14

128 363 208 66

87 83 81 76

60 57 49 45

37 30 25 27

95 424 494

86 80 83

60 53 52

37 27 29

Nutnber of wives A

Age at marriage B

440 600 624 677 572

(20) 27 32 44 46

(19) (24) 31 37 41

—

—

—

—

—

—

—

21 22 26 32

24 28 31

21 25 29

19 20

—

—

—

17 .1 19 .9 23 .9 28 .8

19 26 27 25

36 36 36 31

34 33 40 31

30 30 33 —

97 94 89 83

16 .9 19,.4 21,.2

32 32 31

35 40 45

31 47 48

34 39 42

92 93 94

—

Per cent having 2nd 3rd H I

4th J

—

—

a In this and subsequent tables in this chapter averages are put in parenthesis if the proportion of wives having such an interval (in Cols. G to J here) is considerably smaller for a recent cohort group than for earlier groups because of differences in duration of marriage.

Classifying couples by duration of marriage is helpful in evaluat­ ing intercohort trends of birth spacing but has one shortcoming in this study. In the earlier cohorts the only wives with short marriage durations are those who married relatively late in the childbearing period, while in the more recent cohorts the only wives who were married for ten or more years are those who married relatively young. An extreme example is found among wives married 10-14

TIMING AND SPACING OF BIRTHS

years, with average age at marriage varying from about 17 for the cohorts of 1931-35 to nearly 29 for those of 1916-20 (Table 167, fourth deck, column B). The decrease in fecundability as age increases would be expected to lengthen the birth intervals of wives who marry at later ages compared with those who marry younger, and also to depress the proportion having a first, second, third, or fourth birth (Columns G to J). On the other hand, some of the wives who marry in the 30's or later and want to have children undoubtedly realize that conception may not occur quickly and that fewer years remain in the childbearing period, consequently in this TABLE 168 Average Length of Birth Intervals and Per Cent of Wives Having Births of Given Orders, for Cohorts, by Age at Marriage

Cohorts and age at marriage 1921-25 Under 18 18-19 20-21 22-24 25 + 1916-20 Under 18 18-19 20-21 22-24 25 +

Average years Average number of months Num­ since Mar. 1st 2nd 3rd ber of 1st mar­ to to to to 1st 2nd 3rd wives riage 4th C A D E F B

Per cent having

1st G

2nd H

3rd I

4th J

100 148 165 161 103

20.5 18.6 16.6 14.1 9.9

30 30 33 30 28

45 38 40 37 28

50 44 47 42 33

39 38 39 34 30

94 95 92 89 80

83 82 80 77 62

58 64 48 42 34

30 32 24 22 19

77 115 118 141 121

26.1 23.5 21.4 18.9 13.8

26 28 32 35 27

44 43 47 48 33

43 52 56 40 36

38 42 46 44 30

96 97 97 90 83

82 85 87 82 69

57 58 55 47 40

39 33 32 23 19

group there may be less intentional lengthening of intervals by contraception. Because little information is available about these topics it is worthwhile to consider briefly the experience of the GAF wives. Those in the 1916-20 cohorts, aged 40-44 in 1960, had so nearly completed their childbearing that subsequent births would have no significant effect on the average length of their birth intervals. For this group the relationship between age at marriage and birthspacing appears to be like an inverted U. The length of the first interval is 26 months for those who were brides before age 18, increases to 35 months for those marrying at ages 22-24, and then declines to 27 months for the older brides (Table 168, bottom deck). A small rise and larger fall is shown for second births. Larger increases

FERTILITY AND FAMILY PLANNING IN THE U.S.

and decreases appear for third and fourth births, but the longest intervals are found for women marrying at ages 20-21 instead of for the 22-24 group. The 1921-25 women had given birth to nearly all of their first and second babies by 1960, but there is more opportunity for the average length of their third and fourth birth intervals to change. The relationship between age at marriage and length of the first interval is much the same for these women as for the 1916-20 cohorts, but the amplitude of the variation is smaller. For second, third, and fourth births, the longest intervals are found for brides of under 18 instead of 20-21, or 22-24, but the decline is consistent only among those marrying later. More recent cohorts are not con­ sidered in this connection because they had not passed the middle of the childbearing period. The relatively short intervals of women marrying at ages 25 or older, and the relatively low proportion that had borne a child of a given order, suggest that the chief impact of the decrease in fecundability with age may be in preventing a birth rather than merely postponing it. However, the shorter intervals of these women reflect also the fewer years they have from marriage to menopause, and the lower proportions with a birth of a given order reflect also the inverse relation between age at marriage and number of children wanted. The effect that differences in age at marriage and marriage dura­ tion have on cohort trends in birthspacing may be controlled more effectively by comparing different groups up to given ages but differ­ ent dates. For example, wives in the cohorts of 1921-25 were 35-39 years old on July 1, 1960, as were those of the 1916-20 cohorts of July 1, 1955. The later group had married about half a year younger, but no consistent pattern of differences appears in the num­ ber of months from marriage to the first birth and between subse­ quent births (Table 169, next to last deck). Furthermore, the propor­ tion of wives having a birth of a given order is about the same in the two groups. Three sets of cohorts can be compared up to ages 30-34, and show similar trends. In interpreting the significance of the data on months between births, however, it is necessary to remember that they relate only to part of the childbearing period. This incomplete experience provides an excellent base for judging the trend in age at marriage, since so few single women aged 30-34 will wed later. In contrast, as the 1916-20 cohorts lived from ages 30-34 to 40-44 the percentage of wives who had a second birth rose from 67 to 81, and relatively larger increases occurred for third and fourth births.

TIMING AND SPACING OF BIRTHS

The arrival of these babies in later life increased the average length of the second interval by about five months (10 to 15 per cent) and that of the third and fourth intervals by about ten months (25 to 30 per cent). Changes of this type will be considerably smaller for the cohorts of 1926-30 than the two preceding groups, because so many more of their members had borne third and fourth children by ages 30-34. It is probable that when the women of 1926-30 TABLE 169 Age at Marriage and Length of Birth Interval, for White Wives When at Specified Ages, by Cohort

Ages and cohort Ages 20-24 1936-40 1931-35 1926-30 1921-25 1916-20 Ages 25-29 1931-35 1926-30 1921-25 1916-20 Ages 30-34 1926-30 1921-25 1916-20 Ages 35-39 1921-25 1916-20 Ages 40—44 1916-20

Per cent Average no. of mos. ever Num- Age at Mar. 1st 2nd 3rd mar- ber of marto to to to rieda wives riage 1st 2nd 3rd 4th A C D G B E F 69 69 67 59 53

440 479 459 426 315

19..0 19. 2 19. 3 19. 4 19. 4

17 16 17 17 15

22 23 24 24 26

20 23 18 19 19

19

89 88 87 81

600 592 632 495

20 .1 20 .3 20 .9 21..0

20 24 24 22

29 31 30 33

93 92 91

624 669 561

20 .7 21 .3 22 .0

27 28 27

95 93

677 569

21 .5 22 .1

94

572

22. 2

1st H

Per cent having 2nd 3rd 4th K J I

—

77 71 62 57 63

41 36 27 24 24

14 13 8 5 7

27 26 29 31

24 25 24 25

90 84 79 78

73 59 50 47

43 29 18 18

19 12 7 8

35 37 38

32 40 36

31 30 30

92 87 87

79 72 67

50 38 34

25 15 15

30 30

38 42

44 42

37 38

90 92

78 79

49 46

26 24

30

43

46

41

92

81

51

28

— — —

5 3 2 —

2

a The per cent of women ever married for the cohorts at the ages specified is esti­ mated from census reports and unpublished cohort tables of the Scripps Foundation.

reach middle age (in 1975) the average length of their second and third birth intervals will prove to be at least 10 per cent shorter than those of the two preceding groups. Whether the first and fourth intervals will be reduced remains to be seen. A greater shift toward younger marriage and closer spacing of children appears in comparisons up to ages 25-29, which include the cohorts of 1931-35. Here we find not only larger intercohort increases in the proportion of women having first, second, third,

FERTILITY AND FAMILY PLANNING IN THE U.S.

and fourth births, but also a more consistent pattern of reductions in the intervals between the births that have occurred. For example, the proportion of wives having a third birth by ages 25-29 rose from 18 per cent for the cohorts of 1916-20 to 43 for those of 1931—35 while the average length of the interval decreased from 31 months to 27 (Table 169, second deck). As the 1916-20 group lived from ages 25-29 to 40—44 the percentage having a third birth rose 33 points (to 51), but for the 1931-35 cohorts it probably will rise only about 17 points (to 60). If these additional intervals are as long for the recent group as the earlier one the average length of all second intervals will fall from 46 months to 40. But if the recent cohorts have their future third births somewhat closer together as they have done with their third births by 1960, the average length of all third intervals will be about 37 months. In short, there probably will be a 15 to 20 per cent reduction in the length of the third interval, and in the second as well, from the cohorts of 1916-20 to those of 1931-35. A somewhat smaller reduction (10 to 15 per cent) appears likely for fourth births, and a somewhat larger reduc­ tion (20 to 25 per cent) for first births. Data for the cohorts of 1936-40 are available only up to ages 20-24. Comparisons between these cohorts and those of earlier years will be restricted to first and second births because at least half of the wives in the two early groups of cohorts married after ages 20-24 (Column A) and most of those who married younger had not yet given birth to their third and fourth babies (Columns J and K). When considering trends in spacing up to ages 20-24 it is neces­ sary to take into acount the changes in the proportion of all women, as well as wives, who have had a birth of a given order, because of the large intercohort increase in the proportion of women marrying at the younger ages.4 By ages 20-24 the women in the 1936-40 cohorts had borne a somewhat higher proportion of their probable 4 The proportion of all white women who had a birth of a given order is approximately the product of Columns A and H, I, J, or K. More precise per­ centages by current ages 20-24, from cohort tables for native white women are as follows:

ORDER OF BIRTH

Cohorts 1936-40 1931-35 1926-30 1921-25 1916-20

First 56.9 52.5 46.8 38.1 34.0

Second 31.0 26.3 19.3 14.4 12.7

Third 11.2 8.6 5.2 4.3 3.7

Fourth 3.2 2.3 1.3 1.2 1.0

From unpublished tables of Whelpton and Campbell, Scripps Foundation.

TIMING AND SPACING OF BIRTHS

final number of first and second babies than those in the preceding group, and substantially more than those in earlier groups. The average length of the interval from marriage to first birth for wives in our study by ages 20-24 was fairly constant (between 15 and 17 months) for the different cohorts, but the interval between first and second birth had decreased consistently. By the time the 1936-40 women reach the end of the childbearing period, the proportion having a first birth probably will rise from about 57 per cent to somewhat above 85 per cent (the figure for the 1916-20 cohorts). If the first births of these young women after ages 20-24 come 34 months after marriage (the probable interval for the women of 1931-35) the average length of all first intervals will be about 23 months for the cohorts of 1936-40—the same as for those of 1931—35, and 20 to 25 per cent shorter than for the 1916—20 group. Present indications are that the final figures for second births will show slightly closer spacing for the 1936—40 cohorts than the preceding group, and a larger reduction from the 1916-20 cohorts than for first births. The shorter interval for second births to the 1936-40 women by 1960 (aged 20-24) suggests that there may be a similar speeding up of their second births after I960.5 If so, the final interval for all second births will be about 33 months, a 20 to 25 per cent reduction from the 1916-20 cohorts. The foregoing material shows quite conclusively that there was litde change in the spacing of first births after marriage, second births after first, and so on, between the two oldest groups of white wives in our study. Closer spacing apparently was begun by wives born in 1926-30 and continued by those in the next two groups (i.e., in their 20's in 1960). We cannot be certain as yet how large the reductions will be from the cohorts of 1916-20 to those of 1936-40, but the first and second intervals of the latter are likely to be about 20 to 25 per cent shorter, and the reduction up to the 1931-35 group probably will amount to 15 to 20 per cent for the third interval and 10 to 15 per cent for the fourth. Percentage Distribution of Final Birth Rate—The remaining measure of changes in the timing of childbearing that will be con­ sidered expresses age-specific cumulative birth rates of wives as per­ centages of the final birth rate. Births of different orders will be combined, so nothing will be shown directly about changes in the length of birth intervals. Instead, the percentages will show the speed 5This will occur if two tendencies continue as seems likely, namely, (a) single women 20-24 will marry earlier, when fecundability is higher, and (b) these women, and those married younger who have not had their second birth by ages 20-24, will not postpone it so long.

FERTILITY AND FAMILY PLANNING IN THE U.S.

of the family building process as a whole, taking into account changes in size of a completed family. The percentages in question are not affected by changes in the final rate for marriage or births unless they are accompanied by changes in the timing pattern. For example, if the final birth rate of all women rises from one cohort to another because more angle women marry and more wives have one birth instead of none, two instead of one, etc., but the percentage increases are the same at each age, there is no change in the timing pattern and the distribution of the final rate will remain constant. But if the tendency is greater at the younger ages, as has been the case among recent cohorts, there is a change in the timing pattern and the percentages will increase at younger ages and decrease at older ages.6 Intercohort changes in the percentages reflect the joint influence of changes in age at marriage and in the spacing of births after marriage. Younger marriage, with no change in birth spacing after marriage or in the final marriage or birth rate, obviously will raise the percentage of cohort members who are wives at ages 20-24, for example, and hence the proportion of all the cohort babies that are born up to these ages. Similarly, closer spacing of births after marriage with no change in age at marriage or in the final rates will increase the percentage of cohort babies born at younger ages. Because the wives in our study had not reached the end of the childbearing period, their actual cumulative birth rates will be ex­ pressed as a per cent of their most likely expected final rate. By the time the women born in 1936-40 reached ages 20-24 (in 1960), those currently married had a cumulative birth rate of 140 "For example, if all women in a hypothetical later cohort were to have agespecific birth rates 23 per cent higher than those of the 1910 cohort, the per­ centage distribution of the final rate would remain constant as is shown below. In contrast, the actual 1920 cohort will have a final rate about 23 per cent above that of the 1910 cohort, but the rate at ages 15-19 will be smaller and the increase will be largest at ages 25-29. Consequently, the percentages will show a strong tendency to postpone the starting of a family and to concentrate family building in the later 20's. Ages 13-19 20-24 25-29 30-34 35+ Final Birth rates (per 1,000) 1910 cohort 635 281 560 434 358 2,268 Hypothetical later cohort 778 687 344 532 439 2,780 1920 cohort 273 761 793 564 2,780* 389* Per cent of final rate 1910 cohort 12.4 28.0 24.7 19.1 15.8 100.0 Hypothetical later cohort 1920 cohort 9.8 27.4 28.5 20.3 14.0 100.0 * Estimated. The actual rates for the 1910 and 1920 cohorts are from Whelpton 1960.

TIMING AND SPACING OF BIRTHS

(per 100). Since their most likely expected final rate is 302, they had borne nearly half of all their children (Table 170, first line). In contrast, the study wives in the cohorts of 1916-20 who married by 20-24 (by 1940) had a much lower cumulative birth rate (97 per 100 instead of 140), and had given birth to a much smaller TABLE 170 Age at Marriage, Age-Specific Birth Rates, and These Rates as a Percentage of Most Likely Expected Total Births, for White Wives Married by Specified Ages, by Cohort

Age by which married and cohort® By 20-24 1936-40 1931-35 1926-30 1921-25 1916-20 By 25-29 1931-35 1926-30 1921-25 1916-20 By 30-34 1926-30 1921-25 1916-20 By 35-39 1921-25 1916-20

Per Av. Birth rates per 100 cent age at ages0 Un­ 20-24 25-29 30-34 at ever No. mar­ mar­ to of der to to ried1· wives riage 20-24 25-29 30-34 35-39

69 69 67 59 53

440 479 458 424 316

19.0 19.2 19.3 19.4 19.4

140 124 98 88 97

137 123 102 104

72 62 65

89 88 87 81

600 592 632 494

20.1 20.3 20.9 21.0

99 76 59 62

134 114 98 94

93 92 91

624 669 560

20.7 21.3 22.0

72 56 55

95 93

677 569

21.5 22.1

55 54

Most likely Per cent of most likely total total births at ages0 births Un­ 20-24 25-29 30-3' per der to to to 100 20-24 25-29 30-34 35-3!

34 38

302 340 336 309 325

46 36 29 28 30

40 37 33 32

21 20 20

11 11

86 78 78

44 44

335 329 305 307

30 23 19 20

40 35 32 31

26 26 25

14 15

108 93 82

86 77 80

48 53

324 300 304

22 19 18

33 31 27

27 26 26

16 18

92 81

76 79

48 53

298 301

18 18

31 27

26 26

16 18

Cohort groups are omitted unless some members had attained the upper age limit. The per cent of women ever married for the cohorts at the ages specified is estimated from census reports and unpublished tables of the Scripps Foundation. 0 In this table and in similar tables in this chapter, age refers to the current age on a given date for each cohort in the group. For example, the current age of the 1931-35 group on July 1, 1955 varied from 20 years for the 1940 cohort to 24 for the 1936 cohort; on July 1, 1960 the age of this group varied from 25 to 29 years. β

b

proportion of all their expected babies (30 per cent instead of 46). Comparisons between these cohorts and those of intermediate years show that the most of the rise in wives' birth rates by ages 20-24, and most of the acceleration in family building by these ages, oc­ curred after the cohorts of 1926—30.7 'The difference between the 1936-40 cohorts and those of 1931-35 and earlier is statistically significant. The differences between the 1931-35 cohorts and earlier groups are close to the borderline of statistical significance, some being below and others above.

FERTILITY AND FAMILY PLANNING IN THE U.S.

Average age at marriage of women marrying before 20-24 changed but little from the cohorts of 1916-20 to the 1931-35 group consequently the rise in the percentages from the cohorts of 1926-30 to 1931-35 must be ascribed to closer spacing. In contrast, both younger marriage and closer spacing contributed to the larger rise in the percentages from the wives born in 1931-35 to those in the following group. A comparison of the experience of wives up to ages 25-29 and 30-34 suggests that the trend toward earlier childbearing began with the cohorts of 1926-30. By ages 25-29 the wives in this group had a somewhat larger percentage of all the babies they expected than those in the two preceding groups (58 instead of 51). By ages 30-34 the differential is still larger and is statistically significant. The role of younger marriage in raising the percentage is more important as women are considered up to older ages. Among wives marrying by 25-29 the average age at marriage decreased by nearly a year from the cohorts of 1916-20 to those of 1931-35, and a steeper decline occurred among wives marrying by ages 30-34. THE USE OF THESE MEASURES

Neither mean length of birth intervals nor percentage distribution of final rate can measure precisely the changes that are under way in the timing or spacing of births to members of recent cohorts. With both measures it is necessary to make some sort of estimate about the future fertility of women who have not reached the end of the childbearing period. For mean length of intervals this was done above by estimating the final rate for first to fourth births, obtaining the future rate by subtracting the rate by 1960, and esti­ mating the length of future intervals on the basis of the experience of prior cohorts at comparable ages. For percentage distribution of the final rate it was done by using the most likely final rate expected by the wives interviewed. As was brought out in Chapter 1, there are several reasons for having confidence in the average expectations of wives regarding size of completed family. Moreover, we know the extent to which the wives in different socioeconomic groups raised or lowered their average expectations as they lived from 1955 to 1960; this can be kept in mind when comparing the birth spacing trends and differ­ entials of these groups. In contrast, there are several reasons for having less confidence in the estimates that can be made about the future or final rate for first to fourth births, and the future spacing of

TIMING AND SPACING OF BIRTHS

such births. For example, we know that young wives tend (1) to underestimate the extent to which subfecundity will prevent them from having as many births as they want, and (2) to overestimate their ability to prevent conception after they have as many as they want. As a consequence, even though a group of wives gives a very accurate estimate of the average final rate for all births, their estimates for first births (and probably second births) will be too high, and those for fifth and higher order births too low.8 Basing estimates of the spacing of future births on the experience of past cohorts is not satisfactory when efforts are being made to measure the extent of intercohort changes in spacing. Because the average length of past birth intervals is affected so greatly by age or duration of marriage, and because we lack an adequate basis for estimating the length of future birth intervals for various socioeconomic groups, this measure will be used later in this chapter in discussing differentials, but not trends, in the spacing of births. It is helpful in this connection because it is an everyday concept that is easy to interpret. The discussion of time trends of various groups will be based on the per cent distribution by age of the final birth rate according to the wives' most likely expectations. This rate is available for all groups, and is shown to be fairly stable by the comparisons between given cohorts in the 1955 and 1960 studies. The percentages relating births by 1960 to final expected births should have the same validity as the final rates themselves. However, as shown in the discussion of intercohort trends based on Table 170, it is not appropriate to use this measure without some control for age at marriage. For ex­ ample, the ratio of the birth rate at ages under 25 to the expected final rate is smaller for wives in the 1916-20 cohorts than for those in the 1931-35 cohorts, because some of the former (but none of the latter) married at ages 25 or older, and had few if any births before 25. In discussing intercohort trends for various socioeconomic groups it will not be feasible to subdivide by age at marriage in addition to the cohort and group characteristics. In consequence, that discussion will relate only to wives married before age 25. These wives include a substantial majority of all women in the cohorts of 1916 to 1940 who will marry before middle age. This eliminates " Because of the extent to which such expectations are upset by subfecundity and accidental conceptions, wives were asked only two questions about future birth spacing, namely, Q26. "When do you think you'll have your next child?" and, if they expected two or more, Q27. "How many years from now do you think it will be before you have all the future children that you expect?"

FERTILITY AND FAMILY PLANNING IN THE U.S.

the major part (about 75 to 80 per cent) of the influence of the intercohort reduction in age at marriage, and permits attention to be focused on the tendency toward earlier childbearing after marriage. EFFECT OF YOUNGER MARRIAGE AND CHILDBEARING ON POPULATION GROWTH

The various measures of intercohort changes in age at marriage and the timing of births make clear the fact that there has been a significant trend toward earlier marriage and childbearing among recent cohorts, even though it is impossible as yet to state precisely the size of the changes. We shall now consider briefly the question: How much has younger childbearing affected the rate of population growth? This may seem like a simple task, but in fact it is complicated. As we shall see, it is very important to distinguish between the shortrun effect and the long-run effect. In order to strengthen the discussion we shall use rates from cohort fertility tables for all women in the United States rather than those from the GAF studies. The Long-Run Effect—This will be shown by comparing the intrinsic rate of natural increase of two populations which are alike except that one has the fertility timing patern of the cohorts of 1911-15 and the other the pattern of the cohorts of 1931-35. The latter group is the most recent that can be used with assurance at present, before data on marriages and births are available for 1963 and later years. These women have borne nearly 80 per cent of all the babies they will have, compared with about 50 per cent for the 1936-40 group, consequently there is a much better basis for estimating their final birth rate and timing pattern. The 1911-15 cohorts are chosen because of their relatively late childbearing. These women married and had their first three babies at older ages than the women in prior or later cohorts (Table 165). Moreover, the proportion of the final birth rate occurring before ages 20-24, and during ages 20-24 to 25-29, is relatively low for the 1911-15 group, and the proportion at ages 30-34 and older is relatively high (Table 171). The per cents shown in this table will be used as composite measures of the timing patterns we are comparing. In order to elimi­ nate the effect of differences in mortality the hypothetical populations will be assumed to have the age specific death rates of 1960. A population with the final birth rate and timing pattern of the 1911-15 cohorts would have an intrinsic rate of natural increase of 3.0 per 1,000. If it had the younger timing pattern of the 1931-35 cohorts, the rate would rise to 3.2, a gain of nearly 7 per cent as shown below. If we reverse the comparisons we see that a population with the final birth rate and timing pattern of the cohorts of 1931-35

TIMING AND SPACING OF BIRTHS

TABLE 171 Final Birth Rate per 1,000 Women Living to Age 50, and Percentage Distribution by Age, for All Women in Cohorts of 1891-1895 to 1951-1955'

Cohorts

Final birth rate (per 1,000)

Under 20-24

1951-55 1946-50 1941-45 1936-40 1931-35 1926-30 1921-25 1916-20 1911-15 1906-10 1901-05 1896-1900 1891-95

2,820 2,930 3,020 3,130 3,300 3,170 2,895 2,565 2,315 2,271 2,420 2,672 2,962

33 34 35 35 30 25 22 22 24 28 29 26 24

Percentage distribution1· 20-24 25-29 30-34 to to to 25-29 30-34 35-39

36 36 36 35 32 31 29 27 27 30 31 29

19 19 22 24 25 25 24 21 21 23 24

35-39 to 50

7 9 13 15 16 17 15 12 13 15

4 6 8 8 8 9 8 7 8

" The rates and percentages for the cohorts of 1891-95 to 1911-15 are from Whelpton 1960. Those for more recent cohorts are from the actual births by 1960 the medium projections discussed in Chapter 10 and shown in Table 210. b Per cents below the horizontal lines relate to births that occurred before 1960. Those above the line are medium projections. Per cents are not shown for births that will occur after 1970.

would have an intrinsic rate of natural increase of 17.0, and a change to the later timing pattern of the cohorts of 1911-15 would reduce

Cohorts

1911-15 1931-35

Intrinsic rate of natural increase per 1,000 With timing pattern Actual of the other cohorts

3.0 17.0

3.2 15.9

Difference due to change in timing Rate

+.2 -1.1

the rate to 15.9. Changes of intermediate size would be shown if the timing patterns of the cohorts of 1916-20 and 1926-30 were used in place of those of either the 1911-15 or the 1931-35 groups. Similarly, since the projections for the cohorts of 1936-40 or later reflect earlier timing than that of the 1931-35 group, comparisons between the 1911-15 cohorts and the later groups would show still larger changes in the intrinsic rate. The Short-Run Effect—This is more difficult to measure, because there is no convenient yardstick like the intrinsic rate of natural increase, and it is necessary to resort to models. A few have been prepared in attempts to determine the extent to which the slow popu­ lation growth during the depression years of the 1930's and the

FERTILITY AND FAMILY PLANNING IN THE U.S.

rapid growth during the post war years were due to changes in the final marriage and birth rates as compared with changes in age at marriage and childbearing.9 Unfortunately, none of them focuses on the question raised here—the effect of the recent shift to younger childbearing—consequently another model has been pre­ pared. It shows the number of births and the population growth that would occur from 1945 to 1970 if all women in the cohorts of 1916-20 and later had (1) their actual (or projected) final rates and timing pattern, and (2) the same final rates but the timing pattern of the cohorts of 1911-15 (see Table 171). The differences TABLE 172 Estimated Increase in Births and Population Growth during Five-Year Periods Due to Different Ages at Childbearing: Timing Patterns of 1911-1915 Cohorts Compared with Those of Later Cohorts"

Periodb A 1965-70 1960-65 1955-60 1950-55 1945-50

Timing pattern Actual (or projected) for cohorts of Same as for cohorts of 1911-15 1916-20 and later Popu­ Popu­ No. of lation No. of lation births growth growth births B C D D 22,300 21,450 21,265 19,640 17,700

14,200 14,000 15,400 13,590 11,755

20,240 19,050 18,295 17,530 16,470

12,190 11,660 12,500 11,530 10,555

Increase due to change in timing Population growth No. of No. of Per births persons cent F G H ( B - D ) (C — E) (G -S- E) 2,060 2,400 2,970 2,110 1,230

2,010 2,340 2,900 2,060 1,200

16.5 20.1 23.2 17.9 11.4

a The model is based on the medium projections of birth rates and population growth, described in Chapter 10. Numbers are in thousands. Births are adjusted for incomplete registration. b July 1 through June 30.

between the two sets of results represent the additional births and population growth that occurred by 1960, or will occur during the 1960's according to the medium projection in Chapter 10, because of the trend toward earlier childbearing from the cohorts of 1911-15 to those of 1931-35 and later. Let us look first at the results of 1945-50, on the bottom line of Table 172. The actual number of births during the five years beginning July 1, 1945 was approximately 17,700,000 (Column B). If the women in the cohorts of 1916 and later had had the same pattern of age at marriage and childbearing as those of 1911-15, there would have been about 16,470,000 births (Column D). In * For examples see Whelpton 1954, p. 266, Whelpton 1960, and FPSPG, pp. 368-370.

TIMING AND SPACING OF BIRTHS

other words, the new pattern raised the number of births in 1945-50 by about 1,230,000 (Column F) and population growth by 1,200,000 or 11.4 per cent (Columns G and H). The trend toward younger childbearing was comparatively new in those years; as it continued its impact increased rapidly. The extra babies numbered about 2,110,000 during 1950-55 and 2,970,000 during 1955-60, and swelled population growth by 18 per cent and 23 per cent, respectively. During the current five-year period the medium projection of popu­ lation growth is 14,000,000 (Chapter 10), with about 2,400,000 or 20 per cent resulting from earlier childbearing. The influence of this factor probably will be substantially smaller in 1965-70, be­ cause the births which younger marriage and closer spacing will move into that period from later years will be balanced in greater degree by those previously moved into preceding periods. Eventually, if women have the same pattern of age at childbearing as those in the cohorts of 1931-35, the stimulus to population growth would shrink to about 7 per cent—the long-time effect according to the intrinsic rate of natural increase. Only the model based on the medium projections of age at child­ bearing is discussed above. As far as the 1945-60 period is concerned, the number of births that have occurred to each cohort group is known, consequendy the trend toward earlier timing can be increased only if the final birth rate of the cohort group is reduced. A model based on a greater shift toward earlier timing (and hence on smaller final rates) would show that births and population growth were in­ creased in greater degree than indicated in Table 172. Correspond­ ingly, a model based on a lesser shift toward younger childbearing (and hence on larger final rates) would show a smaller contribution to births and population growth from the timing factor. It would be impossible, however, to vary the trends and results in major degree without using final rates that would be below (or above) the limits that appear reasonable in the light of the expectations of the GAF wives. The same principle applies, though in lesser degree, to the results for the 1960's shown by the medium model in Table 172. In summary, it is evident that younger childbearing has con­ tributed in major degree to the upsurge in births and the size of the United States population since World War II. The influence of this factor increased from 1945-50 to 1955-60, and almost certainly will diminish during the 1960's. The latter needs to be kept in mind when considering the reasons for the decline of the crude birth rate from 25.0 in 1957 to 21.6 in 1963.10 "This topic will be discussed in more detail in Chapter 10 in connection with population projections.

FERTILITY AND FAMILY PLANNING IN THE U.S.

SOCIOECONOMIC DIFFERENCES IN AGE AT MARRIAGE AND THE TIMING OF BIRTHS

As noted earlier, we will study group differences in birth spacing with the use of the average number of months between marriage and the first birth and between successive births up to the fourth. Such intervals are presented in Table 173 for 18-39-year-old white wives classified by several socioeconomic characteristics. In addition, this table also shows the proportion of wives who have had a birth of each order up to the fourth. These percentages can help in evaluat­ ing the eventual accuracy of the averages for length of interval. For example, when we compare average lengths of the interval between marriage and the first birth for the different religious groups, we can see that we are comparing intervals for a large majority of wives in each group (the 87 to 92 per cent who have had a first birth). This assures us that the average length of the first birth interval will change but little by the end of the childbearing period. However, the intervals between the second and third births are based only on the wives who had a third birth. The proportion of such wives varies between 27 per cent for Jews and 46 per cent for Catholics. We do not know how many more women in each group will have a third birth, but it could be fairly substantial. Hence, we have less confidence that the differences shown for the interval between the second and third birth adequately represent the differences that will eventually prevail. We should also keep in mind the fact that the lower the number of women for which an interval is given, the greater the sampling variability. Thus, there are two good reasons for having more con­ fidence in the differentials for first and second birth intervals than in those for third and fourth. Before discussing differences in birth spacing, let us see how age at marriage varies between socioeconomic groups. The classification of wives by religious preference shows that Jew­ ish wives are oldest when first married (22.0 years old, on the aver­ age), Catholic and "other" (mostly Orthodox) wives are somewhat younger (21.0 years old), and Protestant wives are youngest (20.0 years old). The Protestant-Catholic differential noted here was also found in the 1955 Study. Wives classified by education attainment show wider differences in age at marriage—from about 19 years for wives with only a gradeschool or incomplete high school education to 22 years for those with a college education. Part of this difference is doubtless due to a selective factor—the low probability that a woman who marries

TIMING AND SPACING OF BIRTHS TABLE 173 Age at Marriage, Average Years Married, Length of Birth Intervab, Per Cent Having Specified Number of Births, and Number of Births by Interview and Total Expected for Selected Characteristics

Characteristic

Aver. no. of Aver. Aver. Aver. no. of mos. births No. age yrs. Mar. 1st 2nd 3rd Total of at mar- to to to to Per cent having By exwives mar. ried 1st 2nd 3rd 4th 1st 2nd 3rd 4th 1960pected

Total 2,414 Wife's religion Protestant 1,596 Catholic 668 Jewish 106 Other 44 Wife's education College 427 High school, 4 1,153 High school, 1-3 579 Grade school 255 Husband's income $10,000 or more 261 $7,000-59,999 405 $6,000-$6,999 312 $5,000-$5,999 423 $4,000-54,999 380 $3,000-53,999 306 Under $3,000 327 Husband's occupation Upper white collar 725 Lower white collar 312 Upper blue collar 465 Lower blue collar 670 154 Farm Region of residence 581 Northeast 715 North Central 744 South 374 West Size of place of residence 12 central cities 187 Other cities of 50,000 + 427 Rings of 12 central cities 443 Rings of other cities 537 Cities 2,500-49,999 282 Rural nonfarm 335 203 Rural farm

20. 3 10. 2

24

33

34

88

68

40 19

2. 3

3.1

20. 0 10. 3 21. 0 9. 8 22. 0 11. 6 21. 0 9. 3

24 22 31 24

34 30 39 29

34 32 87 34 28 89 42 43 92 — — 89

67 70 81 64

39 18 2. 2 46 26 2. 5 6 2.1 27 41 16 2. 2

3. 0 3. 7 2. 5 3. 0

9. 0 22. 3 9. 5 20. 7 18. 9 11.1 19. 1 13. 0

28 25 22 21

32 34 32 32 33 32 34 35 30 33 35 30

82 86 92 95

65 64 74 81

36 13 2. 0 34 16 2.1 47 24 2. 6 58 38 3. 1

3. 0 3. 0 3. 3 3. 7

21..5 21..1 20..4 20..3 20. 1 19. 8 19. 6

12..0 11..1 11. 0 10..3 10..0 8. 4 8. 4

31 27 25 24 24 19 20

33 34 33 35 31 31 31

37 35 36 38 33 32 27

32 35 32 29 31 28 29

90 87 88 90 91 85 81

76 70 75 70 70 62 55

44 42 44 38 43 35 34

2 .5 2 .3 2. 4 2..3 2. 3 2. 2 2,,1

3 .1 3 .0 3.1 3..1 3. 2 3. 3 3,.2

21..2 19. 7 19,,7 19 .8 20 .2

10..3 9. 8 10..7 10 .3 10 .3

26 25 23 23 21

34 33 34 31 30

35 37 35 33 31

33 86 34 84 32 90 29 91 27 88

68 38 17 2 .2 65 35 16 2..1 71 43 21 2 .4 70 42 22 2 .5 75 53 27 2 .6

3 .0 3 .0 3 .2 3.2 3 .5

21 .2 9 .8 20 .4 10 .1 19 .9 10 .3 19 .9 10 .6

23 23 26 25

32 32 35 33

34 30 90 34 30 88 34 35 84 35 28 90

70 71 63 69

40 43 35 45

20 22 16 20

2 .4 2 .4 2 .1 2 .4

3 .2 3 .3 2 .9 3 .2

9 .8 24 9 .6 24 9 .8 26 10 .7 26 10 .5 23 10 .5 23 10 .3 21

29 32 31 33 38 35 29

41 35 34 32 39 33 31

87 63 86 65 86 64 89 73 87 68 89 72 88 73

33 36 37 43 41 43 48

14 2 .1 18 2 .2 19 2 .2 20 2 .4 18 2 .2 21 2 .4 26 2 .7

3 .0 3 .1 3 .1 3 .2 3 .0 3.1 3 .6

21 .1 20 .6 20 .8 20 .4 20 .0 19 .6 19 .6

31

34 34 29 32 29 33 27

23 21 19 18 17 19 20

FERTILITY AND FAMILY PLANNING IN THE U.S.

at a young age (say 17) will go to college. However, it may also reflect differences in marriage customs for different segments of the population. Age at marriage seems to be directly related to income: the higher the income, the higher the age at marriage. This is probably due at least in part to the greater proportion of college wives in the higher income group. Age at marriage is highest among wives whose husbands have an upper white-collar occupation. The remaining occupation groups show little difference in age at marriage. Age at marriage is higher in the Northeast than in the rest of the country—probably because of the higher proportion of Catholics in this region. Regional differences are not great, however. Differences between the various size-of-place categories are also rather small as far as age at marriage is concerned. In general, average age at marriage appears to be influenced in only minor degree by socioeconomic factors. The average for all white wives 18-39 is 20.3 years, and most groups show averages close to this figure. The major differences are that the better-educated tend to marry later than the less-educated, and that Catholics marry later than Protestants. But even here the differences are not great. The important fact to keep in mind is that age at marriage is so low in all groups that the great majority of couples marry early enough to avoid the fecundity impairments that become prominent at later ages. Consequently, most of them have ample time to have all of the children they want. For the various socioeconomic groups in Table 173, the average number of months between first marriage and first birth varies be­ tween a low of 19 months and a high of 31.11 These averages may be somewhat deceptive if it is not realized that they relate to highly skewed distributions. For example, the average number of months in the first birth interval is 24 for all white wives 18-39 years old. However, two-thirds of the wives with a first birth had it before they had been married 24 months. In fact, 50 per cent of them had a first birth within 15 months after marriage (this is the median value for the first birth interval). The most frequently reported inter11 In discussing the interval between marriage and first birth, it should be noted that the averages include births occurring before marriage, which have negative intervals. Hence, if we were to limit comparisons only to births occurring after marriage, the average number of months from marriage to first birth might be slightly higher than the 24 months now shown. However, only 1.4 per cent of the wives with births reported that their first birth occurred before marriage. Therefore, the influence of negative intervals is negligible. This proportion is greatest in the grade-school group, but even here it is only 5.0 per cent.

TIMING AND SPACING OF BIRTHS

val (i.e., the mode) is even lower: 10 months. The relationship be­ tween these values and the distribution of first birth intervals can be seen more clearly in Figure 36. The reason that the average value is so high is, of course, that some couples had their first birth long after marriage. The proportion of couples who did so is small (e.g., only 7.5 per cent had the first birth after five years of marriage) but it is high enough to keep the average well above the median. MODE(IO) PERCENT

8

1 MEDIAN 115) AVERAGE (24)

7

6

5

4 3

2

0 -IO

20

30

40

50

60

70

80

90

100

IIO

MONTHS

FIGURE 36 Per cent distribution by number of months between first marriage and first birth, white wives, 18-39

The average length of the first birth interval is highest for Jewish wives (32 months). In fact, this is the highest value for this interval shown in Table 173. This is consistent with the high proportion of couples in this group who use contraception before the first pregnancy. Protestants have a slightly higher average first birth interval than Catholics, but the difference is not great (24 versus 22 months). However, the distributions of first birth intervals for the two religious groups show some significant differences (Table 174). For example, the proportion of couples who have a first birth in the first year

FERTILITY AND FAMILY PLANNING IN THE U.S.

of marriage is 37 per cent for Catholics, but only 29 per cent for Protestants. The four education groups show relatively wide differences in the average length of the first birth interval: 28 months for the college wives and 21 months for those with only a grade-school educa­ tion. The distributions of intervals also differ widely: the proportion of wives reporting a first birth in the first year of marriage is 36 per cent for grade-school educated wives, compared with only 23 per cent for those with some college education (Table 174). The classification of wives by their religion and education provides some insight into the spacing differentials for these two variables TABLE 174 Per Cent Distribution by Months between Marriage and First Birth, for Wives, by Religion and Education Years between first marriage and first birth

Characteristic Total Wife's religion Protestant Catholic Other Wife's education College High school, 4 High school, 1—3 Grade school

Before mar­ Under riage 1 year

No. of wives

Total

2,113

100

1

1,382 595 136

100 100 100

2 1

350 988 534 241

100 100 100 100

7 or 3-4.99 5-6.99 more years years years

1-1.99 years

2-2.99 years

31

35

13

13

4

4

—

29 37 22

36 35 31

13 11 15

13 10 24

4 3 5

4 4 4

1 1 2 5

23 31 33 36

36 35 38 29

15 13 10 13

16 14 10 11

4 4 3 3

5 3 3 3

(see Table 175). We noted above that Protestants had slightly longer first birth intervals than Catholics (24 and 22 months, respectively). Table 175 indicates that this differential is due almost entirely to the different spacing patterns of Protestant and Catholic wives with some college education. College-educated Protestant wives report an average of 29 months between marriage and the first birth, and simi­ larly educated Catholic wives report only 19 months. In the lower education groups, the religious differentials are small and irregular. Generally, Table 175 indicates that Catholics have shorter interbirth intervals than Protestants. Of the 15 comparisons that can be made, Catholics have shorter intervals in 10, equal intervals in 2, and longer intervals in only 3. However, the differences are small. It is clear that the two religious groups do not have very different spacing patterns.

TIMING AND SPACING OF BIRTHS

The association between the first birth interval and education is strongest among Protestants. As noted above, college-educated Protestant wives report an average of 29 months between marriage and first birth; those with a grade-school education report only 21 months. Among Catholics, however, there is little association between the first birth interval and wife's education. The averages for the different education groups vary irregularly between 19 and 24 months. Husband's income is related to the average length of the first birth interval (Table 173). The higher the income, the longer the TABLE 175 Age at Marriage, Average Years Married, Length of Birth Intervals, Per Cent Having Specified Number of Births, and Number of Births by Interview and Total Expected for Protestant and Catholic Wives, by Wife's Education Aver. no. of

Wife's education College Protestant Catholic High school, 4 Protestant Catholic High school, 1-3 Protestant Catholic Grade school Protestant Catholic

Aver. Aver. Aver. no. of mos. births No. age yrs. Mar. 1st 2nd 3rd Total of at mar- to to to to Per cent having By ex­ wives mar. ried 1st 2nd 3rd 4th ist 2nd 3rd 4th 1960pected 284 79

22.0 23.2

9.1 7.6

29 19

33 22

34 32

34 *

81 84

64 36 13 63 44 19

2.0 2.3

2.8 4.2

752 341

20.2 21.2

9.7 8.9

25 23

33 31

33 34 34 27

85 86

61 66

33 39

14 20

2.0 2.3

2.8 3.6

392 168

18.5 20.1

11.1 11.2

21 24

35 31

35 34

31 31

91 93

74 74

46 50

21 31

2.6 2.7

3.1 3.6

168 80

18.9 19.6

12.9 13.0

21 21

35 28

35 37

32 28

93 99

81 81

53 66

35 45

2.9 3.5

3.5 4.3

couple waited for their first birth. The average is 31 months if hus­ band's income is $10,000 or more and only 20 months if it is less than $3,000. Husband's occupation is also directly related to the length of the first birth interval: the higher the occupational status, the longer the interval. However, the differences are not large. There are relatively minor variations in the average length of the first interval when couples are classified by region of residence or size of place of residence. In general, then, the length of the first birth interval is clearly related to socioeconomic status: the higher the status, the longer the interval. This is obviously linked with group differences in the timing

FERTILITY AND FAMILY PLANNING IN THE U.S.

of the first use of contraception and the effectiveness with which contraception is used early in marriage. The relationship between the lengths of various birth intervals, the timing of first use of contraception, and socioeconomic status is illustrated in Table 176, where wife's education is used as a measure of socioeconomic status. This table indicates that the timing of the TABLE 176 Age at Marriage, Average Years Married, Length of Birth Intervals, Per Cent Having Specified Number of Births, and Number of Births by Interview and Total Expected, by Wife's Education by Interval When Contraception First Used Aver. no. of

Interval when contraception first used Before 1st preg. College High school, 4 High school, 1-3 Grade school Between 1st & 2nd College High school, 4 High school, 1-3 Grade school Between 2nd & 3rd College High school, 4 High school, 1-3 Grade school After 3rd College High school, 4 High school, 1--3 Grade school Never used College High school, 4 High school, 1- -3 Grade school

Aver. Aver. Aver. no. of mos. births No. age yrs. Mar. 1st 2nd 3rd Total of at mar- to to to to Per cent having By ex­ wives mar. ried 1st 2nd 3rd 4th ist 2nd 3rd 4th 1960pected 259 472 140 35

22 .1 9 .1 33 33 39 33 20 .6 9 .2 28 35 35 34 19 .2 10 .6 27 39 37 — 19 .6 11 .6 24 39 — —

70 270 155 50

22 .1 9 .6 20 32 20..4 9 .3 19 34 18. 7 11..1 22 38 19.2 12.,0 21 34

29 112 83 35

22. 1 9 .7 20. 6 10. 6 18. 5 11. 2 19..0 13..2

79 82 85 94

60 60 61 69

30 8 1 .8 28 10 1 .9 34 11 2 .0 40 11 2 .2

2 .7 2 .8 2 .7 3 .0

30 99 36 38 99 43 30 99 40 — 100

89 71 79 94

56 37 46 46

20 2 .8 17 2 .3 19 2..5 24 2..8

3 .6 3 .2 3. 2 3. 3

28 25 .—. — 100 23 25 31 28 100 23 28 28 32 100 22 24 41 — 100

93 96 92 91

59 48 57 66

24 27 40 40

2 .8 2..9 3. 1 3 .3

3. 8 3. 6 3. 6 3..9

26 26 100 98 81 42 3.7 27 33 100 97 87 56 4.0 32 32 100 98 86 67 4.1

4.3 4.5 4.7

15 84 62 42

19.7 12.4 22 18.4 13.3 18 18.5 14.0 21

30 31 32

50 200 129 87

24.2 6.6 21 21.2 8.9 33 19.6 10.4 19 19.4 13.6 20

31 37 — 31 37 25 34 32 28

—

—

58 60 82 86

34 29 59 67

14 6 18 8 36 19 54 40

1.2 2.9 1.2 2.4 2.2 3.2 3.0 3.6

first use of contraception is not the only factor that causes birth intervals to vary. For example, among couples who used contraception before their first pregnancy, the college wives had longer first births intervals than the grade-school wives (33 versus 24 months). We do not know whether the couples in the grade-school group wanted their first birth sooner than those in the college group or whether they were less effective Users, but we suspect that the latter is the primary cause of the difference.

TIMING AND SPACING OF BIRTHS

It is interesting to note, however, that after the couple begins to use contraception, the interval lengths do not necessarily vary with wife's education. For example, the second and third birth intervals are sometimes longer for the grade-school group than the college group. Much of the apparently random variation may be due simply to sampling variability. On the average, the education groups have about the same numbers of months in the second, third, and fourth intervals (Table 173). The second interval varies between 32 and 34 months, the third interval between 33 and 35 months, and the fourth interval between 30 and 32 months. In general, the second, third, and fourth intervals do not appear to vary with socioeconomic status as much as the first interval does. It is difficult to be sure that this is true because sampling variability is so much greater for the higher-order intervals that we would expect more variation from this source alone. In other words, for the higher birth intervals it is hard to say how much of the variation is due to sampling error and how much is due to real differences between the groups sampled. We can say, however, that the length of the higher-order birth intervals is not as directly related to socioeconomic status as the first birth interval is. That is, there is less indication for the higher order births that higher socioeconomic status is associated with longer intervals between births. It is tempting to theorize that couples in the different socioeco­ nomic groups vary mainly in when they start their families, but have approximately equal intervals between later births. Our data do not permit us to verify this guess, however. TRENDS IN THE TIMING OF BIRTHS BY SOCIOECONOMIC STATUS

As noted earlier in this chapter, we shall study cohort trends in timing for the different socioeconomic groups with the use of tables showing the per cent distribution of the most likely expected total number of births by the age of occurrence of births up to 1960. These data relate only to women married before age 25. The central topic of this section is, then, the proportion of births occurring in certain age intervals. Such proportions are affected not only by age at marriage and the spacing of births, which were dis­ cussed in the preceding section, but also by the final number of births expected. Obviously, couples with larger numbers of births will tend to have them later in life than other couples because high-order births must come at later ages. A couple that married when the wife was 20 and had only one child, for example, could have 100 per cent of their childbearing during ages 20-24. But an­ other couple who married at the same age and had five children would be very likely to distribute them over several five-year periods.

FERTILITY AND FAMILY PLANNING IN THE U.S.

They might have three children, or only 60 per cent of their childbearing, at ages 20-24. In spite of this tendency, couples in the cohorts of 1931-35 expect more births than couples in earlier cohorts, but have had a higher proportion of their expected births before age 25. This compression of childbearing into the younger ages results partly from younger age at marriage and partly from having children with less delay after marriage. The trend toward earlier childbearing has been shared by all socioeconmic groups. Table 177, for example, shows that Protestants have increased the proportion of total births occurring before age TABLE 177 Age at Marriage and Per Cent Distribution of Most Likely Expected Total Births by Age for Protestant and Catholic Wives Marrying before Age 25, by Cohort

Religion and cohort Protestant 1936-40 1931-35 1926-30 1921-25 1916-20 Catholic 1930-40 1931-35 1926-30 1921-25 1916-20

Most likely Per cent of most likely expected Average expected total births occurring at ages No. of age at total Under 20253035- Under wives marriage births 20 24 29 34 39 25 312 375 364 374 297

18.4 18.9 19.3 19.6 19.7

2.85 3.10 3.12 2.94 3.10

20 14 10 10 11

42 36 34 30

32 27 28

112 167 157 156 123

19.1 20.1 19.9 20.4 20.5

3.48 4.04 3.93 3.46 3.41

9 7 6 4 5

36 29 27 29

33 31 30

19 17

21 21

9

56 46 44 41

10

43 35 31 34

25 from 41 per cent in the cohorts of 1916-20 to 56 per cent in the cohorts of 1931—35 (the most recent cohort group for which we can make a good estimate). Catholics also show a rise in this proportion: 34 per cent for the earlier cohort group and 43 per cent for the later. Although the rise for Catholics does not seem as impressive as that for Protestants, we must remember that it was accompanied by a greater increase in most likely expected births. Therefore, it probably represents at least as great a change in the timing of births of a given order. Table 177 also shows that the greatest change in timing patterns is between the cohorts of 1926-30 and 1931-35. This was observed earlier for all couples, but here we can see that it is also true for both Protestants and Catholics. All of the education groups show the trend toward earlier timing of births. This finding may seem somewhat surprising because a priori

T I M I N G AND SPACING OF B I R T H S

reasoning suggests that the trend toward younger childbearing should be concentrated among the better-educated. The argument goes like this: the less-educated have always been reluctant to begin using contraception early in marriage, so they have always had relatively many births by young ages; the better-educated, however, have fre­ quently postponed births; consequently, the major source for a deTABLE 178 Age at Marriage and Per Cent Distribution of Most Likely Expected Total Births by Age, by Wife's Education, by Cohort

Education and cohort College, 4+ 1936-40 1931-35 1926-30 1921-25 1916-20 College, 1-3 1936-40 1931-35 1926-30 1921-25 1916-20 High school, 4 1936-40 1931-35 1926-30 1921-25 1916-20 High school, 1-3 1936-40 1931-35 1926-30 1921-25 1916-20 Grade school 1936-40 1931-35 1926-30 1921-25 1916-20

Most likely Per cent of most likely expected Average expected total births occurring at ages No. of age at total Under 20- 25- 30- 35- Under wives marriage births 20 24 29 34 39 25 16 35 40 31 21

*

*

*

21 .6 21 .6 21 .9 22..2

2.96 3.16 2.88 2.61

1 2

42 67 64 56 37

19. 6 20 .4 20 .2 21 .1 21 .5

2.98 3.36 2.80 2.75 2.27

240 277 256 267 187

19..0 19 .7 19 .9 20..3 20..6

2.87 3.14 3.30 2.84 2.71

114 133 128 138 122

17 .5 17..6 18..4 18..8 19,.2

28 55 72 82 84

16..3 18..5 18..0 18..8 18. 8

26 20 16 15

48 35 46

39 36 26 25

38 36 38

12 7 6 4 5

42 32 32 31

35 32 33

3.11 3.61 3.50 3.25 3.51

29 23 13 15 12

41 37 34 30

25 24 25

3.93 4.12 3.66 3.70 4.03

33 17 17 13 14

37 36 31 28

28 25 25

— —

6 3 2 1 —

30 26

25 27

20 19

15 19

16 16

7

27 22 16 15

6

42 38 27 25

7

49 38 36 36

10

64 50 49 42

11

54 53 44 42

crease in age at childbirth should be the better-educated. The data show, however, that wives with only a grade-school education are having births at younger ages in the more recent cohorts. For example, the grade-school group shows a rise in the proportion of births occur­ ring before 25 from 42 per cent in the 1916-20 cohorts to 54 per cent in the 1931-35 cohorts (Table 178). College graduates show

FERTILITY AND FAMILY PLANNING IN THE U.S.

a rise in this proportion from 15 per cent in the earlier group to 27 per cent in the later. The different magnitudes of these changes are not as important as the fact that they show that all educational groups have participated in the trend toward earlier childbearing. Table 179 shows that all occupation groups, too, have participated in the trend toward earlier childbearing. The percentage of expected TABLE 179 Age at Marriage and Per Cent Distribution of Most Likely Expected Total Births by Age, by Husband's Occupation, by Cohort

Husband's occupation and cohort Upper white collar 1936-40 1931-35 1926-30 1921-25 1916-20 Lower white collar 1936-40 1931-35 1926-30 1921-25 1916-20 Upper blue collar 1936-40 1931-35 1926-30 1921-25 1916-20 Lower blue collar 1936-40 1931-35 1926-30 1921-25 1916-20 Farm 1936-40 1931-35 1926-30 1921-25 1916-20

Most likely Per cent of most likely expected Average expected total births occurring at ages No. of age at total Under 20- 25- 30- 35- Under wives marriage births 20 24 29 34 39 25 92 156 192 185 123

19 .2 19 .7 19 .9 20 .6 21 .0

2 .76 3 .35 3 .10 2..83 2 .47

11 9 6 4 4

37 29 28 27

39 34 34

56 75 67 75 52

18 .9 19 .8 19 .7 20 .2 20 .5

2 .96 3 .28 3 .06 2..72 3..12

12 9 10 6 4

37 34 30 26

32 28 28

93 121 101 120 117

18 .2 19 .0 19,,4 19,.0 19.,4

3..06 3..43 3. 48 3.,04 3. 13

20 14 9 12 13

40 34 35 28

33 23 27

139 157 160 145 106

18,.2 18 .9 19,.1 19..5 19..3

2..99 3,.36 3,.50 3..37 3. 58

22 13 12 10 9

45 35 33 33

27 25 28

28 42 32 35 45

18..4 19 .1 19..2 20 .0 19..6

3,.31 3.,50 3,,81 3.,52 3..84

20 15 5 5 11

37 39 28 29

31 38 28

20 23

23 21

18 18

19 17

15 16

6

46 35 32 31

13

46 44 36 32

9

54 43 47 41

8

58 47 43 42

11

52 44 33 40

births occurring before age 25 has risen by approximately the same amount in each occupation group. For example, the rise in the upper white-collar group is from 31 per cent (1919-20 cohorts) to 46 per cent (1931-35 cohorts), or 15 points. The comparable rise for the lower blue-collar group is from 42 to 58 per cent, or 16 points.

TIMING AND SPACING OF BIRTHS TABLE 180 Age at Marriage and Per Cent Distribution of Most Likely Expected Total Births by Age, by Size of Place of Residence, by Cohort

Size of place of 1960 residence and cohort 12 central cities 1936-40 1931-35 1926-30 1921-25 1916-20 Other cities of 50,000 + 1936-40 1931-35 1926-30 1921-25 1916-20 Rings of 12 central cities 1936-40 1931-35 1926-30 1921-25 1916-20 Rings of other cities 1936-40 1931-35 1926-30 1921-25 1916-20 Cities 2,500-49,999 1936-40 1931-35 1926-30 1921-25 1916-20 Rural nonfarm 1936-+0 1931-35 1926-30 1921-25 1916-20 Rural farm 1936-40 1931-35 1926-30 1921-25 1916-20

Most likely Per cent of most likely expected Average expected total births occurring at ages No. of age at total Under 20- 25- 30- 35- Under wives marriage births 20 24 29 34 39 25 35 35 37 43 47

18.1 19.9 19.9 20.1 20.5

3.29 3.08 3.04 3.06 2.80

18 9 12 4 8

38 31 37 32 28 20 29 33 22

81 101 96 94 87

19.2 19.4 19.6 20.0 20.0

2.84 3.53 3.23 3.03 3.12

11 12 8 6 11

43 30 34 33 29 20 29 29 18

82 98 108 99 77

19.0 19.8 19.8 20.3 20.1

3.01 3.19 3.47 2.90 2.89

74 145 145 132 89

18.3 19.4 19.7 20.1 20.1

59 59 56 78 50

5

47 43 36 37

8

55 38 39 40

12 9 7 5 7

37 34 34 31 32 17 27 27 22 10

46 41 36 34

3.09 3.42 3.25 2.93 3.21

17 9 6 8 8

41 29 36 32 31 19 30 28 20

8

50 35 40 38

18.9 18.8 19.0 19.7 20.2

2.94 3.17 3.18 2.87 2.84

13 17 11 11 8

36 42 28 27 29 21 25 32 20 10

53 53 38 33

66 81 74 82 51

17.8 19.0 18.7 19.2 19.4

2.93 3.47 3.39 3.01 3.15

27 12 14 12 8

43 38 27 29 24 21 32 32 13 11

55 52 41 40

43 48 44 46 50

17.7 18.9 19.4 19.4 19.7

3.26 3.39 3.74 4.18 3.96

26 17 9 8 10

39 35 33 34 28 15 31 27 16 11

56 44 42 41

FERTILITY AND FAMILY PLANNING IN THE U.S.

Tables 180 and 181 indicate that couples living in all size-of-place categories and in most regions of the United States have been concen­ trating their births more and more in the younger age groups. The only possible exception is in the South, where the cohort trend toward earlier childbearing is not as marked as it is in the rest of the country. Even in the South, however, there has been some increase in the proportion of births occurring before age 25. TABLE 181 Age at Marriage and Per Cent Dbtribution of Most Likely Expected Total Births by Age, by Region, by Cohort

Region and cohort Northeast 1936-40 1931-35 1926-30 1921-25 1916-20 North Central 1936-40 1931-35 1926-30 1921-25 1916-20 South 1936-40 1931-35 1926-30 1921-25 1916-20 West 1936-40 1931-35 1926-30 1921-25 1916-20

Most likely Per cent of most likely expected Average expected total births occurring at ages No. of age at total Under 20253035- Under 29 wives marriage births 20 24 34 39 25 104 127 131 139 124

19.0 19.7 20.0 20.5 20.4

3.17 3.46 3.69 2.91 2.84

14 9 8 5 6

40 29 30 28

36 34 32

129 176 176 166 151

18.7 19.4 19.9 20.4 20.2

3.13 3.66 3.45 3.14 3.29

14 11 6 6 7

42 33 29 28

33 32 28

145 182 168 170 107

18.1 19.1 19.1 19.2 19.5

2.85 2.96 2.88 2.94 3.09

22 13 10 12 12

36 35 31 31

31 25 29

62 82 85 99 69

18.4 19.0 18.8 19.3 19.7

2.91 3.42 3.38 3.32 3.38

18 14 13 8 11

43 39 37 29

30 25 29

18 22

20 20

20 13

16 19

6

49 37 35 34

11

53 39 35 35

9

49 45 43 43

9

57 52 45 40

The data presented in this section show conclusively that the cohort trend toward earlier childbearing has not been an isolated phenomenon. Instead, it represents a major change in family planning patterns that has occurred in virtually all segments of the population. SUMMARY

It is difficult to find measures of age at marriage and birth spacing that can be used to make accurate comparisons between different

TIMING AND SPACING OF BIRTHS

groups of couples currently in the reproductive years of life. The difficulty is greatest among younger cohorts because more members of these cohorts will marry in the future and because many of the cohort members have not yet completed their families. Still, with the data available from the Census Bureau and the present study, certain trends are clear. First of all, the median age of the wife at marriage has declined for recent cohorts. It was quite stable at 21 or 22 years from the cohorts of 1880-89 to 1910-19, but then declined to 19.8 for the cohorts of 1935-39. Also, the median age of mother at first birth fell from nearly 24 years to about 21. Still greater reductions occurred in the median age at second and third birth. These changes are due partly to lower age at marriage, and partly to less delay between marriage and birth. The reduction in the median age of mother at childbirth affects population growth in two ways. The long-range effect is to shorten the length of a generation, which raises slightly the intrinsic rate of natural increase. The short-run effect, which is much more dra­ matic, is to cause a temporary upsurge in the birth rate. Age at marriage and the spacing of births are not the same for all groups. Protestants, for example, tend to marry earlier than Catholics, but Catholics tend to have their births sooner after mar­ riage. Also, various measures of socioeconomic status show that the higher the couple's status, the later the age at marriage and the longer the interval from marriage to the first birth. There is some doubt about whether the length of interbirth intervals after the first birth is related to socioeconomic status. The trend toward earlier childbearing has occurred in virtually all socioeconomic groups. It is by no means limited to better-educated or higher-status couples or to couples in any one religious group, size-of-place category, or region of the country.

CHAPTER 9

White—Nonwhite Differences in Fertility Nonwhites in the United States have a substantially higher birth rate than whites. In the period 1951-61, the crude birth rate of the nonwhite population varied between 32 and 34 per 1,000, while the rate for whites was about 10 points lower. The United States nonwhite birth rate resembles rates for countries with moderately high fertility. It is close to Puerto Rico's and slightly below Chile's. Not only do nonwhites have higher birth rates than whites, but during the 1950s the rates for the two population groups moved even farther apart than they had been. For whites, the number of births per 1,000 women 15-44 years old reached a peak in 1947, declined for a few years, and then rose slowly to a second peak, somewhat higher than the first, in 1957. For nonwhites, the compara­ ble fertility rate rose without interruption from 1945 to 1957 (Figure 37). The rise was more rapid than for the white population, so that the percentage difference between white and nonwhite rates increased markedly. Between 1954 and 1961, the difference between white and nonwhite fertility was greater than it ever had been throughout the 1920's or the depression decade of the 1930's. The wider gap between white and nonwhite fertility rates is not consistent with certain social trends in the nonwhite population. It is generally believed that as a population becomes more urban and as its educational level rises, its fertility declines. Among nonwhites, the move from rural areas to cities has proceeded rapidly and educa­ tional attainment has risen. For example, between 1940 and 1960, the proportion of nonwhites living in central cities of 50,000 or more increased from 32 per cent to 51 per cent; whites showed a decline from 35 per cent to 30 per cent in the same period. As far as education is concerned, the proportion of the population with four or more years of high school increased among nonwhites from 9 per cent to 24 per cent between 1940 and 1960; comparable figures for the whites are 29 and 45 per cent. In spite of these changes, the fertility of nonwhites has risen. This is all the more remarkable in view of the crowded and uncom­ fortable housing available to nonwhites when they move into cities. In the present chapter, we cannot answer many of the questions posed by the recent rise in the fertility of the nonwhite population, but we can present the first information available for a national sample of nonwhite married couples on several variables directly re­ lated to their fertility: the number of children they want and expect,

WHITE—NONWHITE DIFFERENCES IN FERTILITY

their ability to have children, the extent to which they use contracep­ tion, and their patterns of family planning. The nonwhite sample consists of 270 women 18-39 years old, married and living with their husbands.1 These women represent 91 per cent of all eligible respondents in the nonwhite households listed for the sample. This response rate is excellent; in fact, it is better than the rate of 88 per cent for comparable white wives. RATE 170 160

150

r"

140

30 •NONWHITE

120 110

/\

100 90 OWHITE 80

70

1920

1930

1940 YEAR

1950

I960

FIGURE 37 Births per 1,000 women 15-44 years old, for whites and nonwhites, United States 1920-1961

In considering fertility data for these couples we shall try to find out to what extent the differences between whites and nonwhites are related to socioeconomic differences between the two groups. For example, we know that the educational level of the nonwhites is 1 Most of the nonwhites are Negroes (256). In addition there were 10 Japanese, one American Indian, one Chinese, one Eskimo, and one native of Guam.

FERTILITY AND FAMILY PLANNING IN THE U.S.

lower than that of whites. Also, we have seen in Chapter 6 that among whites, educational attainment is closely related to the inci­ dence of Excess FertiUty: the less-educated have a higher proportion of couples with unwanted children than the better-educated. There­ fore, we would expect nonwhite couples to have a higher proportion with Excess Fertility than white couples partly because they are less educated. In order to see how such differences in socioeconomic status affect white-nonwhite differences in fertility variables, we shall compare three groups: nonwhites, whites who match the nonwhites in certain important characteristics, and all whites. The method of choosing the sample of matched whites is described in Appendix B, Note 8. In brief, for each of the 270 nonwhite couples, we selected a white couple who resembled them with respect to certain characteristics: duration of marriage, wife's age at marriage, wife's religion, region of residence, size of place of longest residence since marriage, farm background, and husband's occupation. We were able to match nonwhites with whites very closely on four characteris­ tics: duration of marriage, wife's age at marriage, wife's religion, and region of residence. We were less successful in finding white couples to meet the other matching criteria. The chief merit of the matched white group for this analysis is simply that it resembles the nonwhite sample more than the total white sample does. When we use the matched white sample, we know that we are controlling for some of the characteristics that might affect fertility differences between whites and nonwhites. We know, for example, that we are comparing nonwhites with whites who are predominantly Protestant, who are not as well-educated as the total white population, and who are like the nonwhites with respect to current age, age at marriage, duration of marriage, and a few other characteristics. We cannot claim to have controlled for all of the characteristics that might account for fertility differences between the two groups; but we have controlled for some, and this may help us to see more clearly the nature of the factors that cause the fertility of nonwhites and whites to differ so widely. PAST, EXPECTED, AND DESIRED NUMBER OF BIRTHS

Nonwhite wives have had and expect substantially more births than either the white wives chosen to match them on certain charac­ teristics or all white wives. The averages in Table 182 show that the past fertility of nonwhites exceeds that of matched whites by 28 per cent, and of all whites by 19 per cent. Corresponding differen­ tials for total expected births are 24 and 15 per cent, respectively.

WHITE—NONWHITE DIFFERENCES IN FERTILITY

The high fertility of nonwhites is also reflected in high proportions who have had and expect large numbers of births. Twelve per cent of the nonwhite wives had already bome six or more children by the time of interview, compared with only 3 per cent of the matched whites and 4 per cent of all whites (Table 182). In terms of com­ pleted fertility, 18 per cent of the nonwhite wives expected six or more births altogether, as compared with only 6 per cent of the matched whites and 7 per cent of all whites. TABLE 182 Per Cent Distribution by Number of Births by 1960 and Most Likely Expected Total Births, for Nonwhites, Matched Whites, and All Whites

Number of births By 1960: Total 0 1 2 3 4 5 6 or more Average Most likely expected total: Total 0 1 2 3 4 5 6 or more Average

Nonwhite

Matched white

Total white

100 16 19 21 13 11 7 12 2.74

100 14 20 31 19 9 5 3 2.14

100 12 19 28 21 11 5 4 2.31

100 6 11 22 18 14 11 18 3.61

100 5 8 26 33 15 6 6 2.90

100 4 7 25 29 20 8 7 3.15

Nonwhite minus matched white"

—

2 -1 -10 -6 3 3 9 .60

Nonwhite minus total white*

—

3 —

-7 -7 —

2 8 .43

—

—

1 2 -4 -16 -1 5 12 .71

2 4 -3 -11 -6 3 11 .46

a Differences are computed from unrounded percentages; hence, they are not always equal to differences between rounded percentages.

We have some reason for thinking that nonwhite wives will have even more births than they expect. As we observed in Chapter 2, nonwhites have had more births than whites and expect about the same number of additional births, yet they are less able to control their fertility. As a result, it seems likely that nonwhite wives will have more additional births than they expect to have. Perhaps their fertility will reach their maximum expected average of 4.1 births per woman instead of their "most likely" average of 3.6.

FERTILITY AND FAMILY PLANNING IN THE U.S.

It has been pointed out previously that the cultural norm for family size in the United States is two to four children. About threequarters of the white couples expect to have a number in this range. However, the comparable proportion for nonwhites is only 54 per cent. We shall consider later in this chapter the extent to which the difference between the fertility of whites and nonwhites results from the existence of different family-size norms and the extent to which it is due to different degrees of success in controlling fertility. First, however, we shall try to see where the white-nonwhite differen­ tials in fertility are concentrated. One striking feature of the comparisons shown in this section is the fact that the matched white couples have lower past and ex­ pected fertility than all white couples. In other words, the whites who resemble the nonwhites in this study have lower fertility than white couples who have less in common with the nonwhites. This may seem surprising, in view of the fact that the matched whites have lower educational attainment than all whites, and less education generally means more births. Certainly, if we had chosen the matched whites solely on the basis of education, their fertility would have been higher than that of all whites. Their expected number of births, for example, would have averaged 3.3, as compared with 3.1 for all whites. However, the tendency for less-educated whites to have higher fertility than all whites is more than offset by two other criteria used to match the whites and nonwhites. One of these is religion. White Protestants expect fewer births than all whites (2.9 versus 3.1). The other characteristic is region of residence: Southern white Protestants expect slightly fewer births than all white Protestants (2.8 versus 2.9). The net result is an average of 2.9 births expected for matched whites, compared with 3.1 expected for all whites. Because the matched whites have lower fertility than all whites, the white-nonwhite fertility differentials are greater when matched whites are used as the basis of comparison than when all whites are used. At first, this suggests that the socioeconomic characteristics on which the nonwhites and matched whites are alike do not account for any of the differentials in past and expected births. This is not true, however. As we shall see, there are some socioeconomic groups within which whites and nonwhites have very nearly the same fertility. Within others, the two groups differ widely. One of the most important characteristics related to white-non­ white fertility differentials is region of residence. The differences be­ tween whites and nonwhites are much greater in the South than in any other region of the country. Nonwhites in the South already

WHITE—NONWHITE DIFFERENCES IN FERTILITY

had borne about 50 per cent more births than whites in this area (either matched whites or all whites) and expected about 40 per cent more altogether (Table 183). In the rest of the country, the excess of nonwhite over white fertility is small. Outside the South, nonwhite wives had borne 6 per cent more births than the matched white wives and expected 10 per cent more. However, nonwhite wives had actually bome fewer children than all white wives, and expected the same number by the time they complete their families. TABLE 183 Average Number of Births by 1960 and of Most Likely Expected Total Births, for Nonwhites, Matched Whites, and All Whites, by Region of Residence, Southern Farm Residence, and Wife's Education

Characteristic Total Region Northeast North Central South West Southern farm residence None Some, not on farm now" On farm now Wife's education College High school, 4 High school, 1-3 Grade school

Births by 1960 Matched Nonwhite white

All white

Most likely expected total births All Matched white Nonwhite white

2.7

2.1

2.,3

3.6

2.9

3.1

2.3 2.2 3.2

2.3 2.1 2.1

2.9 3.4 4.0

3.1 3.0 2.8

*

*

2..4 2.,4 2. 1 2..4

*

*

3.2 3.3 2.9 3.2

2.4

2.2

2 .3

3.1

3.0

3.2

2.6 4.5

2.2 1.9

2..2 2,.1

3.5 5.9

2.8 2.8

2.9 2.9

1.7 1.9 3.0 3.9

1.9 1.7 2.4 2.5

2 .0 2 .1 2 .6 3..1

2.4 2.9 3.8 4.7

2.8 2.7 3.0 3.1

3.0 3.0 3.3 3.7

* Either husband, wife, or both have lived on a farm sometime.

The 1960 Census confirms the fact that the excess of nonwhite over white fertility is concentrated in the South. It shows that evermarried women 18-39 years old in the South had bome 42 per cent more children than comparable white women. The corresponding nonwhite excess outside the South is only 10 per cent (Table 184).2 2 The data from the present study and those cited from the 1960 Census are not comparable with respect to marital status and they do not show exactly the same differentials. The Census data indicate, for example, that there is still some excess of nonwhite over white fertility outside the South, but data from the present study show no such excess relative to all white wives. Most of this difference between the Census and the present study is probably due to sampling variability. Our sample of nonwhites is small (only 134 of them live outside the

F E R T I L I T Y AND FAMILY PLANNING IN T H E U.S.

Within the South, the greatest fertility difference between nonwhites and whites is found among couples living on farms. Nonwhites living on farms have unusually high fertility. There are only 33 wives in this group in the present study, but their average of 4.5 births by interview is significantly higher than both the 1.9 for matched white wives and the 2.1 for all white wives living on farms in the South (Table 183). Not only had the nonwhite wives on Southern farms bome over twice as many children as comparable white wives, but they expected over twice as many altogether: 5.9 for the nonTABLE 184 Number of Children Ever Born per 1,000 Ever-Married Women 18-39 Years Old, for Nonwhites and Whites, by Region of Residence, by Rural-Urban Residence in South, 1960 Census

Area

Nonwhite

White

Per cent nonwhite over white

Total Northeast North Central South Total Urban Rural nonfarm Rural farm West

2,770 2,244 2,582

2,205 2,092 2,299

26 7 12

3,098 2,743 3,608 4,275 2,479

2,179 2,043 2,354 2,607 2,249

42 34 53 64 10

2,436 2,560

2,216 2,181

10 17

Summary

Outside South Outside rural South

Source: 1960 Census PC(I)-ID1 Table 249.

whites, 2.8 for matched whites, 2.9 for all whites on Southern farms. These and other facts show that Southern farm residence is closely related to white-nonwhite differences in fertility. Here is an example of a Southern Negro farm wife: CASE 9253. This wife was only 24 years old, but had already borne nine children. Her first child was born when she was 14. All of her children survived and are now living with their mother and her husband, who is a tenant farmer in Louisiana. The husband had virtually no education (he completed second grade), and the South), and it is impossible to get reliable indications of differences between groups with so few cases unless the differences under study are very large. There­ fore, we are not justified in saying that there is probably no difference between nonwhite and white fertility outside the South solely on the basis of data from the present study. The Census shows that there is a small difference, and the present study gives no reason to question this finding.

WHITE—NONWHITE DIFFERENCES IN FERTILITY

wife had a seventh-grade education. The total family income in 1959 was less than $1,000. They did not get any relief or welfare payments. The wife said that she thought it was all right for couples to limit the number of children they had, but that she had never tried this. She said that if she could live her life over, she would have only two children. She definitely doesn't want any more children than she has, " 'cause I'm not able to support them." Data from the 1960 Census confirm the fact that white-nonwhite fertility differentials are greatest in the rural South. Again, the Census data are not quite comparable to those used in the present study COUPLE'S SOUTHERN FARM RESIDENCE TOTAL

NONE

ON FARM NOW

ON (SOUTHERN) FARM NOW

AVERAGE MOST LIKELY EXPECTED TOTAL NUMBER OF BIRTHS I NONWHITE

W//ft MATCHED WHITE

Υ//Λ ALL WHITE

FIGURE 38 Average most likely expected total number of births for nonwhites, matched whites, and all whites, by couple's Southern farm residence

and the differentials are not the same, but they are in the same direction. Table 184 shows that for ever-married women 18-39 years old living in the South, the average number of children ever bom is 64 per cent greater for nonwhites than for whites in rural farm areas, 53 per cent greater in rural nonfarm areas, and 34 per cent greater in urban areas. Such differentials suggest that the traditional patterns of marriage and family life that developed in the rural South during and after the period of slavery still influence nonwhite fertility perceptibly. In­ formal and often temporary unions between men and women be-

FERTILITY AND FAMILY PLANNING IN THE U.S.

ginning in adolescence, relatively ineffective sanctions against illegiti­ macy among the economically deprived,8 a genuine affection for chil­ dren, and the existence of customary arrangements to provide for their care in case the mother cannot do so all contribute to high birth rates.4 The high-fertility patterns of Southern rural Negroes seem to be greatly modified as they are carried to towns and cities. Nonwhite couples with previous Southern farm residence (i.e., either husband, wife, or both had lived on a Southern farm at some time) have had an average of 2.6 births and expect 3.5 altogether (Table 183). These figures are well below the corresponding ones for couples still living on Southern farms; but they are above those for nonwhites with no Southern farm background, which suggests that previous Southern farm residence has some influence on the fertility of nonwhite couples. When we come to nonwhite couples without any previous South­ ern farm residence, we find average past and expected numbers of births that do not differ significantly from those of either the matched white or the total white sample. In other words, by the time nonwhite couples are one generation or more removed from the rural South, their fertility is very much like that of the white population. The average number of births by interview for couples with no Southern farm background is 2.4 for nonwhites, 2.2 for matched whites, and 2.3 for all whites. Corresponding averages for births expected are 3.1, 3.0, and 3.2, respectively. The small differences between these averages are not statistically significant. Again, it is important to point out that we cannot demonstrate with our sample that there are no differences between whites and nonwhites without Southern farm background, but we can say that the differences that do exist are probably not large. Early in the nineteenth century, the great majority of Negroes lived in the rural South. The movement to towns and cities has been going on for many years, and by 1960 few nonwhite couples of childbearing age remained on Southern farms. The proportion is 12 per cent, according to the present study, but this may be an overstatement. The 1960 Census shows only 4.6 per cent of the evermarried nonwhite women 18 to 39 years old still living in rural farm areas of the South. In any case, we can be sure that the propor­ tion is small. * The proportion of wives in the present study reporting pregnancies that began before marriage is 38 per cent for nonwhites and 9 per cent for whites. jFor descriptions of the Negro family in the rural South see Frazier 1949, pp. 318-321 and Powdermaker 1939, pp. 143-222. See page xxxi for the meaning of any abbreviated reference.

WHITE—N0NWH1TE DIFFERENCES IN FERTILITY

Although few nonwhites remain on Southern farms, a large num­ ber have Southern rural background. In the present study, nearly half (45 per cent) of the nonwhite wives reported that they or their husbands had lived on a Southern farm at some time in their lives. Hence, the influence of Southern rural ways of life is still widespread among the nonwhite population. Only 43 per cent of the nonwhite wives reported no Southern farm background for themselves or their husbands. The latter propor­ tion will have risen greatly in another generation, and the influence of Southern rural patterns of fertility behavior will inevitably dimin­ ish; but at present it is still strong. The patterns of fertility behavior practiced by Negroes in the rural South are evidently unique to the nonwhite population. We noted previously that white wives living on Southern farms had and expected fewer births than similarly situated nonwhite wives. This is true even of poor whites. For example, among the 27 white couples living on Southern farms and with husband's income below $3,000 a year, the wives had borne an average of 2.3 children and expected 3.0. These are far below the averages of 4.5 and 5.9 for nonwhite Southern farm wives. Another socioeconomic characteristic closely related to white-nonwhite fertility differences is educational attainment. The excess of nonwhites over whites in past and total expected births is greatest among the least-educated wives. Among wives with only a gradeschool education, nonwhites have had and expect about 50 per cent more births than matched whites, and about 25 per cent more than all whites (Table 183 and Figure 39). At the upper end of the educational continuum, however, nonwhite wives with a college edu­ cation have had and expect fewer births than white wives (either matched white or all white) in the same educational category. The 1960 Census presents essentially the same picture. Among ever-married women 15-39 years old,5 the Census shows that nonwhite women with only a grade-school education, had borne 24 per cent more children than white women with as little education. Census data also agree approximately with those from the present study in showing that among the college-educated, nonwhite women have had 9 per cent fewer births than white women (Table 185). What we see in these differentials is the gradual adoption of attitudes and practices favoring small families as Negroes move from farms to cities and as their educational attainment rises. The culmina5The relevant Census data are not available for the age and marital status groups used in the present study.

FERTILITY AND FAMILY PLANNING IN THE U.S. WIFE'S EDUCATION TOTAL

COLLEGE

mmmrm*

HIGH SCHOOL,

4

HIGH SCHOOL,

1-3

GRA DE SCHOOL

Υ7Ζ77ΖΖΖΖΖΖΖΖΖΔ ι 2

Ol

1 3

1 4

1 5

AVERAGE MOST LIKELY EXPECTED TOTAL NUMBER OF BIRTHS ϋϋ1|NONWHITE

Υ////Ά MATCHED WHITE V / Λ

ALL WHITE

FIGURE 39 Average most likely expected total number of births for nonwhites, matched whites, and all whites, by wife's education

tion of this process is a well-educated Negro upper class with relatively few children. Frazier notes, for example: "Among the upper class in the northern city there are many childless couples and relatively few children in families with children. Men and women who have struggled to achieve a high position in the Negro community are not inclined to have the standards which they TABLE 185 Number of Children Ever Born per 1,000 Ever-Married Women 15-39 Years Old, for Nonwhites and Whites, by Wife's Education, 1960 Census

Wife's education

Nonwhite

White

Per cent nonwhite over white

Total College High school, 4 High school, 1-3 Grade school

2,750 1,741 2,176 2,798 3,418

2,186 1,906 2,039 2,283 2,746

26 -9 7 23 24

Source: 1960 Census, PC(2)-3A, Women by Number of Children Ever Born, Table 25.

WHITE—.NONWHITE DIFFERENCES IN FERTILITY

attempt to maintain lowered by the burden of children. Moreover, since these men have often experienced many privations to achieve an education, they want to spare their children similar hardships."6 The following example shows the concern of well-educated Ne­ groes for giving their children a good education: CASE 9027. This 26-year old Negro wife and her husband both had master's degrees. The husband taught in a public school in New York City, where he made nearly $7,000 in 1959. This couple had TABLE 186 Average Number of Births by 1960, for Nonwhites, Matched Whites, and All Whites, by Wife's Education, by Couple's Southern Farm Residence Southern farm residence

Wife's education

Total College or High school, 4 High school, 1-3 Grade school Total College or High school, 4 High school, 1-3 Grade school Total College or High school, 4 High school, 1-3 Grade school

Total

None

Nonwhite 2.7 2.4 1.8 1.8 2.6 3.0 3.5 3.9 Matched white 2.2 2.1 1.8 1.9 2.4 2.4 * 2.5 All white 2.3 2.3 2.1 2.1 2.6 2.6 3.4 3.1

Some, not on farm now"

2 .6 1 .8 3,.1 3 .4 2 .2 1 .8 2..3 2,,4 2..2 1 9 2..5 2. 8

On farm now

4.5 *

U 7 J 1.9 •

I22 55 J 2.1 1.6 I2 4 J

» Either husband, wife, or both have lived on a farm in the South.

one child, and definitely did not want any more. The wife said, "Since we want to give our child a good education and other ad­ vantages, we can afford only one." To what extent are the white-nonwhite fertility differentials by education linked to those by Southern farm residence? For example, are the white-nonwhite differences particularly wide among the lesseducated largely because a high proportion of the less-educated nonwhites live or have lived on Southern farms? Such questions are explored in Tables 186 and 187, but it should be made clear that * Frazier 1949, p. 331.

FERTILITY AND FAMILY PLANNING IN THE U.S.

the results shown here are based on relatively few cases per cell. Hence, the findings are more suggestive than definitive. First of all, these tables show that among nonwhites the fertility differentials by education exist regardless of whether or not the couple has had any Southern farm background. (There are too few couples who live on Southern farms to see whether the educational differences also exist in this setting.) This means that although Southern farm residence increases the fertility of the less-educated, there are still educational differences in fertility where this influence is absent. TABLE 187 Average Most Likely Expected Total Number of Births, for Nonwhites, Matched Whites, and All Whites, by Wife's Education, by Couple's Southern Farm Residence Southern farm residence

Wife's education

Total College or High school, 4 High school, 1-3 Grade school Total College or High school, 4 High school, 1-3 Grade school Total College or High school, 4 High school, 1-3 Grade school a

Total

None

Nonwhite 3.6 3.1 2.7 2.6 3.8 3.2 4.7 4.1 Matched white 2.9 3.0 2.7 2.9 3.0 3.0 * 3.1 All white 3.1 3.2 3.0 3.1 3.3 3.3 3.7 4.0

Some, not on farm now·

3.5 2.8 4.0 4.0

On farm now

5.9 *

UJ

1

2.8 2.5 2.9 3.0

2.8 1¾ ο i3·2

2.9 2.7 3.2 3.3

2.9 2.6 Il O i 3 2

*

Either husband, wife, or both have lived on a farm in the South.

Secondly, within the nonwhite population we cannot be sure whether or not previous Southern farm residence affects fertility inde­ pendently of educational attainment. For example, among wives with a grade-school education, those with no Southern farm background expect 4.1 births, on the average, and those who had previously lived on Southern farms expect nearly the same number—4.0. The differentials among better-educated wives are in the opposite direction, but they are not large enough to confirm the hypothesis that previous Southern farm residence has an effect on fertility that is independent of educational attainment. The only thing we are sure of is that

WHITE—NONWHITE DIFFERENCES IN FERTILITY

current residence on a Southern farm means much higher fertility for the less-educated. When we consider the white-nonwhite fertility differentials, we find that the largest differentials exist among couples who have rela­ tively low educational attainment (grade-school or incomplete high school) and who have lived or are living on Southern farms. Among couples with no Southern farm background and among well-educated couples, whites and nonwhites do not differ greatly in their fertility. We can present these facts more clearly by dividing the nonwhite and white samples into three main groups: I. Those who have completed high school or have had some college education and those with less education who have never lived on a Southern farm. II. Those who have a grade school or incomplete high school education and who have previously lived on a Southern farm. III. Those who have a grade school or incomplete high school education and who were living on a Southern farm when interviewed. Within Group I, nonwhites have very nearly the same past and expected fertility as whites. The nonwhite wives in this group expect an average of 3.0 births, as compared with 2.9 for matched whites, and 3.1 for all whites. The number of births by 1960 are 2.2, 2.0, and 2.3, respectively. Inasmuch as this group contains 63 per cent of the nonwhite sample, we can say that a majority of nonwhite couples have and expect about the same number of children as white couples in similarly defined socioeconomic groups. Again, it should be emphasized that we cannot demonstrate conclusively the absence of fertility differences between the populations sampled, but we can say that they are small. Within Group II, nonwhites have moderately higher past and expected fertility than whites. They expect 4.0 births, as compared with 3.0 for the matched whites and 3.2 for all whites. Births by 1960 are 3.2, 2.3, and 2.6, respectively. This group contains 26 per cent of the nonwhite sample. Finally, within Group III, nonwhites have much higher fertility than whites. They expect 6.1 births, compared to 3.2 for matched whites and all whites. Births by 1960 are 4.7, 2.5, and 2.4, respec­ tively. This group contains 11 per cent of the nonwhite sample. The most important facts brought out here are (a) that the fertility of a majority of nonwhites does not differ widely from that of whites in similar socioeconomic groups, and (b) that the differences that exist for the remainder of the nonwhites are closely associated with characteristics that will have less influence in the future than they do now—previous and current Southern farm residence and

FERTILITY AND FAMILY PLANNING IN THE U.S.

low educational attainment. These findings give us reason to believe that fertility differences between whites and nonwhites will become narrower.7 Although it seems likely that the long-term trend in the completed fertility of nonwhite couples will be downward, the present sample is too small to give reliable evidence about cohort trends in birth expectations. The averages in Table 188 show a rise in total expected TABLE 188 Average Number of Births by 1960 and of Most Likely Expected Total Births, for Nonwhites, Matched Whites, and All Whites, by Cohort, Use of Contraception, and Pregnancy Planning Status Most likely expected total births

Births by 1960

Characteristic

Nonwhite

Matched white

All white

Nonwhite

Matched white

All white

Total Cohort 1936-42 1931-35 1926-30 1921-25 Use of contraception User ODFCa Nonuser Pregnancy planning status1· Completely Planned Partly Planned Partly Unplanned Excess Fertility

2..7

2.1

2 .3

3..6

2.9

3.1

1..9 2..6 3,.2 3..1

1.5 2.0 2.4 2.6

1 .3 2 .3 2 .7 2 .7

3..5 3 .6 3..9 3 .5

2.8 3.1 2.9 2.8

3.0 3.4 3.2 3.0

3..0 2..3 2,.4

2.3

3..7 3..2 3 .6

3.0

1.4

2 .4 1 .8 1 .8

2.3

3.2 2.6 3.0

1.,0° 1 .6 2..0 4.8

1.1 1.7 2.2 3.7

1 .4 1 .8 2 ,5 3..7

2..1° 2 .5 2,.9 5.7

2.2 2.4 3.1 3.9

2.4 2.8 3.4 4.1

*

*

" The wife regularly douches within one-half hour after intercourse for purposes of cleanliness only. b See Chapter 6 for definitions. 0 This average is based on only 19 cases. Usually we do not show averages based on so few cases, but there is particular interest in this group.

births between the cohorts of 1921-25 and 1926-30 (from 3.5 to 3.9) and then a decline to the cohorts of 1936-42 (to 3.5 again). The differences between these averages are not statistically significant, however. Furthermore, if nonwhite expectations are subject to the same kinds of influences as similar data for whites, the average of 3.5 for the 1936-42 cohorts will probably rise. (For evidence that expectations tend to rise with age, see Chapters 1 and 10.) Is there any evidence that nonwhite couples have and expect 'Fertility differentials from the 1950 Census have also been cited as evidence that nonwhite fertility will probably decline in the future. See Lee 1959.

W H I T E — N O N W H I T E DIFFERENCES IN F E R T I L I T Y

more births than whites because relatively fewer nonwhites use contra­ ception? The data in Table 188 show that the excess of nonwhite over white expectations exists regardless of whether the couple has tried to limit fertility. Nonwhite couples who have used contraception have had nearly 30 per cent more births than Users among the matched white or all white samples, and expect nearly 20 per cent more altogether. Thus, the fact that relatively fewer nonwhites have tried to use contraception does not account for all of the whitenonwhite fertility differentials. However, when we look at the white-nonwhite differentials by pregnancy planning status, we find that the excess of nonwhite over white fertility is concentrated among couples with Excess Fertility (Table 188). In fact, nonwhite couples with no unwanted pregnancies had and expected fewer births than the total white sample, and about the same number as the matched white sample. Inasmuch as 69 per cent of the nonwhite couples had not yet had more preg­ nancies than they wanted, this finding again emphasizes the fact that a majority of nonwhite couples have and expect about the same number of births as similar white couples. Actually, their past and expected fertility may be somewhat lower than that of similar whites, but we cannot be sure about this because our sample is small. The minority of 31 per cent of the nonwhite couples who have Excess Fertility have much higher fertility than similar white couples. They had an average of 4.8 births, compared with 3.7 for the matched whites and all whites with Excess Fertility; they expected 5.7, com­ pared with 3.9 for the matched whites and 4.1 for all whites. These differences are statistically significant. In other words, nonwhite couples not only have a higher preva­ lence of Excess Fertility than whites (31 per cent vs. 17 per cent), but their degree of Excess Fertility is more severe. This resembles a similar finding for educational differences in fertility among the white population (see Chapter 6). It is important because it strongly suggests that the higher fertility of nonwhites is neither intended nor desired. The high-fertility minority of the nonwhite population does not want unusually many children; the couples in this group have been unable, for a variety of reasons, to limit their fertility to the extent desired. The average number of children that nonwhite wives want (2.9) is significantly lower than the number wanted by all white wives (3.3), and about the same or slightly lower than the number wanted by matched white wives (3.1). (See Table 189.) Even with the limi­ tations of small sample size and our imperfect measurement of the number of children wanted, these findings allow us to say that non-

FERTILITY AND FAMILY PLANNING IN THE U.S.

white wives do not want more children than white wives. They want about the same number or fewer. In general, the average number of children wanted by nonwhite wives varies with education and Southern farm residence in the same direction that births by 1960 and total expected births vary, but the socioeconomic differentials in children wanted are smaller. For example, we have seen that nonwhite wives living on Southern farms expect 5.9 births, compared with only 3.1 for those with no Southern farm background. In Table 189 we see that wives living on Southern farms want an average of 3.8 children, compared with 2.6 for wives TABLE 189 Average Number of Children Wife Wants, for Nonwhites, Matched Whites, and All Whites, by Region of Residence, Southern Farm Residence, and Wife's Education

Characteristic

Nonwhite

Matched white

AU white

Total Region of residence Northeast North Central South West Southern farm residence None Some, not on farm now® On farm now Wife's education College High school, 4 High school, 1-3 Grade school

2.9

3.1

3.3

2.5 2.8 3.0

3.5 3.3 2.8

*

*

3.5 3.4 3.0 3.3

2.6 2.9 3.8

3.3 2.8 3.1

3.4 3.1 3.1

2.4 2.7 2.7 3.5

3.1 3.0 3.2 3.0

3.3 3.2 3.3 3.5

• Either husband, wife, or both have lived on a farm in the South.

with no Southern farm background. In the case of expectations, the farm-nonfarm differential is 90 per cent, but in the case of number of children wanted it is slightly under 50 per cent. In other words, it is true that nonwhite wives living on Southern farms want more children than nonwhite wives with no Southern farm background, but this does not account for their higher fertility. The Southern rural farm wives expect about 55 per cent more births than children wanted, but those with no Southern farm background expect only about 19 per cent more. It is only among the well-educated nonwhites that we find ap­ proximate equality between the number of births expected and the number of children wanted. In other groups, there is a consistent tendency for nonwhite wives to expect more births than they want.

WHITE—NONWHITE DIFFERENCES IN FERTILITY

Later in this chapter we shall examine the contraceptive and family planning practices of nonwhite couples in an effort to under­ stand why they tend to have more births than they want. In the next section, however, we shall see whether their ability to have children is the same as that of white couples. FECUNDITY

There is reason to believe that in the recent past, impairments of the reproductive system were more common among nonwhites than whites. One item of evidence bearing on this matter is the high proportion of older women who are childless. Among ever-married women 50-54 years old in 1960, for example, the proportion who never had any children was 28 per cent for nonwhites and 20 per cent for whites. It seems likely that most of the childlessness among nonwhites was involuntary (due to physiological impairments), while most of that among whites was due to the voluntary prevention of pregnancy.8 Evidently, widespread childlessness in the nonwhite population was due partly to a relatively high prevalence of venereal disease.® Voluntary childlessness may also have played a role among bettereducated nonwhite couples, but venereal infection appears to have been the more important factor for the rest of the nonwhite population. Since World War II, however, the prevalence of venereal disease has diminished rapidly among nonwhites. In fact, this is thought to be one of the important causes of the increase in their birth rate.10 Also, the 1960 Census shows that the percentage of ever-married women who are childless has declined to 14 per cent for nonwhites 25-29 years old. This compares with 28 per cent for nonwhite women 50-54 years old (cited above) and with 12 per cent for white women 25-29 years old. In other words, if we use the percentage childless as a rough indicator of the presence of fecundity impairments in a population, it appears that young nonwhites are more fecund than nonwhites were a generation earlier, and are now about as fecund as young whites. Data from the present study show no substantial differences in the over-all prevalence of fecundity impairments between nonwhites and whites (either matched whites or all whites). The proportions Subfecund are 33 per cent for nonwhites and 31 per cent for both 8For evidence on the extent of involuntary childlessness among nonwhites, see Kiser 1958, pp. 190-196. For evidence on this matter for white couples in the present study, see Chapter 4. " See ICiser 1958, pp. 190-196 and Lee 1959, p. 228. "This cause is cited in both Kiser 1958 and Lee 1959.

FERTILITY AND FAMILY PLANNING IN THE U.S.

matched whites and all whites (Table 190). The difference is not statistically significant. Again, it must be emphasized that we cannot demonstrate the absence of any difference with a small sample; we can only say that the difference that does exist is probably not large. There seem to be some differences between nonwhites and whites in the kinds of fecundity impairments they have. Fewer nonwhite couples are classified as Definitely Sterile (6 per cent, as compared with 11 per cent for matched whites and all whites). The difference between nonwhites and all whites is barely large enough to be sta­ tistically significant at the confidence level chosen for this study (.95). TABLE 190 Per Cent Distribution by Fecundity, for Nonwhites, Matched Whites, and All Whites

Fecundity

Nonwhite

Matched white

All white

Number of couples Per cent: Total Fecund Subfecund Total Definitely Sterile Total Contraceptive operation Remedial operation Other Probably Sterile Possibly Sterile Possibly Fecund

270

270

2,414

100 67

100 69

100 69

33

31

31

6 3 4

11 6 4

11 6 4 1 4 8 8

—

7 13 6

—

4 7 9

The lower proportion Definitely Sterile among nonwhite couples is due to a lower proportion who have had contraceptive operations (3 per cent for nonwhites and 6 per cent for matched whites and all whites). Apparently nonwhite couples are less likely than whites to try to control fertility by means of operations preventing pregnancy. This may be because medical attention is less available to nonwhite couples. However, nonwhites do not seem to differ from whites in the proportion of wives reporting remedial operations (those intended to correct a pathological condition rather than to prevent pregnancy). The proportion of couples who are unable to have additional children because of remedial operations is 4 per cent for all groups: nonwhites, matched whites, and all whites (Table 190). Another way in which nonwhites differ from whites with respect to fecundity is in the proportion who are classified as Subfecund

WHITE—NONWH1TE DIFFERENCES IN FERTILITY

because they have had long periods of exposure to the risk of preg­ nancy and have not conceived. The proportion of such wives among nonwhites is 16 per cent, compared with 13 per cent for matched whites, and 11 per cent for all whites (Table 191). The difference between nonwhites and all whites is large enough to be statistically significant. We would expect such a difference because nonwhites make less use of contraception than whites, and consequently they have more opportunity to discover fecundity impairments. This finding resembles a differential discussed in Chapter 4: the fact that the proportion with long periods of noncontraceptive ex­ posure without conception was higher for the less-educated white couples—also because they made less use of contraception. TABLE 191 Per Cent Distribution by Evidence of Subfecundity, for Nonwhites, Matched Whites, and All Whites Evidence of subfecundity Number of couples Per cent: Total Fecund Subfecund Total Operation Pathological condition Noncontraceptive exposure

Nonwhite

Matched white

All white

270

270

2,414

100 67

100 69

100 69

33 6 11 16

31 11 8 13

31 10 10 11

The main effect of this difference between nonwhite and white couples is seen in a higher proportion of nonwhites classified as Pos­ sibly Sterile (13 per cent, compared with 7 per cent for matched whites and 8 per cent for all whites). In summary, nonwhites may have a lower proportion classified as Definitely Sterile and a higher proportion classified as Possibly Sterile than whites. These differences tend to balance each other, with the result that the over-all percentage with fecundity impairments is about the same for both groups. THE USE OF CONTRACEPTION

The use of contraception is less common among nonwhites than whites. By the time of interview, 59 per cent of the nonwhite couples had used contraception, as compared with 80 per cent of the matched whites and 81 per cent of all whites (Table 192). The percentages who had used or expected to use contraception were somewhat closer:

FERTILITY AND FAMILY PLANNING IN THE U.S.

76 per cent for nonwhites, 86 per cent for matched whites, and 87 per cent for all whites. The differences are statistically significant. As was explained in Chapter 5, differences between white groups in the extent to which contraception is used can be accounted for mainly by differences in the timing of first use of contraception. Where delay in first use is common (for example, among the lesseducated), a higher proportion of couples find that they have difficulty conceiving before they feel impelled to begin using contraception. TABLE 192 Percentage of Couples Who Have Used or Expect to Use Contraception and Percentage Who Do Not Expect to Use Contraception, by Fecundity, for Nonwhites, Matched Whites, and All Whites, by Southern Farm Residence

Southern farm residence

No. of couples

Total None Some, not on farm now" On farm now

270 116

Total None Some, not on farm now* On farm now

270 150

Total None Some, not on farm now* On farm now

2,414 1,784

a

121 33

90 30

540 90

Have used

Have used or expect to use

Nonwhite 59 76 65 78 60 80 36 52 Matched white 80 86 79 87

Do not expect to use

Total

Subfecund Fecund

24 22

15 17

9 5

20 48

10 27

10 21

14 13

9 8

4 5

80 83 All white 81 81

87 83

13 17

11 10

2 7

87 88

13 12

10 9

3 3

80 86

86 87

14 13

11 10

2 3

Either husband, wife, or both have lived on a farm in the South.

Consequently, many of them will never use it. In contrast, if a high proportion of couples begin using contraception soon after marriage and before their first pregnancy (as occurs among the better-educated), fewer have the opportunity of discovering whether or not contraception is necessary. They have already begun using it, whether they need to or not, and are classified as Users. This relationship helps to account for the lower proportion of nonwhite couples who have used or will use contraception. We can see this in the white-nonwhite differences in the proportion classified as Subfecund Nonusers who expect not to use contraception. (These

WHITE—N0NWH1TE DIFFERENCES IN FERTILITY

are the couples who discover that contraception might not be needed to prevent additional pregnancies.) Couples in this group make up 15 per cent of the nonwhite sample, compared with 9 per cent of the matched white sample and 10 per cent of the total white sample. The difference of 5 or 6 percentages points between whites and nonwhites in these proportions accounts, however, for only about half of the 11 to 12-point spread between the two groups in the proportion who have used or will use contraception. The rest of the difference is due to the higher proportion of nonwhite wives who say that they will not use contraception even though they seem to be quite able to bear more children—in some cases, many more. Such wives constitute 9 per cent of the nonwhite sample, compared with 4 per cent of the matched white and 3 per cent of the total white sample. Although they are a small minority of the nonwhite population, they are important because they contri­ bute disproportionately to the fertility of nonwhites, and because their attitudes toward the control of conception represent an extreme point of view that may still influence other nonwhite couples. Here are some examples: CASE 9254. This 30-year-old wife has five children. She and her family live in a shack on a farm where her husband works as a hired hand. When asked what she thought of couples' using methods to prevent pregnancy, she said, "I just think it's wrong to do anything, but it's hard when you just keep having them. I don't know why I think this." CASE 9298. This 34-year-old wife, also a Southern farm resident, has had 13 pregnancies and nine children. She said that she was "for" family limitation, but had never used anything herself and didn't plan to. The interviewer noted, "For sure she does not want any more children, but she is doing nothing to prevent them. I would say she would be good for six more." Some of these wives seem to rationalize their failure to try to prevent conception by saying that they do not think they will have many children. For example: CASE 9252. This 26-year-old wife lives on a farm in the South. She has had three children during the six years she has been married. She says that she is "for" family limitation, but won't use it herself because she doesn't think she will have children frequently: "I think nature and the Lord helps us out with things like that. Some time I'll get sick and can't have more children."

FERTILITY AND FAMILY PLANNING IN THE U.S.

It is clear that these women and others like them do not want unusually large families, but their culture has accustomed them to frequent childbearing and they have no ready means of learning about methods of family limitation. Also, it is doubtful whether they are strongly motivated to try to limit family size. They seem to regard the burden of numerous children with the same fatalism (and, per­ haps, despair) with which they "accept" their extreme poverty. The use of contraception among nonwhites is closely related to Southern farm residence. Only 36 per cent of the nonwhite wives living on Southern farms reported that they or their husbands had COUPLE'S SOUTHERN FARM RESIDENCE TOTAL

NONE

SOME. NOT

ON (SOUTHERN Y/ FARM NOW 7

0

I

20

I

40

I

60

ι

70

IOO

PERCENT USER

188888¾ NONWHITE

{2¾¾¾ MATCHED WHITE

V///1 ALL WHITE

FIGURE 40 Per cent Users for nonwhites, matched whites, and all whites, by couple's Southern farm residence

tried to limit family size (Table 192 and Figure 40). This is the lowest proportion of Users found for any socioeconomic group in this study. Another 15 per cent of the Southern farm residents ex­ pected to use contraception, so that slightly over half of them (52 per cent, after rounding) had used contraception or expected to do so. Of the 48 per cent who expected never to use contraception, 27 per cent were Subfecund. Most of them would not have to use contraception to prevent additional pregnancies. The remaining 21 per cent were Fecund and said they thought they would never use contraception.

WHITE—NONWHITE DIFFERENCES IN FERTILITY

Nonwhite couples who have lived on a Southern farm, but no longer do so, are much more likely to use contraception. Apparently the move from rural to urban areas does much to change attitudes toward family size and the use of methods to prevent pregnancy. Another possibility is that migration is a selective process—i.e., that the men and women who leave the farms are already more ambitious and more likely to want to plan their families than those who stay. Also, we must remember that some of the husbands and wives with Southern rural backgrounds left these areas when they were children and have grown up among people who knew about and used con­ traception. Whatever the explanation, we find that 60 per cent of the nonwhite couples with previous Southern farm background have used contraception (Table 192). The proportion who have used or expect to use contraception is 80 per cent, which is not very far from the 86-87 per cent for whites. Couples without any previous Southern farm residence differ little from those who come from Southern rural backgrounds. Slightly more of them have used contraception (65 per cent instead of 60 per cent), but slightly fewer have used or expect to use (78 per cent instead of 80 per cent) contraception. None of these differences is statistically significant. Nonwhites are closest to whites in the proportion using contracep­ tion in the Northeast and among well-educated couples. Table 193 shows that in the Northeast, 76 per cent of the nonwhite couples have used contraception and 95 per cent have used or will use. This compares with 77-78 per cent of whites who have used and 85 per cent who have used or will use (the figures for matched whites are nearly the same as those for all whites). Thus, the propor­ tion of nonwhites who have begun use is virtually the same as for whites, and the proportion of past or future Users is actually higher for the nonwhite population. Nonwhite wives with a college education also report a high preva­ lence of use. Altogether, 95 per cent of them have used or expect to use contraception, as compared with 97 per cent for comparable matched whites and 93 per cent for all college-educated whites (Table 194). Again, it appears that better-educated nonwhite couples readily adopt moderate family-size goals and the methods of attaining them. This suggests that continued improvements in the education of nonwhites will bring an increase in the proportion using contraception. This conjecture gains some support from Table 195, which shows that the percentage of nonwhite couples who have used or will use

FERTILITY AND FAMILY PLANNING IN THE U.S.

contraception is considerably greater for recent than earlier cohorts. In the cohorts of 1936^42, 83 per cent of the nonwhite couples are past or future Users; the comparable proportion in the cohorts of 1921-25 is only 58 per cent. Also, the proportion of couples who have actually begun using contraception is greater for the cohorts of 1926-30 and 1931-35 (66 and 65 per cent, respectively) than for the older cohorts of 1921-25 (53 per cent). This too suggests an upward trend. (The low proportion of 52 per cent in the youngest TABLE 193 Percentage of Couples Who Have Used or Expect to Use Contraception and Percentage Who Do Not Expect to Use Contraception, by Fecundity, for Nonwhites, Matched Whites, and All Whites, by Region of Residence

Region of residence Total Northeast North Central South West Total Northeast North Central South West Total Northesist North Central South West

No. of couples 270 41 74 136 19 270 41 74 136 19 2,414 581 715 744 374

Have used

Have used or expect to use

Nonwhite 59 76 59 51 *

Subfecund

Fecund

76 95 76 68

24 5 24 32

15 2 16 20

9 2 8 12

*

*

*

*

14 15 12 15

9 5 9 11

4 10 3 4

*

*

*

13 15 12 12 11

10 10 10 10 9

Matched white 80 86 85 78 78 88 79 85 *

Do not expect to use

Total

*

All white 87 81 77 85 82 88 83 88 80 89

3 6 2 2 2

cohorts is undoubtly due to the presence of many recently-married couples who have not yet begun contraception. It is important to note, however, that this proportion is already just about as high as the 53 per cent of Users in the oldest cohort.) The findings presented in this section show that nonwhite couples are far from uniform in their practices affecting fertility. They range all the way from the near absence of fertility control of the Southern farm couple to the highly deliberate control of the welleducated urban couple. Group differences are much wider than they are among white couples. The proportion of nonwhite couples who use contraception seems likely to rise as the education level of the population increases. Data

WHITE—N0NWH1TE DIFFERENCES IN FERTILITY

TABLE 194 Percentage of Couples Who Have Used or Expect to Use Contraception and Percentage Who Do Not Expect to Use Contraception, by Fecundity, for Nonwhites, Matched Whites, and All Whites, by Wife's Education No. of couples

Wife's education

Have used

Total College High school, 4 High school, 1-3 Grade school

270 37 73 86 74

59 86 67 56 42

Total College High school, 4 High school, 1-3 Grade school

270 37 82 86 65

80 97 82 77 71

Total College High school, 4 High school, 1-3 Grade school

2,414 427 1,153 579 255

81 88 83 78 66

Have used or expect to use

Do not expect to use

Total

Nonwhite 76 24 95 5 19 81 79 21 57 43 Matched white 86 14 97 3 90 10 16 84 78 22 All white 87 13 93 7 90 10 15 85 72 28

Subfecund Fecund 9

15 5 15 14 22

—

4 7 22

9 3 9 9 14

4 —

1 7 8 3 2 2 4 7

10 4 8 12 21

TABLE 195 Percentage of Couples Who Have Used or Expect to Use Contraception and Percentage Who Do Not Expect to Use Contraception, by Fecundity, for Nonwhites, Matched Whites, and All Whites, by Cohort

Cohort

No. of couples

Have used

Total 1936-42 1931-35 1926-30 1921-25

270 60 65 79 66

59 52 65 66 53

Total 1936-42 1931-35 1926-30 1921-25

270 56 74 74 66

80 70 86 81 79

Total 1936-42 1931-35 1926-30 1921-25

2,414 513 600 624 677

81 78 84 83 77

Have used or expect to use

Do not expect to use

Total

Nonwhite 76 24 17 83 15 85 77 23 58 42 Matched white 14 86 87 13 91 9 84 16 17 83 All white 87 13 92 8 91 9 12 88 80 20

Subfecund

Fecund

15 3 9 16 30

9 13 6 6 12

9 7 3 14 14

4 5 7 3 3

10 4 5 10 18

3 4 4 2 2

FERTILITY AND FAMILY PLANNING IN THE U.S.

on past and expected use for the different cohorts also suggest a strong tendency toward more use of contraception. METHODS OF CONTRACEPTION

After nonwhite couples begin to use contraception, they tend to use less effective methods than either matched whites or all whites. This is shown by the comparisons in Table 196. One of the most important differences shown here is the greater reliance of nonwhite couples on douche, a method that is relatively TABLE 196 Percentage of Users Who Have Used Specified Method of Contraception, for Nonwhites, Matched Whites, and All Whites

Nonwhite minus* Total white

Method

Nonwhite

Matched white

No. of Users" Percentage reporting Condom Diaphragm Rhythm Douche Withdrawal Jelly alone Suppositories Abstinence Other Totalb

160

215

1,948

58 30 18 50 21 19 16 2 1 215

53 40 30 28 19 12 6 1

50 38 35 24 17 11 6 4 1 186

—

189

Matched white

Total white

4 -10 -12 22 1 7 11

8 -8 -18 26 4 8 10 -2 1 28

—

1 25

* Includes a small number who did not report kind of method used. b The total exceeds 100 because many couples used two or more methods. 0 Differences are computed from unrounded percentages; hence, they are not always equal to differences between rounded percentages.

low in effectiveness. Half of the nonwhite Users have used this method, as compared with only about a quarter of the white Users. Among nonwhite couples, it is the second most common method, while among whites it ranks fourth. Condom is the most popular method among nonwhites, as it is among whites. In fact, the proportion using this highly effective method may be somewhat greater for nonwhites than for whites, but we cannot be sure because the white-nonwhite differences are not quite large enough to be significant. The proportions of Users using diaphragm and rhythm are both lower for nonwhites than for whites. However, the proportions using jelly alone and vaginal suppositories are greater among nonwhites.

WHITE—NONWHITE DIFFERENCES IN FERTILITY

There are very small and insignificant differences between whites and nonwhites in the proportion using withdrawal. It is important to note that nonwhite couples show higher pro­ portions trying to control fertility with the most readily available "female" methods: douche, jelly, and suppositories. Unlike dia­ phragm, these methods can be obtained without a doctor's prescrpition and, like diaphragm, they are applied by the wife. In other words, it appears that in many nonwhite families, the wife feels that she must take the chief responsibility for contraception, and she tries to do this without medical advice, which is generally less available to nonwhite than to white wives. Another interesting feature of the comparisons shown in Table 196 is the fact that nonwhites show an average of 215 reports of methods per 100 couples, compared with 189 for matched whites and 186 for all whites. This means that the nonwhite Users tend to use more methods than white Users, probably in an effort to find methods that give them satisfactory protection. FERTILITY PLANNING STATUS

Nonwhite couples have been much less successful than whites in planning fertility. Only 7 per cent of the nonwhite couples have Completely Planned Fertility, compared with 19 per cent for matched whites and 21 per cent for all whites (Table 197).11 But the most important contrast is in the proportion with Excess Fertility: 31 per cent for nonwhites and 17 per cent for matched whites and all whites. As was pointed out earlier, not only do the nonwhites have a higher prevalence of Excess Fertility, but within this group they have many more children than comparable white couples (Table 188). The proportions in the intermediate planning groups (Partly Planned and Partly Unplanned) are nearly the same for whites and nonwhites. We may say, then, that the two color groups differ chiefly 11 The planning status groups are defined in detail in Chapter 6. Briefly, the definitions are as follows:

Completely Planned: Users with no pregnancies and other Users who stopped using contraception in order to conceive before every pregnancy. Partly Planned: Couples who had one or more conceptions before starting to use contraception because they wanted these conceptions as soon as possible. Any con­ ception after use began occurred when contraception was stopped in order to conceive. Partly Unplanned: Couples who have had one or more "unplanned" pregnancies, but have not had more pregnancies than wanted. Excess Fertility: Either the husband or the wife or both did not want another child at the time of the last conception.

FERTILITY AND FAMILY PLANNING IN THE U.S.

with respect to the proportions in the two extreme planning groups —Completely Planned and Excess Fertility—and we shall focus atten­ tion mainly on the proportions in these groups. As we might expect from previously described differentials, the white-nonwhite differences in planning status are greatest among couples on farms in the South. Among nonwhites living on Southern farms, nearly half (48 per cent) have had more pregnancies than the husband or the wife wanted (Table 197 and Figure 41). This TABLE 197 Per Cent Distribution by Pregnancy Planning Status, for Nonwhites, Matched Whites, and All Whites, by Southern Farm Residence

Southern farm residence

No. of couples

Total

Totail None Some, not on farm now® On farm now

270 116

100 100

121 33

100 100

Total None Some, not on farm now® On farm now

270 150

100 100

90 30

100 100

Total None Some, not on farm now* On farm now

2,414 1,784 540 90

a Either

Com­ Partly Un­ pletely Partly Excess Planned Planned planned Fertility Nonwhite 7 19 23 9

43 40

31 28

45 42

29 48

42 44

17 15

43 27

19 27

100 100

20 18 27 20 All white 25 21 21 26

37 37

17 17

100 100

22 27

23 24

37 29

18 20

7

18 9 Matched white 19 22 17 24 —

husband, wife, or both have lived on a farm in the South.

is the highest proportion with Excess Fertility found for any socio­ economic group in this study. It compares with 27 per cent for matched whites and 20 per cent for all whites on Southern farms. The only white group that has a proportion with Excess Fertility anywhere near this consists of those with a poor education (the wife did not go beyond grade school) and a low income (the husband made less than $4,000 a year); such couples had 39 per cent with Excess Fertility. Nonwhite couples outside the rural South are more successful family planners than those who remain, but they are still not as successful as whites. A few of them have Completely Planned Fertility

WHITE—NONWHITE DIFFERENCES IN FERTILITY

(about 8 per cent), compared to none in the rural South, but this is well below the proportion of 17-22 per cent for the white population. Also, the prevalence of Excess Fertility is lower among nonwhites not living on Southern farms (29 per cent instead of 48 per cent), but this is still nearly twice as high as the white proportions of 15-18 per cent. Obviously, the move from the Southern farm to other areas has brought with it a substantial improvement in the success with which nonwhite couples plan their families. But they have still not achieved the levels observed for the white population. COUPLE'S SOUTHERN FARM RESIDENCE TOTAL

N ONE

'////////λ SOME 1 NOT ON FARM NOW

ONfSOUTHERN) FARM NOW

"I 20

Γ 30

40

PERCENT EXCESS FERTILITY NONWHITE 1¾¾¾] MATCHED WHITE

X///\ ALL

WHITE

FIGURE 41 Per cent Excess Fertility for nonwhites, matched whites, and all whites, by couple's Southern farm residence

The extent to which nonwhites differ from whites in their family planning practices can also be seen when we compare the different education groups (Table 198 and Figure 42). In every educational category, including the college group, nonwhite couples have shown less success in family planning than white couples. Their proportion with Completely Planned Fertility is lower, and their proportion with Excess Fertility is higher. Apparently, planning status is the one fertility variable in which nonwhites differ consistently from whites, regardless of socioeconomic

FERTILITY

AND FAMILY

PLANNING

IN THE

U.S.

TABLE 198 Per Cent Distribution by Pregnancy Planning Status, for Nonwhites, Matched Whites, and All Whites, by Wife's Education Wife's education

No. of couples

Total

Total College High school, 4 High school, 1-3 Grade school

270 37 73 86 74

100 100 100 100 100

Total College High school, 4 High school, 1-3 Grade school

270 37 82 86 65

100 100 100 100 100

Total College High school, 4 High school, 1-3 Grade school

2,414 427 1,153 579 255

100 100 100 100 100

Completely Planned

Partly UnPartly Planned planned

Nonwhite 7 19 24 14 26 7 22 5 12 1 Matched white 22 19 35 16 24 26 10 22 11 23 All white 21 25 35 18 24 26 13 29 24 6

Excess Fertility

43 43 51 35 43

31 19 16 38 43

42 46 45 45 31

17 3 5 22 35

37 36 37 37 38

17 11 14 21 32

FIGURE 42 Per cent Excess Fertility for nonwhites, matched whites, and all whites, by wife's education

364

WHITE—NONWHITE DIFFERENCES IN FERTILITY

status. The white-nonwhite differentials vary in magnitude, but they are in the same direction in all education groups and all regions of the country (Tables 198 and 199). Nonwhite couples have a higher prevalence of Excess Fertility not only because they have a higher average number of children. TABLE 199 Per Cent Distribution by Pregnancy Planning Status, for Nonwhites, Matched Whites, and All Whites, by Region of Residence

Region of residence

Total Northeast North Central South West Total NortheEist North Central South West Total Northeast North Central South West

No. of couples

270 41 74 136 19 270 41 74 136 19 2,414 581 715 744 374

Total

100 100 100 100 *

100 100 100 100 *

100 100 100 100 100

Com­ pletely Planned

Partly Planned

Partly Un­ Excess planned Fertility

Nonwhite 7 19 7 27 12 26 15 4 *

*

Matched white 19 22 10 20 18 24 21 21 *

*

All white 21 25 19 28 22 25 25 21 16 28

43 41 39 43

31 24 23 38

*

*

42 63 41 37

17 7 18 21

*

*

37 40 36 36 35

17 14 17 18 21

TABLE 200 Percentage of Couples with Excess Fertility, for Nonwhites, Matched Whites, and All Whites, by Number of Births by 1960 Number of births Total 0 1 2 3 4 5 or more

Nonwhite

Matched white

All white

31

17

17

21 39 60 77

12 29 43 60

11 28 41 46

Even when we compare couples with identical numbers of births, we find a higher proportion of nonwhites with Excess Fertility. For example, among couples with only two births, the proportion with Excess Fertility is 21 per cent for nonwhites, compared with 11-12 per cent for whites (Table 200). Similarly, among couples with three

FERTILITY AND FAMILY PLANNING IN THE U.S.

births, the incidence of Excess Fertility is 39 per cent for nonwhites and 28-29 per cent for whites. Apparently so many nonwhite couples want small families that a sizable minority of those with only two or three children think that they have already had too many. In the rest of this section, we shall focus attention on the propor­ tion of couples with Excess Fertility, since this is the most meaningful criterion with which success in family planning is usually judged. Under what contraceptive circumstances do excess (or unwanted) pregnancies occur? In Chapter 6, we classified the contraceptive circumstances of unwanted last conceptions under three main headings:12 a. before contraception began b. when contraception was being used always ("accidental" pregnancies) c. under other circumstances (usually because contraception was being used irregularly). Table 201 presents data on such conceptions for couples classified by wife's education. We note first that nearly the same proportions of whites and nonwhites have had unwanted last conceptions that were accidental. It is important to understand that this does not mean that the two color groups are equally effective users of contraception. It is due simply to the fact that relatively fewer nonwhite couples have used contraception, and hence that a smaller proportion of them have been exposed to the risk of having accidental conceptions. In any case, it is important to emphasize the fact that the reason nonwhites have a higher proportion with Excess Fertility is not be­ cause they have more accidental conceptions. The main reason for nonwhite couples' higher prevalence of Ex­ cess Fertility is that so many of them have unwanted conceptions before they begin to try to control their fertility. The proportion is 13 per cent, compared with 4 per cent for white couples. Even after they have had more children than they want, many nonwhite couples do nothing to prevent additional pregnancies. This kind of behavior is characteristic of the poor and uneducated. Among the wives who went no further than grade school, over a quarter reported one or more unwanted conceptions that they had not tried to prevent. "Last, or most recent, unwanted conceptions are used rather than all un­ wanted conceptions, so that each couple with an unwanted conception is repre­ sented only once in the distribution. If we were to use all unwanted conceptions, some couples would be represented only once (because they had only one un­ wanted conception) but other couples would be represented more than once (be­ cause they had two or more unwanted conceptions).

WHITE—NONWHITE DIFFERENCES IN FERTILITY

The other cause of a high prevalence of Excess Fertility among nonwhite couples is irregular use of contraception. Fourteen per cent of the nonwhite couples reported one or more unwanted pregnancies occurring under such circumstances, compared to 7 per cent for white couples. These findings again emphasize the fact that successful family planning requires that the use of contraception be regarded as an accepted and normal part of married life. Couples should not delay TABLE 201 Percentage Whose Last Conception Was Unwanted and of Specified Contraceptive Status, for Nonwhites, Matched Whites, and All Whites, by Wife's Education

Wife's education

Total unwanted

Total College High school, 4 High school, 1-3 Grade school

31 19 16 38 43

Total College High school, 4 High school, 1-3 Grade school

17 3 5 22 35

Total College High school, 4 High school, 1-3 Grade school

17 11 14 21 32

Before use of contraception began

Accidental

Other unplanned

Nonwhites 13

4 5 3 9

—

7 10 27 — Matched white 4 6 —

2 9 9

7 3 2 9 14

6 7 6 7 8

7 5 6 9 10

—

—

3 12

14 14 7 19 16

All white 4 —

2 5 15

use until they have more children than they want, and they should use contraception regularly, without omission, when they do not want conception to occur. The consequences of long-delayed use and irregu­ lar use are many unwanted children. One of the most interesting features of the comparisons presented in this section is the fact that nonwhite couples have a higher propor­ tion with Excess Fertility than white couples even when we limit comparisons to the better-educated. We would expect the /^-edu­ cated nonwhites to be less successful family planners than comparable whites on the basis of differentials reported earlier in this chapter, and this expectation has been confirmed. But we would also expect the better-educated nonwhite couples to be at least as successful in

FERTILITY AND FAMILY PLANNING IN THE U.S.

controlling fertility as comparable white couples because they have had fewer children by interview, they expect fewer altogether, and a high proportion of them have used or expect to use contraception. Instead, we find that among couples with college-educated wives, 19 per cent of the nonwhites have Excess Fertility compared with 3 per cent of matched whites and 11 per cent of all whites in this education group (Table 198). First, we must ask whether these differences are statistically sig­ nificant at the confidence level chosen for this study (.95). The difference between the 19 per cent for nonwhites and the 11 per cent for all whites is not significant at this level. However, the differ­ ence between the 19 per cent for nonwhites and the 3 per cent for matched whites appears to be significant, even though the number of cases involved is small (37 for both groups). In spite of this finding, however, we suspect that the extremely low proportion of 3 per cent for matched whites may result from sampling variability. It is well below the 11 per cent for all college-educated whites and the 12 per cent for college-educated white Protestants. Consequently, there is a very good possibility that college-educated nonwhites do not have a higher prevalence of Excess Fertility than comparable whites in the population represented by the sample. However, we have other reasons for suspecting that the bettereducated nonwhites may have a higher prevalence of Excess Fertility than comparable whites. We know that they have had fewer births by 1960 than comparable whites (this is confirmed by the 1960 Census) and we have good reason to believe that their family-size goals are lower than those of comparable whites. The latter conjecture is based not only on the findings of this study, but also on the impres­ sions of investigators who have written about the fertility of upperclass nonwhites.13 Because they want relatively few children, perhaps nonwhite couples have a harder time preventing unwanted conceptions than comparable whites. As we learned in Chapter 6, many white couples, including the well-educated, have accidental and other unplanned conceptions that occur before they are wanted. But most of the white couples with such "timing" failures wanted more children eventually. If a well-educated nonwhite couple had an accidental or other un­ planned conception, however, they would be less likely than a com­ parable white couple to have wanted another child. Therefore, they would be more likely to be classified as having Excess Fertility. This hypothesis gains some support from the fact that the propor­ tion with Excess Fertility is higher among nonwhites than whites 13See,

for example, Frazier 1949, pp. 331—332.

WHITE—NONWHITE DIFFERENCES IN FERTILITY

among couples who have had only two or three births (Table 200). Thus, it seems quite likely that some nonwhite couples are classified as having Excess Fertility simply because their low family-size goals are difficult to achieve. SUMMARY

The rapid cultural change that the nonwhite population is under­ going has left its mark on fertility differences between whites and nonwhites in this country. We find that nonwhite couples have had and expect more births than white couples, and that this difference is brought about partly by the unusually high fertility of a minority of nonwhite couples who live in the rural South, and partly by the moderately high fertility of the many nonwhite couples who have Southern farm origins. Nonwhite couples with no Southern farm back­ ground have and expect about the same number of births as similar white couples. These differentials suggest that as the influence of Southern rural patterns of mating and childbearing diminishes, the fertility differences between whites and nonwhites will decline. It is clear that nonwhite wives do not want as many children as they expect to have. They want fewer children than white wives. The white-nonwhite differences in past and expected fertility are greatest among the least-educated. The differences are smallest among couples in the college group. Nonwhite wives with some college educa­ tion had and expected fewer births than comparable whites. If we combine couples in the college group with noncollege couples who have had no Southern farm background, we find that whites and nonwhites have approximately the same past and expected num­ ber of births. Inasmuch as this group contains 63 per cent of the nonwhite couples, we may say that a majority of nonwhite couples have and expect about the same number of births as similar white couples. Although there is some evidence that in the past there was a higher prevalence of fecundity impairments in the nonwhite popula­ tion, the present study gives us no reason to believe that this is still true. Nonwhite couples appear to have had fewer contraceptive opera­ tions than whites, but they also have more fecundity impairments that were discovered through the nonuse of contraception. On bal­ ance, both groups have about the same proportion classified as Subfecund. A lower proportion of nonwhites than of whites have used contra­ ception, and a lower proportion expect to have done so by the end of the childbearing period. Part of this lesser use is due to the fact

FERTILITY AND FAMILY PLANNING IN THE U.S.

that a higher proportion of nonwhites delay using contraception and become Subfecund before making any attempt to control fertility. It is also due in part to the failure of other nonwhite couples to use contraception even though they clearly need it. Nonwhite couples are considerably less successful in controlling fertility than white couples. This seems to be true in all socioeconomic groups. One of the important causes of nonwhite couples' high prev­ alence of Excess FertiUty is the fact that some couples (particularly the poor and uneducated) do not use contraception before they have unwanted pregnancies. The other important cause (which operates in all socioeconomic groups) is the failure of some couples to use contraception regularly.

C H A P T E R 10

Family Size and Population Growth in Future Years The primary purpose of the 1955 and 1960 studies of Growth of American Families was to provide information needed in order to improve population projections for the United States. The use of such projections has been increasing steadily as more planning for the future is done by commercial enterprises and government agencies. Accompanying greater use has been a desire for projections that would be closer to actual developments, especially with respect to the popula­ tion not yet born. It is true that the differences between past projec­ tions of population growth and the actual growth that followed have not been due entirely to differences in fertility. Net immigration and mortality also have departed from the trends outlined, but these differ­ ences usually have been somewhat smaller in the short-run and much smaller in the long-run than those for fertility. Changes in mortality and immigration are likely to have less in­ fluence on the trend of United States population growth in the future than in the past if thermonuclear war does not occur. Age-specific death rates have declined gradually but substantially in the last hundred years. The expectation of life at birth, which summarizes these changes very well, increased from 48.4 in 1910 to 66.6 in 1960 for males and from 51.8 to 73.1 for females. It is now so close to the human life span—often said to be 100 years—that further increases are likely to be small unless revolutionary discoveries make possible the control of cardio-vascular diseases, the retarding of the aging process, or some other way of reducing substantially the death rate of people past 60. Annual net immigration was relatively large during the first four­ teen years of this century, when between 500,000 and 1,300,000 persons were admitted annually. The flood was cut to a trickle by World War I, and then was controlled by the quota laws and ad­ ministrative regulations adopted in 1922 and later years. During the 1950's net immigration amounted to about 300,000 annually. Unless public attitudes change it will continue to be small and have a dimin­ ishing influence on population growth. Changes in fertility have had more effect on the course of the nation's population growth. The crude birth rate fell from about 55 per 1,000 in the early 1800's to 17 in the 1930's—more rapidly than the death rate—and the rate of natural increase tumbled from

FERTILITY AND FAMILY PLANNING IN THE U.S.

about 35 to 6 per 1,000. Then, in contrast to mortality, the downward trend of fertility was reversed. A small rise occurred as the economy recovered from the depression of the 1930's, World War II was followed by the "baby boom," and from 1947 to 1958 the crude rate was about 50 per cent above the low of the 1930's. Since 1957, however, there has been a small decrease. Size of completed family also declined substantially. The number of births per 1,000 ever-married women living to ages 45-49, was well above 4,700 for such women in 1910, but only about 2,500 for corresponding women in the 1950's. During that decade it became clear that the trend was turning and that wives who reach middle age in the late 1960's and early 1970's would have larger families. One question was how much larger their families would be. A more important question was what would happen among still younger women. The 1955 and 1960 studies of Growth of American Families were planned to provide partial answers. Before the 1955 GAF results became available the common pro­ cedure for projecting fertility was to compute period birth rates of various types, fit straight lines or curves, and extrapolate past trends mathematically, graphically, or on a "judgment" basis. In choosing among various procedures, consideration was sometimes given to the effect that changes in social and economic conditions had had on fertility in the past, and might have in the future. The cohort fertility tables developed in the 1950's provided useful information about the past experience of actual groups of women as they lived through the childbearing period, but not about their future births. The first information concerning the birth expectations of a representative sample of a nation's married couples came from the 1955 study. The second study provides similar expectations as of 1960 and some facts regarding the reliability of these expectations. In subsequent sections of this chapter we shall discuss the relia­ bility of the past and expected future birth rates of the wives inter­ viewed; a procedure for using their expectations in projecting final birth rates of cohorts; how to allow for changes in the timing of marriages and births; and the relation between cohort and period birth rates. Finally we shall evaluate the projected period birth rates, and discuss the numbers of births and the population growth that they imply. ARE THE GAF WIVES TYPICAL AS TO PAST CHILDBEARING?

With respect to past fertility the wives in the GAF studies are an excellent sample of the universe from which they were selected. This is shown in Table 202, where cumulative birth rates for the

FUTURE FAMILY SIZE AND POPULATION GROWTH

wives in the 1955 and 1960 studies (adjusted slightly so as to represent not only these wives but also widows, divorcees, and other evermarried women) are compared with rates for all ever-married women from cohort fertility tables and from the 1960 Census. The 1960 GAF rates in Column B are slighdy larger than the census rates (Column C), perhaps because the GAF interviewers always ques­ tioned the wife herself, while the census enumerators were allowed to accept information from another member of the household or a neighbor, who might not have known about children who had TABLE 202 Cumulative Birth Rate by Age in 1955 or 1960, for Wives in GAF Studies, and for Ever-Married Women, According to Cohort Fertility Tables and the 1960 Census 1955 1960

Current age 20-24 25-29 30-34 35-39

Wives in GAF study A 1,460 2,360 2,730 2,740

Ever-married women Estimated Cohort from GAF fertility tables0 wives' Censusb D B C 1,460 2,345 2,715 2,715

1,441 2,241 2,627 2,686

1,579 2,390 2,761 2,824

Wives in GAF study E 1,250 1,950 2,380 2,560

Ever-married white women Estimated Cohort from GAF fertility wives· tables0 G F 1,242 1,927 2,335 2,516

1,303 1,962 2,362 2,478

• The rates in Cols. B and F are obtained from those in Cols. A and E by allowing for the slightly lower fertility (at some ages) of women who were widowed, divorced, separated, or married with husband absent than of wives with husband present. The procedure is explained later in the text. b Census 1960, p. 3. 0 The tables were prepared by the Scripps Foundation to supplement those in Whelpton 1960.

died or were living away from home. In contrast, the rates based on the GAF studies are smaller (with one exception) than those from the cohort fertility tables in Columns D and G. Only two of the differences are large enough to be statistically significant, however, and one of these (at ages 20-24 in 1960) is due in part to the fact that births to single women are included with those to married women in the cohort tables but not in the GAF studies.1 1 The birth rates in Columns B and E, estimated chiefly from the experience of the GAF wives as explained later in this chapter, include premarital births to these and other wives, but not the births to women who had not yet married. In contrast, the rates in the cohort fertility tables are obtained by dividing the cumulative birth rate of all women in a given cohort at a given age by the corresponding cumulative first marriage rate, which means that all births to single women are credited to ever-married women.

FERTILITY AND FAMILY PLANNING IN THE U.S.

While there may be a tendency for the GAF rates to be slightly too low, it is also possible that the cohort table rates are slightly too high because the adjustment for births not registered is too large. THE RELIABILITY OF SHORT-RUN BIRTH EXPECTATIONS

The average number of births that the wives in the 1955 study expected during the following five years agrees very closely with the number that actually occurred during this five-year period to com­ parable wives in the 1960 study, as is shown by the following rates:

Rate (per 1,000 wives): Expected in 1955 Reported in 1960 Difference

Total

1931-37

Cohorts 1926-30

1921-25

1916-20

700 720 20

1,500 1,420 -80

940 860 -80

430 480 50

180 210 30

The comparison is restricted to white wives (because other wives were not interviewed in 1955) and to 1960 wives married in 1955 or earlier. Actual childbearing may have been somewhat below that expected among wives under 30 in 1955, and slightly above among those aged 30-39; a more definite statement is not justified because each of the four differences is too small to be statistically significant. The fairly close agreement for the cohort averages does not mean that each couple in the 1955 study had the number of babies the wife expected during the following five years. On the contrary, those with certain characteristics (e.g., with more education, using contra­ ception) had fewer than expected, and others (e.g., with less educa­ tion, not yet using contraception) had more.2 What it does mean is that the extra births to wives who had more than they expected were approximately balanced by the births that did not occur to wives who had fewer than expected. Whether this balance is a peculiar characteristic of the 1955-60 period cannot be known until similar information regarding future periods is collected and analyzed. The foregoing comparisons support the conclusion that the GAF studies provide valuable information regarding the future fertility of wives aged 18-39. By using this information properly, it should be possible to improve projections of fertility and population for the United States. A procedure for doing so was discussed in detail in the first book, Family Planning, Sterility, and Population Growth, and is now being used again with certain modifications made possible by data from the second study. Let us consider it briefly. 1For a discussion of the differences for couples classified by these and other characteristics, see Chapter 1.

FUTURE FAMILY SIZE AND POPULATION GROWTH

THE FUTURE FERTILITY OF SINGLE WOMEN WHO WILL MARRY

Since both studies are restricted to wives with husband present, it is necessary to make estimates regarding the future fertility of the women who were single in 1955 or 1960 but will marry before they reach the end of the childbearing period. The proportion of single women marrying at later ages was obtained for each group of cohorts by projecting the period and cumulative first marriage rates in cohort marriage tables prepared by the Scripps Foundation on the baas of census data on women by marital status and marriage registration TABLE 203 Expected Final Birth Rate of GAF Wives Marrying at Specified Ages as a Percentage of the Rate for Those Marrying by Younger Ages, and Projected Percentages for Women Who Will Marry after 1960, by Cohort Percentage Cohort

a. b. c. d.

1921-25 1926-30 1931-35 1936-40

a. 1921-25 b. 1926-30 d. 1931-35 & 1936-40 a. 1921-25 d. 1926-30 to 1936-40 d. 1921-25 to 1936-40

Sotirce Wives first 1960 study 1960 study 1960 study Projected1" Wives first 1960study 1960 study Projectedb Wives first 1960 study Projected Wives first Projected

Most likely A marrying at 92.0 90.4 91.0 91.0 marrying at 64.8 68.1 66.0 marrying at 47.4 48.0 marrying at 35.0

Minimum B ages 20-24 90.1 87.8 87.3 88.0 ages 25-29 62.7 64.3 64.0 ages 30-34 44.1 44.0 ages 35-39 25.0

Maximum C

to 25-29" 96.2 93.4 91.9 93.0 to 30-34* 69.8 72.7 71.0 to 35-39" 53.8 54.0 to 45-49» 45.0

a The percentages in the first, second, third, and fourth decks are based, respec­ tively, on the final rates of wives marrying before ages 20-24, 25-29, 30-34, and 35-39. b The projected percentages are based on the preceding lines, with the earlier cohorts given less weight than the later cohorts.

data collected by the National Vital Statistics Division. (Actual and projected marriage rates are shown later in Table 209). The "ex­ pected" final birth rate for each group of future brides is estimated on the basis of the relationships between age at marriage and expected final rate among the wives interviewed. The actual and projected rates conform to the well-know tendency for older brides to have smaller families. For example, the families expected by the wives in each cohort who married at ages 20-24 to 25-29 are slightly smaller (4 to 13 per cent) than those expected by the wives who married before 20-24 (Table 203, top

FERTILITY AND FAMILY PLANNING IN THE U.S.

deck). But the women who remained single until ages 30-34 to 35-39 expect families only about half as large as those marrying before 30-34 (third deck). The relatively small differences between ratios for given age-at-marriage groups in different cohorts suggest that the projected percentages (bottom line of each deck) will yield useful results for the cohort members who will marry after 1960 but before middle age. TABLE 204 Most Likely Final Birth Rate Expected by Wives, and Estimated for Single Women Who Will Marry before Ages 45-49, by Cohort, by Age at Marriage Age at, or date of, marriage8

a. b. c. d. e. f. Sh. 1. 1· k.

Age at marriage Under 15-19 15-19 to 20-24 20-24 to 25-29 25-29 to 30-34 30-34 to 35-39 35-39 to 45-49 Date of marriage Before 1960 study After 1960 study Weighted average of lines g and h Other ever-married women, I9600 All women marrying by 45-49

Cohort 1936-40 1931-35 1926-30 1921-25 1906-1Ob A B C D E 3,400 2,920 2,785 1,980 1,420 1,030

3,640 3,360 3,130 2,235 1,605 1,160

3,580 3,400 3,110 2,290 1,590 1,150

3,510 3,100 2,940 2,020 1,440 1,060

3,060 2,470 2,900 2,870 2,900

3,380 1,830 3,275 3,040 3,255

3,310 1,380 3,245 3,030 3,225

3,020 1,060 2,960 2,700 2,930

3,018 2,077 1,508 930 678 2,458 2,458 2,119 2,364

" In Cols. A to D, the rates on lines a to e above the horizontal lines are the expecta­ tions of GAF wives. Those below are estimated by applying the per cents in Table 203. b The rates in Col. E are for wives 40-44 years of age, from Census, Fertility, 1950, and relate to 1950. They are used instead of rates for women 45-49 because all wives married by ages 25-29 are combined in the latter group. Lines a to i are restricted to wives married once with husband present. No adjustment has been made to include the very few marriages or births after 1950. 0 The rates in Cols. A to D are for ever-married women not included in study, i.e., widowed, divorced, or not living with husband for other reasons. Col. E relates to these women and wives married more than once.

The "most likely" expected final rates of the 1960 GAF wives, classified by cohort and age at marriage, are shown in the upper right triangle of Table 204 (above the underlines) and on line g; the estimates for women marrying after 1960 are shown on lines c to f (below the underlines) and on line h. The estimated final rate for the 1960 wives and future brides combined (line i) differs only slightly from that expected by the wives alone (line g). This is due primarily to two factors. First, most women have married by ages 25—29, consequently so few marry later than their lower fertility

FUTURE FAMILY SIZE AND POPULATION GROWTH

has little influence on the group as a whole. Second, most of the women who marry between ages 20-24 and 25-29 have plenty of time to bear the two to four children wanted (and expected) by most of those who marry younger. Similar conclusions hold true for the 1955 study. A second adjustment is needed to allow for the women who had married but were not included in the GAF studies because their marriage had been ended by death or divorce, or they were not living with husband for other reasons. The number of such women is relatively small, and census data show that they have not been far behind wives with husband present in respect to previous childbearing.3 The estimates for them, based on GAF and census data are on line i of Table 204. Combining the two adjustments gives final birth rates for all ever-married women that are only slightly below those of wives with husband present in 1960. (Compare lines g and k.)4 The methodology outlined above was used for the white wives in the 1955 study and all wives in the 1960 study. The next step in the 1955 study involved estimating the final rates for all wives on the basis of those for white wives. This is no longer necessary, because nonwhite wives were interviewed in 1960. Instead, a new step is desirable, namely, allowing for the changes in final fertility expectations that apparently occur as wives grow older and have been married longer. This was not feasible in the first study because there was no information as to whether such changes took place, to say nothing of their nature or extent if they did. Now we know what happened between 1955 and 1960. THE INCREASE IN SIZE OF EXPECTED FAMILY BETWEEN 1955 AND 1960

The white wives aged 18-39 in the 1955 study expected a fined birth rate of 3,020 per 1,000 according to the most likely series. Five years later comparable wives five years older and married five years longer expected a final rate of 3,170. (The increase of 150 is well above the minimum for statistical significance.) When the four groups of cohorts are considered separately, the size of the increase appears to vary inversely with age and directly with the number of years remaining in the childbearing period, which seems logical. The largest rise is 210 for the cohorts of 1931-35—from 3,190 in 1955 (when the wives were 20-24 years old) to 3,400 8 Census 1950, pp. 5C-45. See page xxxi for the meaning of any abbreviated reference. 4Comparable rates for white women in 1955 are in FPSPG , pp. 486-487.

FERTILITY AND FAMILY PLANNING IN THE U.S.

in 1960 (when they were 25-29). Neeurly as large an increase is found for wives aged 25-29 (cohorts of 1926-30), but those for the two older groups are considerably smaller. It would seem more reasonable to find the least rise among wives 35-39 in 1955, but instead it appears in the 30-34 group. This could be due to sampling error. Increases are larger for the minimum expectations than for the most likely. In theory, the maximum expectations should have risen least. In fact, they rose somewhat more than the most likely, but this may be due to a slight change in the wording of the question regarding the largest number of births expected by wives who were uncertain whether they could have children in the future, or whether they would if they could.8 The increases for the 1926-30 cohorts in all three series are large enough to be statistically significant, as are those for the 1931—35 group in the minimum and maximum series. Although comparisons for more periods are needed in order to be sure whether the number of children expected usually increases with rising age and longer marriage, there is one reason for thinking that it does. The total number of children that wives wanted when they were interviewed in our two studies is substantially larger than the number they recalled wanting when their first baby was a year old: Average number wanted: When first baby was one year old At interview

1960 Study 2.4 2.9

1955 Study 2.6 3.1

Part of this difference probably reflects memory bias, but part may well be real. In spite of the pleasure that the first baby brings most couples, the problem of looking after a lively yearling may well lead many wives to reduce temporarily the number of children they want and expect. Many more of the wives in the 1931-35 cohorts had a first child about a year old in 1955 (when they were 20-24) than in 1960. A temporary tendency for a year-old baby to lower the number of children the mother wants would have been much less important for these cohorts on the later date than the earlier. Similarly, many more of the wives in the 1936-40 cohorts had a first child about a year old in 1960 than will be the case in 1965, 5 In the 1955 study such wives were asked either: 20i. If you can have (more) children, how many more do you want to have? or 27. If you do have (more) children . . . how many do you think you will have? In the 1960 study comparable wives were asked: 31. If you do have (more) children, what is the largest number you are likely to have in all (counting those you now have) ?

FUTURE FAMILY SIZE AND POPULATION GROWTH

consequently the group as a whole may want and expect larger com­ pleted families on the later date than five years earlier. The foregoing evidence of an increase in birth expectations with age between 1955 and 1960 led to the decision to provide for such an increase in the fertility projections for the present study. Un­ fortunately, because of the irregularity of the 1955-60 changes and the absence of information for earlier periods, our treatment of this factor cannot be precise. The medium, low, and high allowances that were chosen appear in Table 205, Columns F to H. The high is quite similar to Column E which shows what would happen if the 1955-60 increases in most likely expectations of wives of different TABLE 205 Increase in Most Likely Expected Final Birth Rate as White Wives Grew Older from 1955 to I960, and Projected Increase for Years after I960, by Cohort

Cohort 1931-35 1926-30 1921-25 1916-20

Age in 1955 A

Most likely final birth rate expected by wives in 1955» 1960 C B

20-24 25-29 30-34 35-39

3,190 3,100 2,980 2,900

3,400 3,290 3,000 3,010

Total increase Increase, from age in CoL A to 45-49 1955 to 1955-60 1960 experi­ Projected, after I960'' ( C - B ) ence1" Medium Low High D F G H E 210 190 20 110

530 320 130 110

350 200 100 25

200 100 50 0

500 300 150 50

» From FPSPG, pp. 486-487, Col. L, Une 2. b Each group is assumed to have the increases shown in Col. D for current and older ages. 0 The basis of projecting is discussed in the text.

ages (and in different cohorts, Column D) were to apply at successive ages in the same cohort in the future. The medium and low series are successively smaller than the high. Using the most likely experience of 1955-60 as a basis for the high projections rather than the medium is a conservative approach, which seems desirable until information for additional periods becomes available. The effect of allowing for single women who will marry after 1960, for widows, divorcees, and other women not living with hus­ band, and for a tendency for expectations to rise as wives grow older is summarized in Table 206. The first two adjustments reduce the final rate in all cases (compare Columns A and D), but by a relatively small amount. (The largest is 165, or 6.1 per cent, for the cohorts of 1936-40, minimum series.) The allowance for an in­ crease in expectations as wives grow older and are married longer

FERTILITY AND FAMILY PLANNING IN THE U.S.

more than balances the foregoing reductions for the most recent group, but fails to do so for the earliest. In each of the three series the upward trend of completed fertility from the cohorts of 1921-25 to 1931-35 is slightly steeper for the estimates in Column E than for those comparable to the 1955 study (Column D), or for the wives expectations in Column A. In contrast, the decline from the TABLE 206 Final Birth Rate Expected by GAF Wives in 1960, Estimated for Single Women Who Will Marry Later and for Other Ever-Married Women, and Adjusted for the Tendency to Increase with Age, by Cohort Estimated for All women marrying by ages 45-49

Series and cohort Most likely 1936-40 1931-35 1926-30 1921-25 Minimum 1936-40 1931-35 1926-30 1921-25 Maximum 1936-40 1931-35 1926-30 1921-25

Expected by GAF wives A

Single women marrying later B

Other evermarried women C

Not ad­ justed for increase with age D

Adjusted for increase with age E

3,060 3,380 3,310 3,020

2,470 1,830 1,380 1,060

2,870 3,040 3,030 2,700

2,900 3,255 3,225 2,930

3,250 3,455 3,325 2,955

2,720 3,070 3,050 2,840

2,070 1,420 1,040 710

2,530 2,760 2,795 2,535

2,555 2,955 2,975 2,750

2,755 3,055 3,025 2,750

3,470 3,760 3,650 3,310

2,945 2,335 1,825 1,490

3,290 3,385 3,345 2,970

3,320 3,620 3,555 3,225

3,820 3,920 3,705 3,275

1931-35 to the 1936-40 group is slightly smaller for the final estimates than for the others. THE PROJECTIONS OF FINAL BIRTH RATES OF EVER-MARRIED WOMEN

The remaining steps in the 1960 procedure are the same as (or very similar to) those of the first study. The final fertility rate of all women marrying by middle age, based on the wives most likely expectations but rounded to end in 50 or 00, is used in the medium projection (Table 207, Column B). It presents the best available information regarding size of completed family for the wives in the four groups of cohorts, and reasonable estimates for the other women

FUTURE FAMILY SIZE AND POPULATION GROWTH

who had married by 1960 or will do so before reaching middle age. As in the first study, the final rates in the low fertility series are above those based on the minimum expectations of the wives, and the rates in the high series are below those based on their maxi­ mum expectations. The chief reason for narrowing the range is the same as before, namely, the minimum and maximum expectations TABLE 207 Projected Final Birth Rate of Ever-Married Women in Cohorts of 1916-1920 and Later, and Actual Rate for Prior Cohorts Year when 15-19 A

Medium B

Low C

High D

Medium E

Low F

High G

1951-55 1946-50 1941-45 1936-40 1931-35 1926-30 1921-25 1916-20

1970 1965 1960 1955 1950 1945 1940 1935

3,000 3,100 3,200 3,300 3,450 3,300 3,000 2,700

2,500 2,700 2,900 3,100 3,300 3,200 2,970 2,680

3,500 3,500 3,500 3,500 3,600 3,400 3,030 2,720

3,050 3,050 3,050 3,050 3,050 3,000 2,950 2,680

2,500 2,500 2,600 2,750 2,850 2,850 2,850 2,660

3,350 3,350 3,350 3,350 3,250 3,150 3,050 2,700

1911-15 1906-10 1901-05 1896-1900 1891-95 1886-90 1881-85 1876-80 1871-75

1930 1925 1920 1915 1910 1905 1900 1895 1890

Actual" 2,481 2,458 2,625 2,917 3,237 3,518 3,719 4,024 4,365

Cohort'

Cumulative birth rate by ages 45-49 1955 projectionsb

1960 projections

" The cohorts between the horizontal lines were included in the GAF studies. The cohorts of 1956-60 to 1966-70 are assumed to have the same rates as those of 1951-55. b FPSPG, p. 340. c From Whelpton 1960. The rates are adjusted to allow for incomplete registration and enumeration.

represent extremes for each couple. If thermo-nuclear war is avoided and other terrible catastrophes do not occur, it is certain that some couples will fail to reach the minimum, others will exceed the maxi­ mum, but an important proportion (perhaps a majority) will be distributed between these extremes. Those who exceed their most likely expectations will tend to cancel those who remain below them; consequently the cohort averages will be much closer to the medium than the low or high. In the 1955 projections the difference between

FERTILITY AND FAMILY PLANNING IN THE U.S.

the low and high final birth rates varied from 40 for the cohorts of 1916-20 (then at ages 35-39) to 850 for the cohorts of 1946-50 (and later). In the 1960 series the range is slightly wider for the later cohorts. In both cases the medium is at (or close to) the mid­ point between the low and high. The new series are above the old for the cohorts of 1921-25 to 1946-50, with the excess largest for the 1931-35 group. Nearly all of the difference results from the new step—allowing for expecta­ tions to increase with age. Little was said above regarding the number of births expected by wives 18 or 19 in the two studies. This is because such wives, in spite of the decline in age at first marriage, composed only about 10 per cent of all women in the cohorts of 1936-40 in 1955, and the same proportion of all women in the cohorts of 1941-45 in 1960. In 1955 the most likely expected number of children averaged 2.9 for the 70 white wives in the cohorts of 1936-37, slightly below the 3.2 figure for the 436 wives in the cohorts of 1931-35. The difference is too small to be statistically significant; consequently, attention was not called to it in the first book. Now the data from the 1960 study show quite convincingly that married women in the 1936-40 cohorts will have smaller families than those in the 1931-35 group. A similar situation may be developing in the second study, for wives in the cohorts of 1941—42 expect smaller completed families than those in the 1936-40 group. Moreover, the white wives aged 18-19 in 1960 expect fewer children than did the wives of these ages in 1955. In the first study the medium projections continued the most likely final birth rate of 3,050 for ever-married women in the latest cohorts interviewed; there was no strong indication of an upturn or downturn. Now, with good evidence that young wives expect smaller families, the medium projection is that the final rate will decrease from 3,450 for the cohorts of 1931-35 to 3,000 for those of 1951-55, which happens to be the average for the cohorts of 1879 to 1925. The low projection in both studies is a decline to 2,500, which is slightly above the rate for the 1906-15 groups. The date when this low might be reached is now postponed ten years, however, because it appears that the rise which has been under way will con­ tinue longer and crest at a higher point than seemed likely when the previous projections were made. The earlier high projection set the final rate of the cohorts of 1936-40 and later at 3,350, slightly above the 3,250 rate for the 1931-35 group (aged 20-24 in 1955). The present scries gives 3,600

FUTURE

FAMILY

SIZE AND POPULATION

GROWTH

as the figure for the latter, and 3,500 for subsequent cohorts in view of the evidence indicating at least a slightly smaller final rate for women 20-24 in 1960 than for those 25-29. Although the final rate of the 1931-35 cohorts may be slightly above 3,500, no other group has had such large families since the cohorts of 1886-91, whose members are now in their 70's.

FIGURE 43 Actual and medium, low and high projected final birth rates, for all evermarried women living to middle age, by groups of cohorts

The differences between the projected final low and high rates are smaller for ever-married women in the cohorts of 1921-25 to 1936-40 in the present series than the first (Figure 43). This merely reflects the fact that the wives in each group were five years further along in the childbearing period at the time the second study was conducted, and the low and high series start to diverge in 1960 383

FERTILITY AND FAMILY PLANNING IN THE U.S.

instead of 1955. It is also worth noting that for these cohorts the excess of the 1960 projected final rates over those for 1955 is smallest for the high series, and largest for the low. This represents the in­ fluence of the foregoing factor and of the new allowance for expecta­ tions to increase with age. The former tends to reduce the high series, but the latter to increase it; the net result is a small increase. In contrast, both factors tend to raise the low series; the net result is substantially larger rates in the 1960 low series even though the allowance for expectations to increase with age is smaller in this series than in the high. PROJECTING THE PACE AT WHICH FAMILIES WILL GROW TO FINAL SIZE

After the final birth rates of ever-married women were projected, the intermediate rates were filled in. As was shown in Chapter 8, the recent trend has been for wives to concentrate more of their childbearing at younger ages and in the earlier years of married life. Average months from marriage to first birth, and from first to second decreased considerably from the cohorts of 1916-20 to those of 1931-35. A similar shortening of the third and fourth birth intervals occurred up to the cohorts of 1926-30, and probably is being continued by the next group. The extent to which such changes have affected population growth in the last four decades has been greatly underrated. Only recently has it been realized that more than half of the upsurge of the crude birth rate after World War II was due to the making up of the births postponed during the war and the advancing of other births from future years because more women married at younger ages and had their babies sooner after marriage.® The increase in the size of family of married couples, which is commonly cited as the cause of the postwar "baby boom," was a secondary factor. The causes of these changes in age at marriage and in child spacing are not well understood. Some studies have shown that economic conditions play an important role, but little is known about other factors. We can be certain that changes in various causal factors will occur in future years and will influence age at marriage and child spacing after marriage. But until more is known about these factors and unless useful projections can be made of their future trends, there is little basis for making more than low and high assump­ tions about changes in the timing of marriage and childbearing—as­ sumptions that probably will bracket actual developments. e See

Whelpton 1960, pp. 37-40 and FPSPG , pp. 263-271.

FUTURE FAMILY SIZE AND POPULATION GROWTH

Let us consider first the timing of births to women who marry before the end of the childbearing period. The measure of timing that will be used for each group of cohorts is the ratio of the cumula­ tive birth rate by specified age group to the final rate, i.e., the rate by ages 45-49. It might be thought that the length of future birth intervals can be based on wives' expectations, as is done for their final birth rate. Unfortunately, it appears that wives' opinions are much less reliable with respect to the spacing of future children and the final distribution of couples by parity than they are with respect to the average birth rate during the next five years or the average size of completed families. The chief reason is that changes of an opposite nature will tend to balance each other with respect to average family size or birth rate in the next five years but not with respect to the other two variables. For example, if 1,000 wives say they expect to have three children, but 200 have two and 200 others have four, the average number will remain three, as expected, but the percentage having three will be 60 instead of 100. Similarly, if 1,000 wives who recently had their second baby say that they expect to have a third two years after the second (and no more subsequently), but 200 are unable to have the third until six years later, and another 200 have the third a year early and an accidental fourth two years after that, the average number of births in the next five years is three, as expected, but the average length of the interval from the second to the third is nearly three years instead of the expected two. The yardstick we shall use to measure changes in the timing of births to ever-married women is not perfect, for it is affected to some extent by changes in final family size. For example, if the average number of births per woman increases from 2.5 in one group of cohorts to three in a later group, but age of mother at first birth, second, etc., remains unchanged, the ratio of the cumulative rate by ages 15-19 or 20-24 to the final rate may be lowered, raised, or left unchanged. If the larger families result primarily from a rise in the rate for third and higher order births, the ratios in question would decrease because these births tend to come at the mid 20's or later. It would then appear that births were being postponed to older ages when in fact there were simply more of the births that usually come at these ages. On the other hand, if the increase in family size occurred chiefly because more couples were having two children instead of none or one, the birth rates by young ages would rise relative to the final rate, and it would appear that births were being advanced

FERTILITY

AND FAMILY PLANNING IN THE U.S.

to younger ages. Actually, both types of change have taken place, and one has tended to balance the other. The medium, low, and high projections of changes in the timing of births to ever-married women, presented in Table 208, are similar to those of the 1955 study. In both cases the medium series shows a slight reversal of the trend toward relatively higher birth rates among young married women, the low series a greater reversal, and the high series a continuation of the upward trend. In

F I G U R E 44 Cumulative medium birth rates of ever-married women by specified ages, in p e r cent of final birth rate, by groups of cohorts

the current medium series the ratio of the cumulative birth rate by ages 15-19 and 20-24 to the final rate is assumed to decrease slowly from current or recent high marks, and approach the ratios of 25.4 and 47.8 for the cohorts of 1936-40 and similar ratios for the 1906-10 group.7 (Figure 44). The low series illustrates a return to a timing pattern much like that of the cohorts of 1891-95 and 1896-1900. The ratio of the * T h e ratios in question are computed from cumulative birth rates in cohort fertility tables for ever-married women. These rates are somewhat inflated at younger ages because the cumulative birth rate of all women of a given age is divided by the cumulative first marriage rate by that age, which means that the births to single women are credited to ever-married women.

386

FUTURE FAMILY SIZE AND POPULATION GROWTH TABLE 208 Projected or Actual Final Birth Rate of Ever-Married Women, and Rate by Younger Ages as a Percentage of That Rate, by Cohort"

Cohort

Rate by 45-49 A

Rate by younger ages as a percentage of rate by 45-49 20-24 15-•19 25-29 30-34 35-39 40-44 C B D G E E

1966-70 1961-65 1956-60 1951-55 1946-50 1941-45 1936-40 1931-35 1926-30 1921-25 1916-20

3,000 3,000 3,000 3,000 3,100 3,200 3,300 3,450 3,300 3,000 2,700

26 .0 26 .3 26.8 27 .3 27 .7 28 .2 25 .4 21 .1 20 .3 22.6 25 .9

1966-70 1961-65 1956-60 1951-55 1946-50 1941-45 1936-40 1931-35 1926-30 1921-25 1916-20

2,500 2,500 2,500 2,500 2,700 2,900 3,100 3,300 3,200 2,970 2,680

23 .0 24 .1 25 .7 27 .6 29 .2 31 .1 27 .0 22 .1 20 .9 22 .9 26..1

1966-70 1961-65 1956-60 1951-55 1946-50 1941-45 1936-40 1931-35 1926-30 1921-25 1916-20

3,500 3,500 3,500 3,500 3,500 3,500 3,500 3,600 3,400 3,030 2,720

29. 0 28.5 27. 9 27.,0 26. 2 25.,7 23. 9 20. 2 19..7 22. 4 25.,7

1911-15 1906-10 1901-05 1896-1900 1891-95 1886-90

2,481 2,458 2,625 2,917 3,237 3,518

27. 5 27. 4 23. 6 20. 9 18. 5

Medium projections 48.0 74.0 90.0 48 .2 74.6 90.6 48 .6 75.5 91.5 48 .9 76.2 92.2 49 .2 76.8 92.8 49 .5 77.3 93.3 47 .8 76.1 92.7 40 .5 69.3 89.1 35 .2 61.9 83.7 35 .4 58.5 81.1 39 .0 59.0 79.1 Low projections 42 .0 67.0 86.0 43 .4 69.0 87.5 45 .5 71.9 89.6 48 .2 75.0 91.5 77.8 50,.7 93.2 53 .4 80.7 95.2 79.4 50 .9 95.2 72.4 42 .3 92.1 36 .3 63.8 86.3 59.1 35 .8 82.0 39,.3 59.4 79.7 High projections 54.0 81.0 94.0 80.2 53. 0 93.7 79.1 51..7 93.4 77.4 49. 6 92.9 47.,7 75.8 92.4 46. 3 74.6 91.7 45. 1 73.1 90.6 38. 8 66.4 86.4 34.2 60.1 81.2 35. 0 58.0 80.3 38..8 58.6 78.5 Actual 44. 5 61.2 80.0 48. 5 68.3 82.5 48. 2 70.2 86.6 43. 6 66.8 84.2 40. 2 63.4 82.1 78.9

98 .0 98 .1 98 .2 98 .3 98 .4 98 .3 98 .2 97 .1 95 .3 94 .1 93 .3

99.8 99.8 99.8 99.8 99.8 99.8 99.8 99.7 99.5 99.3 99.2

97 .0 97,.3 97 .6 98 .0 98.4 98,.6 99,.0 98,.8 97 .5 95 .1 94 .0

99.8 99.8 99.8 99.8 99.8 99.8 99.8 99.8 99.8 99.8 100.0

99.,0 98.9 98. 8 98. 6 98. 4 98. 0 97. 4 95. 6 93. 2 93. 2 92. 6

99.9 99.9 99.9 99.9 99.9 99.9 99.9 99.6 99.3 98.8 98.5

92. 8 93. 6 95. 9 95. 3 94.0 92. 5

99.2 99.2 99.9 100.0 99.4 99.0

» The projected percents below the horizontal lines are based on actual rates by these ages and projected final rates. The high projected final rates exceed the low, hence the projected percentages below the horizontal lines are smaller in the high deck than the low.

FERTILITY AND FAMILY PLANNING IN THE U.S.

birth rate of ever-married women by ages 15-19 to the final rate is projected as falling from a recent high of 31.1 to 23.0, and the corresponding rates for ages 20-24 from a current peak of 53.4 to 42.0. It will be noted that even the low series does not approach closely the extreme pattern of late childbearing set by the cohorts of the 1920's, for whom the cumulative birth rate of ever-married women aged 20-24 was only 35 per cent of the final rate. The high series provides for relatively more births to the young married women of future cohorts than of any past cohorts for which we have data. The rates by ages 15-19 and 20-24 are set at 29 and 54 per cent, respectively, of the final rate, which exceed the high marks of 27.5 and 48.5 per cent set by the cohorts of 1911-15 and 1906-10 (Table 207, bottom deck). The numerical changes in the rates are much larger because these earlier cohorts had smaller families than are projected in the high series. In all three series, higher ratios are projected between the cumula­ tive rate by ages 35-39 and the final rate, reaching 97 in the low and 99 in the high compared with the past peak of 95.9 for the cohorts of 1901-05. This means very few births in the late 30's (or 40's) in the high series and a substantial reduction in the low. The three series are also alike in another respect, namely, the cumula­ tive birth rates by the middle childbearing ages (25-29 and 30-34) are now higher relative to the projected final rate than at any previous time. Little change is anticipated for the medium series in the future, a substantial rise for the high series, and a decline for the low. Even the latter, however, provides for relatively less childbearing after ages 30-34 for future cohorts than for those that have now passed the end of the childbearing period. PROJECTING THE MARRIAGE RATES OF SINGLE WOMEN

Fertility rates for ever-married women are of little value in pro­ jecting the population unless accompanied by first marriage rates. As in the 1955 study, the latter have been projected on the basis of the cohort marriage tables prepared by the Scripps Foundation. The chief difference between the present and previous series is a small increase in the low and medium first marriage rate by ages 45^49 for the cohorts of 1916-20 to 1941-45. This appears desirable in view of the fact that the first marriages during 1955-60 at ages under 25-29 exceeded slightly the high projections of the 1955 study. In the revised medium series, the cumulative rate by middle age crests at 965 per 1,000 for the women of 1921-25 (whose rate had reached 947 by ages 35-39), and declines to 940 for the cohorts

F U T U R E F A M I L Y S I Z E AND POPULATION G R O W T H TABLE 209 Projected or Actual Cumulative Marriage Rate of All Women by Ages 45-49, and Rate by Younger Ages as a Percentage of That Rate, by Cohorta

Cohort

Rate by 45-49 A

Rate by younger ages as a percentage of rate by 45-49 15-19 20-24 25-29 35-39 40-44 30-34 B C D F G E

1951-55 1946-50 1941-45 1936-40 1931-35 1926-30 1921-25

940 945 945 950 955 960 965

16.5 16.9 17.1 18.0 18.1 13.3 13.0

Medium projections 68 .1 90.4 96.3 69 .8 91.0 96.3 71 .4 92.1 96.8 73 .2 92.6 96.8 73..1 93.5 96.9 70 .6 92.0 96.8 62 .0 89.9 95.6

98 .4 98..4 98..4 98,.4 98..4 98,,4 98..1

99.5 99.5 99.5 99.5 99.5 99.5 99.5

1951-55 1946-50 1941—45 1936-40 1931-35 1926-30 1921-25

910 920 925 935 945 955 965

14.3 16.3 17.5 18.3 18.3 13.4 13.0

60..4 65..2 69..7 74..3 73.,9 71..0 62..0

Low projections 85.7 94.5 88.0 95.1 90.3 96.2 92.5 96.3 94.5 96.8 92.5 97.3 89.9 95.6

97,.8 97 .8 97,.8 97.,9 98..4 98 .4 98 .1

99.5 99.5 99.5 99.5 99.5 99.5 99.5

1951-55 1946-50 1941-45 1936-40 1931-35 1926-30 1921-25

970 970 965 965 965 965 965

18.6 17.5 16.8 17.7 17.9 13.3 13.0

75,.3 74 .2 73..1 72..0 72..3 70 .3 62..0

High projections 97.9 94.9 97.4 93.8 97.4 93.8 92.7 97.4 92.5 96.9 91.5 96.3 95.6 89.9

99 .0 99.0 99 .0 99 .0 98..4 98 .4 98 .1

99.5 99.5 99.5 99.5 99.5 99.5 99.5

1916-20 1911-15 1906-10 1901-05 1896-1900 1891-95 1886-90

950 933 924 922 916 915 911

tl.7 13.8 15.0 15.0 14.6 13.8

56..3 54,,3 58..3 60..1 59. 6 58. 1

98..1 98 .6 97..5 96 .4 97.,6 97. 9 97..7

99.2 99.5 99.5 98.7 98.8 99.2 99.3

Actual 85.8 83.1 81.6 84.5 85.9 84.5

95.7 93.7 92.7 92.4 95.0 94.4 94.0

a The per cents below the horizontal lines are based on actual rates by these ages, and projected final rates. The projections for later cohorts are the same as those for 1951-55.

of 1951-55 and later, as in the first study (Table 209, Column A). The other difference is a narrowing of the spread between the low and high projections. The decline of the low is now assumed to stop at 910 instead of 900—to return to the situation of the

FERTILITY AND FAMILY PLANNING IN THE U.S.

cohorts of 1886-90 instead of those of 1876-85. The revised high series levels at 970 instead of 980, in both cases with the cohorts of 1951-55. This may be more reasonable in view of the slowing down of the increase in the cumulative rate by ages 30-34 which is shown by comparisons between 1960 and earlier years. Allowing for changes in age at marriage is done in the same way as for changes in the timing of births. With the annual first marriage rate, as with the first birth rate, there is a direct short-run relationship to changes in economic conditions, but with a shorter lag. Since firm projections of economic conditions are not available for years ahead, low and high allowances are made for changes in age at marriage, which probably will bracket actual developments but do not differ so greatly as to appear unreasonable. When the fertility projections based on the 1955 study were pre­ pared (in early 1958) it appeared that the trend toward younger marriage was being reversed, and that the proportion of women marrying in their teens and early 20's would decline. We know now that this did happen, but on a smaller scale than was anticipated. As a consequence, the present marriage projections start from a slightly higher base than the previous series but are very like the first with respect to timing. As before, the medium series assumes a small rise in age at marriage. For example, the ratio of the marriage rate by ages 15-19 to the final rate has decreased slightly from the peak of about 18 per cent for the cohorts of 1931-35 or 1936-40; in the medium series it continues to decline to about 16.5 per cent for the cohorts of 1951-55 and later (Table 209, Column B).8 Similarly, the proportion of marriages occurring between ages 15-19 and 20-24 has dropped from a high of about 57 per cent for the 1926-30 cohorts, and is assumed to stabilize at about 52 per cent for the 1 9 5 1 - 5 5 cohorts (compare C o l u m n s B a n d C ) . A greater tendency to postpone marriage is assumed in the low series, the proportions marrying by ages 15-19 and 20-24 falling to 14 and 60 per cent (respectively) for the cohorts of 1951-55. In contrast, the high projections envisage a renewal of the postwar trend to younger marriage which raises the percentages marrying by the two younger age groups to about 19 and 75. PROJECTED BIRTH RATES FOR ALL WOMEN

Numbers of births in future years could be computed by sub­ dividing the surviving women of each age into those single and evermarried on the basis of the projected first marriage rates, and multi8For convenience the cumulative first marriage rate by ages 45-49 will be referred to as the final rate. Very few older women marry for the first time, and as a rule such marriages do not affect birth rates.

FUTURE FAMILY SIZE AND POPULATION GROWTH

plying those ever married by the projected birth rates of such women. It is preferable, however, to multiply the cumulative first marriage rate of women of a given age by the cumulative birth rate of those who have married, which gives the cumulative birth rate of all women. The increase of this rate during a given age and time period is then multiplied by the appropriate number of women in order to obtain the number of births that they had. In obtaining birth rates for all women, similar series for marriages and for births to ever-married women have been combined. This means that the low projections (for example) reflect the combined effect of four types of change which would reduce annual numbers of births and the crude birth rate during the 1960's and 1970's, namely, more women remaining single, a rise in age at marriage, smaller completed families, and a tendency for children to come later in marriage. Similarly, the high series combines four changes in the opposite direction. It is very likely that changes in social, economic, and other conditions will bring at least small changes of each type, but quite unlikely that all will reduce, or increase, numbers of births in the near future.9 Nevertheless, it seems desirable to combine them in this way because it helps to widen rapidly the spread between the low and high numbers of births and crude birth rates in the projected population. One of the shortcomings of past population projections is that those reflecting low and high fertility either did not diverge quickly enough to bracket population growth, or the procedure followed in order to obtain rapid divergence involved an unrealistic assumption about changes in size of completed families, the timing of marriages and births, or both.10 None of the assumptions in the present low and high series seems unreasonable on the basis of the performance of cohorts in recent decades, but when combined they set lower and upper limits that are not likely to be passed. Let us examine briefly the cumulative and age-specific birth rates for all women. In the medium series the final rate rises to a high mark of nearly 3,300 for the cohorts of 1931-35 and then declines to 2,820 for those of 1951-55 and later (Table 210 and Figure * In this connection, it should be remembered that a rise in age at marriage, or increases in the length of intervals from marriage to first, second, and other births, decrease numbers of births when they begin. Later, when the new pattern becomes established, the numbers of births return to where they would have been without these changes, unless in the meantime there have been changes in other factors influencing fertility. M It has been obvious for several years that the fertility projections shown in Whelpton 1947 failed to allow adequately for births postponed during the war, or to be advanced by the postwar pattern of younger marriage and childbearing, to say nothing of the increase in family size. It is also becoming clear that Series I in Census 1958, p. 11, were based on an unrealistically high completed fertility rate—4,004 births per 1,000 women living to ages 45-49.

1985 1980 1975 1970 1965 1960 1955 1950 1945 1940 1935 1930 1925 1920 1915 1910 1905

1985 1980 1975

1966-70 1961-65 1956-60 1951-55 1946-50 1941-45 1936-40 1931-35 1926-30 1921-25 1916-20 1911-15 1906-10 1901-05 1896-1900 1891-95 1886-90 1881-85

1966-70 1961-65 1956-60

Cohort

Year reach 15-19

75 78 84

121 122 125 127 137 146 143 126 86 85 78 88 94 85 84 83

Cumu­ lative rate, 15-19

503 519 542

801 803 808 812 869 924 954 849 702 550 486 471 548 615 610 635

15-19 to 20-24

578 597 626

922 925 933 939 1,006 1,070 1,097 975 788 635 564 559 642 700 694 718

Cumu­ lative rate, 20-24

728 749 776

965 977 992 1,004 1,042 1,083 1,112 1,159 1,016 889 734 618 625 j 735 839 869

20-24 to 25-29

Low projections 543 1,849 535 1,881 524 1,926

309 283 246

276 262 241 225 219 194 214 305 407 427 406 385 335 301 348 431 521

Medium projections 1,887 557 2,444 1,902 558 2,460 1,925 559 2,484 1,943 2,503 560 2,048 570 2,618 2,153 578 2,731 2,209 606 2,815 2,134 710 2,844 1,804 761 2,565 1,524 723 2,247 1,298 643 1,941 1,177 557 1,734 1,267 471 1,738 1,435 502 1,937 1,533 603 2,136 1,587 709 2,296 2,376

1,306 1,346 1,402

30-34 to 35-39

25-29 to 30-34

Cumu­ lative rate, 30-34

Cumu­ lative rate, 25-29

2,158 2,164 2,172

2,720 2,722 2,725 2,728 2,837 2,925 3,029 3,149 2,972 2,674 2,347 2,119 2,073 2,238 2,484 2,727 2,897 3,016

Cumu­ lative rate, 35-39

100 94 86

79 77 74 71 71 78 85 119 165 187 177 166 168 149 158 194 256 301

35-39 to 40-44

2,258 2,258 2,258

2,799 2,799 2,799 2,799 2,908 3,003 3,114 3,268 3,137 2,861 2,524 2,285 2,241 2,387 2,642 2,921 3,153 3,317

Cumu­ lative rate, 40-44

16 16 16

21 21 21 21 22 21 21 27 31 34 41 30 30 33 30 41 52 67

40-44 to 45-49

TABLE 210 Projected or Actual Cumulative Birth Rates by Specified Ages, and Rates during Age Periods, for All Women, by Cohorf

2,275 2,275 2,275

2,820 2,820 2,820 2,820 2,930 3,024 3,135 3,295 3,168 2,895 2,565 2,315 2,271 2,420 2,672 2,962 3,205 3,384

Cumu­ lative rate, 45-49

1985 1980 1975 1970 1965 1960 1955 1950 1945 1940 1935

1951-55 1946-50 1941-45 1936-40 1931-35 1926-30 1921-25 1916-20

1966-70 1961-65 1956-60 1951-55 1946-50 1941-45 1936-40 1931-35 1926-30 1921-25 1916-20

183 180 176 170 156 146 143 126 86 85 78

90 118 146 143 126 86 85 78

Cumu­ lative rate, 15-19

1,197 1,175 1,145 1,097 1,046 996 954 849 702 550 486

573 703 854 954 849 702 550 486

15-19 to 20-24

1,380 1,355 1,321 1,267 1,202 1,142 1,097 975 788 635 564

663 821 1,000 1,097 975 788 635 564

Cumu­ lative rate, 20-24

1,228 1,228 1,226 1,225 1,212 1,220 1,194 1,159 1,016 889 734

799 881 954 1,031 1,159 1,016 889 734

20-24 to 25-29 25-29 to 30-34

2,608 2,583 2,547 2,492 2,414 2,362 2,291 2,134 1,804 1,524 1,298

30-34 to 35-39

212 189 132 154 250 368 427 406

200 208 214 225 250 259 277 360 447 427 406

High projections 518 3,126 533 3,116 559 3,106 596 3,088 642 3,056 655 3,017 689 2,980 774 2,908 761 2,565 723 2,247 643 1,941

(Continued)

Cumu­ lative rate, 30-34

1,968 2,202 2,456 2,655 2,782 2,565 2,247 1,941

Low projections 1,462 506 1,702 500 1,954 502 2,128 527 2,134 648 1,804 761 1,524 723 1,298 643

Cumu­ lative rate, 25-29

Rates above the horizontal lines are projected. Other rates are from Whelpton 1960.

1970 1965 1960 1955 1950 1945 1940 1935

Cohort

a

Year reach 15-19

TABLE 210 (Continued)

3,326 3,324 3,320 3,313 3,306 3,276 3,257 3,268 3,012 2,674 2,347

2,180 2,391 2,588 2,809 3,032 2,933 2,674 2,347

Cumu­ lative rate, 35-39

49 51 55 62 69 79 98 174 228 201 177

78 75 75 69 65 102 172 177

35-39 to 40-44

3,375 3,375 3,375 3,375 3,375 3,355 3,355 3,442 3,240 2,875 2,524

2,258 2,466 2,663 2,878 3,097 3,035 2,846 2,524

Cumu­ lative rate, 40-44

20 20 20 20 20 23 23 32 41 49 60

16 18 19 20 21 21 20 22

40-44 to 45-49

3,395 3,395 3,395 3,395 3,395 3,378 3,378 3,474 3,281 2,924 2,584

2,275 2,484 2,682 2,898 3,118 3,056 2,866 2,546

Cumu­ lative rate, 45-49

FERTILITY

AND FAMILY

PLANNING

IN THE

U.S.

45). The latter is almost exactly the average for all cohorts from 1876 to 1920. The low and high series also crest with the women born in 1931-35, but the subsequent decline is considerably larger in the low series (to 2,275) than in high (to 3,395). The low would mean a return to the small families in the 1906-15 cohorts. In each of the series, the final rates for the cohorts of 1926-30 to 1946-50 are well above the corresponding projections based on the first GAF study, primarily because the second showed that white wives who were in their late 20's and early 30's in 1960 wanted

FIGURE 45 Actual and medium projected cumulative birth rates from 15-19 to 45-49, for all women, by groups of cohorts

and expected somewhat larger families than similar wives five years younger in 1955. Unlike the first series, however, the present medium and high decline from the peak that probably will be reached by the cohorts of 1931-35, reflecting the smaller families expected by the wives born in 1936-40. The timing patterns projected for all wives in the current series are very similar to those in the first. In both cases the medium allows for a small shift of childbearing from younger to older ages, the low for a larger shift, and the high for a continuation of the recent tendency toward completing families at younger ages. 394

FUTURE FAMILY SIZE AND POPULATION GROWTH

It is unlikely, of course, that future trends of birth rates from cohort to cohort will be as smooth as those of the medium, low or high projections. Much more probable are fluctuations such as those that occurred in the past decades for which we have a complete (or almost complete) record. How Do CHANGES IN COHORT BIRTH RATES AFFECT PERIOD BIRTH RATES?

So far we have been looking at fertility projections on a cohort basis, because it allows the use of wives' expectations regarding size of completed family, and the past childbearing experience of actual groups of women as they lived from menarche to menopause. In projecting the population from 1960 to 1985, however, it is the agespecific birth rates of this period that are employed. It is desirable, therefore, to consider what the birth rates projected on a cohort basis signify on a period basis. Gross reproduction rates will be helpful in this connection. Birth rates at age groups up to 30-34, and the gross reproduction rate, were substantially higher in 1955-60 than in any period since 1915-20, and probably higher than in 1905-10 or 1910-15. The decline of the crude birth rate and the gross reproduction rate that began early in the nineteenth century and was rapid for several decades, hit bottom in the 1930's, and was followed by a rapid rise. Less is known about trends of age-specific birth rates before 1930, but quite likely all were downward. In contrast, from 1930-35 to 1955-60, the birth rate from ages 15-19 to 20-24 more than doubled, and gains were almost as large at the next younger and older groups (Table 211, bottom deck, Columns A to C). At ages 25-29 and over, however, the downward trend lasted somewhat longer, and the recovery has been slower. The gross reproduction rate, which summarizes all the age-specific rates, climbed from a low of 1.05 in 1935-40 to 1.78 in 1955-60 (Column H).11 In view of this recent rise, it may be surprising to note that in the medium projections the gross reproduction rate is slightly smaller for 1960-65 than 1955-60 and declines to less than 1.40 by the late 1970's. Even in the high series it crests at 1.87 in 1960-65, and then declines gradually to 1.75 or less. In the low series the fall is much steeper, and the figure of about 1.05 in the late 1970's uThe gross reproduction rates presented here are slightly smaller than those in various reports of the NVSD or the April issues of Population Index. The reason is that all three sets of rates are based on births adjusted for incomplete registra­ tion, but only the Scripps rates are based on numbers of women increased to allow for underenumeration.

FERTILITY AND FAMILY PLANNING IN THE U.S.

and early 1980's is the same as the previous low mark established in 1935^40. The projected decreases in the final birth rate of all women from the 1931-35 cohorts to the 1951-55 group—amounting to about 15 per cent in the medium series and 27 per cent in the low12—would be expected to bring similar reductions in period rates. Instead, the TABLE 211 Actual or Projected Age-Specific Birth Rates of All Women, and Gross Reproduction Rate, by Five-Year Periods Series and period"

10-14 to 15-19 A

15-19 to 20-24 B

Birth rates at agesb 20-24 25-29 30-34 to to to 25-29 30-34 35-39 E C D

35-39 40-44 Gross re­ to to production 40-44 45-49 rate F G H

Medium 1980-85 1975-80 1970-75 1965-70 1960-65

121 122 125 127 137

803 808 812 869 924

992 1,004 1,042 1,083 1,112

560 570 578 606 710

219 194 214 305 407

78 85 119 165 187

21 27 31 34 41

1.36 1.37 1.42 1.56 1.72

Low 1980-85 1975-80 1970-75 1965-70 1960-65

75 78 84 90 118

519 542 573 703 854

776 799 881 954 1,031

506 500 502 527 648

189 132 154 250 368

75 69 65 102 172

20 21 21 20 22

1.05 1.04 1.11 1.29 1.57

High 1980-85 1975-80 1970-75 1965-70 1960-65

183 180 176 170 156

1,175 1,145 1,097 1,046 996

1,226 1,225 1,212 1,220 1,194

596 642 655 689 774

250 259 277 360 447

79 98 174 228 201

23 32 41 49 60

1.72 1.75 1.77 1.84 1.87

Actual 1955-60 1950-55 1945-50 1940-45 1935-40 1930-35 1925-30 1920-25

146 143 126 86 85 78 88 94

954 849 702 550 486 471 548 615

1,159 1,016 889 734 618 625 735 839

761 723 643 557 471 502 603 709

427 406 385 335 301 348 431 521

177 166 168 149 158 194 256 301

30 30 33 30 41 52 67 85

1.78 1.62 1.43 1.19 1.05 1.10 1.33 1.54

a The b The

periods begin on July 1 and end on June 30. rates in Table 210 are arranged here on a period basis.

gross reproduction rate falls more rapidly—24 per cent in the medi­ um series and 41 per cent in the low—from 1955-60 to 1980-85 (Table 210, Column H). The difference results chiefly from the changes in age at marriage and in child spacing that have already taken place or are projected for the future. "The final rate declines from 3,295 to 2,820 in the medium series, and from 3,118 to 2,275 in the low (Table 210).

FUTURE FAMILY SIZE AND POPULATION GROWTH

PROJECTIONS OF BIRTHS AND POPULATION 1960 ΤΟ 1985

Once the projections of birth rates have been accepted, the next step is to compute the numbers of births that they will yield. The procedure is simple. Starting with the population by sex and age on 7-1-60 estimated by the Bureau of the Census, the number of survivors on 7-1-65 is computed by applying survival rates and adding an allowance for net immigration during the period.13 The number of births is computed for each cohort group by multiplying the average TABLE 212 Number of Births and Crude Birth Rates: Actual for 1910-1915 to 1955-1960, and Projected for 1960-1965 to 1980-1985 Numbers of Urths b

Crude birth rates

Period»

Medium A

Low B

High C

Medium D

Low E

High F

1980-85 1975-80 1970-75 1965-70 1960-65

27,781 25,927 23,877 22,298 21,446 Actual 21,265 19,637 17,697 14,447 12,180 12,125 13,648 14,747 14,575 14,355

20,712 19,589 18,609 18,442 19,573

36,764 33,391 29,783 26,386 23,349

22.3 22.3 22.1 22.1 22.9 Actual 24.7 25.0 24.5 21.3 18.8 19.4 22.9 26.6 28.1 29.8

18.1 17.9 17.9 18.6 21.0

27.0 27.0 26.5 25.7 24.8

1955-60 1950-55 1945-50 1940-45 1935-40 1930-35 1925-30 1920-25 1915-20 1910-15 a

The periods are from July 1 to June 30. Numbers of births are in thousands, and include an allowance for births not registered. b

number of women in group on 7-1-60 and 7-1-65 by the corre­ sponding period birth rate from Table 211.14 Adding the births, subdividing by sex, applying survival rates, and allowing for immigra­ tion gives the number of children under 5 on 7-1-65. Corresponding steps give the births and population on successive five-year dates. The medium projections show slightly more births in 1960-65 than in 1955-60, somewhat larger increases in the next three periods, and then a smaller increase. The projected number for 1980-85 is 27,800,000, about 31 per cent more than in 1955-60 (Table 212). uSee mortality and migration assumptions, footnotes to Table 213. "For example, the average of the number of women aged 20-24 on 7-1-60 and 25-29 on 7-1-65 (cohorts of 1936-40) is multiplied by the rate for these cohorts from ages 20-24 to 25-29.

FERTILITY

AND FAMILY

PLANNING

IN THE

U.S.

TABLE 213 Projected Population of the U.S., by Age, July 1, 1960 to 1985, with Medium, Low, and High Fertility Trends (in millions)8 Total Age

1960

1965

1970

1975

1980

1985

Medium 0-4 5-9 10-14 15-19 20-24

20. 4 18. 8 16.,9 13 .5 11 .1

20. 9 20.4 18. 9 17 .0 13..6

21. 8 21..0 20. 5 18 .9 17. 1

23. 4 21. 8 21. 0 20 .5 19. 0

25,.4 23,.4 21 .9 21 .1 20. 6

27.2 25.4 23.4 21.9 21.2

25-29 30-34 35-39 40-44 45-49

10 .9 12 .0 12 .5 11 .7 10 .9

11..3 11,.0 12..0 12 .4 11 .5

13..8 11.,4 11. 1 11,.9 12,.2

17..2 13. 9 11..4 11,.0 11,.7

19,.2 17,.3 13,.9 11 .4 10 .8

20.7 19.2 17.3 13.8 11.2

50-54 55-59 60-64 65-69 70-74

9 .7 8 .5 7 .2 6 .3 4.8

10..6 9 .2 7 .8 6 .3 5 .2

11,,1 10,,0 8..4 6 .9 5 .2

11. 8 10,.6 9,.3 7,.5 5,.7

11..4 11..3 9 .8 8 .2 6 .2

10.5 10.8 10.4 8.7 6.9

75-79 80-84 85 +

3 .1 1 .6 .9

3 .6 1 .9 1 .1

3 .9 2 .3 1 .4

4 .0 2 .5 1 .6

4 .4 2 .6 1 .8

4.8 2.8 1.9

Total

180 .7

194 .7

208 .8

223 .9

240 .5

258.3

18,.2 18 .1 19 .2 20 .5 19 .0

19 .2 18 .3 18 .1 19 .3 20 .6

20.3 19.3 18.3 18.2 19.4

213 .2

223 .6

234.5

Low 0-4 5-9 10-14 15-19 20-24

20 .4 18 .8 16 .9 13 .5 11 .1

19 .1 20 .4 18 .9 17 .0 13 .6

18,.0 19 .2 20 .5 18 .9 17 .1

Total

180 .7

192 .8

203 .3 High

0-4 5-9 10-14 15-19 20-24

20 .4 18 .8 16 .9 13 .5 11 .1

22 .8 20 .4 18..9 17 .0 13 .6

25 .8 22 .8 20,.5 18,.9 17 .1

29 .1 25 .8 22,.9 20,.5 19 .0

32 .7 29 .1 25, 8 22,.9 20 .6

36.0 32.7 29.2 25.9 23.0

Total

180 .7

196 .5

214,.7

235..5

259..3

285.8

398

FUTURE FAMILY SIZE AND POPULATION GROWTH TABLE 213 (continued) Male Age

1960

1965

1970

1975

1980

1985

0-4 5-9 10-14 15-19 20-24

10.4 9.6 8.6 6.8 5.6

10.7 10.4 9.6 8.6 6.8

Medium 11.1 11.9 10.7 11.1 10.4 10.7 9.6 10.4 8.6 9.6

13.0 11.9 11.2 10.7 10.4

13.9 13.0 12.0 11.2 10.7

25-29 30-34 35-39 40-44 45-49

5.4 5.9 6.1 5.7 5.4

5.6 5.5 5.9 6.1 5.6

6.9 5.6 5.5 5.8 5.9

8.6 6.9 5.6 5.4 5.7

9.6 8.7 6.9 5.6 5.3

10.4 9.6 8.6 6.8 5.5

50-54 55-59 60-64 65-69 70-74

4.8 4.1 3.4 2.9 2.2

5.1 4.4 3.7 2.9 2.3

5.4 4.8 4.0 3.1 2.3

5.7 5.0 4.3 3.4 2.5

5.5 5.3 4.5 3.6 2.7

5.1 5.1 4.8 3.8 2.9

75-79 80-84 85 +

1.4 .7 .4

1.5 .8 .4

1.6 .9 .5

1.6 1.0 .6

1.8 1.0 .6

1.9 1.1 .7

Total

89.3

95.9

102.7

110.0

118.1

126.9

9.3 9.2 9.8 10.4 9.6

9.8 9.3 9.3 9.8 10.4

10.4 9.8 9.4 9.3 9.8

104.5

109.5

114.8

Low 0-4 5-9 10-14 15-19 20-24

10.4 9.6 8.6 6.8 5.6

9.8 10.4 9.6 8.6 6.8

9.2 9.8 10.4 9.6 8.6

Total

89.3

95.0

99.8 High

0-4 5-9 10-14 15-19 20-24

10.4 9.6 8.6 6.8 5.6

11.6 10.4 9.6 8.6 6.8

13.2 11.6 10.4 9.6 8.6

14.9 13.2 11.7 10.4 9.6

16.7 14.9 13.2 11.7 10.4

18.4 16.7 14.9 13.2 11.6

Total

89.3

96.9

105.7

115.9

127.7

141.0

TABLE 213 (continued) Female 1980

1985

12 .4 11 .5 10..7 10 .3 10 .2

13.3 12.4 11.5 10.8 10.5

8 .6 7 .0 5 .8 5 .6 6.0

9 .6 8 .7 7 .0 5 .8 5 .5

10.3 9.6 8.7 7.0 5.7

5 .8 5 .2 4 .5 3 .7 3 .0

6 .2 5 .6 5 .0 4 .1 3 .3

5 .9 6 .0 5 .3 4 .6 3 .6

5.4 5.7 5.7 4.9 4.0

2.0 1.1 .7

2 .3 1.,4 .8

2 .4 1 .5 1 .0

2 .6 1 .6 1..1

2.9 1.8 1.2

98.7

106..1

113..9

122,,4

131.4

8..9 8 .9 9..4 10.,1 9..4

9,.4 8 .9 8,.9 9..5 10..2

9.9 9.4 9.0 8.9 9.6

108.,7

114.1

119.7

Age

1960

1965

1970

0-4 5-9 10-14 15-19 20-24

10.0 9.3 8.3 6.7 5.6

10.3 10.0 9.3 8.4 6.8

Medium 10 .7 11 .4 10 .3 10 .7 10 .1 10 .3 9 .3 10 .1 9 .4 8 .5

25-29 30-34 35-39 40-44 45-49

5.5 6.1 6.4 5.9 5.5

5.7 5.6 6.1 6.4 5.9

6 .9 5 .8 5 .6 6 .1 6.3

50-54 55-59 60-64 65-69 70-74

4.9 4.3 3.7 3.3 2.6

5.4 4.7 4.1 3.4 2.9

75-79 80-84 85 +

1.7 .9 .6

Total

91.3

1975

Low 0-4 5-9 10-14 15-19 20-24

10.0 9.3 8.3 6.7 5.6

9.4 10.0 9.3 8.4 6.8

8 8 9..4 10. 1 9. 3 8..5

Total

91.3

97.8

103. 4 High

0-4 5-9 10-14 15-19 20-24

10.0 9.3 8.3 6.7 5.6

11.2 10.0 9.3 8.4 6.8

12. 6 11. 2 10.1 9. 3 8. 5

14. 3 12. 6 11. 2 10. 1 9..4

16. 0 14. 3 12.7 11. 2 10. 2

17.6 16.0 14.3 12.7 11.4

Total

91.3

99.6

109. 0

119. 6

131. 6

144.8

* Based on the 1960 Census, not adjusted for underenumeration. The population below the horizontal lines was living in 1960 and is not affected by projected fertility

trends. Each figure is rounded independently. The fertility assumptions are described in the text. The mortality assumptions are those used by the Bureau of the Census. Survival rates for 1960 were computed from Abridged Life Tables for 1960 published in NOVS1960, Vol. II—Section 2, and for 2,000 from the High mortality assumptions in Greville 1957, pp. 19,21. Rates for intervening five-year periods were obtained by straight line interpola­ tion, except for a few ages at which the mortality projected for 2,000 exceeded that of 1960; in these cases the 1960 rates were continued unchanged. Net immigration is assumed to be 300,000 per year, distributed by age and sex like the immigrant aliens of July 1, 1957 to June 30, 1962. This assumption is the same as that used by the Bureau of the Census.

FUTURE FAMILY SIZE AND POPULATION GROWTH

In contrast, the crude birth rate has a downward trend from 24.7 per 1,000 in 1955-60 to 22.3 in 1980-85, with most of the decline occurring during the first interval. If actual fertility were to follow the low projections, the number of births would decrease about 13 per cent from 1955-60 to 1965-70 and then increase at a slower pace. The crude birth rate would fall more rapidly and longer; the low projection for 1980-85 is 18.1, somewhat under the previous low of 18.8 in 1935-40. At the opposite extreme, the high series shows the number of births rising rapidly from 1955-60 to 1980-85, with nearly 37,000,000 births in 1980-85, or 73 per cent more than in 1955-60. But even this would not match the pace of population growth; the crude birth rate would have a gradual upward trend to 27.0 in 1975-80 and 1980-85. The medium fertility projections call for the population to increase from 180,700,000 on 7-1-60 to 258,300,000 in 1985 (Table 213). The absolute gain of nearly 78,000,000 in the next 25 years would be half again as large as that in the past 25 years, but the relative gain (43 per cent) would be only slightly higher. If fertility should follow the low projections, the 1985 population would be 234,500,000 and the increase over 1960 would be only a little larger in absolute numbers than that from 1935 to 1960. At the other extreme, the high fertility series raises the population to 285,800,000 in 1985, and makes the gain over 1960 about twice as large as that of the preceding 25 years. The population projections shown here have also been published in a Census Bureau report (Siegel 1964), where they are compared to the Bureau's own projections. SUMMARY

This chapter shows how sample data on birth expectations may be used to project fertility rates to future years. The major difference between the method presented here and the method developed for the 1955 study is the introduction of an allowance for birth expecta­ tions to rise as wives grow older. Whether or not this allowance will be justified by events remains to be seen, but it is consistent with the changes in expectations that occurred between 1955 and 1960. It represents the kind of experimentation that must be done in the process of developing a fully articulated method of using birth expectations to forecast fertility.

APPENDIX SELECTED BIBLIOGRAPHY INDEX

APPENDIX A

Sampling DEFINITION OF SAMPLE

As noted in Chapter 1, the 1960 study required samples of four dis­ tinct population groups. They are defined as follows: Group A. White women 18 to 39 years old, currently married and living with husband, or husband temporarily absent in armed forces. This group is defined in the same way that the group of wives included in the 1955 study was defined. Group B. White women 40 to 44 years old, currently married and living with husband, or husband temporarily absent in armed forces. This group was needed to represent 1960 survivors of the 35-39-year-old wives represented in the 1955 study. Group C. Nonwhite women 18 to 39 years old, married and living with husband, or husband temporarily absent in armed forces. Group D. White women 23 to 44 years old, not living with husband at the time of the 1960 interview, but married and living with husband five years earlier. This group was needed to represent 1960 survivors of the 18-39-year-old wives rep­ resented in the 1955 study who had since become widowed, divorced, or separated. METHOD OF SAMPLING

The sample was selected by the Survey Research Center of the University of Michigan, under the direction of Leslie Kish, using the method known as "area probability sampling." The procedure involves the following steps:1 1. Selecting a sample of "primary sampling units," which consist of counties or groups of counties. These are chosen so that they are representative of various areas of the United States. 2. Within each primary sampling unit, selecting a probability sample of places or areas (towns, communities, or segments). 3. Within each place or area, selecting a probability sample of blocks (in cities) or areas (in the open country). In the larger cities (with populations of 50,000 or more), the blocks were stratified on the basis of average rental and property values so that those chosen would be representative of the cities. 4. Within each block and geographic area, listing all the addresses 1This method was also used to choose the 1955 sample. It is described in more detail in FPSPO, pp. 444-449.

FERTILITY AND FAMILY PLANNING IN THE U.S.

(or giving dwelling units in open country numerical designations) and choosing a random sample of them. This procedure results in a certain amount of "clustering" of dwelling units. That is, the dwelling units chosen are not randomly distributed throughout the United States, but tend to be clustered together in small groups of three or four. This has some effect on sampling errors, which will be described later. Altogether, 8,426 occupied dwelling units were chosen by this method. Interviewers called at each of the addresses and administered a short screening interview to find out if an eligible respondent lived there. If there was an eligible respondent present, she was asked for an interview. TABLE A-I Response Rates for Women Eligible for the 1960 Study

Sample group A. B. C. D.

Per cent distribution Number Not interviewed Inter­ of viewed Not eligible (response at respondents Total rate) Total Refused Home Other

All groups White wives 18-39 White wives 40—44 Nonwhite wives 18-39 White women 23-44, no longer married

3,782 2,743 662 297

100 100 100 100

88 88 86 91

12 12 14 9

6 6 7 3

6 6 6 5

1 1 1 1

80

100

82

18

4

9

5

Often the interviewer had to make several calls to obtain an interview. In fact, fewer than half of the interviews finally obtained were taken on the first call—only 43 per cent. The majority of the remaining interviews were obtained on the second or third call, but in a few cases the interviewer had to visit the house seven or more times to interview an eligible respondent. RESPONSE RATES

The response rate is the proportion of eligible respondents who granted an interview. Only under very unusual circumstances is the response rate as high as 100 per cent. A few eligible respondents are never found at home, regardless of how often the interviewer calls back; some are sick, some are mentally incompetent, a few cannot speak English or a language for which an interpreter is readily available, and some simply refuse to be interviewed. The response rate for the present study is 88 per cent, which is about average for national surveys taken by the Survey Research Center. Table A-I shows how the response rate varies for the different

APPENDIX A

groups sampled. It is highest (91 per cent) for the nonwhite wives, which is an excellent result in view of the fact that the census undercount is generally greater for nonwhites than whites. It is lowest (82 per cent) for the white wives who were no longer married. The rates for white married women are close to the average (88 per cent for those 18-39 years old, and 86 per cent for those 40-44). Table A-I also shows that the main reasons for non-response were the refusal to be interviewed and the failure to find the re­ spondent at home, even after repeated calls. COMPARISON OF THE SAMPLE WITH INDEPENDENT ESTIMATES

How well do the various samples represent all of the women in the population groups from which they were selected? We cannot answer this question with precision, but we can make rough compari­ sons to see whether the sample and the total population have approxi­ mately the same percentage distributions on certain important char­ acteristics for which we can make independent estimates. These com­ parisons are shown in Table A-2 for white wives 18-39 and nonwhite wives 18-39, the two most important components of our sample. The footnotes to this table show the sources of the independent esti­ mates (the 1960 Census for all characteristics except religion) and the population groups to which they relate. Unfortunately, it was not possible to obtain census data for currendy married women with husband present or with husband tem­ porarily absent in the armed forces. However, we tried to come as close as possible to the sample definition. Consequently, the inde­ pendent estimates relate to various groups: ever-married women, cur­ rently married women with husband present, currently married women with husband present or husband temporarily absent, and, in the case of religion, all women 14 years of age and over. In spite of the deficiencies of this procedure, the comparisons are capable of bringing to our attention gross differences between sample groups and the women they are supposed to represent. For convenience, we will refer to the population represented by the sample as the "universe." First, let us see how representative the sample is with respect to a very important variable—past fertility. The top panel of Table A-2 shows good agreement between sample and universe distributions of children ever born—both for white and nonwhite wives. However, there is a slightly lower proportion of white women with no children in the sample (12 per cent) than in the universe (15 per cent). Actually, these estimates are rounded from 12.5 and 14.5, respectively,

FERTILITY AND FAMILY PLANNING IN THE U.S. TABLE A-2 Per Cent Distribution by Selected Characteristics, for 1960 Sample of White and Nonwhite Wives 18-39 Years Old, and Independent Estimates for the Same Population Groups

Characteristic Total Number of births 0 1 2 3 4 5 or more Wife's age 18-19 20-24 25-29 30-34 35-39 Wife's religion Protestant Catholic Other Wife's education College High school, 4 High school, 1-3 Grade school Region of residence Northeast North Central South West

White wives 18—39 Sample Independent estimate® (2,414)

Nonwhite wives 18-39 Sample Independent (270) estimate"

100

100

100

100

12 19 28 21 11 9

15 20 28 19 10 8

16 19 21 13 11 19

17 19 19 14 10 21

3 18 25 26 28

4 19 23 27 28

4 19 24 29 24

4 20 24 26 25

66 28 6

65 28 7

90 7 3

89 7 4

18 48 24 11

17 45 23 15

14 27 32 27

10 27 30 33

24 30 31 15

24 30 29 17

15 27 50 7

17 19 51 13

a Sources

of independent estimates: Number of births: 1960 Census data for ever-married women. Wife's age: 1960 Census data for currently married women, husband present or "husband absent, other" (in order to approximate the inclusion of wives with husbands in armed forces). Wife's religion: 1957 data from Current Population Survey (Census Series P-20, No. 79) relating to the female population 14 years old and over. Wife's education: 1960 Census data for women married, husband present Region of residence: 1960 Census data for women married, husband present or "husband absent, other."

so that the difference is only 2.0 percentage points. This is just about large enough to be statistically significant at the .05 level. This difference is probably due in part to the fact that the census distribution is based on ever-married women, because such women have a slightly higher proportion with no previous births than do cur­ rently married women. The 1960 Census shows that ever-married

APPENDIX A

white women 20-39 years old had 13.4 per cent with no children and that married women with husband present had 12.8 per cent, about half a percentage point less. Such a difference could explain part of the 2.0 spread between sample and universe found in 1960, and the rest could be due to sampling variability. In any case, the divergence of sample from universe is a very minor one. Sample and universe show excellent agreement with respect to wife's age (second panel, Table A-2). None of the differences is significant. Table A-2 also shows that the sample and universe agree re­ markably well on wife's religion, one of the most important inde­ pendent variables used in this study. The definitions of the sample and universe populations are quite different, however (the universe estimate is for all women 14 years old and over in 1957), so that we cannot be sure that there is equally good agreement between sample and universe for women who meet the sample definition. But we can be reasonably certain that the sample does not grossly misrepresent religious preference. Another important independent variable—wife's education—is less well represented by the sample (fourth panel, Table A-2). The sample has a smaller proportion of wives with a grade-school education than the universe. The comparison for whites is 11 per cent for the sample and 15 per cent for the universe; for nonwhites, the percentages are 27 and 33, respectively. Although these are not large differences, they are statistically significant at the .05 level. We can only guess at their causes: response error, differences in the marital status of the sample and universe groups, etc. It is interesting to note, however, that a slight difference in the same direction was found in the 1955 study (FPSPG, pp. 452-453). The sample also shows a slight over­ statement of the proportion of nonwhite wives who have attended college. The sample represents the regional distribution of white wives quite adequately (bottom panel, Table A-2). However, the sample and universe differ substantially on region of residence for nonwhite wives. The greatest difference is for the proportion living in the North Central states: 27 per cent for the sample and 19 per cent for the universe. The difference is statistically significant. Fortunately, how­ ever, the sample agrees very well with the universe with respect to the proportion of nonwhite wives living in the South. As Chapter 9 indicates, the widest fertility differentials between whites and nonwhites are found in the South, so it is important that the sample represent the proportion in this region adequately. The over-representation of nonwhite wives in the North Central

FERTILITY AND FAMILY PLANNING IN THE U.S.

states is balanced by an under-representation of such wives in the West. We do not know the reasons for these differences, but we must remember that within the various sub-areas that make up the sample, residential segregation prevents nonwhites from being dis­ tributed in the same way that whites are distributed. Therefore, any given sample block or small area is Ukely to be either predominantly white or predominantly nonwhite. In other words, color is not as randomly distributed as other characteristics (age or sex, for ex­ ample). Therefore, it would not be difficult to select at random too many or too few nonwhite sub-areas within a larger area, and thereby misrepresent the color distribution of the larger area. In other words, sampling errors are relatively high for the geographic distribu­ tion of nonwhites. The observed differences may, therefore, be due simply to sampling variation. Fortunately, 1960 Census data show that nonwhite wives in the North Central states and the West have approximately the same average number of children ever born—2.5 and 2.6, respectively (Table 184). So the geographic difference apparently does not result in a serious misrepresentation of factors affecting the fertility of nonwhite wives. In summary, we have shown that the sample and universe agree very well with respect to past fertility, age, and religion. The sample understates somewhat the proportion of white and nonwhite wives with a grade-school education. Also, it overstates the proportion of nonwhite wives living in the North Central states and understates the proportion living in the West. None of the differences noted is large enough to affect seriously the results of the study. SAMPLING ERRORS

As we have seen above, sample and universe values do not always agree. Some differences arise from response errors (inaccurate replies), the failure of a few respondents to answer certain questions, and various errors that can occur in the process of transferring information from the interview schedule to the punch cards used in tabulating the results. In this section, however, we are concerned with sampling errors—that is, errors arising because the sample does not (and can­ not, even under the best of circumstances) represent the universe with complete accuracy in all respects. A sampling error is the amount by which a sample value for a particular statistic can reasonably be expected (because of random factors in choosing the sample) to differ from the universe value for that statistic. A sampling error must be associated with a statement of the probability that the sample value is within sampling error

APPENDIX A

of the universe value. In the present study, as in the 1955 study, we have chosen the 95 per cent "level of confidence" for sampling errors. For example, we find that 80.7 per cent of the 2,414 white 18-39-year-old wives report that they or their husbands have used contraception. The sampling error for this particular statistic is 2.3 per cent at the 95 per cent level of confidence. This means that the chances are 95 in 100 that the value for the proportion of Users in the entire population represented by these wives is within 2.3 percentage points of 80.7—i.e., that it is in the range between 78.4 and 83.0. If we wanted a higher level of confidence, the sampling error would have to be higher. For example, if we wanted to define the range of universe values with a 99 per cent probability of being correct, the sampling error for the proportion of Users would be 3.0 per cent, rather than 2.3 per cent. After choosing a particular confidence level, the value of a sampling error depends on several factors: the number of cases on which the sample value is based, the distribution within the sample of the characteristic to which the value relates, and the degree of clustering involved in the method of choosing the sample. One of the most important factors is the number of cases on which the statistic is based. The fewer the cases, the higher the sampling error (other things being equal). For example, we find that among 35-39-year-old white wives, 77.4 per cent report some use of contraception. There are 677 such wives, and the sampling error for this statistic is 3.9 percentage points, compared with 2.3 for all white wives 18-39 years old. Sampling errors also vary with the value of the statistic to which they relate. For proportions, sampling errors are greater for those close to 50 per cent than for smaller or larger percentages. Table A-3 gives some impression of how sampling errors vary for different statistics and for different subgroups of the sample. All of these sampling errors were computed by the Survey Research Center's Sampling Section, under the direction of Leslie Kish. The method used to compute them is so complex, however, that it would not be feasible to compute a sampling error for every statistic cited in this book. Therefore, we present approximate, sampling errors for different percentages and different numbers of cases in Table A-4. These are central estimates based on many individual sampling errors, such as those shown in Table A-3. To get some idea of how much the approximate sampling errors in Table A-4 differ from the specific estimates in Table A-3, let us compare sampling errors from both sources for a few statistics.

FERTILITY AND FAMILY PLANNING IN THE U.S.

First, the sampling error for the percentage of Users for the sample of white 18-39-year-old wives is 2.3 according to Table A-3, and about the same according to Table A-4. However, the sampling error for the percentage Fecund for the entire sample is 1.9 according TABLE A-3 Sample Values and Sampling Errors for Per Cent Fecund, Per Cent Users, and Per Cent with Completely Planned Fertility, for White Wives, 18-39 Years Old, by Selected Characteristics

Group Total Husband's income Under $4,000 $7,000 or more Wife's age 18-24 35-39 Wife's education College Grade school Wife's religion Protestant Catholic

No. of wives

Per cent with Completely Planned Per cent Fecund Per cent Users Fertility Per Sampling Per Sampling Per Sampling cent error cent error cent error

2,414

69..3

1.9

80.7

2.3

21.3

1.7

633 666

72 .2 68. 9

3.2 2.3

73.8 85.7

4.2 2.4

20.2 29.9

3.5 3.1 1

513 677

86 .9 52 .7

3.0 3.7

78.2 77.4

4.1 3.9

26.5 17.6

4.0 3.4

427 255

78 .5 60 .4

3.9 5.9

88.3 65.9

3.8 9.0

34.9 5.9

3.2 3.2

1,596 668

68 .7 69 .6

2.5 3.8

84.4 69.8

2.4 4.3

23.4 11.5

2.3 2.7

TABLE A-4 Approximate Sampling Error of Percentagesa (The figures in this table represent two standard errors) Estimated percentages

2,400

2,000

1,000

50 30 or 70 20 or 80 10 or 90 5 or 95

2.8 2.5 2.3 1.7 1.2

3.1 2.8 2.5 1.8 1.3

3.8 3.5 3.0 2.3 1.6

Number of cases 700 500 300

200

100

50

4.3 4.0 3.5 2.6 1.9

7.3 6.7 5.9 4.4 3.2

10.2 9.3 8.1 6.1 4.4

14.4 13.2 11.5 8.6

4.9 4.5 3.9 3.0 2.1

6.1 5.6 4.9 3.7 2.7

a For most items, the chances are 95 in 100 that the percentage for the population from which the sample was drawn lies within a range equal to the estimated percentage plus or minus the sampling error.

to Table A-3 and about 2.5 according to Table A-4. We find an even greater difference between the two tables in the sampling error for the percentage of Users among wives with a grade-school educa­ tion: 9.0 per cent from Table A-3 and about 6.3 per cent interpolated from Table A-4. This example should serve to warn the reader that

APPENDIX A TABLE A-5 Approximate Sampling Errors of Differences between Percentages* (The figures in this table represent two standard errors) Number of cases 2,400

2,000

1,000 4.8 4.9 5.4

700

500

300

200

For percentages from 35 to 65 5.1 5.7 6.7 7.9 6.7 8.0 5.2 5.7 6.2 7.2 8.3 5.8 7.4 8.5 6.0 6.6 7.9 8.9 7.0 8.7 9.6 10.4

2,400 2,000 1,000 700 500 300 200 100 50

4.0

4.2 4.4

2,400 2,000 1,000 700 500 300 200 100 50

3.2

3.4 3.5

2,400 2,000 1,000 700 500 300 200 100

2.4

2.5 2.6

For 2.8 2.9 3.2

percentages around 10 and 90 3.1 3.4 4.1 4.8 4.1 4.8 3.1 3.4 3.5 3.7 4.3 5.0 4.4 5.1 3.6 3.9 4.7 5.3 4.2 5.7 5.2 6.2

2,400 2,000 1,000 700 500 300 200

1.7

1.8 1.9

For 2.0 2.1 2.3

percentages around 5 and 95 2.3 2.5 3.0 3.5 2.3 3.0 3.5 2.5 3.1 3.6 2.5 2.7 2.6 3.2 3.7 2.9 3.4 3.0 3.9 3.8 4.2 4.5

For percentages around 20 and 80 3.8 4.1 4.5 5.4 6.3 6.4 3.9 4.2 4.6 5.5 4.9 5.8 6.6 4.3 4.6 5.9 6.8 4.8 5.3 7.1 5.6 6.3 7.7 7.0 8.3

100

50

10.7 10.7 10.9 11.1 11.3 11.9 12.6 14.4

14.9 14.9 15.1 15.2 15.3 15.7 16.3 17.6 20.4

8.5 8.5 8.7 8.9 9.1 9.5 10.1 11.5

11.9 11.9 12.0 12.2 12.3 12.6 13.0 14.1 16.3

6.4 6.4 6.5 6.6 6.8 7.1 7.5 8.7

a These values show how large a difference between two percentages for two different subgroups of the sample must be for statistical significance at the 95 per cent probability level.

the values shown in Table A-4 are approximate; they may deviate rather widely from estimates computed for a specific statistic. A very important kind of sampling error expresses the extent to which two values for different subgroups of the sample can be expected to differ because of random factors. For example, we find that the proportion of Users is 84.4 per cent for Protestants and

FERTILITY AND FAMILY PLANNING IN THE U.S.

69.8 per cent for Catholics. Could we reasonably expect such a differ­ ence to have arisen because of chance factors in the selection of the sample? Again, we need a probability level that will state how sure we want to be, and again we have chosen 95 per cent. At this level of confidence, the sampling error for the difference between the two percentages cited is 4.9. This means that if the difference between the two percentages is 4.9 or greater, we can be reasonably certain (with at least a 95 per cent probability of being right) that Protestants have a higher proportion of Users than Catholics. ActuTABLE A-6 Approximate Sampling Errors of Averages and of Differences between Two Averages, for Number of Births by Interview and Total Number of Births Expected (The figures in this table represent two standard errors) Number of cases

2,400 .07

2.400 2,000 1,000 700 500 300 200 100 50

.10

2,000

1,000

Number of cases 700 500 300

200

100

Sampling errors of averages" .08 .11 .16 .21 .25 .36 .12 Sampling error of differences between two averages1b .11 .14 .18 .22 .26 .37 .15 .18 .37 .11 .14 .16 .22 .27 .16 .20 .24 .28 .38 .18 .21 .25 .38 .19 .29 .23 .26 .39 .30 .42 .29 .33 .44 .36 .51

50 .51 .51 .52 .52 .53 .53 .55 .57 .62 .72

* For most items, the chances are 95 in 100 that the average for the population from which the sample was drawn lies within a range equal to the estimated average plus or minus the sampling error. b These values show how large a difference between two averages for two different subgroups of the sample must be for statistical significance at the 95 per cent probability level.

ally, the difference is 14.6 percentage points, so we can be virtually certain that there is a real difference between Protestants and Catholics in the universe from which the sample was drawn. Again, it would not be feasible to compute sampling errors for the differences between percentages for every comparison that has been made in the book, so we present approximate values in Table A-5. These values vary with the level of the percentage and the number of cases on which the percentages are based. If we interpolate between the values shown on this table, we find that the sampling error of the difference between the percentages cited in the preceding paragraph is about 4.4.

APPENDIX A

In addition to sampling errors for percentages and differences between percentages, discussed above, we need to know the sampling errors for the average number of births by 1960 and average total number of births expected. It was found that the sampling errors for births by 1960 differ very little from those for total expected births, so we present approximate sampling errors for both statistics in Table A-6. Again, these sampling errors are central estimates based on specific values calculated for many different groups. The following example illustrates the use of Table A-6. The sampling error of the difference between the number of births ex­ pected by 1,596 Protestant and 668 Catholic wives can be found by interpolating between various values shown. It is about .17. The average numbers of births expected by Protestant and Catholic wives are 2.94 and 3.71, respectively. The difference of .77 between them is much greater than the estimated sampling error of .17, so we can be virtually certain that Protestants and Catholics differ in this respect in the entire population from which the sample was drawn. In order to find approximate sampling errors with the use of Tables A-4, A-5, and A-6, it is necessary to know the number of cases on which the percentages or averages are based. In many tables, these numbers are shown. In others, however, it did not seem worth while to take up the large amount of space that would be needed to show the number of cases on which the statistic presented in each cell was based. In order to make these numbers available to interested readers, they are presented in a mimeographed appendix, which may be obtained by writing to the Scripps Foundation, Miami University, Oxford, Ohio.

APPENDIX B

Technical Notes NOTE 1. METHOD OF ESTIMATING NUMBER OF BIRTHS EXPECTED IN 1955-1960

Wives interviewed for the 1955 study were asked several questions to determine when they expected their future children to be born. If a wife said that she expected to have more children, she was asked the following questions: 34. We are interested in when your children might come. Do you expect to have a child in the next 12 months? (If yes) 34a. Are you pregnant now? (If yes) 34b. When is the baby due? 35. About when do you expect to have your next child? 35a. And about how many years will there be between that child and the next? (Repeat question until account for all children expected) (If can't give timing above) 35b. Well, we'd just like to have your general idea of when your additional children might come. The number of children the wife was asked to account for was the number she gave as her "most likely" total number of children minus the number she already had by interview. On the baas of her replies to these questions, each of the wife's future children was assigned a minimum and a maximum number of years between interview and birth. The maximum number of births expected in 1955-60 (i.e., during the five years following interview) was determined by counting the number of children the wife expected according to her minimum timing estimates. The minimum number of births expected in 1955-60 was determined from her maximum timing estimates. For example, if a wife indicated by her answers to questions 34 through 35a that she expected three additional births and that the third might occur as early as four years after interview or as late as six years after interview, her minimum number of births expected in 1955-60 was two and her maximum was three. It is important to note that the minimum and maximum numbers of births expected in 1955-60 are consistent with the wife's most likely estimate of the number of additional births she will have, and not with her minimum and maximum estimates of additional births.

APPENDIX B

Hence, there is a much narrower range between minimum and maxi­ mum births expected in 1955-60 than there is between the minimum and maximum numbers of additional births expected after 1955. This can be seen in the following table: Births in 1955-60 All births after 1955

Minimum Most likely .69 .70 .96

Maximum .71 1.24

The .69 to .71 births expected in 1955-60 are consistent with the average of .96 for "most likely" additional births expected. They are not consistent with the averages of .70 and 1.24 for minimum and maximum additional births expected. NOTE 2. METHOD OF DETERMINING TOTAL NUMBER OF BIRTHS EXPECTED

Three sets of expected numbers of births were developed: mini­ mum, most likely, and maximum. These numbers are based on the replies to one or more questions, which are described below. The procedure used to establish the wife's expectations depends first of all on the couple's fecundity classification. Three fecundity groups were used for this purpose: 1. The Definitely Sterile—All wives but one in this group were assigned no additional births. The one exception was a currently pregnant wife whose husband had had a vasectomy after she con­ ceived. 2. The Probably Sterile—If a wife in this group wanted at least one more child and if she gave evidence that pregnancy would not endanger her health, her most likely additional number of births was assumed to be .3. In effect, this means that we assumed that 30 per cent of such women would each have one additional birth. Her minimum additional was set at zero, and her maximum addi­ tional was the same as the number she would have if she could. 3. The Possibly Sterile, Possibly Fecund, and Fecund—This group accounts for nearly all of the expected additional births (99.2 per cent of the most likely). The remainder of this technical note describes the procedure followed to establish their expectations. The large majority of wives in this category were asked the follow­ ing question: Q. 22. You now have—living children (or soon will have after the baby comes). Do you expect to have (more) children (after this baby)? Would you say definitely yes, probably yes, you're uncertain, probably no, or definitely no?1 1 The phrases in parentheses are optional; the interviewer may or may not use them, depending on the situation.

FERTILITY AND FAMILY PLANNING IN THE U.S.

If the wife said definitely or probably yes, she was asked: Q.23. How many children do you expect to have in all (count­ ing those you now have) ? (The interviewer probed further if the wife gave an indeterminate answer, such as "It depends on God's will.") Question 23 is the main question about expected numbers of chil­ dren in this study. Altogether, 33 per cent of the white 18-39-year-old wives were asked this question. This is a majority of the 42 per cent who expected to have births in the future. The reply to Question 23 was taken as the "most likely" total number of children expected. Children who had died were added to this number to obtain births expected. After question 23, the interviewer asked: Q.24. How sure are you that you really will have — children in all . . . would you say that you are very sure, fairly sure, or not sure? If the wife said fairly sure or not sure, she was asked: Q.24a. What do you think is the largest number of children you will probably have in all? and Q.24b. You've said that you expect to have (more) children . . . now, what do you think is the smallest number of children you will probably have in all? Questions 24a and 24b were added in order to get some idea of the range of possible numbers of children that the less certain wives had in mind. If the wife said that she was very sure or fairly sure of her reply to Question 23, it was used to establish her minimum and maximum expectations. If she gave a single number of children in reply to Question 23, this was taken as the minimum and maximum number expected, as well as the most likely. If she gave a range (such as three or four) in reply to question 23, the high and low were taken as maximum and minimum, respectively, and the average was taken as the most likely. If the wife said that she was "not sure" of her reply to question 23, her replies to questions 24a and 24b were taken as her maximum and minimum expectations, respectively. The procedure used to establish most likely expectations means that some wives were assigned a fractional number of children, such as 3.5. This practice is justified by the assumption that half of the women represented by this individual will have three children, and half will have four. Again, it is important to remember that we are thinking in terms of groups of women, not individual women.

APPENDIX B

The above questions were asked only of those women who thought they could and probably would have more children. Those who thought they could have more, but were uncertain whether or not they would were asked a different question. If the wife said in reply to question 22 that she was uncertain or probably would not have any more children, she was asked: Q.31. If you do have (more) children, what is the largest number you are likely to have in all (counting those you now have)? If the wife said that she probably would not have any more children, her reply to question 31 was ignored, and it was assumed that she would have no more children. However, if she was uncertain, her minimum number of children was assumed to be the number she already had, her maximum number was her reply to question 31, and her most likely was the average of minimum and maximum. The procedure just described was used for wives who said they thought they could have more births. In some cases, their replies had to be modified in view of evidence that came out later in the interview that they might be unable to have any more births, but such exceptions were rare. In addition, there were some wives who were unsure whether they could have more births in the future, but to whom we assigned some additional births. Most of them are classified as Possibly Sterile or Possibly Fecund, although a few are classified as Fecund. Their expectations were based on their replies to the following questions: Q.47. If something happens so that you can have a child later on, do you expect that you will? Would you say definitely yes, probably yes, you're uncertain, probably no, or definitely no? If the wife said definitely or probably yes, she was asked: Q.47a. If you can have (more) children, how many do you ex­ pect to have in all (counting those you now have) ? If the wife gave a single number in response to the latter question, it was taken as her minimum, most likely, and maximum expected number of births. If she gave a range, the high was taken as the maximum, the low as the minimum, and the average as the most likely. Only 13 per cent of the most likely expected additional births were assigned to wives in the Possibly Sterile or Possible Fecund categories. Although it might seem advisable to modify their replies in view of the evidence that their fecundity is below normal, we have some justification for not doing so from the 1955-60 comparisons of women who were classified as Subfecund in 1955. As noted in

FERTILITY AND FAMILY PLANNING IN THE U.S.

Chapter 1, such wives had more births in 1955-60 than they thought they would. Some wives failed to answer the questions about the number of children expected. In the white 18-39-year-old group, such wives made up 4 per cent of the total. In many cases they said in response to Question 22 that they were uncertain whether they would have more children, and in response to Question 31 that they did not know what the largest number was that they would have. Other wives said that they definitely or probably expected more, but refused to say how many more. In spite of their failure to answer these questions, we wanted to assign expected numbers of births to them, so that women like them would be represented in our findings. The number of additional births assigned to these women was based on their age, use of contraception, religious preference, and the number of children they had already borne. A table was prepared showing the average number of additional births expected for each combination of these characteristics from the 1955 study, and this was used as a guide to assigning the expectations for "unknowns" in the 1960 study. NOTE 3. COMPARISON OF 1955 AND 1960 FECUNDITY CLASSIFICATIONS

The method of classifying couples by fecundity for the 1960 study differs from that for the 1955 study in several respects. The 1960 scheme represents an attempt to rank couples by the severity of their impairments in greater degree than did the 1955 scheme. The 1960 system of classification does not include an indeterminate category, as did the 1955 system. Also, more information relating to fecundity was obtained in the 1960 study than in the 1955 study. In order to allow for these changes in comparing the 1955 and 1960 findings, an effort was made to classify the couples in the 1955 study according to the 1960 system. These estimates are necessarily rough because the 1955 study did not provide the same kind or amount of informa­ tion on fecundity that the 1960 study did, but they do indicate the kinds and approximate magnitudes of differences between the two methods of classification. Couples who were classified as Definitely or Probably Sterile in the 1955 study were reclassified on the basis of a review of their interview schedules. For other couples (i.e., those classified as Semifecund, Indeterminate, and Fecund), assumptions were made about their distribution among the 1960 fecundity categories. The Semifecund and Indeterminate were assumed to be distributed among the 1960 fecundity categories according to the proportions found for similar 1960 couples when they were classified according to their

APPENDIX B

status in 1955. Couples classified as Fecund in the 1955 study were assumed to be Fecund according to the 1960 method of classification. Although the latter assumption appears to be one that biases the results of the comparison, no alternative assumption was possible. For example, a review of the schedules for the 1955 Fecund cases would not have helped because such schedules contain virtually no information about the presence of fecundity impairments. A comparison of the percentage distributions of the two samples by fecundity (Table B-I) shows that both give similar pictures of the prevalence of fecundity impairments in the United States. The proportion of couples classified as Subfecund is 34 per cent according to the 1955 scheme, and 31 per cent according to the 1960 scheme. The difference arises almost entirely from the assumption that over half of the couples classified as Indeterminate in 1955 would be classified as Fecund in the 1960 system. A more detailed look at the reclassification shows the following differences between the two systems: Definitely Sterile—The definitions of the Definitely Sterile cate­ gory are identical in both studies—that is, a couple in this group cannot have children in the future. However, this definition was ap­ plied more strictly in the 1960 study, partly because more information was available than in 1955. This reduced the proportion of couples classified as Definitely Sterile, from 10.4 per cent to an estimated 8.9 per cent. This comparison should not be taken as an exact measure of the difference between the two Definitely Sterile categories because, as noted previously, the 1955 schedules gave less information than did the 1960 schedules, and it is not always possible to be certain that if more information were available a particular case would not have been classified as Definitely Sterile. An example of a reclassified case is the following: CASE 2435 (1955 study). This 34-year-old woman had been married for over 10 years and had never been pregnant. Neither she nor her husband had ever used contraception. According to the inter­ viewer, "She and her husband had been taking treatments and she had an operation. The doctor doesn't seem to know at this point why she can't have a baby. They are still hoping, however." In the 1960 study, such a case would be classified as Probably Sterile. Probably Sterile—Here we are referring to the category Probably Sterile as it was defined in the 1955 study: "Probably Sterile couples are those for whom a birth is considered improbable (rather then impossible) in the future on the basis of specific medical evidence."2 ' FPSPG,

p. 22.

FERTILITY AND FAMILY PLANNING IN THE U.S.

There is no comparable category in the 1960 study, inasmuch as wives reporting the kind of medical evidence referred to in the above definition were distributed among several categories. The severity of the reported impairment determined which category was used to de­ scribe a couple's fecundity. A review of the 1955 Probably Sterile TABLE B-I Estimated Distribution of Couples in the 1955 Study, by 1955 Fecundity Classification, by 1960 Fecundity Classification

I960 Fecundity classification

Total

Total Fecund Subfecund Total Definitely Sterile Probably Sterile Possibly Sterile Possibly Fecund

2,713 1,865

Total Fecund Subfecund Total Definitely Sterile Probably Sterile Possibly Sterile Possibly Fecund

100.0 100.0

Total Fecund Subfecund Total Definitely Sterile Probably Sterile Possibly Sterile Possibly Fecund

848 241 135 322 150

100.0 100.0 100.0 100.0 100.0 100.0 68.7 31.3 8.9 5.0 11.9 5.5

1955 Fecundity classification Subfecund Defi­ Proba­ nitely bly Semi- Indeter­ sterile sterile fecund minate Fecund Total

1,794 1,794

Number of couples 283 187 919 — 71 5

328 —

848 283 182 328 240 — 241 1 — 135 36 45 47 — 7 322 94 194 150 87 — — 42 Per cent distribution by 1955 classification 66.1 33.9 10.4 6.9 12.1 — — 3.8 96.2 .3 •— —

100.0 33.4 21.5 38.7 — 100.0 99.6 .4 — 26.7 33.3 100.0 34.8 •— 100.0 2.2 29.2 60.2 100.0 — — 28.0 58.0 Per cent distribution by 1960 classification 100.0 100.0 100.0 100.0 100.0 — •— 100.0 7.7 2.7 —•

—

— — — •— —

91.3 26.2 14.7 35.0 16.3

100.0 84.8 12.7 2.5 —

97.3 .5 24.1 50.3 22.5

100.0 —

14.3 59.1 26.5

121 66 55 —

7 27 21 4.5 3.5 6.5 —

5.2 8.4 14.0 100.0 54.5 45.5 —

5.8 22.3 17.4

cases suggests that half of these couples would have been classified as Possibly Sterile in the 1960 study, a quarter as Probably Sterile, and a quarter Possibly Fecund. A very few might have been classified as Fecund. Again, it should be understood that these are rough approximations at best. Semifecund —In both the 1955 and 1960 studies, couples were classified as Subfecund if they had long periods during which contra-

APPENDIX B

ception was not used and pregnancy did not occur. The length of such periods was two more years if the wife had never been preg­ nant, and three or more years if the wife had ever been pregnant. However, if the couple had had more than one birth for every three years of married life, they were considered to be Fecund, even if they had had a long noncontraceptive period. The same criteria were used in the 1960 study, but instead of putting all such couples into one group, as was done in 1955, they were assigned to different groups on the basis of the length of the period of noncon­ traceptive exposure and on whether or not they had had a birth after such a long period. It is estimated that most of the 1955 Semifecund couples would be classified as Possibly Sterile in the 1960 classifica­ tion; about a quarter would be Possibly Fecund, and a relatively small proportion (about one-seventh) would be Probably Sterile. These estimates are not based on a review of the 1955 Semifecund cases, but are derived from data relating to 1955 for women interviewed in 1960 who were eligible for the 1955 sample. Indeterminate—This category was used in the 1955 study for wives who reported long periods of noncontraceptive exposure (as defined above) and who also reported that they douched for cleanli­ ness soon after intercourse. Since douching reduces the likelihood of conception if done immediately after intercourse, it was uncertain whether the long periods without pregnancy were due to douching or to subnormal fecundity. Therefore, the couple's fecundity was de­ scribed as "Indeterminate." In the 1955 study, the wives were not asked how soon after intercourse they douched, but in the 1960 study they were. If they usually douched less than half an hour after intercourse, it was assumed that douching reduced the proba­ bility of conception. Later use was assumed to have no effect on the probability of conception. These assumptions permitted us to divide what would have been the Indeterminate couples into those without impairments (i.e., those whose long noncontraceptive periods were assumed to be the result of douching less than half an hour after intercourse) and those assumed to have impairments. The latter couples were assigned to the various fecundity categories according to the same criteria used for other couples with long periods of noncontraceptive exposure. A slight majority of the wives in the Inde­ terminate category (55 per cent) were found to be douching less than one-half hour after intercourse (again, using 1955 data for women in the 1960 sample who were eligible for the 1955 study). The other couples classified as Indeterminate in 1955 were assumed to be distributed among the Probably Sterile, Possibly Sterile, and Possibly Fecund as were similar couples in the 1960 sample.

FERTILITY AND FAMILY PLANNING IN THE U.S.

To summarize the above discussion, the most important quantita­ tive differences between the 1955 and 1960 systems of classifying couples by fecundity are: first, the new system is somewhat more strict with respect to the Definitely Sterile category and shows a slightly lower proportion in this group than the 1955 system; second, the new system classifies as Fecund slightly over half of the couples who would have been designated Indeterminate under the old system. The result of the latter change raises the proportion Fecund and lowers the proportion Subfecund by about 3 percentage points. In brief, the 1955 and 1960 systems of fecundity classification give us very similar views of the prevalence of fecundity impairments among white couples in the reproductive ages. The 1960 system of classification represents a refinement of that used in 1955, but the broad picture it gives is very much the same. NOTE 4. THE CLASSIFICATION OF PROTESTANT DENOMINATIONS

The denominations included in each category are shown below: Category

Liberal

Intermediate

Fundamentalist

Other and NA

Denominations

Community Church (no denomination) Congregational Episcopal Presbyterian Quaker Reformed, Dutch Reformed, or Christian Reformed Unitarian or Universalist African Methodist Episcopal Baptist Christian Christian Scientist Disciples of Christ Evangelical and Reformed Lutheran Methodist United Brethern or Evangelical Brethern United Church of Christ Church of Christ Church of God Church of God and Christ Jehovah's Witness Latter Day Saints, Mormon Nazarene or Free Methodist Pentecostal or Assembly of God Plymouth Brethern Primitive Baptist or Free Will Baptist Salvation Army Seventh Day Adventist Southern Baptist United Missionary or Protestant Missionary No denomination specified Spiritualist Unity

APPENDIX B

NOTE 5. ESTIMATES OF THE PROPORTION OF COUPLES WHO WOULD NOT NEED TO USE CONTRACEPTION TO PREVENT UNWANTED PREGNANCIES

Socioeconomic differentials in the proportion of couples who have ever used contraception are said to result primarily from differ­ ences in the proportion who delay using contraception. Some of the couples who delay use discover fecundity impairments before they need to begin using contraception and consequently never become Users. Similarly, the upward cohort trend in the proportion of couples using contraception is said to have occurred primarily because more of the couples in the earlier cohorts delayed use, discovered impaired fecundity, and never needed to begin. This explanation implies that a substantial proportion of couples encounter fecundity impairments before they have too many children. Does this appear to be true? For example, what proportion of white couples become Subfecund before, or just after, a third birth occurs? (Three is chosen because it is the average number of children wanted by the wives interviewed.) A minimum estimate of 11 per cent for the United States white population is suggested by the proportion of couples in the 1916-20 cohorts who have been married 10 or more years, are Subfecund, have never used contraception, and have had three births or fewer. This is a minimum estimate because it does not include Users who would have become Subfecund before having four births even if they hadn't used contraception. We get a maximum estimate of 38 per cent from the proportion of couples in the 1916-20 cohorts who are Subfecund and have had three births or fewer. This greatly over­ states the proportion of Nonusers who would encounter fecundity im­ pairments before having four births because it includes couples who would have had four or more births if they hadn't used contraception. We can get some idea of the proportion of couples who can't have more than three births from statistics for groups that make little or no use of contraception. For example, 21 per cent of the white ever-married women 50-74 years old in Utah in 1910 reported that they had had three births or fewer. Among ever-married women who were aged 65 or over in 1960 and lived in rural-farm areas of Utah, the proportion reporting that they had had three births or fewer was 19 per cent. These figures probably overstate the propor­ tion of couples who cannot have more than three births because some of the low fertility of these women was due to early widowhood rather than subfecundity, and some of it may have been due to the use of contraception.

FERTILITY AND FAMILY PLANNING IN THE U.S.

The Hutterites, a contemporary religious sect living in the United States and Canada, have unusually high fertility. In 1950, married Hutterite women 45 years of age and over reported that they had had 10.4 births, on the average. The proportion of these women re­ porting three births or fewer was only 10 per cent. This proportion may be close to the biological minimum. We know that it does not adequately represent the proportion of couples in the United States who could have no more than three births because the prevalence of childlessness is lower among Hutterites (3 per cent) than it is in the United States (6 to 8 per cent for recent cohorts). It appears, then, that about 15 ± 5 per cent of white married couples in the United States cannot have more than three births even if they never use contraception. This is a very rough approxi­ mation and should certainly not be considered definitive in any sense. However, its magnitude does suggest that it is possible to account for most of the differences between groups in the proportion ever using contraception by differences in the proportion of couples who delay use and then discover that contraception is not needed to pre­ vent unwanted births. NOTE 6. THE 1955-60 RISE IN THE USE OF CONTRACEPTION FOR FECUND AND SUBFECUND COUPLES Between 1955 and 1960, the proportion of Users increased more for Subfecund than for Fecund couples, as the following data show: Couple's fecundity Total Fecund Subfecund

Per cent 1955 70 83 45

User 1960 81 89 62

I960 minus 1955

11 6 18

One of the reasons for the greater increase in use among the Subfecund is the fact that the "Indeterminate" category, which was in­ cluded with the Subfecund in 1955, was eliminated from the fecundity classification used in 1960. This category comprised wives who regularly douched after intercourse, for purposes of cleanliness only and not for contraception, and who had borne children relatively infrequently. Since the 1955 schedule lacked questions which made it possible to tell whether douching or subfecundity had reduced their fertility, couples with such wives were classified as "Indetermi­ nate" with respect to fecundity. In 1960, however, such couples were classified as Fecund if the wife douched within a half hour after intercourse, and as Subfecund if she douched later. As a result, there are fewer such couples among the Subfecund in 1960 than there were in 1955. Since such couples are classified as Nonusers, the new method

APPENDIX B

of classification tends to make the proportion of Users higher among the Subfecund and lower among the Fecund (i.e., it shifts Nonusers from the Subfecund to the Fecund). This tendency, however, cannot account for all of the difference between the two fecundity groups in the increased use of contraception. If we reclassified half of the Indeterminate couples of the 1955 study as Fecund (approximately the proportion of such couples who were classified as Fecund in the 1960 study), we would have the following proportions of Users, by fecundity, for 1955: Couple's fecundity (assuming that one-half of the Indeterminate couples were Fecund) Total Fecund Subfecund

Per cent User 1960 minus 1955 1955 70 11 81 8 48 14

Thus, even with a fecundity classification that is more comparable to that used in 1960, we find that the rise in the proportion of Users has been greater for Subfecund than for Fecund couples. NOTE 7. METHOD OF DETERMINING THE CONTRACEPTIVE STATUS OF USER'S PREGNANCIES

The pregnancies of Users have been classified as beginning under the four following circumstances: 1. Before the couple began using contraception. 2. When contraception was stopped in order to conceive. 3. When contraception was being used always. 4. When contraception was stopped for reasons other than the desire to conceive. First, the interviewer determined when the couple began using contraception (Q80, 80a, and 80e). All conceptions occurring before use of contraception began were, of course, assigned to the first cate­ gory. After determining the interpregnancy interval of first use, the wife was asked whether she (and/or her husband) had used a method in each of the subsequent interpregnancy intervals (Q81). If they had not used contraception in the interval before a given pregnancy, this pregnancy was assigned to the fourth category (contraception stopped for reasons other than the desire to conceive). If contraception was used in the interval before a given preg­ nancy, the wife was asked whether she had stopped using contra­ ception (Q81b). If she had stopped, she was asked whether she wanted to get pregnant then (Q81c). If she said she did, then the

FERTILITY AND FAMILY PLANNING IN THE U.S.

conception was assigned to the second category (contraception stopped in order to conceive). If the wife said that contraception was stopped but that she did not want to get pregnant then, the pregnancy was assigned to the fourth category. If the wife said that contraception was not stopped before a given pregnancy, she was asked whether the method was being used "always, or only part of the time." (Q81d) If she said "always," the pregnancy was classified under the third category. (Often such conceptions are referred to as accidents.) If she said, "part of the time," the pregnancy was assigned to the fourth category. NOTE 8. METHOD OF CHOOSING THE MATCHED WHITE SAMPLE

In order to make the fertility comparisons shown in Chapter 9, we had to choose 270 white couples who would match the 270 nonwhite couples on a number of characteristics closely related to fertility. The difficulty in finding comparable white couples increases as the list of characteristics lengthens, so we had to keep it fairly short. The matching criteria finally selected are as follows: For all couples: 1. 2. 3. 4. 5. 6.

Duration of marriage Wife's age at marriage Wife's religion Region of residence Wife's education Size of place of longest residence

For couples who did not live longest on a farm: 7. Farm background 8. Husband's occupation Success in finding white wives who are comparable to nonwhites on these characteristics also depends on the number of categories within each characteristic. For example, it would be harder to find matched whites if we tried to match on single years of age at marriage than it would if we simply tried to match on two broad age-atmarriage groups. The categories finally selected are presented in Tables B-2 and B-3, which also show the per cent distributions by all matching criteria for nonwhites, matched whites, and all whites. These tables show that we were able to find whites to match nonwhites perfectly on four characteristics: duration of marriage, age at marriage, wife's religion, and region of residence. Matching was almost as good on wife's education; there are not quite as many

APPENDIX B

wives with a grade-school education in the matched white sample as in the nonwhite sample, but the difference is small. The characteristic that was most difficult to match was size of longest place of residence since marriage. The majority of nonwhites TABLE B-2 Per Cent Distribution by Characterbtics on Which White and Nonwhite Couples Were Matched, for Nonwhites, Matched Whites, and All Whites Nonwhite Nonwhite minus minus matched total white» white»

Characterbtic

Nonwhite

Matched white

Total white

Number of couples Per cent: Total Years married Under 5 5-9 10-14 15 or more Age at marriage Under 20 20 or over Wife's religion Protestant Catholic Other Region of residence Northeeist North Central South West Wife's education College High school, 4 High school, 1-3 Grade school Size of place of longest resi­ dence since marriage City 100,000 or more Smaller place Farm

270

270

2,414

100

100

100

—

26 27 24 23

26 27 24 23

23 27 29 21

—

54 46

54 46

51 49

—

90 7 3

90 7 3

66 28 6

15 27 50 7

15 27 50 7

24 30 31 15

14 27 32 27

14 30 32 24

18 48 24 11

53 33 15

36 51 13

26 64 10

—

4

—

—

-

-5 2

—

— —

—

— — — —

4 -4 24 —21 -3 -9 -2 20 -8

3

-4 -21 8 17

16 -18 2

27 -31 4

—

-3 —

"Differences are computed from unrounded percentages; hence, they are not always equal to differences between rounded percentages.

(53 per cent) had lived longest in cities of 100,000 or more, and it was impossible to find as many whites who had also lived in large cities and who matched the nonwhites as nearly as possible on other characteristics. Consequently, the matched whites tend to have lived longest in smaller places than the nonwhites.

F E R T I L I T Y AND F A M I L Y PLANNING IN T H E U . S .

Among couples who had not lived longest on farms, the matched whites have a higher proportion with previous farm experience than do the nonwhites, according to Table B-3. This table also shows that the matched whites have somewhat higher occupational status than the nonwhites. In summary, the 270 white couples who were chosen to match the nonwhite sample tend to have lived longest in smaller cities than the nonwhites, they have more previous farm experience, and higher occupational status. In spite of these differences, the characteristics of the nonwhite population resemble those of the matched whites TABLE B-3 Per Cent Distribution by Characteristics on Which White and Nonwhite Couples Not Living Longest on Farm Were Matched, for Nonwhites, Matched Whites, and All Whites Nonwhite minus matched white"

Nonwhite minus total white·

—

—

Characteristic

Nonwhite

Matched white

Total white

Number of couples Per cent: Total Farm background Some None Husband's occupation Upper white collar Lower white collar Upper blue collar Lower blue collar Farm Other

230

235

2,162

100

100

100

55 45

62 38

51 49

-7 7

4 -4

9 11 14 57 1 7

14 10 22 50 1 4

33 14 20 29 1 4

-5 1 -8 8

-24 -2 -6 29

—

•—

4

4

a Differences are computed from unrounded percentages; hence, they are not always equal to differences between rounded percentages.

more closely than they do those of all whites. Farm background represents a slight, but unimportant, exception to this generalization. There are two important characteristics on which we did not try to match whites and nonwhites, but which are related to some fertility variables. One of these is the wife's birth cohort. However, we automatically chose white wives who belonged to nearly the same cohorts as nonwhites when we selected those with the same duration of marriage and age at marriage. The assumption that this would result in good matching on birth cohort is verified in Table B-4. The other characteristic on which we felt it might have been desirable to match whites and nonwhites is husband's income. However, we knew that attempts to match on income were bound

APPENDIX B

to be futile if we also wanted to match other important characteristics, because of the great income differentials between whites and nonwhites who are comparable in certain other respects. We might have found enough whites to match the nonwhites if we had had many more whites from which to choose, but our sample of whites was limited. The income distributions in Table B-4 show that the nonwhites resemble the matched whites more than they do all whites, TABLE B-4 Per Cent Distribution by Cohort, Husband's Income, and Couple's Southern Farm Residence, for Nonwhites, Matched Whites, and All Whites Nonwhite Nonwhite minus minus matched total white" white"

Characteristic

Nonwhite

Matched white

Total white

Number of couples Per cent: Total Cohort 1936-42 1931-35 1926-30 1921-25 Husband's income $7,000 or more $6,000-86,999 $5,000-85,999 $4,000-$4,999 83,000-83,999 $2,000-82,999 Under $2,000 Southern farm residence None Some, not on farm nowb On farm now

270

270

2,414

100

100

100

22 24 29 24

21 27 27 24

21 25 26 28

—

3 6 12 19 17 21 22

19 13 19 13 14 11 10

28 13 18 16 13 8 6

-16 -7 -7 6 2 10 12

-24 -7 -6 3 4 13 16

43 45 12

56 33 11

74 22 4

-13 12 1

-31 22 8

—

2 -3 2

—

1 -1 4 -4

a Differences are computed from unrounded percentages; hence, they are not always equal to differences between rounded percentages. b Either husband, wife, or both have lived on a farm in the South.

with respect to husband's income. However, the income differences between nonwhites and matched whites are still wide. To summarize them: the nonwhites have only 21 per cent with incomes of $5,000 or more, the matched whites have 51 per cent in this income group, and all whites have 59 per cent. There is a good argument, however, for not trying to match whites and nonwhites on income. First of all, we have shown in earlier chapters that most of the fertility variables under study are more closely related to education than they are to income. This means

FERTILITY AND FAMILY PLANNING IN THE U.S.

that it is more important to match whites and nonwhites on education than income. Secondly, if we were able to find whites of equivalent education with income low enough to match the nonwhites, we would probably be selecting a very special group of whites, not at all representative of their educational category. Therefore, we should match on education but not income in order to avoid choosing highly unrepresentative whites. Although nonwhites and matched whites have similar percentage distributions on a number of characteristics that may affect their fertility, it would be a mistake to assume that the two groups are almost exacdy alike in these respects. This is because a given char­ acteristic, defined in the same way for both groups, may actually describe two different sets of circumstances for whites and nonwhites. Just because the same descriptive phrase is used in both cases does not mean that the conditions it describes are the same. Take Southern farm residence, for example. (As shown in Chapter 9, this is quite an important characteristic for the nonwhites.) The percentages of nonwhites and matched whites living on Southern farms are very similar: 12 per cent and 11 per cent, respectively (Table B-4). How­ ever, for the nonwhites, living on a Southern farm typically means living in a rural slum; the average income of the husbands in this group is only $1,900. Southern farm residence for whites, in contrast, is accompanied by much better living conditions. The average husband's income for white farm dwellers in the South is $4,000, over twice that for nonwhites.

SELECTED BIBLIOGRAPHY

INTERVIEW SURVEYS ON FAMILY PLANNING AND RELATED TOPICS IN THE UNITED STATES1

This bibliography lists the books and articles reporting the results of surveys. Articles reporting only plans for the surveys or commentaries on them are omitted. Where appropriate, we first show the publica­ tions giving the principal results of each survey; other reports are listed alphabetically by author. INDIANAPOLIS STUDY

Pascal K. Whelpton and Clyde V. Kiser, eds., Social and Psychological Factors Affecting Fertility, Milbank Memorial Fund, New York; five volumes published in 1946, 1950, 1952, 1954, and 1958. DETROIT AREA SURVEYS

Ronald Freedman, David Goldberg, and Harry Sharp, " 'Ideals' About Family Size in the Detroit Metropolitan Area: 1954," Milbank Memorial Fund Quarterly, vol. 33, no. 2, pp. 187-197, April, 1955. Ronald Freedman and Harry Sharp, "Correlates of Values About Ideal Family Size in the Detroit Metropolitan Area," Population Studies, vol. 8, no. 1, pp. 35-45, July, 1954. David Goldberg, Family Role Structure and Fertility. Ann Arbor, Michigan, University Microfilms, 1958. Microfilm copy of type­ script: ix, 156, 157-184, 185-197. Abstracted in Dissertation Abstracts, vol. 19, no. 3, pp. 595-596, 1958. David Goldberg, "Fertility of Two-Generation Urbanites," Population Studies, vol. 12, no. 3, pp. 214-222, March, 1959. David Goldberg and Clyde H. Coombs, "Some Application of Un­ folding Theory to Fertility Analysis," in Emerging Techniques in Population Research, Milbank Memorial Fund, New York, 1963, pp. 105-129. David Goldberg, Harry Sharp, and Ronald Freedman, "The Stability and Reliability of Expected Family Size Data," Milbank Memo­ rial Fund Quarterly, vol. 37, no. 4, pp. 369-385, October, 1959. 1 For references to surveys in other countries, see Ronald Freedman, "The Sociology of Human Fertility," Current Sociology, vol. 10/11, no. 2, 1961-1962.

FERTILITY AND FAMILY PLANNING IN THE U.S. GROWTH OF AMERICAN FAMILIES STUDY, 1955

Principal report Ronald Freedman, Pascal K. Whelpton, and Arthur A. Campbell, Family Planning, Sterility, and Population Growth, McGraw-Hill Book Company, New York, 1959. Other reports Arthur A. Campbell, "Socioeconomic Correlates of Fertility and Fertility Expectations in a Cross Section of White Married Couples in the United States in 1955," in Thirty Years of Re­ search in Human Fertility: Retrospect and Prospect, Milbank MemorialFundjNew York, 1959, pp. 107-116. Arthur A. Campbell, Pascal K. Whelpton, and Ronald Freedman, "The Growth of American Families: Results of a National Sur­ vey," in American Statistical Association, Proceedings of the Social Statistics Section, 1958, American Statistical Association, Washington, 1959, pp. 128-132. Deborah S. Freedman, "The Relation of Economic Status to Fertility," American Economic Review, vol. 53, no. 3, pp. 414-426, June, 1963. Deborah S. Freedman, Ronald Freedman, and Pascal K. Whelpton, "Size of Family and Preference for Children of Each Sex," American Journal of Sociology, vol. 66, no. 2, pp. 141-146, September, 1960. Ronald Freedman, "Fecundity and Family Planning in the White Population of the United States: 1955," in Thirty Years of Re­ search in Human Fertility: Retrospect and Prospect, Milbank Memorial Fund, New York, 1959, pp. 61-73. Ronald Freedman, "Social Values About Family Size in the United States," Proceedings of the International Population Conference, Vienna, 1959, pp. 173-183. Ronald Freedman and Doris P. Slesinger, "Fertility Differentials for the Indigenous Non-farm Population of the United States," Population Studies, vol. 15, no. 2, pp. 161-173, November, 1961. Ronald Freedman, Pascal K. Whelpton, and Arthur A. Campbell, "Family Planning in the U.S.," Scientific American, vol. 200, no. 4, pp. 50-55, April, 1959. Ronald Freedman, Pascal K. Whelpton, and John W. Smit, "SocioEconomic Factors in Religious Differentials in Fertility," Ameri­ can Sociological Review, vol. 26, no. 4, pp. 608-614, August, 1961.

SELECTED BIBLIOGRAPHY

Jeanne Clare Ridley, "Number of Children Expected in Relation to Non-familial Activities of the Wife," Milbank Memorial Fund Quarterly, vol. 37, no. 3, pp. 277-296, July, 1959. Jeanne Clare Ridley, The Relationship of Non-familial Activities to Fertility Behavior, Ph.D. dissertation in sociology, University of Michigan, 1957. GROWTH OF AMERICAN FAMILIES STUDY, 1960

Principal report Pascal K. Whelpton, Arthur A. Campbell, and John E. Patterson, Fertility and Family Planning in the United States, Princeton University Press, Princeton, N.J., 1965. Other reports Arthur A. Campbell, "Concepts and Techniques Used in Fertility Surveys," in Emerging Techniques in Population Research, Milbank Memorial Fund, New York, 1963, pp. 17-38. Arthur A. Campbell, "The Incidence of Operations That Prevent Conception," American Journal of Obstetrics and Gynecology, Vol. 89, No. 5, pp. 694-700, July 1, 1964. Arthur A. Campbell, Pascal K. Whelpton, and Richard F. Tomasson, "The Reliability of Birth Expectations of U.S. Wives," in Pro­ ceedings of the International Population Conference, New York, 1961, vol. 1, John Wright and Sons, Ltd., London, 1963, pp. 49-58. Pascal K. Whelpton, "Cohort Analysis and Fertility Projections," in Emerging Techniques in Population Research, Milbank Me­ morial Fund, New York, 1963, pp. 39-64. Pascal K. Whelpton, "Trends and Differentials in the Spacing of Births," Demography, vol. 1, no. 1, pp. 83-93, 1964. Pascal K. Whelpton, Arthur A. Campbell, and John E. Patterson, "Trends and Determinants of Family Size in the United States," American Journal of Public Health, vol. 54, no. 11, pp. 1834— 1840, November, 1964. PRINCETON STUDY

Principal reports Charles F. Westoff, Robert G. Potter, Jr., Philip C. Sagi, and Elliot G. Mishler, Family Growth in Metropolitan America, Princeton University Press, Princeton, N.J., 1961. Charles F. Westoff, Robert G. Potter, Jr., and Philip C. Sagi, The Third Child, Princeton University Press, Princeton, N.J., 1963.

FERTILITY AND FAMILY PLANNING IN THE U.S.

Other reports Frank W. Notestein, Elliot G. Mishler, Robert G. Potter, Jr., and Charles F. Westoff, "Pretest Results of a New Study of Fertility in the United States," in Bulletin de I3Institut International de Statistique, Stockholm, 1958, vol. 36, no. 2, pp. 154-164. Robert G. Potter, Jr., "Contraceptive Practice and Birth Intervals Among Two-Child White Couples in Metropolitan America," in Thirty Years of Research in Human Fertility: Retrospect and Prospect, Milbank Memorial Fund, New York, 1959, pp. 74-92. Robert G. Potter, Jr., Charles F. Westoff, and Philip C. Sagi, "Delays in Conception: A Discrepancy Re-examined," Eugenics Quar­ terly, vol. 10, no. 2, pp. 53-58, June, 1963. Robert G. Potter, Jr., Philip C. Sagi, and Charles F. Westoff, "Im­ provement of Contraception During the Course of Marriage," Population Studies, vol. 16, no. 2, pp. 160-174, November, 1962. Robert G. Potter, Jr., Philip C. Sagi, and Charles F. Westoff, "Knowledge of the Ovulatory Cycle and Coital Frequency as Factors Affecting Conception and Contraception," Milbank Memorial Fund Quarterly, vol. 40, no. 1, pp. 46-58, January, 1962. Robert G. Potter, Jr., Philip C. Sagi, and Charles F. Westoff, "Some Neglected Factors Pertaining to Fertility Control," Fertility and Sterility, vol. 13, no. 3, pp. 259-264, May-June, 1962. Philip C. Sagi, "A Component Analysis of Birth Intervals Among Two-Child White Couples in Metropolitan America," in Thirty Years of Research in Human Fertility: Retrospect and Prospect, Milbank Memorial Fund, New York, 1959, pp. 135-148. Philip C. Sagi, "Interim Report on the Study of Future Fertility of Two-Child Families in Metropolitan America," in American Statistical Association, Proceedings of the Social Statistics Section, 1958, American Statistical Association, Washington, 1959, pp. 133-137. Philip C. Sagi, Robert G. Potter, Jr., and Charles F. Westoff, "Con­ traceptive Effectiveness as a Function of Desired Family Size," Population Studies, vol. 15, no. 3, pp. 291-296, March, 1962. Philip C. Sagi and Charles F. Westoff, "An Exercise in Partitioning Some Components of the Variance of Family Size," in Emerging Techniques in Population Research, Milbank Memorial Fund, New York, 1936, pp. 130-140. Charles F. Westoff, "The 'Family Growth in Metropolitan America' Study: A Progress Report," in Clyde V. Kiser (ed.), Research

SELECTED BIBLIOGRAPHY

in Family Planning, Princeton University Press, Princeton, N.J., 1962, pp. 185-191. Charles F. Westoff, "Religion and Fertility in Metropolitan America," in Thirty Years of Research in Human Fertility: Retrospect and Prospect, Milbank Memorial Fund, New York, 1959, pp. 117134. Charles F. Westoff, "The Social-Psychological Structure of Fertility," Proceedings of the International Population Conference, Vienna, 1959, pp. 355—366. Charles F. Westoff, Robert G. Potter, Jr., and Philip C. Sagi, "Some Estimates of the Reliability of Survey Data on Family Planning," Population Studies, vol. 15, no. 1, pp. 52-69, July, 1961. Charles F. Westoff, Robert G. Potter, Jr., and Philip C. Sagi, "Some Selected Findings of the Princeton Fertility Survey," Demog­ raphy, vol. 1, no. 1, pp. 13-135, 1964. NATIONAL SURVEYS OF THE POPULATION STUDIES CENTER

Morris Axelrod, Ronald Freedman, David Goldberg, and Doris Slesinger, "Fertility Expectations of the United States Popula­ tion: A Time Series," Population Index, vol. 29, no. 1, pp. 2531, January, 1963. Ronald Freedman, David Goldberg, and Doris Slesinger, "Current Fertility Expectations of Married Couples in the United States," Population Index, vol. 29, no. 4, pp. 366-391, October, 1963. Ronald Freedman, David Goldberg, and Doris Slesinger, "Fertility Expectations in the United States: 1963," Population Index, vol. 30, no. 2, pp. 171-175, April, 1964. SOCIAL RESEARCH, INC. STUDIES

Lee Rainwater, And the Poor Get Children, Quadrangle Books, Chicago, 1960. Lee Rainwater, Family Design: Marital Sexuality, Family Planning, and Family Limitation, Aldine Publishing Co., Chicago, 1965.

INDEX Accidental pregnancies, see Fertility planning status Age at marriage, factors affecting, 322 childlessness, affected by, 165 trends in, 322 American Institute of Public Opinion, cited 34 Axelrod, cited, 48 Badenhorst, cited, 44 Births, actual, 1955-1960, 76 by age at first birth, by cohort, 311 by age at marriage, trends in, by cohorts, 314 by age of mother at birth of child­ ren, birth order, 302 compared with actual rates, 19 declining, 309 by education of wife, 18 by education of wife and husband, 93, 94 by employment of wife, and fertility, 107, 109 expected in 1955-1960, 16 Catholic, 18 by farm wives, 105 by less educated, 17 and number of children wanted, by income, 106, 171 Protestants, 18 expected, total by age of wife, 57-58 by age at marriage, 59 compared with births wanted, 64 by duration of marriage, 61 by fecundity, 56, 67, 167, 170, 171 by frequency of sacraments for Catholics, 83 by income of husband, 26, 28, 32 1955 compared with 1960, 64 nonwhite compared with white, 334, 336 by occupation of husband, 28, 114 reliability of, 10, 13, 25-26 by religion, 72, 73, 113, 114 by size of place of residence, 28 family size, 111 general trends in spacing of, 300 by importance of religious activities in daily life of family, 89 by income, 103 by income of husband, 104 inter-birth intervals, 324 by education, region, and location, 332

trends in length of, 307 by wife's education, 327 interval between first marriage and first birth, 323 education and, 324 religion and, 324-325 by median age at birth as means of estimating trends, 305 by 1960, total expected, by religion, 71, 12 projection of, 1960-1985, 397 by reasons given for, 102 by region, 121 spacing, problems of measurement, 301 age at marriage, decline in, 309 long-time trends in, 302 reasons for, 102 summary, 333 subfecundity, as reason for low birth expectations, 92 by income, 106 vs. wanted births, 100 by wife's work history, 107 and fertility, 107-109 and religion, 110 Birth expectations, see Births Birth intervals, trends in, length of, 307 Birth rate cumulative, by cohort, 312 nonwhite, 334 by place of residence, 342-343 projections, 379, 397 cohorts compared, 382 crude rate trends, 1955-1985, 401 effects of cohort rates on, 395 rate for all women, 390 by region, 339-340 Bureau of the Census, Current Popula­ tion Survey, March, 1957, cited, 77 1950 Census, cited, 377 1960 Census, cited, 339 Catholic Catholic-Protestant differences in contraceptive operations, 141-142, 145 family size, trends, 77, 79 remedial operations, 152 use of contraception, 188, 217, 278 expected births, 72, 113 by age, 21 by church attendance, 82

INDEX Catholic (cont.) expected births (cont.) by education, 84—86, 96-97 by frequency of sacraments, 83 by marriage in Church, 73 by religious activities in child­ hood, 87 trends in, 98 teachings adherence to, by regularity of church attendance, 285 re contraception, 70 re family size, 69 re operations preventing conception, 145 Childless couples by fecundity and by wife's age, 165 Childlessness, defined, 12 attitudes toward, 162 among Hutterites, 164 Conception timing and reasons for, 196-198 unwanted, cases cited, 229, 249 wanted, 222 Contraception attitudes toward, 174-175 cases cited, 187 by education, 177 by income, religion, and occupa­ tion, 202-203 effectiveness of, 288, 298-299 Enovid, licensed by Food and Drug Administration, 277 and exposure to risk of pregnancy, method of computing, 290 failure rate, 291-294 of condom, 297 Princeton findings compared with present study, 293-295 of rhythm, 297 trends in, 295 methods of, 276-277, 298-299 nonwhite compared with white, 360 pill, adoption of, 285 attitudes toward, 286 cases cited, 287 by religion, 282-283 rhythm, 175-176 trends in, 283, 291, 297 use of, 209-210 and accidental conceptions, 140, 289 defined, 289-290 by age, 184 by cohort, 207

discrepancies between 1955 and 1960 studies, 212 by fecundity, 156, 216 by income of husband, 185 nonwhite compared with white, 353-354, 367 by number of births, 195 by religion, 188, 217, 278 trends in, 207, 295 Contraceptive operations, see Opera­ tions preventing conception Desired family size, 37, 166, 192 age at marriage, 65, 66 cases cited, 38, 39, 40 Desired vs. expected family size, 64 Detroit Area Survey, cited, 8, 29, 38 Education effect on ability to estimate births, 24 effect on attitudes toward birth control, 177 Educational attainment and reasons cited for contraceptive operations, 148 Excess fertility, see Fertility planning status Expected family size, see Births expected, total Family planning, see Fertility planning status Family size. See also Births attitudes of husbands: changes between 1955 and 1960, 181 birth rates, 1955 compared with 1960, 373 birth timing, 222 completed families, 372 contraception, varying use of, 195 effect of early marriage on, 270 expected, 377 expected, by income, 26 future predicted, 387-388 of completed family, by religion, 113 increase between 1955 and 1960, 377 ideal, 33 by sex and income, 35 in 60's and 40's, 34 by wife's work experience, 111 ideal and desired as affected by religion, 90 reasons for, 91 low size of family, reasons given, 68

INDEX

method of estimating, 385 nonwhite compared with white, 54, 56 reasons given for, preference, 53 by religion, 69 by religion and occupation, 114 urban-rural outside Standard Metropolitan Area, 116 wishes concerning, rural-farm wives, 119 Fecundity and subfecundity by age, 161, 165 cases cited, 126, 129-131 and childlessness, 165 classification system, 128 comparison of 1955 and 1960 studies, 131 degrees of, 132-133 by education, 158-159 effects of, summarized, 172-174 nonwhite compared with white, 134, 351-352 by number of births expected, 56, 67, 92, 168 compared with births wanted, 100, 167, 169-171 by income, 106 by religion, 158 subfecundity permanance of, 217 reports of, 125 trends in, 155, 160 by use of contraception, 156, 206, 216 by wife's work experience, 258 Fertility completely planned, comparison of 1955 with 1960, 264 cases cited, 228 size and place of residence, 259 region, 259 differentials declining, 116 urban-rural, 116 by religion, 70 future trends among less educated, 27 by husband's occupation, 112 husband's changes in attitudes between 1955 and 1960, 181 nonwhite compared with whites, 363-365 contraception, use among, 353-354 cases cited, 355 methods, 360 planning status, 361

by education, 343-357 excess fertility, 349-350 fecundity impairment, 351 fertility in northern cities, 344 cases cited, 345-346 parity, inversely related to fertility, 262 by place of origin, 346 by place of residence, 362 planning, irregular use of, 367 pregnancy, planning, 349-350 proportion of couples having excess fertility, 366 by region, 356, 357 sterility, 352 subfecundity among, 351 summarized, 369 overestimation of, 21 partly planned, 225 by place of residence, 115 planned trends in cohorts, 261 planning status trends, 260 successful planning defined, 226 summarized, 369 variations by length of residence, 120 wives attitudes, changes between 1955 and 1960 studies, 180 working wives, affect on, 258 Fertilityplanning status, 221 and accidental conceptions, 223 characteristic patterns, 221 completely planned fertility cases cited, 228 comparison of 1955 and 1960 studies, 264 by region, 259 by size of place of residence, 259 by duration of marriage, 243 by education, 246 excess fertility, 225 cases cited, 235-237 by education, 27, 247 among nonwhites, 349-350, 366 compared with whites, 363-365 by place of residence, 362 and number of births, 238 by occupation of husband, 272 reasons for, 243 trends in, 267 1955-1960 increase explained, 268-269 and irregular use of contraception, 246 long-run effects, 394 nonwhite compared with white, 361

INDEX Fertility planning status (cont.) by number of births, 230 partly planned fertility, 225 cases cited, 232-233 partly unplanned fertility, 234-235 by region, 259 by religion, 252-253, 255, 266 by size of place of residence, 259 socioeconomic characteristics sum­ marized, 256-258 summary of findings, 272-275 trends in, 260 Frazer, cited, 342, 344, 368 Freedman, 48, 50 Fundamentalists, births expected by, 74 Goldberg, cited, 29, 300, 303, 304 Grabill, cited, 300, 303, 304 Hysterectomies, number of, distribution of incidence of, 151 Ideal vs. expected family size, 50-53, 64 Ideal family size, 33 in 40's and 60's, 34 by sex and income, 35 Immigration, effects on population growth, 371 Indianapolis Study of 1941, 7, 163, 239 Interviews, described, 5 Jewish couples birth expectation, 72 contraceptive operations among, 141142 remedial operations, socioeconomic characteristics related to, 153 Kiser, cited, 351 Kish, cited, 5 Lee, cited, 348, 351 Lenski, cited, 78 Low-size family expectation, reasons given, 68 Marriage rates, method of predicting, 388 Number of children desired changes with wife's age, 45 by husband, 41 at marriage, compared with present, 44 by wife, by race and income, 41 Number of children born, 46

Occupation and birth expectation, 28 birth intervals, 325 Operations preventing conception contraceptive operations cases cited, 138 Catholic teachings re, 145 by education, 147, 148 by region, 147 by size of place of residence, 147 by wife's work experience, 146 in Puerto Rico, 139 reasons given for, 136, 137, 139, 148 by region, 144, 147, 149 by religion, 141—142, 145 by size of place of residence, 147 vasectomy, 149 by wife's work experience, 146 remedial operations effects on fertility, 151 hysterectomy, 151 number of, 150 by religion, 152 by socioeconomic characteristics, 153 trends in, 154 Place of residence and birth expecta­ tion, 28 Planned families. See also Fertility planning status changing status of plans, 243 characteristics of, Indianapolis Study, 239-241 characteristics, summarized, 232 completely planned among farm wives, 266 trends by religion, 266 partly planned, cases cited, 232-233 size compared with unplanned, 230 socioeconomic characteristics, sum­ marized, 256-258 unplanned, cases cited, 232, 234, 235 accidents in, 234 characteristics, summarized, 234235 Population growth effects of immigration on, 371 effects of younger marriages on long-run, 316 short-run, 317 Population Studies Center, University of Michigan, cited 49, 81 Potter, cited 289, 295 Powdermaker, cited 342 Pregnancies unwanted, means of prevention, 250

INDEX variation with religion, 254 Princeton Study, cited, 3, 7, 29, 37, 42, 255, 265, 292 compared with present study, 131 family planning failure rates com­ pared with present study, 293, 295 Projected population growth, sum­ marized, 401 Protestant attitude toward birth control, 178 contraceptive operations among, 141, 142 expected births, 72 expected size of family, 113 teaching about birth control, 70 wives, effects of childhood religious activities, 88 Puerto Rico, contraceptive operations in, 139 Rainwater, cited, 199 Region of residence, effects on use of contraceptives, 185 Reliability of expectations by nonwhite wives, 50 Religion and inter-birth intervals, 325 Religious preference effects on ability to estimate birth, 23 Remedial operations, see Operations preventing conception Sample, 1955 and 1960 compared, 4 Sample of present study described, 3

Schachter, cited, 300 Social Research, Incorporated, 8 Spacing of births, see Births Sterility. See also Fecundity and subfecundity cases cited, 129-130 caused by operation, 134 defined, 128 subfecundity, distribution of, 190 nonwhite, 352 Subfecundity, defined, 14. See also Fecundity and subfecundity by education, 159 effects on wife's employment, 258 expected births, 56 expected family size, 67 extent of, 14-15 by wife's age, 161 number of conceptions before re­ ported, 156 reason for low birth expectations, 92 trends in, 155, 160 use of contraceptive and subfecundity, 156, 206 Tietze, cited childlessness among Hutterites, 164 Timing of births, see Births

Westoff, cited, 29, 265, 280; 1963 29, 87 Whelpton, cited 50, 384, 391