From the Ground Up: Translating Geography into Community through Neighbor Networks [Course Book ed.] 9781400830572

Citation preview

From the Ground Up

From the Ground Up T R A N S L AT I N G G E O G R A P H Y I N T O COMMUNITY THROUGH N E I G H B O R N E T WO R K S

Rick Grannis

PRINCETON UNIVERSITY PRESS PRINCETON AND OXFORD

Copyright  2009 by Princeton University Press Published by Princeton University Press, 41 William Street, Princeton, New Jersey 08540 In the United Kingdom: Princeton University Press, 6 Oxford Street, Woodstock, Oxfordshire OX20 1TW All Rights Reserved Library of Congress Cataloging-in-Publication Data Grannis, Rick, 1965– From the ground up : translating geography into community through neighbor networks / Rick Grannis. p. cm. Includes bibliographical references and index. ISBN 978-0-691-14025-4 (acid-free paper) 1. Community life. 2. Neighborhood. 3. Communities. 4. Ecology. I. Title. HM761.G73 2009 307.33620973—dc22 2009003739 British Library Cataloging-in-Publication Data is available This book has been composed in Minion Printed on acid-free paper.  press.princeton.edu Printed in the United States of America 10 9 8 7 6 5 4 3 2 1

Contents

List of Illustrations and Tables

ix

Prologue

xv

CHAPTER ONE Neighborhoods and Neighboring

1

Geography and Community It’s the Kids, Stupid! Overview of the Book

1 4 8

CHAPTER TWO The Stages of Neighboring

17

Neighboring: A Superposed Relation Stage 1 Neighboring Stage 2 Neighboring Stage 3 Neighboring Stage 4 Neighboring Main Points in Review

17 20 20 23 25 27

CHAPTER THREE Reconceptualizing Stage 1 Neighboring

28

Proximity Boundaries Face Blocks Tertiary Face Blocks Intersections Main Points in Review CHAPTER FOUR Reconceptualizing Stage 1 Neighbor Networks Layers of Complex Network Structures T-Communities and Islands Main Points in Review

28 29 31 32 34 35 37 37 42 46 v

CONTENTS

CHAPTER FIVE Selection and Influence Selecting Homophilous Immediate Neighbors Influence Homophily and Influence Acting on Different Stages of Neighboring Main Points in Review CHAPTER SIX Respondents, Interviews, and Other Data Gang Neighborhood Ethnography and Interviews Overview of the Other Data Collection Events Structured Interviews Cognitive Mapping and Alternatives Data Collection in 68 Los Angeles Neighborhoods Adaptive Link-Tracing The Second Los Angeles Data Collection College Town Census and Resample Administrative Data Main Points in Review CHAPTER SEVEN Selecting Stage 1 Neighbors Selecting Racially Homophilous Tertiary Street Neighbors Accepting Heterogenous Higher-Stage Neighbors A Dialogue with Administrative Data Segregating Tertiary Street Networks Tertiary Street Network Borders The Impact of a Single Tertiary Street Connection Main Points in Review CHAPTER EIGHT Unintentional Encounters The Substantive Reality of Passive Contacts The “Lived” Experience of Tertiary Street Networks A Note about Large, Multiunit Complexes Main Points in Review vi

48 48 52 56 57 59 60 61 61 62 65 66 67 68 70 72 73 73 76 78 79 84 89 90 93 93 96 105 107

CONTENTS

CHAPTER NINE Stage 3 Neighbors and Tertiary Streets Tertiary Street Proximity and Stage 3 Neighbors Tertiary Street Networks and Stage 3 Neighbor Networks More Than Proximity Main Points in Review CHAPTER TEN The Importance of Neighbor Networks Three Degrees of Neighboring A Note about the Exhaustive Census Neighboring Is a Family Relation The Importance of Convenient Availability Main Points in Review CHAPTER ELEVEN Network Influence Theory Social Influence Network Theory Beyond Density The Horizon of Observability Structural Cohesion Merely a Mechanism? Main Points in Review CHAPTER TWELVE Influence Networks in a College Town T-Communities, Children, and the Horizon of Observability T-Communities and Social Control Neighbor Influence and T-Community Culture Main Points in Review CHAPTER THIRTEEN Influence Networks in a Gang Barrio Geographic Neighborhood and Sociological Neighborhood Neighborhood Community and Tertiary Street Networks An Efficacious Neighborhood Neighborhood Efficacy as a Function of Influence Networks

109 109 113 119 127 129 129 134 135 139 144 148 148 151 155 158 159 161 162 162 164 166 176 178 178 180 182 184 vii

CONTENTS

Influence Networks as a Function of Tertiary Street Networks Main Points in Review

187 190

CHAPTER FOURTEEN Implications

192

Summary What It All Means

192 197

APPENDIX Survey Instrument

201

Notes

207

References

219

Index

237

viii

Illustrations and Tables

ILLUSTRATIONS FIGURE 3.1: Face block FIGURE 4.1: Stage 4 neighbor network FIGURE 4.2: Stage 3 neighbor network, which induces stage 4 neighbor network FIGURE 4.3: Stage 2 neighbor network, which induces stage 3 neighbor network FIGURE 4.4: Stage 1 neighbor network, which induces stage 2 neighbor network FIGURE 4.5: T-communities FIGURE 4.6: Tertiary street island FIGURE 6.1: Distribution of t-community sizes FIGURE 7.1: Pasadena’s tertiary streets by residential demographics FIGURE 7.2: Enlarged subsection of figure 7.1 FIGURE 7.3: For spatially adjacent block groups, distribution of absolute difference in population that is white FIGURE 7.4: For spatially adjacent block groups, distribution of absolute difference in population that is Asian FIGURE 7.5: For spatially adjacent block groups, distribution of absolute difference in population that is Hispanic FIGURE 7.6: For spatially adjacent block groups, distribution of absolute difference in population that is black FIGURE 7.7: Demographic differences between spatially adjacent block groups by number of shared tertiary streets FIGURE 8.1: Passive contacts by type FIGURE 8.2: Sample cognitive map 1 FIGURE 8.3: Sample cognitive map 2 FIGURE 8.4: Cognitive maps compared to tertiary street island map

32 38 39 40 41 44 46 71 80 81 86 86 87 87 91 94 99 101 104 ix

I L LU S T R AT I O N S A N D TA B L E S

FIGURE 8.5: Cognitive maps compared to tertiary street island map (continued) 105 FIGURE 8.6: Cognitive maps compared to tertiary street island map (continued) 106 FIGURE 8.7: Composite of cognitive maps compared to tertiary street island map 107 FIGURE 9.1: Neighbors known as a function of distance (in house-steps) 110 FIGURE 9.2: Adaptive link-tracing network mapped onto t-community 114 FIGURE 9.3: Complete census of neighborhood network 117 FIGURE 9.4: Complete census of neighborhood network mapped onto census tracts 118 FIGURE 9.5: Complete census of neighborhood network mapped onto elementary school catchment areas 119 FIGURE 9.6: Complete census of neighborhood network mapped onto t-community 120 FIGURE 10.1: Sample adaptive link-tracing network 130 FIGURE 10.2: Distribution of path lengths in samples of size 10 132 FIGURE 10.3: Sample adaptive link-tracing network, respondents only 133 FIGURE 10.4: Why your neighbors know more neighbors than you do 135 FIGURE 10.5: Household neighbor network sizes, by presence of children 136 FIGURE 10.6: Neighboring relations, by presence of children in either household 138 FIGURE 10.7: Percentage of households reachable at each neighbor network step, by presence of children 140 FIGURE 10.8: Frequency of monitoring neighbors’ children in spontaneous playgroups 143 FIGURE 10.9: Frequency of socialization with neighbors 145 FIGURE 11.1: Example networks with identical order and size but disparate quantity and length of paths 152–53 FIGURE 12.1: “My neighbors share my values” at times 1 and 2 172 FIGURE 12.2: Comparison of models’ predictive power for “My neighbors share my values” 173 x

I L LU S T R AT I O N S A N D TA B L E S

FIGURE 12.3: “People are the best part of this neighborhood” at times 1 and 2 FIGURE 12.4: Comparison of models’ predictive power for “People are the best part of this neighborhood” FIGURE 13.1: Gang neighborhood tertiary street network FIGURE 13.2: Number of alters respondents believed to be “for” them FIGURE 13.3: Alters identified as “for” respondents by distance between alters’ and respondent’s residences

174 175 180 185 189

TABLES TABLE 2.1: Guttman Scale of Neighboring Relations TABLE 6.1: Overview of Data Collections TABLE 6.2: Neighbors Known, by Data Collection TABLE 7.1: Similarity of Neighbors in Previous Neighborhood, by Previous Neighbors’ Influence on Decision to Relocate TABLE 7.2: Similarity of Neighbors in Current Neighborhood, by Influence on Relocation Decision, Desire to Remain in Neighborhood, and Awareness of Different-Race Neighbor TABLE 7.3: Demographic Variability Accounted for, by Island or T-community TABLE 7.4: Block Groups’ Demographics as a Function of Demographics of Island or T-community TABLE 7.5: Block Groups’ Demographics as a Function of Demographics of T-community, by Presence of Children TABLE 7.6: Percentage Racial Difference between Spatially Adjacent Block Groups, by Whether They Are Barriers, Borders, or Internal Adjacencies TABLE 8.1: How Respondent Met Neighbor, by Percentage Identified as Passive Contact TABLE 8.2: Overview of Cognitive Maps, by Data Collection TABLE 9.1: Identified Neighbors by Same T-community or Census Tract, by Data Collection TABLE 9.2: Neighbors Met by Passive Contacts, by Data Collection

19 62 63 74

75 82 83 84

88 97 102 112 113 xi

I L LU S T R AT I O N S A N D TA B L E S

TABLE 9.3: Neighbors Met by Passive Contacts, by Same Census Tract or T-community, by Data Collection 113 TABLE 9.4: Adaptive Link-Tracing Sequences, by Same Census Tract or T-community, by Data Collection 116 TABLE 9.5: Probability of Those Identified Living in Same T-community or Census Tract, by Distance from Respondent, College Town Census, All Neighbors 122 TABLE 9.6: Expected and Actual Number of Neighbors, by Same T-community or Census Tract, College Town Census, All Neighbors 123 TABLE 9.7: Probability of Those Identified Living in Same T-community or Census Tract, by Distance from Respondent, College Town Census, Farthest Neighbors Only 124 TABLE 9.8: Expected and Actual Number of Neighbors, by Same T-community or Census Tract, College Town Census, Farthest Neighbors Only 124 TABLE 9.9: Probability of Those Identified Living in Same T-community or Census Tract, by Distance from Respondent, College Town Follow-up, All Neighbors 125 TABLE 9.10: Expected and Actual Number of Neighbors, by Same T-community or Census Tract, College Town Follow-up, All Neighbors 126 TABLE 9.11: Probability of Those Identified Living in Same T-community or Census Tract, by Distance from Respondent, College Town Follow-up, Farthest Neighbors Only 127 TABLE 9.12: Expected and Actual Number of Neighbors, by Same T-community or Census Tract, College Town Follow-up, Farthest Neighbors Only 127 TABLE 10.1: Distribution of Neighboring Relations, by Presence of Children in Either Household 137 TABLE 10.2: Proportion of Neighboring Relations Providing Specific Services, Households with Children Only 142 TABLE 10.3: Proportion of Neighboring Relations Providing Specific Services, All Households 144 TABLE 10.4: Proportion of Neighboring Relations Providing Services Who Were Met through Passive Contact, All Households 146 xii

I L LU S T R AT I O N S A N D TA B L E S

TABLE 10.5: Proportion of Neighboring Relations Providing Services Who Were Met through Passive Contact, Households with Children Only TABLE 12.1: Similarity in Residents’ Beliefs about Neighborhood, by Neighbor Network Steps Separating Them TABLE 12.2: Neighborly Interactions and Perception of Neighborhood, All Households TABLE 12.3: Neighborly Interactions and Perception of Neighborhood, Households with Children Only

147 167 168 169

xiii

Prologue

IT WAS 1995 and I sat in a Catholic church in a large city in Southern California. The mayor and the entire city council were there to respond to local residents, who were angry about a stop sign that had been placed at an intersection in their community a few months earlier. For three exhausting hours, mothers and fathers pleaded with their political leaders to remove the stop sign, which they claimed had destroyed their neighborhood community. Unfortunately, their lay presentations were obviously incoherent. They clearly had no idea why the stop sign had wrecked their lives. Person after person came to the microphone and tried to offer up reasons, but they all sounded hollow. Everyone agreed, however, that neighborhood life had been much better before the stop sign was put in. The politicians had come armed with environmental impact reports. They had also brought an urban planner as a hired gun. Slowly, patiently, and methodically, the mayor, the city council, and their urban planner showed that, not only had traffic decreased substantially on the east-west street traversing the neighborhood where the stop sign had been placed, but it had even decreased slightly on nearby north-south cross streets. After about three hours, the meeting really began to fall apart. A curious city council member began pressing those who came to the microphone for their addresses. It soon became clear that individuals from several blocks away were there protesting. The city council member asked that only those who had been directly affected by the stop sign be allowed to speak. Many individuals said that they and their families had been affected even if they didn’t live on the street with the stop sign. Another city council member retorted that they were just trying to just stir things up; perhaps they were really upset over completely unrelated issues. This suggestion divided the crowd of protesters. After a few last enraged yells, many left and went home. xv

PROLOGUE

From that moment of the meeting on, a much smaller audience stayed “on topic.” Only those whose addresses fronted the street with the stop sign were allowed to speak. As the meeting pressed on into the night, the politicians and their aids “proved” over and over to these misguided constituents that what they believed was true could not be true. That one little stop sign could not have had the impact they claimed it did. The meeting remained calm. Children played in the aisles, and babies cried. The city council thanked those who attended for their participation and reassured them that they were always ready to listen to their constituents. When the meeting finally ended, a few in the crowd politely clapped. Most just went home. · · · · · It was my ethnographic experience in this community in Southern California that first made me aware of the importance of both small streets and neighbor networks, although long before this event I had been fascinated by the large gang that dominated this city. From my first encounter with them while teaching at a local middle school (where the vast majority of the student population claimed some affiliation with them), I saw that the gang was well received, or at least highly tolerated, in their neighborhood. The school even offered an optional after-hours class, officially titled Practical Citizenship, but unofficially titled How to Be a Positive Gang Member. The school’s administrators had decided, if you can’t beat them, join them. The members of the gang in the neighborhood covered a wide spectrum of ages. While many were in their twenties or thirties, with a few even older veteranos hanging around, no one joined the gang at that age. They joined when they were in middle school, before they could drive. They had contact with gang members at school, but unless that contact was reinforced back home, as they walked the streets of their neighborhood, the draw of the gang didn’t usually take. Seeing this pattern play out, I suspected the importance of pedestrian access, and I began to explore its relationship to the gang. A vice-principal who became interested in my study helped me do a rough analysis, and we found out that the one-third or so of the student body who had little or no affiliation with the gang lived in areas that were not conxvi

PROLOGUE

nected to the rest of the school’s catchment area by walkable streets. This made perfect sense to both the vice-principal and me. If they were going to join the gang, these middle school recruits had to be able to casually walk to meet gang members. · · · · · A few days after the stop sign meeting, I scheduled an appointment with one of the city council members and suggested that the stop sign, by slowing traffic on the north-south street, had actually made it easier for members of the gang who lived to the east to cross the street. A wall had come down, and the gang’s neighborhood was now expanding to include new neighbors, new juvenile neighbors, to the west. The city council member thanked me for my time but suggested that mine was an “ivory tower” theory. He had to live in the real world, where a stop sign couldn’t stop—or help—the spread of such a powerful gang. The gang had been kept out of the neighborhood before by the vigilance of the upstanding residents who lived there. If the gang was moving in now, it was because these good citizens had moved out and had been replaced by less worthwhile people. A week later, I offered my idea about the stop sign to the community rights group that had organized the parents’ meeting. As with the city council member, I suggested that the slowing of traffic on the northsouth street had made it possible for the gang to move west. The unwanted changes that people had noticed in their community reflected its silent merger with the gang’s territory. The community rights activists were less cordial than the professionally smooth councilman had been. They were not concerned about what the gang did over in that other neighborhood. They were concerned about the increase in drinking and delinquency among local youth. Sure, some of the middle-schoolers were sporting the gang’s insignia, but those were the bad apples you have in every neighborhood. The community activists went back to studying their environmental impact reports. · · · · · That meeting occurred 14 years ago. Now the neighborhood in question is a well-established area of the gang. xvii

PROLOGUE

Ever since then, I’ve watched for instances of minor physical changes that affect communities. I’ve watched communities plant hedges and put up temporary barriers on streets. I’ve watched bike paths go in and come out. I’ve watched communities change stoplights and stop signs and bus stops. I’ve always watched for the sociological impact of these changes. In many cases, it has been apparent. Sometimes it has troubled local residents. In no case, however, have residents attributed improvement or deterioration in their neighborhood to a hedge or a bike path or a bus stop. They don’t believe that in our modern world of the Internet and cell phones, hedges and bike paths and bus stops could possibly have an impact. It’s true they have little impact on adults, especially adults in cars or those easily navigating public transportation systems. They don’t have too much effect on high school students either. But sixth-graders who are potential gang recruits and preschool children who grow up without seeing a face of another race don’t typically move about that way, and they don’t live exclusively in a cyberworld, especially if they are poor. They play with the next-door neighbors or the kids down the block; and, while they play and learn what the world is like, their parents often stand on the street, watching them and talking to each other about trivia and work and sometimes how we raise our children and what is good and bad in the world. Unplanned relationships form, neighbor networks grow, and community is born from the ground up.

xviii

From the Ground Up

C H A P T E R

O N E

Neighborhoods and Neighboring

GEOGRAPHY AND COMMUNITY Human behavior necessarily occurs within (or must transcend) physical space. Nowhere is this truer than in residential life. As real-estate agents and homeowners (especially those with children) often declare, where one makes one’s home matters almost as much as what one does inside it. In the rapidly shrinking world of the twenty-first century, psychologists, economists, political scientists, and sociologists still acknowledge the importance of the neighborhood context. Not all neighborhoods are alike, however. Some neighborhoods are characterized by high levels of effective community. They offer social capital to their residents, a social organization that facilitates and coordinates cooperative action for mutual benefit, which allows them to deal with daily life, seize opportunities, reduce uncertainties, and achieve ends that would not otherwise be possible.1 This social organization is a resource that is not individually attainable because social capital is not a characteristic of individuals; it is a supraindividual property of social structure, and it seems to be particularly well grounded in neighborhood communities.2 Sources of social capital tied to the neighborhood community are analytically distinct from, and are as consequential as, the more proximate family processes and relationships occurring in the home. Some neighborhoods develop a further layer of mutual trust and shared norms, values, and expectations,3 beyond the resource potential of neighbor networks, which allows them to use these networks to achieve desired outcomes. Collective efficacy occurs when members of a collectivity, with social capital resources, believe they are mutually able and willing to use them to achieve an intended outcome.4 The distinction is a subtle, but important, one. A neighborhood may have social capital resources available for its constituent residents to use, but they may not trust the willingness or ability 1

CHAPTER ONE

of their fellow residents to use these resource networks for the collective good, or they may not even be certain that they agree on what the collective good is. From a less positive perspective, neighborhoods show remarkable continuities in patterns of criminal activity. For decades, criminological research in the ecological tradition has confirmed the concentration of interpersonal violence in certain neighborhoods, especially those characterized by poverty, the racial segregation of minority groups, and the concentration of single-parent families. Even in neighborhoods with less socioeconomic or racial isolation, crime rates persist despite the demographic replacement of neighborhood populations.5 In addition, neighborhoods not only determine one’s exposure to crime and violence,6 but also a host of less tangible deleterious factors7 that contribute to the development of an urban underclass, signs of social disorder that lead residents to perceive their neighbors as threats rather than as sources of support or assistance.8 Researchers have taken a growing interest in the role of neighborhoods in shaping outcomes for children, families, and neighborhood residents in general.9 These “effects” have included phenomena ranging from child and adolescent development10 (e.g., abuse and maltreatment, school completion11 and achievement,12 drug use,13 deviant peer affiliation, delinquency14 and gangs,15 adolescent sexual activity16 and pregnancy,17 childbearing18 and parenting behaviors,19 etc.) to concentrated disadvantage and its many corollaries (restricted economic attainment20 and labor market failure, crime21 and violence,22 physical disorder,23 the perpetuation of racism,24 to name just a few). The conclusion reached by all of these studies is that neighborhoods influence our behavior, attitudes, and values.25 They shape the types of people we will become and expose us to or shield us from early hazards that might restrict the opportunities available to us later in life. After our homes, and in conjunction with them, neighborhoods are where we first learn whether the world is safe and cooperative or inchoate and menacing. The neighborhood one lives in matters. Neighborhoods matter, but different neighborhoods matter in different ways. Different neighborhoods have different effects, of different magnitudes. Some neighborhoods have almost no effect. For the researcher, neighborhoods cluster outcomes that cannot be accounted for 2

NEIGHBORHOODS AND NEIGHBORING

in terms of the characteristics of the individuals or households currently residing in them. It is as if neighborhoods have personalities, enduring characteristics that survive the replacement of their constituent residents.26 These neighborhood effects, however, necessarily involve a geographic context. Thus, to analyze and understand them, neighborhoods necessarily require a geographic equivalent. Researchers have used a wide variety of such equivalents. In fact, “urban social scientists have treated ‘neighborhood’ in much the same way as courts of law have treated pornography: a term that is hard to define precisely, but everyone knows it when they see it.”27 Apparently, however, researchers often don’t know it when they see it. Miller’s (1999) survey suggests that the modifiable areal unit problem (MAUP)28 exists primarily because analysts decide beforehand on the spatial units they will use when they study a phenomenon.29 Having done so, they reach conclusions about the phenomenon that are hopelessly prejudiced by their choice of spatial unit. While many statistical techniques and error-modeling approaches have been used to counteract, reduce, or remove the effects of MAUP, Miller argues that the ultimate solution has to involve a behaviorally oriented definition of neighborhood for use in the practical measurement of neighborhood factors. One needs better intuitions about the general nature of neighborhoods, not better statistical methods. The very existence of the modifiable areal unit problem evidences that theory has taken a back seat. Those researchers30 who have developed methods for creating optimal analytic units with respect to predefined objective functions note correctly that MAUP would be irrelevant if neighborhood equivalents were chosen for theoretical reasons rather than administrative convenience.31 Despite this need for a conception of neighborhoods that is tied to the behaviors and interactions of residents that produce these effects, however, when a geographic definition of neighborhood is required for the purpose of quantitative analysis, “most social scientists and virtually all studies of neighborhoods . . . rely on geographic boundaries defined by the Census Bureau or other administrative agencies . . . [that] offer imperfect operational definitions of neighborhoods for research and policy.”32 Administratively defined units such as census tracts and block groups do not directly measure, nor were they designed to measure, 3

CHAPTER ONE

the potential for interaction among residents, the primary process hypothesized to produce neighborhood communities and their effects. In most cases, the sheer ubiquity of data gathered by the Census Bureau or other administrative agencies (e.g., school districts, police districts) proves an overwhelming temptation for researchers. Theory succumbs to the preponderance of data. As a result, sociologists often treat neighborhoods as if they were only colored boxes on a map or sets of geo-referenced variables for use in a geographic information system (GIS).33 This approach often proves productive, but, like all plans, it emphasizes some aspects of what we are studying and de-emphasizes others. A focus on maps, especially maps based on census or administrative geography, emphasizes those aspects of neighborhoods and their residents that can be effectively displayed or associated with administratively defined polygons and ignores those that cannot. To understand the social-interactional aspect of neighborhoods, we may not have to think outside the box, but we do have to think about what’s inside it. In this book, I explore neighborhood communities and attempt to develop a more theoretically grounded neighborhood equivalent. Undoubtedly, neighborhood effects involve a geographic context. Neighborhood effects, however, are not produced by neighborhood geography. Nor are they—at least most of them—merely spatial effects, a byproduct or spurious confound of the geographic location of residents with particular demographic characteristics or psychological profiles. I argue that cataloguing neighborhood effects, by definition, hypothesizes that there exists a thing, a social entity, a neighborhood community, that has effects. Neighborhood effects are the product of these neighborhood communities. I argue that neighborhood communities and their effects emerge from neighboring interactions among their constituent residents.

IT’S THE KIDS, STUPID! Neighborhood communities and their effects involve children (e.g., child development and abuse, school achievement, delinquency, the development of racist attitudes) or adolescents (e.g., gangs, sexual activity 4

NEIGHBORHOODS AND NEIGHBORING

and pregnancy, drug use) either exclusively or primarily. For example, collective efficacy and segregation most powerfully affect children and adolescents. Furthermore, most discussion of neighborhood effects is developmental in nature, focusing on how neighborhoods, in addition to households, may manifest to us a world that is predictable and helpful or one that is capricious and dangerous. In doing so, they help mold the character of the adults we become. The relationship between neighborhood communities and children is a problematic topic, however, because social researchers are adults and, despite their attempts at objectivity, view neighborhoods first through their own adult eyes. Another obstacle is that protocols for treatment of human subjects typically prevent researchers from interacting with non-adults. Thus, the scholarly view of neighborhoods often reflects adults theorizing about neighborhoods, adults observing neighborhoods, and adults talking to other adults. Children, however, do not relate to neighborhoods in the same ways that adults do. Neighborhood communities are more relevant for households with children for at least three important reasons. First, households with children constitute about half of the population of American neighborhoods. According to the 2004 American Community Survey, conducted by the Census Bureau, slightly over half of all persons reside in households with children under 18 living in them.34 Of these households with children, almost half35 have very small children under six living in them. Thus, a majority of Americans live in households with minor children in them, and about a quarter of all Americans live in households with preschool children in them. Furthermore, researchers36 have consistently found that the number of neighbors known is higher for households with children. Thus, these households with children are involved in a much larger majority of neighborly interaction. Second, neighborhoods are especially important for households with children because children are less mobile, and thus more geographically dependent, than adults. Children and their playful interactions depend upon proximity much more than do adults and their interactions. Since children cannot drive and have little, if any, voice in decisions on where to live, they are forced to share lives with neighboring children even more than are their parents. For children, the street in front of their home is “the mediator between the wider community and the private 5

CHAPTER ONE

world of the family.”37 This is where children first learn about the world. They often play games in the middle of these streets38 and use them to walk pets and to ride bicycles, and the majority of their recreational activity occurs there.39 Sidewalks provide access between residence and schools and parks. As a result, the relationships children form primarily depend upon the opportunities to interact provided by walking arenas immediately surrounding them.40 Especially for young children, neighboring children are the most likely to become their playmates.41 Children are even affected by the extremely subtle geography of rain gutters and hedges.42 Thus, the networks of relationships they form will be much more dependent upon passive contacts occurring along them.43 Unlike children, adults have many venues for social relationships beyond their neighborhood, including work and voluntary activities. School-age children may have some of these opportunities, to the extent their parents allow. Preschool children, however, have few, if any, of these alternative social venues. Their lives are tightly bound by geography.44 Households with children are far more influenced by the norms and values of surrounding households with children than households in general are influenced by the norms and values of their surrounding neighbors. Your neighbors’ children are predisposed to become your children’s playmates and friends,45 your neighbors may become some of the role models they emulate,46 and thus the character of those living in neighboring households is a potentially powerful influence on your children. Neighboring parents may become intimately involved in the socialization of your children. Neighbors rear children side by side47 and together have the potential to co-create a safe and value-laden environment. Parents monitor their own children as well as those of their neighbors.48 Some neighborhoods expect residents to share values and to be willing and able to intervene on behalf of children. In these neighborhoods, residents expect each other to actively cooperate in the support and social control of children.49 Parents get to know the parents and families of their children’s friends, they observe children’s actions, both their own and their neighbors, in a variety of circumstances, they talk with other parents about their children, and they establish norms.50 Such structural and normative adult-child closure gives children social 6

NEIGHBORHOODS AND NEIGHBORING

support, provides parents with information, and facilitates control.51 The choice to live in a neighborhood is to some extent a choice to rear children together with one’s neighbors.52 Ultimately, a community of parents may develop around the community of children, mirroring it. People whose children play together form friendship relations based in part on that fact.53 While it is the children who are immobile, confined to neighborhoods, and most immediately affected by them, children’s geographic dependence encumbers their parents as well. A third reason neighborhoods are important for households with children is that most school-age children attend schools in their neighborhood. This pattern affects households because school quality plays an important role in the decision on where to live, both for families who currently have children and for those who think they might some day. Spatially defined neighborhoods typically determine the quality of the public schools one’s children have access to.54 For households with children, the quality of its school district may be one of the most important aspects of a residence under consideration.55 Parents often choose their neighborhood (and even pay more in housing and taxes)56 to gain access to particular school districts.57 Furthermore, school catchment areas may complement any effect of walking arenas, onto which they may be intentionally mapped, since children often walk to school. In review, when one considers neighborhood communities, most of the effects researchers concern themselves with involve children or adolescents. This results in large part because households with children constitute the majority of American households; because children are much less mobile than adults and this affects both them directly, their parents, and their families; and because most school-age children attend schools in their neighborhood. Children and their families are the quintessence of neighborhood life. During Bill Clinton’s 1992 presidential campaign, Democratic Party strategist James Carville hung a sign with three bullet points on it in Clinton’s Little Rock campaign office to keep everybody “on message.” The most famous, reminiscent of the KISS58 principle, was “It’s the economy, stupid!” When we study neighborhood communities and 7

CHAPTER ONE

their effects, it is worth hanging an imaginary sign in front of us to keep us on track. “It’s the kids, stupid!”

OVERVIEW OF THE BOOK The neighborhood one lives in matters. Neighborhoods influence behaviors, attitudes, and values; they shape outcomes for families, and they provide (or fail to provide) resources for residents to achieve or avoid outcomes collectively. While it is the community aspect of neighborhoods that influences norms and values and that generates social capital and collective efficacy, to analyze and understand neighborhoods requires a geographic equivalent for them. Many current neighborhood equivalents, however, imperfectly map onto the interactional processes generating the geographic outcomes being measured. In this book, I attempt to develop a more theoretically grounded neighborhood equivalent, mapping the neighboring interactions that produce neighborhood communities. I argue that neighborhood communities are geographically constrained because the interactions that produce them are geographically constrained. In fact, because children are much more geographically constrained than adults, children and their families are the quintessence of neighboring and neighborhood communities. More importantly, I argue that neighborhood communities are both geographically identifiable and have effects that persist through the replacement of their constituent residents because the networks of interactions that produce them, that translate neighbor-level interactions into neighborhood communities, are constrained by predictable urban geographic substrates. Finally, I show that commonly used administrative units are not those substrates. In chapter 2, I focus on the neighboring relation that forms the basis of neighborhood communities. I argue that the neighboring relationship develops in stages, each stage superimposed on the previous one. In the definitions used in this book, a stage 1 neighboring relation exists between two individuals if they are geographically available to each other. A stage 2 neighboring relation exists between two residents when 8

NEIGHBORHOODS AND NEIGHBORING

their lifestyles cause them to casually and unintentionally encounter each other and thus to have the opportunity to learn about each other through observation and to acknowledge each other’s presence or choose not to. A stage 3 neighboring relation exists between two residents if they have intentionally initiated contact. A stage 4 neighboring relation exists when two residents engage in a substantial activity that indicates mutual trust or a realization of shared norms and values, when they share a belief in each other’s willingness and ability to act together to achieve a common goal, when they influence each other, either actively or passively. Neither stage 3 nor stage 4 implies that the involved parties understand their relationship to be intimate or strong in the sense of having a friendship or an affective bond. These stages of neighboring develop in a logical order, with lower stages necessarily preceding higher stages. Two people cannot be neighbors in any sense if they are not geographically available to each other. While two people can be geographically available to each other and have no passive contacts, they cannot have such unintentional encounters unless they are geographically available to each other. Similarly, while two people can have passive contacts and choose to ignore each other or to actively avoid such contacts, they cannot interact without having encountered each other. Finally, while two people can interact at a superficial level only, they cannot develop mutual trust without some interaction. In chapter 3, I revisit stage 1 neighboring in more detail. This initial stage occurs when we are geographically available to each other. While this availability is often conceptualized in terms of neighborhood-sized distances and the absence of neighborhood-sized boundaries, I conceptualize stage 1 neighboring in terms of neighbor-sized distances and the absence of neighbor-sized boundaries. Neighboring is primarily dependent upon extremely short distances, walking arenas such as tertiary face blocks and tertiary intersections,59 because stage 2 neighboring, that is, passive or unintentional contact, relies upon pedestrian encounters. In chapter 4, I turn to the networks formed by the concatenation of these neighboring relations. Some (perhaps all, perhaps none) of stage 3 neighbor networks translate into stage 4 neighbor networks. Some (perhaps all, perhaps none) of stage 2 neighbor networks translate into 9

CHAPTER ONE

stage 3 neighbor networks. Some (perhaps all, perhaps none) of stage 1 neighbor networks translate into stage 2 neighbor networks. No neighbor networks, however, develop where there are not already stage 1 networks in place. This is why an accurate definition of stage 1 neighbor relations is so important. Most sociological studies of neighborhoods use administrative geography that implicitly defines two households to be stage 1 neighbors if they live in the same administratively defined area. It is not clear, however, whether residents of these spatially defined analytic units are geographically available to each other. In place of administratively defined areas, I define two new neighborhood equivalents, in terms of the concatenated network of walking arenas as represented by tertiary face blocks. These neighborhood equivalents differ only by the intersections they allow to connect face blocks with each other. The first, t-communities, uses only tertiary intersections, while the second, islands, uses all intersections.60 While I expect t-communities to have more pronounced effects, I include islands to measure the potency of nontertiary intersections. Both these new neighborhood equivalents focus on the potential for passive contacts, or unintentional encounters, and thus the interactional aspect of neighborhoods. In chapter 5, I conceptualize the foundations of neighborhood communities in terms of two forces: selection and influence. Households relocate, at least in part, to choose the type of households they want to have as stage 1 neighbors, relocating in favor of homophilous immediate neighbors, not homophilous neighborhoods. Since neighbors respond to household changes along their tertiary streets, the concatenation of these relocation events is necessarily delimited by the tertiary street network, and thus segregation patterns reflect it. Homophilous locational choice, however, cannot account for the entirety of neighborhood communities and their effects. A second community-generating force within neighborhoods consists of the flow and exchange of norms, values, beliefs, and influences among neighbors along their stage 4 networks. Neighborhood communities result from both the concatenation of homophilous locational choices and the exchange of norms, values, and beliefs among neighbors. Their correspondence is not additive, but rather sequential. Relocation, which is responsible for residential demographic differentiation, determines stage 1 neighbors and thus, of necessity, the higher stages of neighboring among which norms, values, 10

NEIGHBORHOODS AND NEIGHBORING

and beliefs flow. Locational-based neighborhood community effects such as segregation correspond to influence-based neighborhood community effects such as social capital and collective efficacy because, while each emerges from different stages of neighboring, the same concatenated, multistage processes guide both. In chapter 6, I review the data used in this book, most of which are original. The data were collected in several distinct settings, an ethnographic study of a gang barrio, and four large collections of structured interviews in 68 Los Angeles neighborhoods and a college town (a total of 70 neighborhoods). The 68 Los Angeles neighborhoods, 20 of which were revisited several years later, added statistical robustness to my study and used an adaptive link-tracing methodology to generate an interview chain that would spread out spatially great distances in order to determine what constrained neighboring relations. A region in the college town was the site of an exhaustive census that fully mapped the geographically embedded neighbor networks. This same region was revisited three years later to discover how these same neighbor networks had evolved. In chapter 6, I discuss these studies in detail, reviewing both the interviews and how neighborhoods and respondents were sampled. Finally, I discuss the administrative data I used to explore the same 70 neighborhoods in which I collected interviews and conducted ethnography. In chapter 7, I explore stage 1 neighboring relations and show that households in the study did indeed relocate so that their stage 1 tertiary street neighbors would be homophilous. They sometimes decided to move from their previous home if those who shared their tertiary streets were different from themselves; they considered with whom they would share tertiary streets in potential future residences; and, if their attempts at homophily proved unsuccessful, they desired to move once again. More than any other factor, respondents correlated racial similarity with homophily. If residents of different races did settle near each other, however, higher stages of neighbor networks generally developed without further reference to racial disparities. In other words, when residents racially segregated their neighbor networks, they typically did it by restricting their geographic availability, by segregating their stage 1 neighbor networks rather than higher stages of neighboring. However, while racial differences did not impede the translation of stage 1 tertiary 11

CHAPTER ONE

street neighbors into stage 3 actualized neighbors, linguistic differences did. To translate stage 1 neighboring relations into stage 3 neighboring relations, it helps to speak the same first language. In that same chapter, I use administrative data to examine racial segregation across the 70 neighborhoods I explore in my interviews and ethnography. I show that as residents racially segregate their stage 1 neighbor networks, discontinuities in the distribution of racial demographics map onto discontinuities in the tertiary street system, especially for the racial distribution of households with children. T-communities and islands have clear “borders” where sharp discontinuities occur in the distribution of racial groups. Furthermore, “invisible” discontinuities in the network of tertiary streets are just as disruptive to population distributions as natural barriers are. Finally, while sharing more tertiary streets is related to greater demographic similarity, the most substantial distinction occurs between those who live in the same tertiary street network and those who do not. A single trivial tertiary street connection may profoundly affect the demographic composition of two otherwise disconnected neighborhood communities. In chapter 8, I proceed to stage 2 neighboring relations and show that passive contacts are sociologically real phenomena, not merely theoretical constructs. Respondents had no difficulty identifying whether or not an activity was an unintentional meeting resulting from the mere fact of being neighbors. The correlation between stage 2 neighboring and children is evidenced by the fact that most passive contacts began when children casually played together; in general, meetings involving children were identified as passive, and meetings not involving them were not. Individual respondents’ stage 2 neighboring relations, as evidenced by their cognitive understandings of their neighborhoods, did not typically reflect formal neighborhood equivalents such as real-estate neighborhoods or school districts but rather the “lived” experience of interconnected tertiary face blocks. Furthermore, residents’ conceptualizations of their neighborhood aggregated to form cognitive neighborhoods that were typically identical to the network of tertiary streets. No one’s cognitive understanding of their neighborhood escaped tertiary street networks. In chapter 9, I show that the actualized stage 3 neighbor networks, which emerge from stage 2 neighbor networks, did not escape tertiary street networks either. Stage 1 neighboring 12

NEIGHBORHOODS AND NEIGHBORING

relations, when measured in terms of house-steps (the number of houses) along a tertiary street, powerfully related to higher stages of neighboring. The latter did not, however, correlate strongly to raw distance “as the crow flies.” The tertiary street network not only constrained individual residents’ interaction patterns but also the networks of neighbors they concatenated into; however, neighborhoods defined by shared boundaries such as census geography or elementary school catchments did not constrain interactions. Furthermore, the effects of t-communities on neighboring relations are not merely spurious confounds of geographic distance. At any distance, neighboring relations are restricted to the shared tertiary street network, but not to shared administrative geography. In chapter 10, I show that a neighborhood network is not typically identical to any individual resident’s neighbor network; it is a true social entity, beyond any individual. Each resident’s neighbor network connects with the neighbor networks of other residents, who connect to still other residents, concatenating and aggregating, neighbor to neighbor to neighbor, and especially child to child to child, to form a network that extends farther geographically and socially than any one resident’s neighbor network. Significantly, however, these aggregated neighborhood community networks maintain relatively short internal path lengths among residents. In chapter 10, I also show that households with children are far more involved in neighborhood life than households without children. They know almost three times as many neighbors and are known by more neighbors than households without children. These differences compound, so that the vast majority (85 percent) of all neighboring relations are between two households with children and only 6 percent of all neighboring relationships involve two households neither of which has children. Furthermore, neighbor-to-neighbor paths among households with children are half as long as those among households without children. Finally, in chapter 10, I show that most residents attributed great value to their neighboring relations. Neighbors performed important services for each other. Not surprisingly, the most important of these services related to children. Furthermore, passively generated contacts proved even more likely than nonpassively generated ones to result in 13

CHAPTER ONE

substantively important neighboring relations. The choice to live in a neighborhood is to some extent a choice about who we would like to be conveniently available to us and to whom we would like to be conveniently available. Chapter 11 formally revisits the discussion of how efficacious neighborhood communities emerge from the flow and exchange of norms, values, and beliefs among stage 4 neighbors. Social influence network theory mathematically models the process of community norm and value evolution. A simple focus on neighbor network density, treating neighbor networks as if all of the information about them was contained within the relations of individual neighbors, ignores the informational content captured in their larger networks. For influence to occur, residents must be within each other’s horizon of observability. The number of distinct paths transmitting norms and values between residents also affects the degree of influence; secondhand knowledge may be less valuable than firsthand knowledge, but what it lacks in immediate value it can make up for in volume. In chapters 12 and 13, I explore influence networks and the neighborhood-level outcomes they relate to in two distinct insular settings, a college town and a gang barrio. I begin with the college town in chapter 12 and explore one particular neighboring relation, trusting each other to watch over children in spontaneous playgroups, and show that this relation is both dense and short enough to be within the horizon of observability, allowing the behavior of neighborhood children and those who monitor them to be observable to most of the other households with children in the t-community. I provide an example of a particular criminal incident, where the observation of the illicit behavior, the parental response, and the evaluation of these behaviors by the neighborhood community was “observed” through influence networks by the residents throughout the t-community, but nowhere else. Shared tertiary streets, but not shared elementary school catchments, circumscribed neighborhood collective memory and produced collective efficacy for children. I then use the longitudinal nature of the college town study to show that neighbors influence each other’s beliefs both by their actions and by their interactions. One’s perceptions of one’s neighborhood’s 14

NEIGHBORHOODS AND NEIGHBORING

values are both similar to those of one’s neighbors and directly related to one’s interactions with one’s neighbors. The beliefs and values foundational to neighborhood effects, such as the working trust necessary for the development of collective efficacy, emerge from these networked interactions. The structure of influence networks, which is heavily determined by the structure of the tertiary street network, powerfully affected residents’ beliefs about their neighbors’ values and utility. The norms and values that emerged within one t-community, while internally consistent, differed from those that emerged in neighboring t-communities. In chapter 13, I explore these neighborhood community processes from a different direction. Instead of identifying a geographic area and asking to what extent it relates to some reasonable facsimile of community, I identify a well-established community that provided identification, social capital, and efficacy for its members and attempt to understand why it was associated with a particular geography. I show that the geography identified by residents of this community perfectly coincided with a tertiary street network but not with school catchment areas or parish boundaries or other potentially competing neighborhood foci. Within this neighborhood community, the neighbor influence network generated an enormous amount of social capital and collective efficacy, including actively preventing the sale of drugs within the neighborhood amid a city rife with the drug trade. More importantly, I show that the neighborhood community took its powerful norms and values from those most intimately involved in the network of trust and loyalty, but that who was most intimately involved in the trust and loyalty network was determined by where they lived in the tertiary street network. Chapter 14 concludes the book. I review my findings about neighborhood communities emerging from the network of interactions of neighbors, networks that concatenate from neighbor-level availabilities and interactions, not neighborhood-level processes. I argue that to properly investigate emergent neighborhood-level outcomes we must focus on communities that could have been produced by neighbor-level interactions. By precisely identifying latent social ties, tertiary street networks provide us with a lens to focus more closely on agentive 15

CHAPTER ONE

social capital and collective efficacy. Furthermore, studying neighborhoods precisely as networks rather than vaguely as diffuse entities highlights their nonlinear response to apparently similar conditions. Relatively minor modifications in the urban ecological environment that mediates individual-level interactions can result in disproportionate sociological outcomes.

16

C H A P T E R T W O

The Stages of Neighboring

NEIGHBORING: A SUPERPOSED RELATION I argue that neighborhood communities are produced by networks of neighbors. So, what does it mean to be neighbors? At one level, neighboring reflects convenience and availability. At another level, it is an intentionally chosen relation. It can be superficial and passive, or it can engage some of the most important norms and values in our lives. How do these different aspects of neighboring relate to each other? Neighboring is a specific type of social relation, emerging in a specific fashion from other social relations, forming networks that are distinguishable from other social networks. These networks have effects that are distinct from other sociological phenomena. Not only is neighboring an identifiable type of relation, it is itself the superposition of several distinct stages. I use the term superposition in both geometric and geological senses. Geometrically, superposition refers to the placement of one object over another so that like parts coincide. If we imagine a network of relations, we can superpose a second network on top of it, so that like parts coincide. The superposed network cannot exist where the previous one does not. The geological usage of superposition derives from the 17th-century Danish scientist Nicolas Steno. Geologists identify strata as layers of rock or soil, each with internally consistent characteristics that distinguish it from contiguous layers. Steno proposed the law of superposition as one of the axiomatic bases of geology. Essentially, this law states that strata are arranged in a time sequence, with the oldest on the bottom and the youngest on the top.1 Thus, the earth’s crust can be thought of as a set of layers laid down each one on top of the previous one. 17

CHAPTER TWO

Subsequent layers cannot emerge until previous layers are formed. What we see on the surface is actually a stacking of many layers. Both of these usages of superposition can be useful for understanding neighboring. Following the geological metaphor, neighboring is the layering of distinct relations, each with internally consistent characteristics that distinguish it from other relations. These relations can be thought of as developing in a logical order; one necessarily precedes another. Subsequent relations cannot emerge until previous ones exist. The geometric metaphor is similar: subsequent layers or stages of neighboring, and thus neighbor networks, cannot exist where previous ones did not. At its simplest, neighboring is the superposition of choice on convenience. Choice is made possible by convenience. Both convenience and choice are necessary to produce what we perceive as a neighborly relation. One could refer to those who are convenient, and those who are chosen, as potential neighbors and actual neighbors, respectively. Before I can choose someone as my neighbor, she must be available to me as a potential neighbor. These two relations build upon each other, literally, from the ground up. While this dependency or superposition is not a terribly profound insight and other researchers have noted it, I highlight it because it will become important when I discuss the concatenation of each of these relations into networks. Of course, we could define many more strata in neighborly relations. We could subdivide convenient availability between neighbors into geographic availability, functional availability, and conscious availability (awareness of each other as geographically and functionally available). We could subdivide neighborly relations actualized through choice into acknowledging each other, on the one hand, and more meaningful interactions such as exchanging of norms and values, trusting each other, and acting together for a common good, on the other. In this book, I identify four stages of neighboring. My stages are somewhat artificial since the quantity and quality of interactions with neighbors, like all human interactions, exist on a continuum. The four stages are not completely arbitrary, however. I will show that residents typically identify their relationships with neighbors in a way that corresponds to these stages, which appear to represent levels of cognitive awareness of neighboring relationships. I argue that they are also theoretically useful for understanding neighborhood effects. 18

T H E S TAG E S O F

NEIGHBORING

TABLE 2.1 Guttman Scale of Neighboring Relations 0

No availability

1

Geographic availability

2

Passive contact

3

Intentional contact

4

Mutual trust

Stage 1 neighboring occurs when we live near each other, however one defines near. A neighboring relationship not only depends upon context, as all relations do, but is also specifically defined by its context, by the latent tie at its foundation. Geographic proximity is essential to the very definition of neighboring. Not all social relations are geographically constrained; but neighboring is so by definition. Stage 1 neighbors are often described as living down the street or around the corner, whether or not they are seen much. This is perhaps the most common definition of neighboring used in research literature. Stage 2 neighboring occurs when we both live near each other and our lifestyles and routines create passive contacts between us, that is, we unintentionally encounter each other on a regular basis. These types of neighbors are often described as being seen around the neighborhood. Stage 3 neighboring occurs when we initiate contact. These are neighbors we say hello to or chat with from time to time; or perhaps we interact more intensely, going inside each other’s houses or sharing things or activities. The distinguishing feature of this stage is initiated contact. We have chosen to interact with these neighbors. Stage 4 neighboring occurs when we engage in activity that indicates trust or a realization of shared norms and values. This interaction can take a variety of forms. An example might be one resident allowing his children to play in the house of a neighbor and trusting that neighbor to supervise them and keep them safe. Thus, we can model neighboring as a Guttman-type scale with the values shown in table 2.1. It is important to note that none of these stages, including stage 4, requires or even presumes that the relation is strong or intimate in the 19

CHAPTER TWO

affective sense. A mutual awareness of shared norms and values can happen through observation in public places, either direct or by third parties, especially if it occurs over a long period of time. I now discuss each of these stages in more detail.

STAGE 1 NEIGHBORING While not all social relations are geographically dependent, neighboring is. Neighboring is a relationship defined by its context within residential geography.2 If someone moves out of a neighborhood and maintains a relationship with the person she lived next door to, they are “friends” or “former neighbors,” but they are no longer neighbors. Therefore, the first, most fundamental, stage of the multistage neighboring relation must be a geographic proximity that makes two residents available to each other. Definition. A stage 1 neighboring relation exists between i and j residents if they are geographically available to each other. How do I define and understand geographic availability? The geographic component underlying neighboring has been typically modeled in terms of proximity and boundaries, usually neighborhood-sized boundaries. I discuss these features subsequently, as well as some of the more subtle geography that affects neighborly relations.

STAGE 2 NEIGHBORING Almost a century ago, Robert Park (1984) told us that physical distances appear to index social distances. Decades later but still decades ago, Tobler (1970) coined the first law of geography, stating that near things are more related to each other than distant things. Both of these assertions are true in the case of neighboring and neighborhoods, but why? Is it really just living in physical proximity that makes people similar to one another? Do people really care who lives next to them just because they live next to them? What is ultimately behind the effects of small distances, tertiary face blocks, and tertiary intersections? 20

T H E S TAG E S O F

NEIGHBORING

A science fiction thought experiment highlights the importance of physical proximity. Imagine we lived in a Star Trek world and everyone simply “beamed” from home directly to wherever they wanted to go, and then beamed back. Houses wouldn’t need doors, and we could all have screens, equivalent to windows, that displayed whatever we wanted to visualize. In such a world, would we care who lived next door (assuming these sophisticated houses were sufficiently soundproof)? Probably not. What we care about is not that someone lives next door or close by but that we may see them when we look through our window, step outside our door, or drive out of our garage. When we drive or walk home, we pass their residence and may see the person as well. Our neighbors’ lives might spill over into our own; what if our neighbors are criminal or obnoxious or merely too friendly? It is not really physical closeness that makes us concerned about our neighbors but rather the possibility of contact. Research in the field of communication has suggested that certain media or contexts serve as “latent ties,”3 possibilities for social interaction that though available have not yet been activated. Latent ties can connect previously unconnected individuals and provide channels for information and resources to flow among them.4 While latent ties were originally conceptualized to understand patterns of communication using electronic technology, for my purposes here a medium is anything that increases or decreases opportunities for contact.5 One type of latent tie, a passive contact, has been identified in studies of neighborhoods (although it has not been labeled as a latent tie). Passive contacts are unintentional encounters that present people the opportunity to acknowledge each other’s presence and to learn about each other through observation and conversation. They may choose to interact, or they may not. The more possibilities people have for passive contacts, or the closer they are in terms of “functional distance,” the greater the chance they will have of meeting one another and interacting socially.6 Passive contacts form a relational substrate of “casual, public contact at a local level—most of it fortuitous, most of it associated with errands, all of it metered by the person concerned and not thrust upon him by anyone.” They are a basis for the development of relations of trust with people who are potential resources.7 A passive contact occurs between two residents of a neighborhood when they 21

CHAPTER TWO

unintentionally encounter each other and thus have the opportunity to acknowledge each other’s presence, to observe each other, and to initiate a conversation. A stage 2 neighboring relation exists between two residents who share passive contacts. Definition. A stage 2 neighboring relation exists between residents i and j when they unintentionally encounter each other and thus have the opportunity to acknowledge each other’s presence, to observe each other, and to initiate conversation. Thus, the importance of one’s “immediate” neighbors, those who live only a few households away, results from the abundance of passive contacts. Passive contacts are typically limited to a walking arena;8 therefore, their influence is quite limited geographically, and pedestrianoriented face blocks provide the primary opportunities for passive contacts.9 Residents’ passive contacts and thus their conceptualizations of their potential interactions, some of which they may choose to actualize, are constrained by the configuration of walking arenas.10 It is important to distinguish geographic availability from functional availability, as evidenced by passive contacts. Once we have settled in a residence, geographic availability is generally beyond our control. We can only change it by moving to a different residence. Functional availability, however, can change at different stages of life, through choices of action, and even seasonally. Those who are available to us as a result of both geography and lifestyle are potential neighbors. Being geographically available is only part of what makes residents functionally available to each other, or likely to be cognitively aware of each other and to have unintentional encounters that provide the opportunity to observe each other and interact. Our personal attributes determine if, for example, we are linguistically compatible—if we speak the same language—or if we have low social distance. Similarities between households in composition, income, or race create attractiveness, both practical and emotional, and thus facilitate interpersonal contact.11 Our lifestyles also determine the temporal pattern of our activities in neighborhood public areas, how much we are inside our residence, how much of our residential life is reclusive or gregarious, and a host of other factors that affect our possibilities to observe each other and 22

T H E S TAG E S O F

NEIGHBORING

our probabilities for contact. In general, the more time people spend in their neighborhood, especially if this time is spent outside, the more opportunities they have for passive contacts. This pattern has a pronounced effect on families with young children. Furthermore, the greater the synchronization of people’s time schedules, the greater their chance for interaction.12 One of the most important factors relating passive contacts to actualized relationships is the length of time two people have lived near each other. The effect of physical proximity becomes more pronounced when combined with time; the longer one has lived in an area, the longer two individuals have been available to each other, the more likely they are to interact; thus, the opportunities for face-to-face interaction afforded by contacts with neighbors are multiplied by the effects of residential stability.13 Residential longevity decreases effective functional distance and increases the number of passive contacts, obliging residents either to interact or to actively avoid doing so. As a result, the number of neighbors one knows, how far they are spread out geographically, and the proportion they make up of one’s total social contacts all increase as a function of time spent in a neighborhood. The longer two individuals share the same geographic and functional space, the greater their passive contacts and the greater the probability they will develop a neighborly relationship. This pattern probably also accounts for at least part of residential longevity’s noted relationship to neighborhood effects14 and civic engagement.15 Finally, functional distance is affected not just by the past. The choice to invest in a relation with another always takes into account the “shadow of the future.”16 If a person intends to stay in the neighborhood and believes another person does as well—as evidenced, for example, by home ownership—their incentives to invest in their relationship increase.17

STAGE 3 NEIGHBORING Functional distance, and the passive contacts it engenders, defines the set of potential neighbors we choose from and, in doing so, bounds the set of actual neighbors. Our passive contacts define a set of individuals 23

CHAPTER TWO

whom we choose to have neighborly relations with, or whom we actively avoid. Although we typically don’t choose to have neighborly relations with all the individuals we have passive contacts with, we can only choose from those persons available to us. The development of neighborly relations is a process whereby contact (“ecology”) establishes acquaintanceship, but mutual interest (“choice”) is required to produce a relation.18 Social scientists have long debated the relative importance of convenience and choice in the formation of neighboring relationships.19 Gans (1961, 1967) argued that people choose their neighborly relations carefully from among the available options and that they only choose to develop neighborly relations with people who are similar to themselves. Keller (1968) found that neighbors avoid dissimilar others, reducing their overall investment in neighboring activity if necessary (but not as severely as Gans suggested), and that a small amount of heterogeneous neighboring results even among dissimilar neighbors, due solely to the overwhelming impact of convenience. In contrast, other researchers20 have concluded that contact more or less automatically results in the formation of relationships, and Wellman, Carrington, and Hall (1983) suggested that people may be relatively easily satisfied with whoever is living nearby. This debate, however, should not ignore the nested nature of these two factors. One must first be a potential neighbor before one can be chosen as an actual neighbor, and my potential neighbors are themselves a limited set. For any given person, virtually all of humanity is unavailable for neighborly interaction. I may choose to pursue a relation with only a tiny fraction of those whose paths I regularly cross in my neighborhood, and I may choose to avoid most of those with whom I have recurring encounters; but, whether I choose 20 percent or 5 percent or 1 percent, this is a less substantial cut than the first one done by convenience. The limitations of convenient access also work upon those whom one may choose as neighbors. Neighboring, like all relations, requires two people to make the relation work. Both my neighbor and I have a limited number of people with whom we can engage in neighborly relations, and we must negotiate our relationship. If my neighbor desires interaction, I am likely to be high on the list of those with whom he interacts, whether I choose this relationship or not. Thus, while the 24

T H E S TAG E S O F

NEIGHBORING

set of individuals with whom I am eligible to engage in neighborly behaviors is quite limited, being defined by my stage 2 neighborly relations, it is obviously quite limited for my neighbors as well. Generally, with the exception of obnoxious behavior on the part of one of them, two people cannot engage in neighborly relations without each having at least tacit approval from the other. While in the case of obnoxious behavior, I may reject interaction entirely, there is usually a wide range between the minimum relation people desire and the maximum they will tolerate.21 My neighbors and I must negotiate our relationship. Ultimately, convenience motivates, guides, and constrains choice. The real question is this: when is a relation of convenience sufficient to allow choice to create a relation, or more appropriately to allow two choices to create a relation. In this book I explore when convenience is sufficient for choice to occur. A stage 3 neighboring relation actually occurs, however, when two stage 2 neighbors choose to initiate contact. Definition. A stage 3 neighboring relation exists between residents i and j if they have intentionally initiated contact.

STAGE 4 NEIGHBORING While two people could live near each other and not have passive contacts or unintentional encounters, they could not have such encounters unless they lived near each other. Similarly, while two people could have unintentional encounters and choose to ignore each other (or to avoid such contacts, e.g., by changing commuting routes or times), they cannot choose to interact without having encountered each other. The choice would not be available. Finally, while two people could interact only superficially, they cannot exchange norms and values and thus develop mutual trust without interaction. A stage 4 neighboring relation occurs when two residents share a belief in each other’s willingness and ability to act together to achieve a common goal. They need not understand their relationship to be intimate or strong in the sense of friendship or an affective bond. As Granovetter argued in his seminal essay, “weak ties,” or less intimate connections between people based on more infrequent social interac25

CHAPTER TWO

tion, may be critical for establishing social resources and integrating the community. In fact, weak ties have been shown to be as significant as (or perhaps more significant than) strong ties for the generation of community social capital.22 While a weak tie may lead to a strong tie, it may lead to an efficacious tie, a mutual expectation to work collectively for the common good. Of course, it may lead to both, but it need not. Definition. A stage 4 neighboring relation exists between residents i and j if they engage in one or more activities indicating mutual trust or a realization of shared norms and values. Is it possible to work together for a common good if you don’t trust each other? Perhaps individuals can work together for a common good under dire circumstances. In general, however, two people cannot choose to work together for a common good unless they have mutual trust, developed by the exchange of norms and values. While two people could exchange norms and values, develop mutual trust, and still not choose to act for the common good, the converse is not true; the final relation is a logical subset of the previous one. Is it possible to trust each other unless you know something about each other’s norms and values? Clearly not. Few if any people trust everyone else (most are a reserved in their trust of terrorists and pedophiles, for example), so to trust you I must know enough about you to locate you within the range of people I am willing to trust. When many neighbors share this type of relation, collective efficacy, the ability of a neighborhood to act for a common good, results. Distinguishing between stage 3 neighborly ties, on the one hand, and the shared expectations for action represented by collective efficacy, on the other, helps clarify the paradox of dense networks: while stage 3 networks foster the conditions under which collective efficacy may flourish, they are not sufficient. I want to reiterate that an efficacious neighborly relation may not be a socially intimate one. Many researchers have dismissed neighboring because it may not involve substantial affect or contribute substantially to an individual’s social support. However, that is an outcome different from the one I am concerned with here. I am interested in neighbor26

T H E S TAG E S O F

NEIGHBORING

hood effects, not individual social support. While community efficacy may depend upon neighborly relations involving working trust and social interaction, it does not require that my neighbor be my friend.

MAIN POINTS IN REVIEW To recap, the neighboring relationship is a multiple-stage relationship in which each stage is superimposed on the previous stage. A stage 1 neighboring relation exists between two individuals if they are geographically available to each other, when they are proximal to each other, and when no boundaries inhibit their contact. A stage 2 neighboring relation exists between two residents when their lifestyles cause them to casually and unintentionally encounter each other and thus have the opportunity to discover each other’s nature through observation and acknowledge each other’s presence. A stage 3 neighboring relation exists between two residents if they have intentionally initiated contact. A stage 4 neighboring relation exists between two residents who engage in substantial activity that indicates mutual trust or a realization of shared norms and values, when they share a belief in each other’s willingness and ability to act together to achieve a common goal, when they influence each other, either actively or passively. Neither stage 3 nor stage 4 implies that the involved parties understand their relationship to be intimate or strong in the sense of friendship or an affective tie. These stages of neighboring develop in a logical order, with lower stages necessarily preceding higher stages. Two people cannot be neighbors in any sense if they are not geographically available to each other. While two people can live near each other with no intervening boundaries and have no passive contacts, because of lifestyle or other personal reasons, they cannot have unintentional neighborly encounters unless they live near each other. Similarly, while two people can have passive contacts and choose to ignore each other, they cannot choose to interact without having encountered each other. Finally, while two people can limit their interaction to a superficial level, they cannot develop a deeper relation of mutual trust without some interaction. 27

C H A P T E R T H R E E

Reconceptualizing Stage 1 Neighboring

PROXIMITY While it may seem obvious, it is worth highlighting that, at its most fundamental level, neighboring is a proximity-dependent relation. When I say that someone is my neighbor, I am always making a statement about his being proximal to me. Neighbors must, by definition, live close to each other; but what constitutes the geographic proximity or availability that defines neighboring? Many studies have called attention to the strong role of extremely short distances in neighborly contacts.1 At least from an individual household’s perspective, the distances associated with neighboring are often effectively measured in feet and yards.2 Residential propinquity’s influence on social interaction is typically limited to those who live within a few households away.3 What is most important is who lives next door, not who lives in the same census tract; who lives a few houses away, not who lives a few blocks away. Beyond that distance, neighbors are rarely known. What criteria do people use to decide who lives close enough to be considered a neighbor? Residents offer a variety of definitions to translate propinquity into neighboring. One of the most common is that neighbors are people whose home you can see from your home (more often than not, this means that a place exists from which both your home and your neighbor’s are simultaneously visible).4 Either way, this definition implies a shared space. Thus, by this definition, neighbors are people you see often from your home or from the area immediately around your home. They are the people you necessarily are aware of because of the simple fact of where you live. A second common definition is that neighbors are people who live within walking distance.5 Walking distance, of course, varies with the 28

R E C O N C E P T UA L I Z I N G S TAG E 1 N E I G H B O R I N G

age and physical fitness of the person walking, and changes seasonally in some areas. No matter how far walking distance is, however, the probability that someone will be identified as a neighbor declines rapidly with increased distance from one’s home. This second definition highlights the casualness of neighboring. Neighbors are people I meet when I casually stroll; they are people I meet accidentally, not on planned journeys. Especially noteworthy in both of these definitions is that not everyone can be my neighbor. In fact, virtually all of humanity cannot. Thus, neighboring distinguishes itself by its origin. It arises among residents who live near each other in the course of recurring contacts in the locality that they jointly inhabit.6

BOUNDARIES A substantial literature demonstrates that residents perceive major roads, railway lines, rivers, and lakes as social boundaries, beginning with the study of “natural areas”7 and “ecological units.”8 Residents use major physical features such as railroad tracks, parks, landmarks, and arterial streets to help identify their neighborhoods.9 Most residents can quickly identify their neighborhood with some associated boundary.10 The relationship between major boundaries and neighboring seems intuitively obvious. Everyone would agree that two households located a hundred meters from me do not have the same impact if one of them is on the other side of a freeway. Boundaries are especially relevant when one considers the segregation of neighborhoods. Highways have historically formed effective boundaries between white and minority residential areas in most U.S. cities.11 Major streets may have been noted for their sociological effects as well. For example, in a particular Philadelphia neighborhood, Bellwether Street formed an effective barrier between black and white communities, referred to by local residents as “the edge.”12 Similarly, Chicago’s 47th Street has at times served as a “racial no-man’s land.”13 Some researchers have argued that boundaries, in spite of their association with segregation, are healthy for neighborhoods and neigh29

CHAPTER THREE

boring. Much of the urban planning literature has stressed the importance of identifiable boundaries to the development of stable neighborhoods. In fact, many modern neighborhoods are built with boundaries defined a priori.14 Boundaries are certainly useful for social researchers as well. In order to perform quantitative analysis on neighborhoods, one must collect and tabulate information for discrete areas. Thus, one must have a way of organizing cities into “statistical neighborhoods.” These statistical neighborhoods depend upon the delimitation of mutually exclusive and exhaustive boundaries. The census tract is the best-known type of statistical neighborhood, developed in the early part of the twentieth century to collect local community statistics in a systematic way. Major institutions such as elementary school districts or church parishes provide additional sources of delimitation. Once the statistical geography is established, however, it becomes a sort of de facto social system, since all subsequent analysis proceeds from this set of boundaries. Often the boundaries are accepted by researchers without much thought about how they separate people, which relations and interactions they facilitate and which ones they hinder. Social researchers seem strangely silent about the social logic that underlies the statistical neighborhood. They rarely discuss what they believe a particular boundary is doing sociologically or what sociological mechanisms or forces it facilitates and which it impedes. Thus, while geographic areas defined by the Census Bureau or other administrative agencies (e.g., school districts, police districts) may meet the requirements for statistical neighborhoods and provide a way to categorize and analyze information on the attributes of residents, they may imperfectly operationalize neighborhoods for purposes of understanding their internal dynamic processes. It is commonly acknowledged that census-defined boundaries delineate neighborhoods spatially but not socially.15 Furthermore, defining neighborhoods only by their boundaries, such as through “natural areas,”16 “ecological units,”17 or census tracts, but not with respect to their propensity for social interaction, does not guarantee that concatenated and consolidated relations can permeate them and thus logically disconnects neighborhood definitions from many of the effects researchers concern themselves with.18 30

R E C O N C E P T UA L I Z I N G S TAG E 1 N E I G H B O R I N G

Boundaries, at least major boundaries, and internal access are not logical inverses of each other. Conjecture. Two households can share the same set of boundaries without being accessible to each other.

FACE BLOCKS Many researchers have found definite links between the built environment and neighboring.19 Often, neighborhood studies focus on dramatic geographic features, such as major boundaries and barriers. While these undoubtedly restrict the availability of others as neighbors, much more subtle geographic features may have the same effect. While we can’t ignore the role of dramatic geographic features, it is important not to overlook subtler geographic features that also force residents to share lives. I have stated that two households located a hundred meters from me do not have the same impact on me if one is on the other side of a freeway. As we change the barrier from freeway to highway to primary to secondary street, it seems to become weaker. However, I intend to argue for the extreme dependence of neighboring on weak microgeography. Rain gutters and hedges have profound effects, not just freeways. Because of this variety of possible barriers, the relevant geography that impacts neighboring does not relate to physical distance in any easily linearizable way. Neighboring depends upon subtle geographic features. Besides focusing on the number of houses or yards separating two households, a natural division, in both cognition and behavior, occurs at the face block. The face block includes all of the dwellings that front on the same street and are situated between the first cross streets, of any type, encountered in both directions away from the respondent’s house. The face block has been found to be an important socio-spatial unit.20 Definition. A face block includes all of the dwellings that front on the same segment of the same street situated between any two cross streets. A face block terminates at an intersection. 31

CHAPTER THREE

FIGURE 3.1. Face block Rectangles represent residences, and the empty channels separating the groups of rectangles represent streets. The shaded residences are part of the same face block.

The face block includes virtually all neighbors who live either next door or directly across the street and most of those within a few house lengths; therefore we might assume that it is just another way of talking about proximity. However, studies have shown that residents have more interaction with those on the same face block than they do with residents beyond an intersection, even if they are spatially closer to the latter group.21

TERTIARY FACE BLOCKS Not all face blocks are oriented toward pedestrians. Some front on large arterial streets that provide access for travelers, while others front on smaller streets devoted to local living space. Studies of perceptions of street life22 reveal that residents clearly perceive the difference between 32

R E C O N C E P T UA L I Z I N G S TAG E 1 N E I G H B O R I N G

(a) heavy traffic face blocks “continuously filled with strangers”23 that are used solely as thoroughfares and corridors between the local neighborhood and the outside world, and (b) light traffic face blocks that form the basis of lively, close-knit communities24 where everyone knows each other25 and residents consider the boundaries between house and street space to be quite permeable.26 This is not a spurious relationship. Residents are perceiving patterns that have been created by urban planners. Cities intentionally design, build, maintain, and “renew” streets to serve different purposes. Some roads are built to be highways moving populations across a city, while others are local, allowing access within a neighborhood. Some have multiple lanes going each way, indicating a commitment to more automobile traffic, and some have a median indicating the need to protect pedestrians from increased traffic. Some have sidewalks and some do not. Some of these roads share a right-of-way with a railroad, a streetcar line, or some other rail line, again indicating a large degree of traffic that would discourage socializing in the middle of the street. Finally, some are on cul-de-sacs, while others go through an underpass, and some go through a tunnel. Passing under a freeway or underground typically creates an impassible psychological barrier for social contact. Readers can imagine a plethora of street types and variations, all of which have different effects on the emergence of the interaction foundational to neighboring. What is important to focus on, however, is that the overwhelming majority of face blocks in most cities are strictly oriented to pedestrian traffic. I will refer to them as tertiary face blocks. Tertiary face blocks are defined by urban planners as portions of local neighborhood roads that are not thoroughfares. I operationally define these tertiary streets as those roads designated as Census Feature Classification Code27 (CFCC) A-41. Described as “local neighborhood roads,” by definition they have one lane on each side, no median or divider, do not share right-of-way with a rail line (e.g., railroad, streetcar line), and do not go through an underpass or a tunnel. In general, they are defined by their simplicity. Even though not all streets so classified are identical (e.g., not all have sidewalks, some are cul-de-sacs, etc.), this is not an arbitrary category. Tertiary streets are designed, built, maintained, and renewed explicitly to provide local and pedestrian traf33

CHAPTER THREE

fic, not to be used as thoroughfares. Nontertiary streets, in contrast, are typically designed as thoroughfares to provide access for automobiles across a city. Definition. A tertiary face block includes all of the dwellings that front on the same segment of the same street, which has been designed and maintained by governing authorities to promote local and pedestrian traffic, situated between any two cross streets. It has only one lane on each side and no median. It does not share right-of-way with a rail line (e.g., railroad, streetcar line) and does not go through an underpass or a tunnel. The tertiary face block is a more or less “natural” unit of face-to-face neighborly interaction.28 Tertiary face blocks are oriented toward pedestrian travel and local residents, rather than outsiders who arrive by automobile or mass transit. Thus, tertiary face blocks are the types of face blocks most likely to give rise to social interactions.29 Tertiary face blocks provide a meeting place for neighbors.30 People use them to walk pets, ride bicycles, and chat with neighbors.31 Houses on a tertiary face block have porches or yards where people may spend a considerable amount of time, and nearby neighbors have frequent visual contact of each other.32 Shared tertiary face blocks provide a “permeable boundary” between households’ private spaces.33 Tertiary face block neighbors are “used for easy sociability and assistance when quick physical accessibility is an important consideration.”34

INTERSECTIONS Face blocks include all of the dwellings that front on the same street and are situated between the first cross street in either direction from a person’s house. Thus, face blocks terminate at intersections. An alternative way of thinking about this, of course, is that intersections connect face blocks with each other. The important question then becomes: Do they also connect neighbors? Do neighbor networks terminate at intersections, or do they bridge them to form larger structures? Intersections form a different metric than face blocks for measuring functional distance. 34

R E C O N C E P T UA L I Z I N G S TAG E 1 N E I G H B O R I N G

Just as different types of streets differentially induce or fail to induce neighborly relations, so different types of intersections may induce or inhibit neighborly relations from bridging them. Fortunately, the intersections of streets can also be operationally defined in a convenient and meaningful way. When streets of different classifications intersect, planners consider the intersection to be of the higher classification.35 For example, if a larger street intersects a tertiary street, planners consider the intersection to be part of the larger street, but not part of the tertiary street. This makes intuitive sense, especially for our purposes, because the inhibiting effects of the larger street dominate.36 The nature of intersections either facilitates or impedes pedestrian-based neighborly interaction. Definition. An intersection is tertiary if all of the face blocks contiguous with it are tertiary. Definition. An intersection is nontertiary if at least one of the face blocks contiguous with it is nontertiary. Studies have shown that tertiary intersections and the tertiary face blocks contiguous with them can serve as a “pedestrian circulation system.”37 Sidewalks provide access between residence and parks, churches, and neighborhood shops. Neighborly relations bridging face blocks occur “through the routes people take in meeting an average day’s basic needs and desires. The newsstand where one buys the Sunday paper, the store one runs to for a quart of milk, and the streets one travels on to visit a friend.”38 People come to envision their neighborhoods as networks of paths and channels along which they move.39 Tertiary intersections guide this “natural movement”40 within a city.

MAIN POINTS IN REVIEW In this chapter, I revisited stage 1 neighboring in more detail. This initial stage of neighboring occurs when we are geographically available to each other. While this availability is often conceptualized in terms of neighborhood-size distances and the absence of neighborhood-sized 35

CHAPTER THREE

boundaries, I conceptualize stage 1 neighboring in terms of neighborsized distances and the absence of neighbor-sized boundaries. Neighboring is primarily dependent upon extremely short distances, walking arenas such as tertiary face blocks and tertiary intersections, because stage 2 neighboring, passive contacts, or unintentional encounters, relies upon pedestrian encounters.

36

C H A P T E R F O U R

Reconceptualizing Stage 1 Neighbor Networks

LAYERS OF COMPLEX NETWORK STRUCTURES I have argued that there are at least four distinct stages of neighborly relations in which individuals may be involved. Each stage is embedded in or superposed on the previous stage. An individual’s stage 4 neighbor network is a subset of her stage 3 neighbor network, which is a subset of her stage 2 neighbor network, which is a subset of her stage 1 neighbor network. These individual neighbor networks evolve into neighborhood community networks through a process of concatenation. Residents have relations with their neighbors who interact with other neighbors, and so on. These neighborly relations concatenate and consolidate neighbor to neighbor to neighbor. The idea of concatenation gets at an important point: It emphasizes the fact that any network is a product of relation built upon relation built upon relation. How the relations fit together matters. There are several important corollaries of this fact. First, the resultant network is as far-reaching as its most extensive ramification. Relations concatenate to form a network typically larger, both in the number of interpersonal relations and in the size of the area it extends to, than any individual’s relations. Thus, relatively micro-level relations result in a macro-level structure. Second, the resultant network is as fragile as its weakest link. Anything that can cause a relation not to form, no matter how trivial, delimits the network. In contrast to the first corollary, micro-level fragilities can inhibit a macro-level structure. Third, the characteristics of the resultant network are not readily predictable from the characteristics of the local networks that concatenate to form it. Only as sets of individual networks concatenate do the characteristics of this aggregated network emerge. 37

CHAPTER FOUR

FIGURE 4.1. Stage 4 neighbor network A hypothetical stage 4 neighbor network. Dots represent households. A line connects two dots if the represented households influence each other, if they share norms and values and are developing working trust.

Therefore, the four distinct stages of neighborly relations concatenate into four distinct stages of neighborhood community networks: a network of concatenated geographic availability, a network of passive contacts, a network of actualized neighborly relations, and an influence network, with its inherent potential for social capital and collective efficacy. I now explore how they relate to each other. Networks of influence concatenate from stage 4 relations in which residents exchange norms and values and expectations and develop trust with their neighbors, whether or not they have developed an intimate, affective relationship. It is these stage 4 neighbor networks that potentially generate a sense of community, social capital, and collective efficacy. These stage 4 neighbor networks do not exist among all residents of a city. They are often quite delimited. Figure 4.1 shows a hypothetical 38

R E C O N C E P T UA L I Z I N G S TAG E 1 N E I G H B O R

NETWORKS

FIGURE 4.2. Stage 3 neighbor network, which induces stage 4 neighbor network Dots represent households. A line connects two dots if the represented households have initiated contact. The stage 4 neighbor network is an edge-induced, rather than node-induced, subgraph of stage 3 neighbor network.

stage 4 neighborhood community network among a tiny neighborhood of 20 households. In this illustration, the dots represent households, and a line connecting them indicates that they influence each other, that they have shared norms and values and are developing working trust. Three distinct influence clusters are apparent. The source of the constraint illustrated in figure 4.1 is the delimited stage 3 neighbor networks, which form the necessary substrate for the stage 3 neighbor networks. Stage 4 neighbor networks cannot extend anywhere stage 3 neighbor networks have not already extended. Figure 4.2 shows a hypothetical stage 3 neighbor network that might have led to the stage 4 neighbor network in figure 4.1. Stage 3 actualized neighbor interaction networks themselves emerge from the substrate of concatenated stage 2 passive contacts and thus stage 2 neighbor networks delimit stage 3 networks of actualized neigh39

CHAPTER FOUR

FIGURE 4.3. Stage 2 neighbor network, which induces stage 3 neighbor network Dots represent households. A line connects two dots if the represented households are functionally available to each other, if they unintentionally encounter each other and thus share passive contacts. The stage 3 neighbor network is an edge-induced, rather than node-induced, subgraph of Stage 2 neighbor network.

borly relations. While individuals’ choices, their failures or refusals to actualize potential neighbors, may make actual neighbor networks smaller than potential neighbor networks, they can be no larger. Figure 4.3 shows a hypothetical stage 2 neighbor network that might have produced the hypothetical stage 3 neighbor network in figure 4.2. Finally, because a passive contact cannot exist unless residents are geographically available to each other, the network of passive contacts cannot transcend the network of geographic availability; it is logically impossible. While individuals’ lifestyles and habits may prevent them from having passive contacts with those who are geographically available to them, their behavior cannot cause them to have passive contacts with those who are unavailable. Figure 4.4 shows a hypothetical stage 40

R E C O N C E P T UA L I Z I N G S TAG E 1 N E I G H B O R

NETWORKS

FIGURE 4.4. Stage 1 neighbor network, which induces stage 2 neighbor network Dots represent households. A line connects two dots if the represented households are geographically available to each other. The stage 2 neighbor network is an edge-induced, rather than node-induced, subgraph of Stage 1 neighbor network.

1 neighbor network that might lead to the hypothetical stage 2 neighbor network in figure 4.3. What is important to note in this stage 1 figure is that, while certainly not all stage 1 relations translate into stage 2, 3, and 4 relations, only existing stage 1 relations do. Stage 4, stage 3, and stage 2 neighborly relations cannot exist where stage 1 neighborly relations do not already exist. For example, it is impossible for stage 4, or stage 3, or stage 2 neighborly relations to exist between the community on the right and the community on the left because no stage 1 relations exist between them. To study efficacious neighborhood communities emerging from neighbor networks, therefore, we need a definition of a neighborhood community whose importance is derived from the potential for influ41

CHAPTER FOUR

ence networks to emerge from within a community, from the potential for neighbor networks to concatenate within it, from the consolidation of stage 2 passive contacts it allows. A neighborhood equivalent that maps the maximal extent of the overlapping circles of passive contacts among residents likewise maps the maximal extent of potential neighbor networks and thus the maximal extent of actual neighbor networks, and thus ultimately maps the largest potentially efficacious neighborhood community.

T-COMMUNITIES AND ISLANDS It is imperative, therefore, that we properly identify these stage 1 relations, this geographic availability foundational to the emergence of passive contacts, neighborly interactions, trust, and the realization of shared norms and values among neighbors. In chapter 3, I defined this geographic availability in terms of shared walking arenas that mediate, guide, and constrain passive contacts, or unintentional encounters. To the extent that this is accurate, then the concatenated network of overlapping passive contacts can be no larger than the concatenated network of walking arenas; conversely, the network of potential neighborly relations, based on concatenated passive contacts, is a subset of the concatenation of these walking arenas. I have argued that tertiary face blocks effectively proxy walking arenas in urban areas. In this study, therefore, the maximal concatenation of contiguous tertiary face blocks, of walking arenas, represents the maximal consolidation of individual residents’ potential contact with each other. Use of this neighborhood equivalent signifies internal access. All residents within it have a potential for neighborly relations using walking arenas. While it is unlikely that all, or even any, residents would traverse the entirety of this neighborhood equivalent, its internal contiguity allows residents to interact with their neighbors down the street, who interact with other neighbors farther down the street, and so on throughout the network. 42

R E C O N C E P T UA L I Z I N G S TAG E 1 N E I G H B O R

NETWORKS

Such a neighborhood equivalent also signifies constraint. To the extent to which passive contacts depend upon walking arenas, the neighborhood so specified defines the limit to the concatenation of neighbor relations. Finally, because passive contacts are necessary for the development of higher-stage neighbor networks, the concatenated network of tertiary face blocks serves as an effective surrogate for networks of potential neighbors, for networks of actualized neighbors, and for influence networks from which emerge the neighborhood effects that researchers concern themselves with. How do I define the maximal concatenation of tertiary face blocks? Face blocks do not actually touch each other; they are separated by intersections. While the distinction between tertiary and nontertiary face blocks is determinative, the distinction between tertiary and nontertiary intersections is not entirely so.1 It may be the case that nontertiary intersections only inhibit, rather than entirely disrupt, the development of passive contacts. Therefore, I will define two types of neighborhood equivalents, one connecting tertiary face blocks using only tertiary intersections and the other connecting tertiary face blocks using all intersections. Definition. A t-community2 is a maximal contiguous network of tertiary face blocks and tertiary intersections. Figure 4.5 illustrates t-communities. In the figure, lines represent streets. Bold lines represent nontertiary streets, and nonbold lines represent tertiary streets. Thus, all crossings of nonbold and nonbold lines are tertiary intersections and all crossings of bold and nonbold lines are nontertiary intersections. Four distinct t-communities exist in the figure, labeled 1, 2, 3, and 4. They are easily identifiable by considering the maximal contiguous set of nonbold lines. Boundaries and internal access are not logical inverses of each other. While the bounded area A is coterminous with t-community 1, the bounded area B is coterminous with three distinct t-communities, 2, 3, and 4. This is quite common. It is tempting to combine t-communities 2, 3, and 4, which are bounded by the same set of nontertiary streets into a single neighborhood equivalent. In fact, this melding is exactly 43

CHAPTER FOUR

FIGURE 4.5. T-communities

what many neighborhood equivalents would do, despite the impossibility of overlapping circles of tertiary street-based passive contacts connecting their residents. Clearly, however, no access is available among residents of t-communities 2, 3, and 4 via tertiary streets. If we were to relax the definition of the concatenated network of walking arenas that circumscribe the overlapping networks of passive contacts and thus the higher stages of potential neighborly relations, it would not be by including long stretches of nontertiary streets that act as boundaries. The next logical relaxation would be to allow for the possibility that neighborly relations might cross nontertiary intersections. Therefore, I define my second neighborhood equivalent by relaxing the above definition to include nontertiary intersections. Definition. An island3 is a maximal contiguous network of tertiary face blocks and any intersections. Consider again figure 4.5. In this figure, there are four t-communities but three islands. T-communities 1 and 2 form part of the same island because they are the maximal contiguous set of tertiary face blocks and any intersections. Islands highlight the fact that residents of t-communities 2, 3, and 4 do not have access to each other, even if we allow them to cross nontertiary intersections. It is highly unlikely that overlapping circles of passive contacts will connect their residents. If any t-communities were to be grouped by traditional analyses, 44

R E C O N C E P T UA L I Z I N G S TAG E 1 N E I G H B O R

NETWORKS

they would almost certainly be t-communities 2, 3, and 4. In contrast, I argue that it is far more likely that overlapping circles of passive contacts, if they connect any two t-communities, will connect tcommunities 1 and 2. Since islands are maximal networks of tertiary face blocks and any intersections, while t-communities are maximal networks of tertiary face blocks and tertiary intersections, t-communities are necessarily subsets (although not necessarily proper subsets) of islands. While I expect t-communities to have more pronounced effects, islands will measure the potency of nontertiary intersections. I offer one further illustration to help readers understand islands and t-communities. Imagine if all of the nontertiary face blocks (but not nontertiary intersections) were removed from a city. In a few cities, a gridlike pattern of tertiary streets would remain. In most, however, multiple independent networks of tertiary streets would now exist, effectively isolated from each other.4 These are what I am terming islands. Households in one island would not be able to reach other, often nearby, households in another island. Figure 4.6 illustrates this point. It maps an area in Los Angeles. Lines represent only tertiary streets. Arrows indicate streets that continue beyond the edge of the map. In the center of the map is a large set of tertiary streets, indicated by bold lines, which, while connected to each other, do not connect to any of the tertiary streets extending from the various edges of the map. This set of tertiary streets forms an island. Note that the map is drawn to scale. The distances separating the tertiary streets forming the central island and the tertiary streets near the edges are in most cases quite substantial, often measured in hundreds of feet. This analysis may seem counterintuitive to some readers, who would imagine that the network of tertiary streets would ultimately connect all households within a city, although perhaps at a great distance. This very idea was at the heart of human ecologists’5 arguments that physical distance was a crude index for functional distance. In the latter half of the twentieth century, however, many street systems have been designed (or redesigned) to create disconnected networks of tertiary streets. Urban renewal has transplanted these patterns into much older cities as well.6 Consequently, in most cities, the tertiary street system is not continuous, even if it once was. 45

CHAPTER FOUR

FIGURE 4.6. Tertiary street island Lines represent Tertiary Streets. Streets that comprise Island are in bold. Arrows represent Streets which continue beyond the edge of the figure. Note: Map is drawn to scale.

Now, if one imagines overlaying these islands with the grid of nontertiary streets they would be dissected into t-communities. T-communities are defined both by their internal connection via tertiary streets (as islands are) and by being bounded by nontertiary streets.

MAIN POINTS IN REVIEW In this chapter, I turned my focus to the networks formed by the concatenation of neighboring relations. Some (perhaps all, perhaps none) stage 3 neighbor networks translate into stage 4 neighbor networks. Some (perhaps all, perhaps none) stage 2 neighbor networks translate into stage 3 neighbor networks. Some (perhaps all, perhaps none) stage 46

R E C O N C E P T UA L I Z I N G S TAG E 1 N E I G H B O R

NETWORKS

1 neighbor networks translate into stage 2 neighbor networks. No neighbor networks, however, develop where there were not already stage 1 networks in place. This is why an accurate definition of stage 1 neighboring relations is so important. Most sociological studies of neighborhoods use administrative geography that implicitly defines two households as stage 1 neighbors if they are in the same administratively defined area. However, residents of these spatially defined analytic units may not be geographically available to each other. In contrast, I define two new neighborhood equivalents in terms of the concatenated network of walking arenas as represented by tertiary face blocks. These two neighborhood equivalents are distinguished by the intersections that connect face blocks within them. The first neighborhood equivalent, t-communities, includes only tertiary intersections, while the second, islands, includes all intersections. While I expect t-communities to have more pronounced effects on neighborhoods as social entities, I include islands to measure the potency of nontertiary intersections. Both of these new neighborhood equivalents are meant to focus on the potential for passive contacts, or unintentional encounters, between neighbors, and thus on the interactional aspect of neighborhoods.

47

C H A P T E R F I V E

Selection and Influence

I HAVE ARGUED that neighborhood communities, which are necessarily embedded within geographic contexts, are produced by networks of neighbors interacting within them. I have argued that the relationship between these neighborhood communities and their geographical contexts results from the relationship between geography and the individual neighboring relations that concatenate to produce neighborhood communities. I have argued that the geography most relevant to neighboring relations involves walking arenas and is best proxied by tertiary streets. Up to now, I have focused primarily on the static portion of this process. I have focused on neighboring relations, geographic and other constraints on neighborly interactions, and the networks neighboring relations aggregate into. In this chapter, I focus on the dynamic processes that correspond to these neighbor networks, primarily two forces: selecting homophilous stage 1 tertiary street neighbors and influencing stage 4 neighbors through our norms and values. I will argue that these forces work together, sequentially, to create neighborhood communities with their demographic distinctiveness and their varying degrees of social capital and collective efficacy.

SELECTING HOMOPHILOUS IMMEDIATE NEIGHBORS A rigorous debate has focused on whether neighborhood communities represent anything more than visible evidence of homophily, the tendency of like-minded people to congregate in the same places. Is there anything to neighborhood effects beyond the clustering of people with resources and people with needs? Is it just a fact that households with similar values and norms prefer to live side by side?1 48

SELECTION AND INFLUENCE

Historically, sociological studies of neighborhood communities were guided conceptually by Robert Park’s (1916, 1984) and Ernest Burgess’s (1984) economic definition of neighborhoods. They laid the foundation for urban sociology by defining local communities as “natural areas” that developed as a result of free-market competition between businesses for land use and between population groups for affordable housing. This competition for scarce neighborhood resources, for efficient locations among businesses and individuals, does cause similar types of residents to share this “complex commodity,” which is a bundle of spatially based attributes, including structural, infrastructural, demographic, environmental, political, social, and sentimental characteristics, including class status, a package of taxes and public services, and proximity to schools, churches, stores, and labor markets.2 A neighborhood, according to this view, is a subsection of a larger community—a collection of both people and institutions occupying a spatially defined area influenced by ecological, cultural, and sometimes political forces.3 This view seems quite reasonable. We do indeed choose our neighborhoods (and thus our neighbors) in part because of their nonneighbor-based features (commuting time, availability of markets, churches, and other institutions, etc.). Thus neighborhood effects might be latent side-effects of the clustering of people with similar needs, desires, and resources. Households have been shown to prefer easy access to shopping and other commercial activities.4 Households are willing to pay more to locate in areas with lower crime rates.5 Some households choose a residence so as to minimize commuting time to the workplace.6 Commute time has also been found to have more effect on low-income than high-income households.7 In fact, commuting time forms the core of most models of urban residential structure and successfully accounts for part of the distribution of demographic characteristics within urban areas.8 I argue that these market-based models are insufficient, however. One of the primary motivating forces behind households’ locational choices is whom they will have as neighbors. This influence is often missed by researchers because they perceive neighboring as entirely volitional, or believe that its nonvolitional aspects can be separated out from its volitional aspects. I argue, however, that geography forces cer49

CHAPTER FIVE

tain people to share lives, to become potential neighbors. Furthermore, it excludes other people as potential neighbors. If one desired neighborly relations, but not with current neighbors, one could move or encourage the unwanted neighbors to do so. Social distance theory argues that households with similar cultures based on similar ethnic and status backgrounds will have low social distance from each other, while households that have dissimilar cultures based on different ethnic and status backgrounds will have high social distance from each other. An impressive volume of literature shows that this social distance translates into spatial distance. This is especially true for higher-status households that seek to separate themselves from lower-status households, minimizing interaction and association with perceived social inferiors and thereby demonstrate and consolidate their position in the status hierarchy. When considering a new residence, how do households choose among potential neighbors whom they’ve never met? Numerous studies have shown that people base relocation decisions on easily discernable characteristics. For example, households tend to settle near other households of similar age and family structure.9 The clearest evidence of homophily and social distance, however, is racial residential segregation. By controlling for socioeconomic and lifestyle factors, researchers have shown that residential racial segregation evidences households choosing neighbors, not just neighborhoods. It is the “exception” to urban ecological models.10 Race is arguably the most highly visible characteristic of potential new neighbors to which households can respond. Potential neighbors often use it as evidence of exposure to crime and violence11 and other less tangible factors12 that contribute to the development of an urban underclass, signs of social disorder that lead residents to perceive their neighbors as threats rather than as sources of support or assistance.13 Neighbors may respond to demographic changes in their neighborhoods by exiting. Home-seekers choose a new neighborhood and reject alternatives, in part, because of who their potential neighbors will be. They often reject potential neighbors based on characteristics that are readily observable when searching for a new residence, with race (and status, crime, and other factors it is often perceived to proxy) being one of the most dramatic bases for making decisions. 50

SELECTION AND INFLUENCE

So how does this relate to my stages of neighboring? Residential homogeneity is most commonly discussed in terms of neighborhoods, but residents don’t typically respond to neighborhoods. Residents move out of a residence, or fail to do so, in response to those whom they meet and interact with in essence because of who their immediate stage 1 neighbors are or would be. The key word is immediate. Segregation studies typically measure whether residents settle in homogeneous neighborhoods, represented often by census geography. Thus, a researcher might argue that segregation occurs as a result of households choosing whom they will have in their census tract. I’m arguing, by contrast, that what is important is who lives a few houses away, not who lives a few blocks away. Research based on the 1992 Detroit Area Study found that whites and, to a more limited extent, blacks claimed they would either move out of a residence or be unwilling to move into a residence based solely on the racial composition of the immediately adjacent residences.14 Respondents were presented with a sequence of five cards depicting a focal home surrounded by 14 other homes, with the number of black households among the 14 increasing through the sequence. When one of the 14 surrounding residences was occupied by a black household, 4 percent of white respondents would try to move out, and 13 percent would be unwilling to move in. These percentages increased steadily as the number of residences occupied by black households among the 14 increased, so that when eight of the 14 surrounding residences were occupied by black households, 53 percent of white respondents would try to move out and 71 percent would be unwilling to move in. The important point for my argument is that these cards did not characterize nebulous neighborhoods but rather immediate neighbors, specifically the 14 nearest residences. I argue that households make decisions on their residence based on who their immediate neighbors will be— who lives a few houses away, not who lives a few blocks away. When households choose a new residence, they do not examine administrative reports of neighborhood demographics; they look at the home’s immediate surroundings. Typically, they do so in a way that focuses on tertiary face blocks. Since Schelling (1971), researchers have shown that even relatively small tendencies toward homophily can create dramatically segregated 51

CHAPTER FIVE

neighborhoods, by affecting the relocation decisions of one household at a time. Individual households decide who will and who will not be their immediate neighbors by making locational decisions. Their new neighbors have made and continue to make decisions about whom they desire for their immediate neighbors. These decisions concatenate into larger-scale patterns. If neighbors make similar decisions, these decisions concatenate to produce segregated communities. Most of these studies, however, have assumed a somewhat undifferentiated gridlike structure. Instead, I argue that the structure in which segregation operates is a network of tertiary face blocks and intersections. As a result, t-communities are a distinct metric to measure segregation and any other social phenomena associated with neighborhoods. They not only provide a more accurate index of the forces other neighborhood equivalents attempt to measure, they also measure distinct forces. Being defined by their internal access, rather than boundaries, t-communities both enhance and delimit the overlapping circles of passive contacts among residents and thus potential neighbor networks. T-communities are important because of their potential for social interactions to concatenate within them. If people choose a home where their immediate neighbors are similar to themselves, and if people tend to make these choices using parameters such as race, then the segregated networks of neighborly relations that emerge will be delimited by t-communities. T-communities, therefore, provide a template for social distance to act on and can serve as an effective geographic marker for neighbor networks. T-communities directly measure the concatenation of social forces that are delimited to act along tertiary streets. Tcommunities measure the process of households choosing their immediate neighbors, not the ecological niceties of their neighborhoods.

INFLUENCE Locational choice and homophily account for some of the effects of neighborhoods, but numerous studies have shown that neighborhoods with similar population demographics—race, socioeconomic status, family structure, and a host of other characteristics—often yield different outcomes for their residents. Market explanations, while intuitively 52

SELECTION AND INFLUENCE

appealing, have failed to account for the richness and complexity of these effects. There is another distinct process at work creating what are typically referred to as neighborhood effects. This process involves households interacting with neighbors or failing to interact with them, influencing each other or failing to do so, and creating or failing to create various neighborhood effects. Some two decades ago, Jencks and Mayer (1990) argued that to the extent that neighborhood effects exist, they are constituted from social processes that involve collective aspects of community life, and these intervening processes are therefore at least part, if not most, of the reason for their potency. Theories of social-interactional processes or mechanisms are concerned with, not the more static features of sociodemographic composition, but with how neighborhoods bring about a change in a given phenomenon of interest and how neighborhood effects are transmitted. Jencks and Mayer (1990) hypothesized that the causal mechanisms underlying neighborhood effects included an epidemic or contagion model, which emphasizes the role of peer influence, and a collective socialization model, which emphasizes the positive impact of successful adult role models. Neighborhoods are more than just neighbors residing nearby each other. They are vital entities—or at least they have the potential to be. Even when it appears static, neighborhood social life generally is not; it is a stable equilibrium reached amid the strife of social flows. The vibrant, living part of neighborhoods consists of the flow and exchange, both spoken and silently modeled, of norms, values, identities, symbols, ideas, affect, sentiment, and other social and cultural goods and resources among neighbors along the conduits provided by neighbor networks. This flow pressures neighbors toward conformity. This flow of norms and values acts most profoundly on children: “For example, when parents know the parents of their children’s friends, they have the potential to observe the child’s actions in different circumstances, talk to each other about the child, compare notes, and establish norms. Such intergenerational closure of local networks provides the child with social capital of a collective nature.”15 This social flow happens verbally, through the exchange of personal information, life histories, and stories, as well as through establishing and enforcing rules for neighborhood children; however, it happens 53

CHAPTER FIVE

nonverbally as well, and in some ways more potently. Community members model and enforce “appropriate behaviors” in their daily interactions with each other, especially with children.16 In addition to such modeling, neighbors may use rewards to encourage normative behaviors or sanctions to discourage nonnormative behaviors. These sanctions or rewards vary greatly with the nature of the society. In communities that emphasize social control and social cohesiveness, sanctions may be enforced by direct social pressure for conformity.17 As norms, values, ideas, and other social goods and resources traverse and commingle along neighbor networks, they can engender a sense of community and identity, social capital, mutual trust, social control, collective efficacy, and many other important facets of neighborhood life. The collectively efficacious community network that emerges, or that fails to emerge, however, is embedded in the network of interactions that is embedded in the network of passive contacts that is embedded in the network of geographic availability. Social disorganization theory suggests that if the network of geographic availability brings together residents who espouse a clear set of community norms, who model appropriate behaviors, especially for children, and who encourage or discourage behaviors they witness modeled by others, socially well-integrated neighborhoods are likely to result. Modeling works both through vertical, or intergenerational, influence and through horizontal, or reciprocal, influence, in which neighbors simultaneously exert influences on each other’s lives.18 Together, neighbors’ actions create a set of acceptable behaviors from which community residents, especially children, can choose their own courses of action.19 The more homogeneous the set of neighbors and thus the more homogeneous the set of “acceptable behaviors,” the fewer alternatives are available for residents, especially children and youth, to choose from.20 Multiple exposures to similar ideas and values from multiple sources and provided by dense neighbor networks lead to greater consistency in normative behavior,21 greater consensus in beliefs,22 and ultimately to emergent community.23 In contrast, neighborhoods with high residential turnover, low rates of home ownership, and a concentration of unfamiliar residents make it difficult for residents to establish the social ties and shared values needed to exercise social control and to achieve common goals. Young individuals living in so54

SELECTION AND INFLUENCE

cially disorganized neighborhoods are at higher risk of nonconventional outcomes (dropping out of school, premarital sex, early childbearing, use of illegal substances) at least in part because their immediate community does not have consensus on, or is unwilling or unable to enforce, values and norms.24 When neighborhood adolescents commit crimes or violence or use drugs or engage in imprudent sexual activity, they reflect an understanding that such activities are normative and condoned, if not validated, by their neighborhood community. Furthermore, when other neighborhood community members witness them engaging in such activity and fail to respond, they reflect an understanding that they are not accountable to each other and that intervention is neither expected nor welcome. Similarly, when children and adolescents believe that such social ills as racism, unemployment, poverty, or child abuse are normative, it is because such values have been reinforced in their neighborhood community. Some neighborhood effects, such as acts of crime and violence or abuse, appear almost instantaneously. Other neighborhood effects, such as underachieving at school, seem to be longer-term outcomes. All these neighborhood effects, however, emerge from the consciousness of neighborhood residents. Neighborhood effects occur, or fail to occur, because neighbors believe, or do not believe, that it is their responsibility to act cooperatively on behalf of the neighborhood community and that such action will be welcomed and is expected. In general, this is what Sampson terms collective efficacy and what Putnam terms social capital. While they may disagree on the specifics of each term, they would agree that residents of some neighborhoods have the capacity to act cooperatively for the common good and, in fact, do so in a coordinated and meaningful way. Collective efficacy and social capital emerge from the understandings of individual residents. The real question, therefore, is where this understanding comes from. Some of it, to be sure, is brought by residents from their experiences of previous neighborhoods. Some residents have been fortunate and have lived in efficacious neighborhoods. Other residents have been less fortunate and have lived in neighborhoods that are not efficacious. Still others have had a mixed experience. 55

CHAPTER FIVE

Upon entering a neighborhood, individuals observe each other. They do so casually by watching each other from the safety of their household. They do so more actively by engaging in conversation or mutual activity. The latter occurs more quickly in households with children. In some cases, they observe a welcome absence of undesirable activity (no police cars arriving at residences, no loud parties, no screaming, etc.). Through time, they begin to develop an opinion of those they observe as being menacing, helpful, apathetic, and so on. They engage in conversation and develop their opinions further. Some neighbors are friendlier than others. Some neighbors share one’s values more than others. All neighbors, however, to a greater or lesser extent, contribute to a neighborhood community. As a person interacts with each neighbor, she becomes more cooperative or less so, more guarded or less so. While each household brings a history of experiences into a new residence, they further develop that history, reinforcing or challenging past conclusions about the nature of neighborhood life. While at any given time the likelihood of a neighborhood responding efficaciously to a particular event might be determinable as a function of its residents and their attitudes, both the residents and their attitudes are constantly in flux.

HOMOPHILY AND INFLUENCE ACTING ON DIFFERENT STAGES OF NEIGHBORING Locational-based neighborhood effects such as residential differentiation and segregation correspond to influence-based neighborhood effects such as social capital and collective efficacy because in choosing to move away from dissimilar households, residents are implicitly choosing to segregate their networks of potential neighborly interactions as well. Since contact is a necessary prerequisite for interaction, if households settle in such a way that their immediate neighbors are similar to themselves, they have settled so as not to have neighborly interactions with those different from themselves. In resettling thus, they have segregated their phase 1 neighbor network. To the extent that they have settled randomly with respect to the other determinants of 56

SELECTION AND INFLUENCE

phase 2 neighbor networks, these networks will segregate as well. If the household characteristics that influence the transition of phase 2 neighbor networks into phase 3 networks and of phase 3 networks into phase 4 networks do not also covary with phase 1 neighbor networks, which is unlikely, phase 2 neighbor networks will translate into phase 3 and phase 4 neighbor networks in a straightforward way. Thus, segregation patterns will correspond to neighborhood effects because the same concatenated, multistage processes are guiding both. Locationalbased neighborhood effects such as residential differentiation and segregation correspond to the influence-based neighborhood effects such as social capital and collective efficacy because each acts on different stages of neighboring. Relocation determines stage 1 neighbors and thus, of necessity, the higher stages that influence works upon. Neighborhood communities result from both the concatenation of homophilous locational choices and the flow and exchange of norms, values, and beliefs among neighbors. Their correspondence is not additive, as in a regression model, but rather sequential. Relocation, which is responsible for residential differentiation and segregation, determines stage 1 neighbors and thus, of necessity, the higher stages that influence works upon to create social capital, collective efficacy, and other important neighborhood effects. Simultaneously, neighbors influence each other, for example in how welcome one is in the neighborhood, and in all that determines the neighborhood’s character. Which neighbors influence which other neighbors, and how much, makes all the difference, however.

MAIN POINTS IN REVIEW In this chapter, I conceptualized the foundations of neighborhood communities in terms of two forces: selection and influence. Households relocate, at least in part, to choose the type of households they want to have as stage 1 neighbors, relocating in favor of homophilous immediate neighbors, not homophilous neighborhoods. Since neighbors respond to household changes along their tertiary streets, the concatenation of these relocation events is necessarily delimited by the tertiary street network, and thus segregation patterns reflect it. Homophilous 57

CHAPTER FIVE

locational choice, however, cannot account for the entirety of neighborhood communities and their effects. A second community-generating force within neighborhoods consists of the flow and exchange of norms, values, beliefs, and influences among neighbors along their stage 4 networks. The structure of stage 4 influence networks can predispose neighborhood communities toward various collective norms and values, and either increases or decreases their inclination to produce social capital and collective efficacy. Neighborhood communities result from both the concatenation of homophilous locational choices and the subsequent sharing of norms, values, and beliefs among neighbors. The structure of stage 4 influence networks is heavily determined by the structure of the stage 1 tertiary street networks. Relocation, which is responsible for residential demographic differentiation, determines stage 1 neighbors and thus, of necessity, the higher stages of neighboring among which norms, values, and beliefs flow. Locational-based neighborhood community effects such as segregation correspond to the influence-based neighborhood community effects such as social capital and collective efficacy because, while each acts on different stages of neighboring, the same concatenated, multistage processes guide both. Selection and influence work together, sequentially, to produce homophilous neighborhood communities with varying degrees of social capital and collective efficacy.

58

C H A P T E R

S I X

Respondents, Interviews, and Other Data

I HAVE MADE many theoretical arguments in this book so far. I have argued that neighborhood communities and their effects are produced by networks of neighbors who interact. I have argued that the neighboring relationship has four stages, geographic availability, unintentional encounters, intentional contact, and influence. I have argued that the relevant geographic availability involves walking arenas such as tertiary streets. I have argued that, because children are much more geographically constrained to walking arenas than adults, they and their families are far more involved in neighborhood life. I have argued that selection and influence work together, sequentially, to produce homophilous neighborhood communities with varying degrees of social capital and collective efficacy. I have not yet empirically defended any of these arguments, however. To do that, I need relevant data. In this chapter, I review the data used to defend these arguments. Most of the data used in this book are original, deriving from ethnographic experience and four large collections of structured interviews. The data were collected in several distinct settings, a gang barrio, 68 Los Angeles neighborhoods, and a college town. The gang barrio was the site of the ethnographic study. The 68 Los Angeles neighborhoods, 20 of which were revisited several years later, added statistical robustness to my study and used an adaptive link-tracing to generate an interview chain that spread out spatially great distances in order to determine what factors constrained neighborly relations. A region in the college town was the site of an exhaustive census that fully mapped the geographically embedded neighbor networks. I revisited this same region three years later to discover how these same neighbor networks had evolved. I discuss these studies in detail below. 59

CHAPTER SIX

GANG NEIGHBORHOOD ETHNOGRAPHY AND INTERVIEWS My study of the gang barrio began as ethnography, although it did come to incorporate a set of structured interviews. I spent two years in ethnographic research in this neighborhood, attempting to understand the relationship between the gang and the neighborhood community in which it was situated. In addition to extensive participant observation, I conducted 184 structured interviews. These interviews included 13 veteranos, or gang elders, 42 active gang members, 67 other self-identified gang members, and 62 members of the local neighborhood community who claimed not to be gang members. The age of interviewees ranged from teenagers to senior citizens, although it was heavily skewed toward the younger members of the neighborhood. Respondents for the interviews were identified from a variety of sources that I had developed through my ethnographic experience as well as snowball samples1 begun from a local middle school, a Catholic parish, several small Protestant congregations, two gang outreach groups, and a newsletter operated by and for gang members. The initial respondents were asked to identify other residents whom they knew personally and believed to be “important” or “influential.” Other were identified in response to our other regular interview questions. This process was repeated for a period of two years (although by the second year very few new names were generated). Studies of networks have found that central (socially involved, boundary spanning, or highly visible)2 individuals are more accurate in their reports and are recalled more by others. Furthermore, recall errors tend to be biased in favor of more common and long-term, routine, and typical interaction,3 that is, the more a pair interacts, the more they recall each other and agree about their interactions with a third person.4 Therefore, respondents’ reports will be biased toward persons who are central to normative interaction patterns. The interviews focused on a wide variety of topics concerning gang life. For the purposes of this book, they included many questions about different types of relations individuals had with each other and the conditions surrounding their interaction. Respondents identified a total 60

R E S P O N D E N T S A N D OT H E R DATA

of 2,880 others with whom they had different types of neighborly relations. These relations ranged from simple acquaintance to economic exchange to loyalty. The interviews also concerned residents’ identifications and sympathies with various neighborhood entities.

OVERVIEW OF THE OTHER DATA COLLECTION EVENTS The other four data collections all differed substantially from the gang ethnography but were quite similar to each other in the format of the interviews and the content of the questions. All interviews were conducted at respondents’ residences by me and by teams of graduate and undergraduates working under my direct supervision. In these four data collections, 2,577 respondents identified 22,481 other residents with whom they claimed to share neighborly relations. These interviews were collected in four separate studies. In the first, 68 t-communities in Los Angeles were sampled. In the second, a complete census was conducted of a college town. In the third, the same college town neighborhood, along with some surrounding areas, was resampled. In the fourth, we resampled 20 of the original 68 Los Angeles neighborhoods. The Los Angeles data collections tested whether t-communities delimited neighbor networks. The exhaustive census explored whether multiple neighbor networks existed in a t-community, side by side, or whether t-communities typically corresponded to a single network of neighbors. The second college town data collection intentionally added both a longitudinal perspective as well as a comparative aspect by including surrounding areas.

STRUCTURED INTERVIEWS In all of these interviews, respondents were asked a standard set of questions about their neighbors and the geographic constraints affecting their meeting and interaction. An interview guide is included in the appendix. Specifically, respondents were asked to identify any neighbors that they “knew personally” and were subsequently asked a series 61

CHAPTER SIX

TABLE 6.1 Overview of Data Collections

Data collection

Dates interviews were conducted

Number of completed interviews

Number of neighbors identified

First Los Angeles data collection

January 1996 to June 1997

642

5,092

College town census

May to September 2000

998

10,883

College town resample

May to July 2003

582

3,578

Los Angeles resample

May 2004 to October 2005

355

2,928

2,577

22,481

Total

of questions about what kinds of activities they and their children (if they had children) engaged in with these neighbors (and their children). All questions concerning interactions with neighbors were asked about each neighbor who had been identified, one at a time. After these questions had been asked about each identified neighbor, respondents were asked if they engaged in these activities with any other neighbors. This frequently prompted people to recall neighbors they had not previously identified. If they did so, the series of questions was repeated about their interactions with these neighbors. At the very end of the interview, respondents were asked, one by one for each person they identified, to state how they had met the person and then whether or not they believed this occasion was an “unintentional meeting resulting from the mere fact of being neighbors.” Overall, respondents identified a median of nine neighbors and a maximum of 57. These numbers varied somewhat through the different data collections.

COGNITIVE MAPPING AND ALTERNATIVES The first Los Angeles sample and the college town census had a unique aspect that resulted from a limitation due to protocols on the treatment 62

R E S P O N D E N T S A N D OT H E R DATA

TABLE 6.2 Neighbors Known, by Data Collection L.A. L.A. College town College town time 1 time 2 census resample

Data collection

All

Number of people Maximum identified by respondents Median Minimum

48

48

46

57

37

9 0

11 0

9 0

9 0

7 0

of human subjects, but which I tried to convert to useful data. Protocols prevented me from asking respondents to provide full names and addresses for people who had not consented to be interviewed but about whom we were asking questions (i.e., about the relationships they had with the interviewee). To deal with these restrictions, I borrowed a cognitive mapping technique from geography.5 Respondents were given a blank piece of paper and a pencil and asked to “draw the streets in your neighborhood.” If they asked for further clarification (most did), they were told to visualize their neighborhood and to draw the streets in it. No other assistance was provided. The word neighborhood itself was intentionally not defined. Respondents were then asked to identify their residence on the map they had just created as well as the residences of the neighbors they had identified. Often respondents simply pointed to the residence or identified it as “the fourth house down” (pointing in a direction) or “the yellow house on the corner.” In all cases, however, interviewers collected sufficient information to allow residences to be unambiguously identified. Many respondents realized that the maps they had drawn did not include all of the neighbors they had identified, and most of them felt the need to make additions to their maps in order to include these households. This addition was never suggested to them, but it was allowed. Despite the fact that no one suggested to respondents that the neighbors they identified be limited to their maps or that the maps be limited to the neighbors, the fact that the cognitive mapping exercise preceded the neighbor identification in the same interview may have introduced 63

CHAPTER SIX

some bias. The second Los Angeles data collection and the college town resample attempted to deal with this possible bias by not requesting cognitive maps. For these data collections, I devised an alternative solution to determine whether neighbors (whether identified or not) lived in the same t-community, elementary school catchment, or census tract. At the end of the interview, after all other questions had been asked and long after neighbors had been identified, we presented the respondents with a map of the t-community and, one by one for each neighbor identified, asked whether the neighbor lived in this t-community. We also presented the respondents with a map of their elementary school catchment (which differed for different portions of the t-community) and, one by one for each neighbor identified, asked whether the neighbor lived in this elementary school catchment. Finally, we also presented the respondents with a map of their census tract (which again differed for different portions of the t-community) and, one by one for each neighbor identified, asked whether the neighbor lived in this census tract. We alternated the order in which the t-community map, school catchment map, and census tract map were presented to mitigate any order effects. Thus, no representation of a neighborhood, either drawn by residents or mentioned by interviewers, was introduced until the very end of the survey. All questions concerning neighbors were conducted without reference to geography. Again, this was done to control for the possibility that the cognitive mapping exercise might introduce bias. Because we needed to unambiguously identify the residences of those identified as neighbors (whose addresses we were not allowed to ask for), we asked residents to identify their residence as well as the residences of all the neighbors they had identified on whichever map was presented to them last (e.g., t-community, elementary school catchment, or census tract), which, as stated above, varied in order. If the identified neighbor’s residence was not on the last map, we allowed respondents to use either of the other maps to identify the residence. We were prepared with fold-out maps of their portion of the city to identify any who might live beyond all of the maps, but these maps were never required.6 64

R E S P O N D E N T S A N D OT H E R DATA

DATA COLLECTION IN 68 LOS ANGELES NEIGHBORHOODS The first Los Angeles data collection interviewed 642 residents sampled from a broad array of t-communities, 68 in total, from the Los Angeles Primary Metropolitan Statistical Area (PMSA). All t-communities in the Los Angeles PMSA were first identified.7 This initial sampling frame was stratified to yield t-communities of different size and population density, both because t-communities ranged in size from a few hundred to almost a half million residents and because t-communities ranged in density from those composed entirely of single-family homes with large lots to those composed entirely of large apartment buildings. The unusually large t-community with nearly half a million residents is in south central Los Angeles. The second largest t-community is substantially smaller, having a population of only 31,918. The largest sampled t-community had a population of 24,763, the smallest sampled t-community a population of 1,059. The mean population size for t-communities that could have been sampled was 14,351. The mean population size of actually sampled tcommunities was 13,546. All identified t-communities were partitioned into four equal groups by population density and then independently partitioned into four equal groups by number of residents. Thus, we classified all t-communities according to population and density as belonging to one of 16 distinct sets. From each of these 16 sets, four t-communities were selected at random, to yield 64 t-communities. The number of t-communities sampled increased to 68 because four of the t-communities originally targeted for interviews were located in what their assigned interviewer considered a “high crime” area and they asked to be reassigned. We chose four new t-communities for these interviewers, and I interviewed the original four t-communities. These sampled t-communities were surrounded by several other residential t-communities, generally being separated only by a nonresidential street. Specifically, an average of 5.82 other t-communities was immediately adjacent to each sampled t-community. While we did not control for this factor, 91 percent of these adjacent t-communities were residential in nature (rather than industrial or commercial zones). 65

CHAPTER SIX

Again, while not controlling for this factor, we verified that “boundaries” between the sampled t-communities and the t-communities immediately adjacent to them were typically (88 percent) no more substantial than nonresidential streets. Having selected the 68 t-communities, geographic coordinates were generated randomly to select the location of the initial household to sample. If the location was an apartment building or other multiunit dwelling, the interviewer used a random number generator to select the unit number. If the initially chosen household, either singlehousehold or multihousehold dwelling, was either unavailable or unwilling to participate, the household to its immediate right was chosen. If that household was unavailable or unwilling, the household to its right was chosen, and so on. An interview was requested from the first adult resident contacted within each selected dwelling unit. The scheduling of interviews was dispersed as much as possible to attempt to cover each day of the week equally as well as every hour between 8:00 a.m. and 7:00 p.m. Interviews themselves were conducted either at the door or in the respondent’s home if he or she so chose.

ADAPTIVE LINK-TRACING I have described how the initial household was sampled in Los Angeles. In both Los Angeles samples, subsequent households to interview were identified using an adaptive link-tracing protocol to generate an interview chain that traversed each of these neighborhoods thoroughly. After interviewing the starting household, the next household to interview was chosen according to the following criteria: it was identified as occupied by someone the current interviewee knew personally, its occupants had not yet been interviewed, and, of all such households, it was farthest away from the household that identified it. We chose the household that was farthest away to maximize the possibility that networks would spread out spatially in a “small world” effect8 and cross into other t-communities because we were interested in determining what constrained neighborly relations. In a few cases, the respondent identified no one he or she knew personally, or the person identified was not home or declined to participate. In these cases, we chose the next per66

R E S P O N D E N T S A N D OT H E R DATA

son from among those the previous interviewee knew personally and who satisfied our criteria; that is, we had not yet interviewed the person and, of all such households, this was the one farthest away from the household that identified it. We repeated the process with each successive household until we interviewed the tenth household or no households that satisfied the criteria remained. Fifty-five of the 68 chains terminated because 10 interviews were completed. One interview chain terminated after 10 interviews but no other households satisfied the criteria, two terminated after nine interviews, four terminated after eight interviews, one terminated after seven interviews, three terminated after six interviews, one terminated after five interviews, and one terminated after two interviews. Thus, the interview chains were quite long and could potentially traverse great distances. In total, surveys were requested from 998 persons. While 36 percent of the initial households declined to participate, only 8 percent of the successive households declined to participate. This change in rate of participation may have resulted because the interviewers told the potential interviewee that they had just finished interviewing a household that identified the potential interviewee (although they didn’t say which household).

THE SECOND LOS ANGELES DATA COLLECTION The second Los Angeles data collection was a sample of 20 of the original 68 t-communities in the Los Angeles PMSA, including one randomly selected t-community from each from the original 16 population-density strata as well as the four neighborhoods that had been added to replace the “high crime” neighborhoods in the first Los Angeles data collection because they had appeared to exhibit some theoretically interesting peculiarities. As with the first Los Angeles sample, the initial household to interview was selected using randomly generated coordinates, and subsequent interviewees were identified using the same adaptive link-tracing protocol as the first in order to generate an interview chain that traversed each of these neighborhoods thoroughly. All of the other 67

CHAPTER SIX

protocols were identical to the first data collection in treatment of multiunit dwellings, unavailable or unwilling respondents, interviewing the first adult resident contacted, scheduling interviews to equally cover each day of the week as well as every hour between 8:00 a.m. and 7:00 p.m., and conducting interviews either at the door or in the respondent’s home. The only distinction between the first and second sample was that I did not terminate the chain of interviewees until we interviewed the twentieth household, instead of the tenth. Fifteen of the 20 chains terminated because 20 interviews were completed, one terminated after three interviews, one terminated after seven interviews, one terminated after 11 interviews, one terminated after 15 interviews, and one terminated after 19 interviews. Thus, the interview chains were even longer than the first data collection and could have traversed great distances if no social forces prevented them from doing so. It needs to be noted, however, that this second set of interviews in some of the same t-communities several years later was not intended to be longitudinal but rather to verify the results by using both the longer sampling chains and by omitting the cognitive mapping exercise.

COLLEGE TOWN CENSUS AND RESAMPLE Two data collections occurred in the insular setting of a college town. The intent of the first data collection in the college town was to be an exhaustive census of a single t-community. We identified 1,214 residences in this t-community, and we attempted to interview each and every one of them. By the time we terminated the study, representatives of 998 households had agreed to interviews, yielding a response rate of 82 percent. Thus, our goal of an exhaustive census was reasonably successful. While the Los Angeles data collections tested whether t-communities delimited neighbor networks, this exhaustive census focused on the extent to which neighbor networks permeated t-communities. The exhaustive census explored whether multiple neighborhood community networks existed in a t-community side by side, or whether t68

R E S P O N D E N T S A N D OT H E R DATA

communities typically corresponded to a single network of neighbors. This was inspired by findings from the first Los Angeles data collection concerning the segregation of residents within t-communities by language. In the first Los Angeles data collection, all successive households spoke the same language as the household that identified them as “known personally.” Since it appeared that the neighbor networks of those who do not speak the dominant language within a t-community might be quite distinct, it was possible that other distinct networks might exist coextensively within a t-community. The college town was a reasonable community to test this hypothesis within because both census data and the interviewing experience confirmed that a substantial minority of residents did not speak English. Census data indicated that a large proportion (28 percent) of the t-community’s residents did not speak English even as a secondary language, a rate we confirmed when we interviewed households. Thus, this t-community offered the opportunity to investigate whether multiple distinct sociological communities existed. The t-community that formed the basis of this study was part of a middle-class residential neighborhood in a northeastern college town. It was bordered by four other t-communities, all residential in nature, and only nontertiary streets separated three of them from the sampled t-community. A highway separated the fourth. The second college town data collection sampled the same tcommunity three years later, and thus added a longitudinal element, taking a second snapshot of the same community. This data collection included in its entirety the same t-community as the exhaustive census, but for comparative purposes it also included some surrounding areas. Specifically, three elementary school catchment areas overlapped this tcommunity, one entirely contained within it and the other two sharing it with two other t-communities. For comparative purposes, our sampling frame was actually these three elementary school catchments, which included the original t-community sampled three years earlier in its entirety. The second data collection was not an exhaustive census attempting to interview every household, but instead a random sample of 30 percent of all households (363 households) was drawn. Of these, represen69

CHAPTER SIX

tatives of 303 households actually agreed to interviews, yielding a response rate of 84 percent. Two hundred and thirteen (213) of these respondents had been participants in our exhaustive census conducted three years earlier. To match the number of households sampled from the focal t-community, an additional 363 households were sampled from the two overlapping elementary school catchments in regions that were not part of the originally surveyed t-community. Of these, representatives of 279 households actually agreed to interviews, yielding a response rate of 77 percent. Thus, 582 respondents were interviewed in total. Among them, they identified 3,578 neighbors. The second college town data collection added a longitudinal perspective. The second college town data collection also included surrounding areas to add a comparative aspect.

ADMINISTRATIVE DATA Finally, I used administrative data to explore the same 70 neighborhoods I explored in my interviews and ethnography.9 For my administrative data, I used U.S. census data both because they are by far the most commonly used form of administrative data and because the census collects both geographic information, including data on streets, and demographic information. Specifically, I used census block group data because, in virtually all cases, tertiary streets internally connect them and thus they can be associated with t-communities with relative ease. Therefore, if two block groups can access each other via tertiary streets, all of the households within either block group can access all of the households in the pair of block groups via tertiary streets. This is often not the case for residents of larger units, such as tracts, which are often not internally connected by tertiary streets. I used 2000 data because the dates they were gathered most closely coincide with the dates of the interviews. It should be noted, however, that the t-communities were originally identified using 1990 data. Fortunately, no tertiary streets were changed in the sampled areas, so that t-communities were identical in 1990 and 2000. Furthermore, I performed all calculations using both 1990 data and 2000 data. No significant differences resulted. 70

R E S P O N D E N T S A N D OT H E R DATA

FIGURE 6.1. Distribution of t-community sizes (in number of households)

To match street data with geocoded block group data, I used Geographic Data Technology’s Pro/Filer product. The Pro/Filer product also provided geocoded shapefiles for elementary school attendance zones. As stated earlier, the advantage of block group data is that, in virtually all cases, tertiary streets internally connect the block groups. There were exceptions, however. Two block groups comprising the 70 t-communities I studied were internally disconnected via tertiary streets, and thus households within them could not access all of the households via these walking arenas. Thus, two t-communities were not exact combinations of census block groups but rather contained only part of a block group. For these t-communities, I estimated their populations using the populations of their constituent block groups weighted by the proportion of city blocks within those block group that were also in the t-community. In the end, the distribution of sizes of t-communities (displayed in figure 6.1) ranged from 221 households to 10,364 households with a mean of 1,890 households and a median of 1,670. The distribution of the number of block groups in a t-community ranged from 1 to 26 with a mean of 4.7 and a median of 4. 71

CHAPTER SIX

MAIN POINTS IN REVIEW To recap, most of the data used in this book are original, collected in several distinct settings at several distinct times, yet each of these distinct data sets contributes to understanding neighboring and how neighbor networks influence neighborhood outcomes in a unique way. For example, in addition to being useful for identifying the 70 tcommunities serving as the sampling frames for the interviews (and the overlapping neighborhood equivalents), the administrative data also allow me to examine racial segregation across these same neighborhoods. The interviews allow me to explore each of the stages of neighboring specifically, from the geographic substrate to the reality of passive contacts to the importance of households with children to the development of trust, norms, and values. The 68 Los Angeles neighborhoods add statistical robustness to my study, while the adaptive linktracing samples generated in these neighborhoods directly model the process of concatenation, uncovering the underlying neighbor network structures, and naturally testing the constraints of neighboring at all of its stages. The complete census of the college town allows me explore in depth the relationship between the various stages of neighboring, while its longitudinal nature allows me to explore how neighbors influence each other’s beliefs over time, how this process is guided by the structure of the tertiary street network, and how this relates to neighborhood-level outcomes. Finally, since the gang neighborhood was a well-established community providing identification, social capital, and efficacy in a way that surrounding neighborhoods did not, this allowed me to explore why this very real community was guided and constrained to associate with a particular geographic neighborhood.

72

C H A P T E R

S E V E N

Selecting Stage 1 Neighbors

SELECTING RACIALLY HOMOPHILOUS TERTIARY STREET NEIGHBORS How do we segregate our neighboring relations? We could segregate at any stage, at the stage 4 influence level, at the stage 3 interaction level, or at the stage 2 passive contact level. We could interact with those different from ourselves but only superficially, or we could have unintentional encounters with those different from ourselves but choose not to initiate contact, or we could be geographically available to those different from ourselves but arrange not to have encounters with them. All of these are possible ways we could segregate our neighboring relations. In this chapter, however, I show that the segregation of neighboring relations occurs at the most fundamental level. When residents segregate their neighbor network, they typically do so by segregating their stage 1 tertiary street neighbor network, rather than higher stages of neighboring. I will show that households relocate so that those households with whom they share tertiary streets are similar to themselves. To achieve this goal, they sometimes decide to move from their home, if those who share their tertiary streets are sufficiently different from themselves. They consider with whom they may share tertiary streets in potential future residences, and, if their attempts at homophily prove unsuccessful, they may move once again. My interviews showed that households did indeed relocate so that their neighbors would be similar to themselves. For 24 percent of those interviewed, the primary factor in the decision to move from their previous home was the desire to separate themselves from the people who lived in their neighborhood.1 Furthermore, if neighbors influenced the decision to move, it was because the household that relocated considered their previous neighbors too dissimilar from themselves. Respon73

CHAPTER SEVEN

TABLE 7.1 Similarity of Neighbors in Previous Neighborhood, by Previous Neighbors’ Influence on Decision to Relocate On a scale from 0 to 10, with “10” being virtually identical and “0” being completely different, how much were your neighbors in your previous neighborhood like you? Did your previous neighbors influence your decision to move from that area?

Yes (n = 630) No (n = 1941)

4.60*** (.195) 6.52*** (.131)

Note: Standard errors in parentheses. *** p  .001, one-tailed test that “Yes” is different from “No.”

dents who claimed that their previous neighbors influenced the decision to move judged their previous neighbors to be substantially less similar to themselves than did respondents whose previous neighbors did not influence their decision to move (4.60 vs. 6.52 on a 10-point scale, p  .001). Upon deciding to leave their previous residence, households considered who their new neighbors would be. For almost half of residents (44 percent), the type of people who lived in their current neighborhood was an important reason they chose to live there. As with the decision to leave their previous residence, respondents who considered who their current neighbors would be before moving to their current residence determined their neighbors to be more similar to themselves than respondents who did not consider who their neighbors would be (7.40 vs. 4.52 on a 10-point scale, p  .001). In addition to these two decision points, the decision to stay in a residence is influenced by one’s current neighbors. Those who desired to stay for many more years rated their neighbors’ similarity much higher than those who did not desire to do so (7.27 vs. 4.65 on a 10point scale, p  .001). Finally, what does it mean for neighbors to be similar? The interviews asked this open-ended question and received a myriad of distinct answers, too divergent to code meaningfully. I did note one interesting fact, however. Respondents claimed that virtually everyone (99 percent) 74

S E L E C T I N G S TAG E 1 N E I G H B O R S

TABLE 7.2 Similarity of Neighbors in Current Neighborhood, by Influence on Relocation Decision, Desire to Remain in Neighborhood, and Awareness of Different-Race Neighbor On a scale from 0 to 10, with “10” being virtually identical and “0” being completely different, how much are your neighbors in your current neighborhood like you? Were the people who live here an important reason for you to move into this area?

Yes (n = 1,140)

7.40*** (.160)

No (n = 1,437)

4.52*** (.134)

Would you like to live in this area for many more years?

Yes (n = 1,569)

7.27*** (.144)

No (n = 1,008)

4.65*** (.159)

Respondents identifying at least one current neighbor as being of a different race

Yes (n = 153)

3.29*** (.332)

No (n = 2,424)

6.44*** (.104)

Note: Standard errors in parentheses. *** p  .001 one-tailed test that, for each question, “Yes” is different from “No.”

they identified was of the same race as the respondent. Respondents who identified just one neighbor as racially different from themselves, however, felt much less similar to their neighbors than those who did not (3.29 vs. 6.44 on a 10-point scale, p  .001). Finally, these racial similarities maintained themselves and did not dissipate across concatenations of neighboring relations. This is shown clearly in the adaptive link-tracing samples. In the first Los Angeles data collection, every person interviewed in 45 of the 68 10-person interview chains claimed to be of the same race as the person who identified them, meaning that the entire chain claimed to be of the same race as every other member of the 10-person interview chain; and in the second Los Angeles data collection, every person interviewed in 11 of the 20 20-person interview chains claimed to be of the same race as the person who identified them, meaning that the entire chain claimed to be of the same race as every other member of the 20-person interview chain. Thus, the racial homogeneity of stage 1 neighboring, which results from households relocating so that their immediate tertiary street 75

CHAPTER SEVEN

neighbors are similar to themselves, remains robust in the aggregated neighbor networks.

ACCEPTING HETEROGENOUS HIGHER-STAGE NEIGHBORS It appears that residents choose to racially segregate their stage 1 tertiary street neighbors, but do they racially segregate their neighboring relations at higher stages as well? The college town census specifically allowed me to test this question about segregation by race. The neighborhood within which we conducted our interviews had a population that was primarily white and Asian. In the neighborhood community that was the subject of the second college town study, the population was 34 percent Asian (according to census records, and consistent with our interview findings). If I focused only on white households, using only respondents who were white and identified neighbors who were white, how did characteristic path lengths2 differ? While the characteristic path length for all neighbors was 2.67 steps in this setting, the characteristic path length between two white households was 2.60 steps. The characteristic path length between two Asian households was 2.44 steps. The characteristic path length between a white household and an Asian household was 2.85 steps. An Asian household was, on average, 0.41 steps closer to another Asian household than to a white household. A white household was, on average, 0.25 steps closer to a white household than to an Asian household. Given the size of the population, all of these differences are statistically significant beyond any conventional criteria, but they are not quite the radical findings one might expect from real segregation. One group was, however, substantially more internally focused than others. Before asking households to identify themselves with standard racial categories, the survey asked them to identify their ethnicity using any category they saw fit. One such self-identified group, Koreans, which included almost half of all Asians in the sample, had a remarkably short characteristic path length of 1.59 steps, almost a full step shorter than others, indicating that whatever cultural values flowed along neighbor networks, they flowed much more immediately amid 76

S E L E C T I N G S TAG E 1 N E I G H B O R S

Koreans than they did among other groups. This group was not only ethnically Korean but linguistically as well, and most interviews required Korean-speaking interviewers to complete. Interestingly, when I divided the class of Asians into non-Koreans and Koreans, the nonKoreans had an average characteristic path length among each other of 2.63 and an average characteristic path length with whites of 2.68. NonKorean Asians, however, had an average characteristic path length with Korean Asians of 2.95. Clearly, whatever segregation occurred in this community occurred between Koreans and everyone else. I do not wish to imply that this is a special characteristic of Korean households in general. It was only a special characteristic of most Korean households in this particular neighborhood setting. The college town consisted of many neighborhoods; however, the neighborhood in which we interviewed was the only one with a substantial Asian population. Several other neighborhoods with very similar socioeconomic characteristics existed, but they were uniformly homes to white populations. Thus, the choice of whether or not to interact as neighbors occurred primarily at the stage 1 level in this setting as well. Once white residents had settled near Asian residents, once they had chosen to share stage 1 neighboring relationships, higher stages developed quite readily with little consideration of race. The distinction we did find, however, was segregation of higher-stage neighboring relations by language. Recall from chapter 6, which discussed the data used in this book, that the very possibility of linguistic segregation raised in the first Los Angeles data collection was one of the reasons for the college town census. In the first Los Angeles data collection, all successive households in the sampled chains spoke the same language as the household that identified them as “known personally.” It raised the possibility that distinct neighbor networks of those who do not speak the dominant language within a t-community might exist coextensively within a t-community. While the neighbor network of Korean speakers we discovered was not completely isolated, they did cluster somewhat separately from the rest of the neighborhood community. This indicates that while racial differences may not impede the translation of stage 1 tertiary street neighbors into stage 3 actualized neighbors, linguistic differences may. 77

CHAPTER SEVEN

A DIALOGUE WITH ADMINISTRATIVE DATA Previously, I argued that sociologists have treated neighborhoods as if they were only colored boxes on a map or sets of geo-referenced variables for use in a geographic information system (GIS) and that this approach inappropriately emphasizes those aspects of neighborhoods and their residents that can be effectively displayed or referenced in a space and ignores those that cannot, such as social-interactional processes and dynamics. In the first part of this chapter, I examined one of the most commonly studied geo-referenced variables, race, and showed that households do indeed relocate so that their stage 1 tertiary street neighbors are racially homophilous. I argued that, because neighbors are forced to share lives, each new household affects the potential neighbor networks of those sharing its tertiary streets. Neighbors potentially respond to the changes along their tertiary streets by exiting their old neighborhood. Home-seekers choose their new neighborhood and reject alternatives, in part, because of who their potential neighbors will be. Furthermore, home-seekers often reject potential neighbors based on characteristics that are readily observable when searching for a new residence, with race being one of the most dramatic bases for rejecting neighbors, as it is often perceived to indicate status, crime rate, and other concerns. These local relocation events concatenate. Since neighbors are responding to household changes along their tertiary streets, the concatenation of these relocation events is necessarily delimited by the tertiary street network and thus reflects it. Segregation patterns will reflect tertiary streets because urbanites relocate in part to control who share these streets. While these conclusions are based on both what I believe to be sound logic and thousands of interviews in both the college town census and the Los Angeles adaptive link-tracing samples, I would be remiss if I assumed that most quantitative analysis of neighborhoods will not continue to rely on geographic entities defined by the Census Bureau or other administrative agencies. In addition to methodological inertia, the enormity of available data will likely prove ample motivation to continue this approach. 78

S E L E C T I N G S TAG E 1 N E I G H B O R S

Thus, while administratively defined neighborhood equivalents such as census tracts and block groups were not designed to measure potential interaction among resident members, very little research will build upon my model unless I show how to proxy the less visible characteristics of my model using features that are conspicuous to these massive databases.

SEGREGATING TERTIARY STREET NETWORKS While I will certainly include the precision of numerical statistics, I would be remiss if I did not begin a geo-statistical analysis with a map. Specifically, I begin by illustrating what I would expect tertiary streetbased segregation patterns to look like on a map. The Los Angeles PMSA, from which I sampled my 68 neighborhoods, contains hundreds of thousands of streets and is far too large to be meaningfully displayed on a map. Instead, I display one of its larger and more famous secondary cities, Pasadena. Pasadena is more than a century old, and the majority of its street system was originally built by the early part of the twentieth century. I say “originally” because many older cities (Pasadena included) have renewed their street systems in the last few decades. Pasadena has a fair amount of racial and economic diversity, from estatelike homes owned primarily by whites to primarily black ghettos. Figure 7.1 displays Pasadena’s tertiary streets. The bold lines represent tertiary streets with majority nonwhite populations, while the nonbold lines represent tertiary streets with majority white populations. Disconnected networks of tertiary streets are immediately visible. The large nonwhite region occupies the eastern half of a single large network. As for the smaller, nonwhite regions in the east, nonwhites occupy tertiary street networks distinct from those occupied by whites. Thus, whites and nonwhites have settled in such a way as to minimize contact along pedestrian-oriented tertiary streets rather than to minimize spatial proximity. Without even noting the presence of the larger thoroughfares that dissect tertiary streets into t-communities, tertiary street networks can account for most of the white-nonwhite segregation in Pasadena. 79

CHAPTER SEVEN

FIGURE 7.1. Pasadena’s tertiary streets by residential demographics Lines represent tertiary streets. Bold lines represent tertiary streets with majority nonwhite populations. Non-bold lines represent tertiary streets with majority white populations.

Figure 7.2 is a “close up” of the “frontier” between the white and nonwhite populations on the large, northern network of tertiary streets. Lines represent tertiary streets. The bolder lines represent tertiary streets with majority white populations, while the less bold lines represent tertiary streets with majority nonwhite populations. The gaps between tertiary streets represent nontertiary intersections. Two tcommunities are visible, one on the right and one on the left. The one on the right is composed of tertiary streets that are all majority white, while the one on the left is composed of tertiary streets that are 80

S E L E C T I N G S TAG E 1 N E I G H B O R S

FIGURE 7.2. Enlarged subsection of figure 7.1

all majority nonwhite. Census geography subdivides the t-community on the right into three separate block groups and subdivides the tcommunity on the left into five separate block groups within two distinct tracts. In Pasadena, discontinuities in racial demographics appear to map onto discontinuities in the tertiary street system. I now statistically analyze the 70 neighborhoods within which I conducted my data collections. I ask whether, across these 70 neighborhoods, residents of the same t-community or same island are more demographically similar to each other than they are to residents of different t-communities or different islands. To what extent can the demographic variability among block groups be accounted for by the t-community or island they are in? To answer these questions, I begin with table 7.3 displaying the adjusted eta-squared statistic, which is analogous to R2 for nominal data, which indicates what proportion of the demographic variability between block groups can be accounted for by either the island (column 1) or t-community (column 2) they are a part of. For non-Hispanic whites, about two-thirds of the variation in their distribution among block groups can be accounted for by assigning their block group the average of its island. T-communities do not offer much additional explanatory power, indicating that most of what is occurring happens at the larger island level. For the other racial 81

CHAPTER SEVEN

TABLE 7.3 Demographic Variability Accounted for, by Island or T-community Racial category

Island

T-community

Non-Hispanic white

0.59***

0.64***

Black

0.39***

0.81***

Hispanic

0.47***

0.62***

Asian

0.48***

0.51***

*** p  .001

categories, about half of the variation in their distribution among block groups can be accounted for by assigning their block group the average of its island. In all three cases, t-communities offer much more additional explanatory power than they did with non-Hispanic whites and, in the case of the distribution of black population members among block groups, it doubles the explanatory power, accounting for almost all of the variation. This reflects what we observed in figure 7.2, in which the large island was divided among the white and nonwhite population but the t-communities clearly were not. It appears that, for nonwhites and especially blacks, who is down the street matters less if a nontertiary intersection intervenes. While the eta-squared statistic indicated that, compared to islands, t-communities offered much more explanatory power for nonwhites than whites, it would be useful to tease apart the relative contribution of tertiary and nontertiary intersections as barriers to racial homogeneity. Table 7.4 displays the results from ordinary least squares (OLS) regression of the proportion of a block group’s racial proportion for the four racial populations (non-Hispanic white, non-Hispanic black, Hispanic, and non-Hispanic Asian) on the racial proportion of the tcommunity and island a block group is in. For all racial populations, the t-community variable is significant and substantial; in the case of the white and Hispanic populations islands are significantly and substantially related to additional racial variability, while in the case of the black and Asian populations, this is not true. It appears that nontertiary intersections form more effective social barriers for the black and Asian populations than for the white and Hispanic populations. 82

S E L E C T I N G S TAG E 1 N E I G H B O R S

TABLE 7.4 Block Groups’ Demographics as a Function of Demographics of Island or T-community Independent variable proportions

White

Black

Hispanic

Asian

T-community

.860**

1.005**

.871**

1.004**

Island

.545**

.067

.517*

.117

.003

.141*

.011

Constant Adj. R2

.170* .627**

.798**

.578**

.477**

Note: N = 327 block groups. * p .05 ** p  .01.

Given the importance of children, I also explore how much of the block group level racial variability among only those households that have children can be accounted for by the t-community their block group is a part of. While I have shown that tertiary streets are more important for children, they are also more important for them because of the presence of other children. Therefore, I have two independent variables in my model, the average population proportion by race of the t-community a block group is in and the t-community level average racial proportion of the households with children. Using the distribution of black populations as an example, we have the following: Yi = ß1X1i + ß2X2i + i, where i indexes block groups and Yi represents the proportion of households with children in block group i that are black. X1i represents the proportion of all households in the t-community that block group i is in that are black (excluding those households actually in block group i itself). I exclude block group i because the typical t-community contains only a few block groups; therefore each block group was compared to the average of the “other” block groups (not including itself) in its t-community to avoid creating a built-in statistical dependency. X2i represents the proportion of households with children in the t-community that block group i is in that are black (excluding those households with children actually in block group i itself). 83

CHAPTER SEVEN

TABLE 7.5 Block Groups’ Demographics as a Function of Demographics of T-community, by Presence of Children Independent variable proportions

White

Black

Hispanic

Asian

T-community average of all households

.446***

.219***

.340***

.353***

T-community average of households with children

0.556***

.779***

.673***

.652***

Constant

.009

2

Adj. R

.005

.772***

.894***

.036 .696***

.004 .487***

* p  .05 ** p  .01 *** p  .001.

Table 7.5 shows that variability in the racial proportion of households with children in a block group is more substantially related to the racial variability of other households with children in the same t-community than to the racial variability of households in general. This is truer for nonwhite populations than for white populations and especially for black populations. The adjusted R2 statistics show that, for most racial groups, much more of the racial variability of households with children is accounted for than households in general. The exception is Asian households, in which only slightly more variability among households with children is accounted for than among households in general. To avoid the possibility that a few t-communities with low proportions of households with children might skew the results, I reran them several ways using various proportions of households with children as cut-offs. All results were similar.

TERTIARY STREET NETWORK BORDERS I also ask to what extent t-communities create sharp discontinuities in demographic patterning. If my model is correct, one would expect demographic differences along borders between t-communities to be more extreme than internal differences within t-communities. I opera84

S E L E C T I N G S TAG E 1 N E I G H B O R S

tionalize this possibility by examining whether the difference between a pair of spatially adjacent block groups is likely to be more extreme if the pair exists along a border between different t-communities or if the spatially adjacent block group is interior to a t-community. What do I mean by border and interior? The 70 t-communities I examine in this book are composed of 327 block groups. Each of these block groups is spatially adjacent to multiple others; that is, it shares a geographic border with them. I refer to these spatially adjacent block groups as pairs of spatially adjacent block groups. Some of these block groups are within the same t-community, and some are immediately adjacent to one of the block groups in the t-community. In all cases, at least one block group in the pair is in one of the 70 focal tcommunities. In total, 679 such pairs exist in which one of the 327 block groups composing my 70 t-communities is spatially adjacent to another block group. Given that there is demographic variation in urban areas, not all of these block groups are demographically identical to their adjacent block groups. In fact, very few of them are. Most differ by about 10 or 20 percent in the racial composition relative to the block groups that border them. Figures 7.3–7.6 graph the distributions in absolute percentage difference between the 327 block groups in our 70 t-communities and the 679 other block groups they are adjacent to. Clearly, there is variation. To what extent does that variation reflect the border between two t-communities rather than internal variation within a t-community? In some of the 679 pairs of spatially adjacent block groups, both members of the pair are within the same tcommunity, 141 in total. I refer to these pairs of spatially adjacent block groups in the same t-community as internal adjacencies. Some of these pairs of spatially adjacent block groups are in two different tcommunities, 538 in total. I refer to these pairs of spatially adjacent block groups, where one is in one t-community and the other member of the pair is in a distinct t-community, as borders. My question then becomes, Are pairs of internally adjacent block groups more similar to each other than pairs of border block groups? Table 7.6 shows that, for all four racial groups, pairs of spatially adjacent block groups that formed borders between t-communities were much more different from each other than were spatially adjacent block 85

CHAPTER SEVEN

FIGURE 7.3. For spatially adjacent block groups, distribution of absolute differences in population that is white

FIGURE 7.4. For spatially adjacent block groups, distribution of absolute differences in population that is Asian

86

S E L E C T I N G S TAG E 1 N E I G H B O R S

FIGURE 7.5. For spatially adjacent block groups, distribution of absolute differences in population that is Hispanic

FIGURE 7.6. For spatially adjacent block groups, distribution of absolute differences in population that is black

87

CHAPTER SEVEN

TABLE 7.6 Percentage Racial Difference between Spatially Adjacent Block Groups, by Whether They Are Barriers, Borders, or Internal Adjacencies Percentage difference

White

Hispanic

Asian

Black

Barrier

Mean Standard error

26.95 0.93

22.10 0.91

14.18 0.63

9.13 0.23

Border

Mean Standard error

27.40 0.45

22.30 0.46

14.22 0.32

9.20 0.11

Internal

Mean Standard error

17.10 0.24

12.10 0.21

7.73 0.14

3.96 0.09

.282

.249

.182

.486

Adjusted eta-squared

groups that were internal to the same t-community. For example, the average absolute difference in white population between two spatially adjacent block groups that were borders was 27 percent, while the average absolute difference in white population between two spatially adjacent block groups that were internal to the same t-community was 17 percent. The statistic eta-squared shows that from 18 percent to 49 percent of all the variation between spatially adjacent block groups was accounted for by their being borders or not. Of course, it could be the case that the subtle geography of tertiary streets and intersections is confounded with the more pronounced geography of freeways and other types of major boundaries, and that they are the more likely cause of this apparent effect. I need to verify that the two do not coincide. To test this possibility, I compared the set of borders that were confounded with natural barriers (e.g., freeways, highways, rivers, hills) with those that were not. Of the 538 pairs of border block groups, 103 were confounded with natural barriers. Table 7.6 summarizes the distributions of the absolute percentage racial differences for three different types of spatially adjacent block group pairs, those that are internal adjacencies, those that are borders but are not confounded with natural barriers, and those that are borders and are confounded by natural barriers. What is clearly apparent is that both types of borders are nearly indistinguishable from each other in terms of their racial disparity. What is 88

S E L E C T I N G S TAG E 1 N E I G H B O R S

also apparent is that both types of borders are distinct from internal adjacencies in terms of their racial disparity. In these cases, at least, sharp demographic discontinuities relate to sharp discontinuities in the tertiary street network, but these discontinuities do not typically result from “natural barriers.” In fact, the eta-square is unchanged to three significant digits among all nonwhite racial categories and to two digits among whites, indicating that these natural barriers do not add any explanatory power to borders. I need to note that one should not interpret this lack of difference to mean that these natural barriers do not create discontinuities in the distribution of population demographics. They certainly do. Instead, these results show that “invisible” discontinuities in the network of tertiary streets are just as disruptive to population distributions as natural barriers are. It does not require a freeway to disconnect a neighborhood.

THE IMPACT OF A SINGLE TERTIARY STREET CONNECTION I identified t-communities as maximal contiguous sets of tertiary streets and tertiary intersections. This only requires that a single tertiary street path be available within them along which neighborly relations can concatenate and through which norms and values and influence can flow. While I have shown that a “border effect” exists, distinguishing urban areas that have no tertiary street connection from those that have at least one tertiary street connection, it is possible that this result is artificially truncating the data. Perhaps, tertiary street networks only become relevant when many internal tertiary street pathways are available. This has obvious policy implications. If adding a single street to connect two otherwise disconnected communities was all that really mattered, then we would expect that the difference between those adjacent block groups connected by no tertiary streets and those adjacent block groups connected by only one tertiary street to account for most of the difference. If more involved cohesiveness was required, I would expect that there would be a gradual progression with the most extreme differences between block groups that were not connected by tertiary streets, 89

CHAPTER SEVEN

slowly diminishing until the most subtle differences occurred between block groups that were connected by many tertiary streets. To test this possibility, I compare the difference between spatially adjacent block groups connected by only one tertiary street to those connected by many, and of course to those connected by none for a baseline. Thus, I partition the racial differences between spatially adjacent block groups into six categories: those that are not connected by tertiary streets (n = 538), which I have labeled borders to indicate their discontinuity, those that are connected by only a single tertiary street (n = 39), those that are connected by exactly two tertiary streets (n = 35), those that are connected by exactly three tertiary streets (n = 32), those that are connected by exactly four tertiary streets (n = 19), and those that are connected by five or more tertiary streets (n = 16). Figure 7.7 displays the average of these absolute differences. In all five cases (the four racial categories and households with children), each set of rows represents a different demographic category. The rows are labeled as representing the average absolute difference between internal adjacencies sharing one, two, three, four, or five or more tertiary streets. A row labeled “borders” that represents adjacent t-communities that are not connected by any tertiary streets, is included for comparison. While sharing more streets is related to demographic similarity, clearly the most substantial division is between those that are in the same t-community and those that are not. This should not be interpreted to mean that tertiary street networks with more tertiary streets interconnecting them are not more substantial than those with fewer, only that whatever stage 1 neighbor networks are allowed for by a single tertiary street connection are sufficiently robust to allow for segregation.

MAIN POINTS IN REVIEW In the first part of this chapter, I explored stage 1 neighboring relations and showed that households do indeed relocate so that their stage 1 tertiary street neighbors are homophilous. They sometimes decide to move if those who share their tertiary streets are different from themselves; they consider with whom they would share tertiary streets in 90

S E L E C T I N G S TAG E 1 N E I G H B O R S

FIGURE 7.7. Demographic differences between spatially adjacent block groups by number of shared tertiary streets

91

CHAPTER SEVEN

potential future residences, and, if their attempts at homophily prove unsuccessful, they desire to move once again. More than any other factor, respondents correlated racial similarity with homophily. If residents of different races did settle near each other, however, higher stages of neighbor networks generally developed without further reference to racial disparities. When residents racially segregated their neighbor networks, they typically did it by restricting their geographic availability, by segregating their stage 1 neighbor network, rather than higher stages of neighboring. However, while racial differences did not impede the translation of stage 1 tertiary street neighbors into stage 3 actualized neighbors, linguistic differences did. To translate stage 1 neighboring relations into stage 3 neighboring relations, it helped to speak the same first language. In the second part of this chapter, I used administrative data to examine racial segregation across the 70 neighborhoods I explore in my interviews and ethnography. I showed that discontinuities in the distribution of racial demographics mapped onto discontinuities in the tertiary street system, especially for the racial distribution of households with children. T-communities and islands have clear “borders” where sharp discontinuities occur in the distribution of racial groups. Furthermore, “invisible” discontinuities in the network of tertiary streets are just as disruptive to population distributions as natural barriers are. Finally, while sharing more tertiary streets related to greater demographic similarity, the most substantial distinction occurs between those that live in the same tertiary street network and those that do not. A single trivial tertiary street connection may profoundly affect the demographic composition of two previously disconnected neighborhood communities.

92

C H A P T E R

E I G H T

Unintentional Encounters

THE SUBSTANTIVE REALITY OF PASSIVE CONTACTS In my survey of households, respondents desired that their neighbors be similar to themselves. Some of them, to achieve this aim, had moved from their previous home. They considered who their new neighbors would be when choosing their current residence, and some of them, if their attempts at homophily had proved unsuccessful, desired to move once again. Why do households care so much about their neighbors? Because, in contrast to my Star Trek thought experiment where everyone “beamed” from home directly to a destination, there is a high probability of contact between neighbors through convenient access. In fact, I have argued that this convenient necessity is the essence of being neighbors. Neighbors are those we meet unintentionally because of their geographic availability to our residence. To demonstrate this point more fully, I return to the interviews. As I have described, respondents were asked to identify neighbors whom they “knew personally.” For each neighbor, one at a time, they were asked to state how they had met each neighbor so identified. After doing so for each and every neighbor, they were asked, again for each neighbor one at a time, whether or not the way they had met the neighbor was an “unintentional meeting resulting from the mere fact of being neighbors,” what I have termed a passive contact. The interviews revealed that passive contacts are a real phenomenon, not merely a theoretical construct. Respondents had no difficulty stating whether meeting someone was the unintentional result of their being neighbors. As the pie chart in figure 8.1 makes clear, a few types of events made up the overwhelming majority of passive contacts, and primarily involved children. First and foremost, 61 percent of all passive contacts began when children casually played together. This is the dom93

CHAPTER EIGHT

FIGURE 8.1. Passive contacts by type

inant type of passive contact. The geographic dependence of the relationship between children became apparent when different respondents in the same neighborhood mentioned the same boundary beyond which their children were not supposed to play (“Don’t cross [name] street,” or “Don’t go beyond the bend”). Having restricted their children to the same quite finite play space, parents essentially chose their playmates as well. Another 21 percent of passive contacts began when parents from two households walked their children to school together or waited with other parents and their children at the school bus stop or took turns doing so. Interestingly, 10 percent of all passive contacts occurred when 94

UNINTENTIONAL ENCOUNTERS

at least one of the parties was walking a dog through the neighborhood. In fact, for those who had dogs but did not have children, these meetings accounted for slightly less than half (46 percent) of all passive contacts. An additional 4 percent of passive contacts took place at neighborhood functions such as block parties or local school events. The residual category “Other” in figure 8.1 included such events as meeting parents when children went door-to-door during fund-raising campaigns or looking for work, or being introduced to one neighbor by another, or encountering someone already known from church or another outside activity, or a number of similar occasions. While passive contacts generally consisted of only a few types of activities, it is worthwhile to turn the question around and ask, Were these activities generally considered passive contacts? While agreement was not universal, respondents shared similar notions of when a meeting was a passive contact. As I have noted, by far the most likely way for two neighbors to meet was through children casually playing together. Respondents universally agreed that this was passive contact. In the four structured interview data collections, 8,841 neighborly relations were identified as having begun when the children of two households casually played together, and every single one of them was identified by the respondent as an unintentional meeting resulting from the mere fact of being neighbors. Almost as unanimous a result, 99 percent of neighbors who reported meeting through walking children to school together or waiting at the school bus stop reported this contact as passive. These activities are also mediated by urban geography, and specifically by tertiary street networks. Again, as I have stated, meetings involving children were generally identified as passive contacts. and meetings not involving them generally were not. There were two clear exceptions to the connection between meeting through one’s children and passive contacts. First, 99 percent of neighbors who mentioned meeting while walking dogs reported this contact as passive, despite the absence of children. In addition, about threequarters of all neighbors who reported meeting through a child seeking work (e.g., babysitting, yard work) or fund-raising (e.g., Girl Scout cookies, raffle tickets) reported the contact as nonpassive, despite the presence of children. 95

CHAPTER EIGHT

Other activities, not very common but nearly universally identified as passive contacts included meeting through a neighborhood incident (e.g., helping to pull a neighbor’s car out of a ditch, calling the police about a neighbor’s loud party). In contrast, some activities were almost never identified as passive contacts. Neighbors who knew each other through work or business relations (e.g., one person is another’s insurance agent) universally agreed that their contact was not passive. Respondents who knew their neighbors from activities outside of the neighborhood (such as through mutual friends) or through organized recreation (e.g., dance class) rarely reported these meetings as passive contacts. Some respondents knew their neighbors because they sought out someone of the same ethnicity or who attended the same church. Only a small fraction of these contacts were reported as passive. In fact, several of the respondents who did report such events as a passive contact felt the need to account for their description and made such statements as “I just kept my eyes open.” About a quarter of all those who reported meeting at neighborhood functions (e.g., block parties, neighborhood watch) labeled this a passive contact, while threequarters did not. Table 8.1 itemizes some of the ways respondents identified meeting their neighbors, the number met that way, and the number (and percentage of the total) who identified the occasion as a passive contact. Clearly, for most types of relationships, there was very little ambiguity about whether the event was a passive contact.

THE “LIVED” EXPERIENCE OF TERTIARY STREET NETWORKS In the next chapter, I will show that actualized stage 3 neighborly relations are a subset both of stage 1 neighbor networks, represented by tertiary streets and intersections, and of stage 2 neighbor networks, represented by passive contacts. If both of these claims are true, then it seems reasonable that intermediate stage 2 neighborly relations must therefore be a subset of tertiary streets and intersections. I would like to do more than infer this result, however. I would like to demonstrate it empirically. 96

UNINTENTIONAL ENCOUNTERS

TABLE 8.1 How Respondent Met Neighbor, by Percentage Identified as Passive Contact Number Number met identified as a How respondent met neighbor this way passive contact

Percentage identified as a passive contact

Children casually playing together

8,841

8,841

100

Walking children to school together or waiting with other parents and their children at the school bus stop

3,081

3,051

99

Walking dogs

1,444

1,429

99

34

32

94

Neighborhood functions

2,338

565

24

Neighborhood children seeking work or selling things

1,504

358

24

Organized recreation (e.g., dance class)

134

16

12

Intentionally seeking a neighbor of the same ethnicity

774

59

8

Intentionally seeking a neighbor who the same church

492

30

6

1,041

45

4

Patronizing a business

164

0

0

Working together

283

0

0

Neighborhood incident

Mutual friends

Doing so is more difficult than it would seem. Passive contacts are quite elusive to measurement. When they have been translated into stage 3 actualized relations, residents can discuss them; but residents are not always fully aware of passive contacts that have not resulted in such relations. I tried to determine where residents’ passive contacts, with people whom they had not yet initiated interaction, occurred in their neighborhood, and had some limited success. It proved impossi97

CHAPTER EIGHT

ble, however, to ascertain whom respondents had such passive contacts with, since they did not yet know the person, much less where he or she lived. The cognitive mapping exercise described in chapter 6 provided an opportunity to explore the potential range of residents’ passive contacts and to test whether my theory that stage 2 neighborly relations were circumscribed by tertiary streets and intersections was reliable. Recall that respondents were asked to draw the streets in their neighborhood and to locate on those maps their own residence and those of any neighbors they knew personally, and that respondents were given no guidelines on constructing their cognitive maps. As a result, there was a tremendous variety in size, style, and detail, not to mention legibility. In what follows I explore two sample cognitive maps and discuss what they tell us, along with their 1,638 companion maps. The first cognitive map (fig. 8.2) includes 18 face blocks and 11 intersections. While the respondent’s map doesn’t show the fact, all of them are tertiary. Thirty-five residences are identified, all on a single face block, 18 on the right side and 17 on the left side. One, identified by the number “142” that was scratched out, relabeled, scratched out, and relabeled again, is the respondent’s house. Seventeen other households are identified solely by numbers indicating that they are “known well,” a designation created by this respondent. Seven households are identified by circles, indicating that they are known but not well, again a designation created by this respondent. Two households are identified by circled numbers; two more are identified by numbers scratched out and replaced by circles, indicating that the respondent felt uncertain about knowing these neighbors well. Four households are identified by X’s, indicating that they are known only a little, again a designation added by this particular respondent. One household is identified by both an X and a circle, indicating that they also were in some intermediate stage between known a little and known. One residence is identified by the word “vacant.” This was one of 29 maps that were difficult to process because the respondent added her own categories. Did she identify 17 households she knew personally, or 33, or some number between? I determined the number to be 33. However, for the 29 maps on which respondents identified multiple categories of neighbors they knew personally, I re98

UNINTENTIONAL ENCOUNTERS

FIGURE 8.2. Sample cognitive map 1

calculated all analyses using their largest and smallest numbers. No significant deviations from what is presented in this book resulted. What is unambiguous is the fact that all of the neighbors she knew personally, whether 17, 33, or somewhere in between, lived on her tertiary face block. What does such a map tell us? The respondent’s cognitive understanding of her neighborhood was limited to a small set of intercon99

CHAPTER EIGHT

nected tertiary face blocks and tertiary intersections (this map was more detailed than most). Her conception of her neighborhood did not reflect the name a real-estate agent might give it, or a school district or church parish. It reflected a “lived” experience. While her experiential neighborhood was a subset of her t-community, the neighbors she knew personally were a subset of this experiential neighborhood. Most of the neighbors the respondent “knew well” lived closer to her, while those she did not know well or knew a little lived farther away. The second cognitive map is different in style. While the previous map uses only lines, numbers, circles, and X’s, this one draws streets with width and rectangular houses. Overall, respondents varied dramatically in how elaborately they drew their streets. Some drew houses with roofs as if they were on their sides, and some drew trees, stepping stones, and other lawn features. In the cognitive map in figure 8.3, there are 12 face blocks and seven intersections, again all tertiary. Eighteen residences are identified, spread throughout the face blocks. One is identified as “My house.” Four other households are identified by the names of residents. The neighbors identified in this map are spread out farther than those in the previous map. One neighbor lived beyond two intersections, one beyond three, one beyond five, and one beyond six. While the respondent’s cognitive world is again a subset of his t-community, his world of identified neighbors approximates this cognitive world in geographic size. The neighbors he knew were spread across the neighborhood he was aware of. While it would be best to find a way to test for actual passive contact, these drawings suggest how residents understand their neighborhoods. If one assumes that residents are aware of the places where recurring passive contacts occur, then the location of these stage 2 neighborly relations would be delimited by these maps. The cognitive mapping exercise was part of the first set of Los Angeles interviews, the 68 neighborhoods, as well as the first set of interviews in the college town, the exhaustive census. In total, these interviews generated 1,640 cognitive maps. In the Los Angeles interviews, respondents’ cognitive maps of the “streets in their neighborhood” ranged widely in size. Several people drew only a single street, while one respondent retrieved a Thomas Brother’s Street Guide and declared that 100

UNINTENTIONAL ENCOUNTERS

FIGURE 8.3. Sample cognitive map 2

the entire world was his neighborhood (when the interviewer reminded him that he needed to draw the streets in his neighborhood, he sketched a more modest neighborhood). In the college town interviews, respondents’ maps did not range in size nearly as much. The largest maps contained 11 streets, while many contained only one or two. Apartment dwellers often drew only their apartment complex and parking lot.1 All together, the 1,640 cognitive maps displayed 15,533 face blocks and 12,517 intersections. Of the intersections, 12,353 (99 percent) were tertiary. Of the face blocks, 15,531 were tertiary and two were not. These 101

CHAPTER EIGHT

TABLE 8.2 Overview of Cognitive Maps, by Data Collection

Data collection Number of cognitive maps

1st Los Angeles data collection

1st college town data collection

642

998

Number of tertiary face blocks

9,070

6,461

Number of tertiary intersections

8,807

3,546

102

62

2

0

Number of nontertiary intersections Number of nontertiary face blocks Greatest number of face blocks Least number of face blocks

100+

11

1

1

Number including entire island

12

0

Number including entire t-community

33

0

numbers bear emphasis. Ninety-nine percent of all intersections identified by respondents were tertiary, and 99.99 percent of all face blocks identified by respondents were tertiary. Residents’ cognitive awareness of the streets in their neighborhood was a subset of t-communities in 99 percent of all cases and of islands in essentially all of them. Furthermore, these were almost always proper subsets. Only 33 of the 1,640 cognitive maps included the entire t-community the respondent lived in, and only 12 included the entire island. The average cognitive map—if an average has any meaning—included nine tertiary face blocks and eight tertiary intersections. If we can assume that residents are aware of the places where recurring passive contacts occur, then these maps suggest that residents’ passive contacts are limited to a small number of tertiary face blocks and tertiary intersections. While perception of one’s neighborhood is an imperfect measure of passive contacts, it does hint at one’s passive awareness of the neighborhood, if not neighbors themselves. What happens when these stage 2 neighborly relations concatenate, neighbor to neighbor to neighbor? If residents conceive of the same local neighborhood, either in whole or in part, the implication is that they might share this neighborhood 102

UNINTENTIONAL ENCOUNTERS

and have a heightened probability for passive contacts. In contrast, if a person’s conceptualization of the neighborhood does not intersect with others’ conceptualization, the implication is that they are unlikely to share passive contacts. To provide a glimpse of how the stage 2 neighbor networks that bridge the gap between stage 1 neighbor networks and stage 3 neighbor networks might concatenate, I use an example of 10 respondents interviewed within a single island in Los Angeles and the cognitive maps that they generated. Figures 8.4, 8.5, and 8.6 show maps of the same area in Los Angeles. The gray lines in each figure consist of a contiguous set of tertiary face blocks connected by tertiary intersections. Superimposed on them are sets of darker lines, each near a numbered star. Each star, numbered 1 through 10, represents the location of the home of a respondent we interviewed in this area, and the set of dark lines is the respondent’s map of “the streets in my neighborhood.” Figure 8.4 displays the maps drawn by respondents 1 and 6. Figure 8.5 displays the maps drawn by respondents 2, 4, 7, 8, and 10. Figure 8.6 displays the maps drawn by respondents 3, 5, and 9. Some maps are quite large (e.g., respondents 1, 3, 5, and 6), while others are rather small, including only a single block face (e.g., respondents 2, 4, and 10). Respondent 1 was the original interviewee in this set, and the other nine were subsequently identified for interviews according to the process described in chapter 6. These maps illustrate a crucial point: while the respondents were located in various places about this t-community, some near a terminal edge, some near the center, none of the maps of “the streets in my neighborhood” crossed a nontertiary intersection. Figure 8.7 is a composite of the 10 respondents’ cognitive maps shown in figures 8.4–8.6. I created it by aligning identical streets in each map. It is placed against a backdrop of the actual tertiary street island (which was shown previously in figure 4.6). As in the above figures, bold lines represent the cognitive maps created by residents and numbered stars represent respondents’ homes. With two small exceptions, the composite of the 10 cognitive maps perfectly matches the actual island of tertiary streets. Respondent 3 identified the thin line labeled “A” as a street in his neighborhood, but it did not actually exist. The thin line labeled “B” actually existed and was a part of the island, but none of the 10 interviewees identified it as “a street in their neighborhood.” 103

CHAPTER EIGHT

FIGURE 8.4. Cognitive maps compared to tertiary street island map

Clearly, these residents’ conceptualizations of their neighborhood overlap to form a “superneighborhood” that is equivalent to the network of tertiary streets. I would argue that their probabilities for passive contacts, or stage 2 neighborly relations, would likewise overlap to form a neighbor network similarly equivalent to the network of tertiary streets. Overall, the concatenation of these 642 cognitive maps in the 68 distinct neighborhoods during the initial Los Angeles chain referral sample included almost every tertiary street within a respondent’s tcommunity on at least one of the maps. One (and only one) of the 642 respondents on one of the 68 chain referral samples included two nontertiary face blocks on her cognitive map. Not one of the 641 other respondents did so. Thus, the 68 concatenations of the cognitive maps in the Los Angeles chain referral sample reflected their 68 tertiary street networks almost perfectly. Similarly, the concatenation of the 998 cog104

UNINTENTIONAL ENCOUNTERS

FIGURE 8.5. Cognitive maps compared to tertiary street island map (continued)

nitive maps in the initial college town census did in fact include each and every tertiary street within the tertiary street network on at least one of the cognitive maps. Furthermore, these cognitive maps did not include any nontertiary face blocks. Thus, the concatenation of the 998 cognitive maps in the college town survey perfectly reflected its tertiary street network.

A NOTE ABOUT LARGE, MULTIUNIT COMPLEXES Our surveys within apartment buildings and other multiunit complexes confirmed that the proximity of residents’ dwellings to each other, measured in feet, was equally foundational for the actualization of neighborly relations. While there was not enough consistency in size and design of apartment buildings to draw definitive conclusions, like other 105

CHAPTER EIGHT

FIGURE 8.6. Cognitive maps compared to tertiary street island map (continued)

studies2 our surveys in larger apartment complexes suggested that residents who shared floors, staircases, or elevators were more likely to identify each other as neighbors of various types. Furthermore, at least cognitively, tertiary “streets” may exist within these complexes. In the cognitive mapping exercise in both Los Angeles and the college town, when asked to draw the “streets in their neighborhoods,” apartment dwellers often drew only their apartment complex and parking lot, and many drew large corridors internal to the complexes, whether or not they could be accessed by any type of motorized vehicle. One other subtle distinction suggested by our surveys within large apartment complexes in Los Angeles was the role of laundry rooms and mailboxes. Whenever multiple laundry rooms or mailboxes existed, residents were overwhelmingly biased in their identification of neighbors toward those who shared the same facilities. It appears that, within these large multi106

UNINTENTIONAL ENCOUNTERS

FIGURE 8.7. Composite of cognitive maps compared to tertiary street island map The thin line labeled “A” was identified by respondent 3 as a street in his neighborhood but it did not actually exist. The thin line labeled “B” actually existed and was a part of the island, but none of the ten interviewees identified it as “a street in their neighborhood.”

dwelling complexes, walkways, staircases, elevators, laundry rooms, mailboxes, and similar locales may engender unintentional encounters that provide residents with the opportunity to observe and evaluate each other’s behaviors in the same way that tertiary streets and intersections do outside these complexes. This needs further study.

MAIN POINTS IN REVIEW In this chapter, I investigated stage 2 neighboring relations and showed that passive contacts are real phenomena, not merely theoretical con107

CHAPTER EIGHT

structs. Respondents had no difficulty stating whether meeting someone was the casual result of being neighbors. The correlation between stage 2 neighbors and children is evidenced by the fact that most passive contacts began when children casually played together. In general, meetings involving children were identified as passive, and meetings not involving them were not. Individual respondents’ stage 2 neighboring relations, as evidenced by their cognitive understandings of their neighborhoods, did not typically reflect formal neighborhood equivalents such as real-estate designations or school districts but rather the lived experience of interconnected tertiary face blocks. Furthermore, residents’ conceptualizations of their neighborhood aggregated to form cognitive neighborhoods that were typically identical to the network of tertiary streets.

108

C H A P T E R

N I N E

Stage 3 Neighbors and Tertiary Streets

TERTIARY STREET PROXIMITY AND STAGE 3 NEIGHBORS The effects of tertiary street proximity on cognitive maps were quite pronounced and quite short-distance. I suggest that this implies that the effects of tertiary street proximity on passive contacts are quite pronounced and quite short-distance. The interviews clearly demonstrated that the effects of tertiary street proximity on the selection of neighborhood acquaintances are quite pronounced and quite short-distance as well, perhaps even more than they were on cognitive maps. Overall, 86 percent of respondents knew at least one of their nextdoor neighbors; 77 percent of respondents knew someone in both of the residences next door; and 79 percent of respondents knew someone in the residence directly facing them across the street. Knowledge of one’s neighbors decayed rapidly as distance from them increased, however. A disproportionate number of relationships occurred between people who lived only a few houses away from each other. On average, respondents knew someone in a little over half (53 percent) of the residences either two or three houses away. In the residences between four and six houses away, the median respondent knew two neighbors (the mean was 1.84). In the residences between seven and nine houses away, the median respondent knew only one neighbor (the mean was 1.1). In the residences between nine and twelve houses away, the median respondent did not know any neighbors (the mean was 0.6). Beyond that distance, few neighbors were known. The proportion of neighbors known is inversely exponentially related to number of house-steps away, approaching zero at distances of double-digit house-steps. np = 0.96 · e0.29·hs, where np is the proportion of neighbors known and hs is the number of house-steps from the respondent. Thus, stage 1 neighboring rela109

CHAPTER NINE

FIGURE 9.1. Neighbors known as a function of distance (in house-steps)

tions, when measured in terms of house-steps along a tertiary street, were powerfully related to higher stages of neighboring. It is important to note that one counts house-steps not “as the crow flies” but as the tertiary street runs. Of the 2,577 standardized interviews reporting 22,481 neighbors who were known personally, only 12 respondents identified 44 others who lived in a t-community other than the person who identified them. In contrast, 370 distinct respondents identified 1,415 distinct neighbors who lived in a different census tract. Overall, only 0.5 percent of residents identified someone who lived in a t-community than their own; in sharp contrast, 14 percent identified someone who lived in a different census tract. Respondents were 30 times more likely to identify someone in a different census tract than a different t-community. There were differences among these data collections that I will explore in subsequent chapters, but for now, it suffices to say that even in the least convincing data collection, the college town resample, only 2 percent of all residents identified someone who lived in a different t-community, while 17 percent identified someone who lived in a different census tract. Furthermore, only 1 percent of those identified in this data collection lived outside 110

S TAG E 3 N E I G H B O R S

the t-community of the person who identified them, while 13 percent lived outside the census tract of that same person. Clearly, the stage 1 neighbor networks that circumscribed the stage 3 neighbor networks consisted of shared tertiary streets and intersections, but not shared census geography. Recall that the resamples in both Los Angeles and the college town differed from the original data collections by removing the cognitive mapping exercise, which potentially biased results in favor of tcommunities. The only incidence of neighbors being identified that lived outside a t-community occurs in the college town resample, which eliminated this exercise, where 12 respondents identified 44 others not within their own t-community. However, 104 respondents identified 467 other residents who lived in a different census tract. Thus, while the cognitive mapping bias might have favored t-communities, it appears to have favored census tracts much more. Furthermore, the cognitive mapping exercise was not used in any of the 20 neighborhoods in the Los Angeles resample, and yet no one in those neighborhoods identified neighbors who did not live in their t-community, while 108 respondents identified 369 other residents who lived in a different census tract. Not only were stage 3 actualized neighboring relations dependent upon tertiary streets, but they depended upon passive contacts as well. As table 9.2 shows, across all data collections, the vast majority of those known in the neighborhood were first met in the neighborhood through passive contacts, or unintentional encounters, as identified by the respondents. Six out of seven pairs of neighbors who knew each other met because their children played together, or they walked their children to school together, or they walked their dogs or engaged in some other activity that brought neighbors into contact because they shared the same network of tertiary streets. Furthermore, some types of neighborhood functions typically labeled by respondents as nonpassive (e.g., block parties, neighborhood watch, neighborhood children seeking work) are certainly influenced by the convenience of local geography, typically the subtle geography of the tertiary street network (i.e., whom you invite to your block party depends on the geography of your face block). If these geographically 111

CHAPTER NINE

TABLE 9.1 Identified Neighbors by Same T-community or Census Tract, by Data Collection Los Angeles Data collection

All

Number of respondents Number of respondents who identified a neighbor who lived in a different census tract

2,577 370 (14.36%)

Number of respondents who identified a neighbor who lived in a different t-community

12 (0.47%)

Total number of neighbors identified as known personally

Original 642 118 (18.4%)

Resample 355 108 (30.4%)

College town Census 998 46 (4.6%)

0

0

0

22,481

5,092

2,928

10,883

Number of neighbors identified as known personally who lived in a different census tract than the person who identified them

1,415 (6.294%)

441 (8.66%)

369 (12.60%)

138 (1.268%)

Number of neighbors identified as known personally who lived in a different t-community than the person who identified them

44 (0.196%)

0

0

0

Resample 582 104 (17.9%)

12 (2.0%)

3,578 467 (13.05%)

44 (1.23%)

influenced contacts were counted with passive contacts, the proportion of neighbors known would be even greater. Finally, the correspondence between stage 1, stage 2, and stage 3 neighboring becomes even clearer when you examine the relationship between neighbors met by passive contacts and geography. As table 9.3 shows, across all data collections, only a tiny fraction of neighbors met by passive contacts (who themselves constituted over 86 percent of all neighbors known) lived outside the respondent’s census tract. Even more striking was that, of the nearly 20,000 neighbors met by passive contacts, only two lived outside the respondent’s t-community. 112

S TAG E 3 N E I G H B O R S

TABLE 9.2 Neighbors Met by Passive Contacts, by Data Collection Los Angeles Data collection Percentage of all neighbors known who were first met in the neighborhood through passive contacts

All 86.598%

College town

Original

Resample

Census

Resample

89.55%

86.65%

86.00%

84.18%

TABLE 9.3 Neighbors Met by Passive Contacts, by Same Census Tract or T-community, by Data Collection Los Angeles Data collection

All

Original

College town

Resample

Census

Resample

Total number of neighbors identified as known personally who were met by passive contacts

19,468

4,560

2,537

9,359

3,012

Number of neighbors met by passive contacts who lived in a different census tract than the person who identified them

114 (0.586%)

37 (0.81%)

22 (0.87%)

28 (0.30%)

27 (0.90%)

Number of neighbors met by passive contacts who lived in a different t-community than the person who identified them

2 (0.010%)

0

0

0

2 (0.07%)

TERTIARY STREET NETWORKS AND STAGE 3 NEIGHBOR NETWORKS What happens when the stage 3 neighbor networks concatenate neighbor to neighbor to neighbor? Just like the stage 2 neighbor networks that give rise to them, these concatenated stage 3 neighbor networks 113

CHAPTER NINE

FIGURE 9.2. Adaptive link-tracing network mapped onto t-community. Adaptive Link Tracing mapped onto T-community (light gray lines). Black nodes represent residents. Black lines indicate that one resident reported knowing another. The location of each node represents the resident’s geographic location in space.

are a subset of stage 1 neighbor networks when those stage 1 neighbor networks are defined by shared tertiary face blocks and shared tertiary intersections. I illustrate the relationship between neighbor networks and tertiary street islands in figure 9.2. This figure shows a neighborhood community network from the initial Los Angeles chain referral sample 114

S TAG E 3 N E I G H B O R S

displayed in figures 8.4–8.7 mapped onto its tertiary street island (displayed as a gray background) with the location of each node representing the households’ geographic location. Black lines indicate identified neighborly relations. Note that there are no arrowheads in this diagram, as they cluttered it too much and made it difficult to understand. One will note that this is the same tertiary street network used in the cognitive mapping example. Just as cognitive maps overlapped and consolidated to form a larger structure, so this chain of interviews extends farther than any individual resident’s neighbor networks. Just as the cognitive maps were limited to the tertiary street network, so this network was as well. It is clear in figure 9.2 that rather than extending beyond the limits of the tertiary street network, the link of interviewees “doubled back” on itself (i.e., the households identified by the current respondent as containing people they she personally were all in the spatial direction of the previous respondent). How common was this pattern? While the concatenations of these networks were typically much more geographically expansive than any of their constituent residents’ social worlds were, similar forces guided them to create a predictable pattern. In both the initial and follow-up Los Angeles chain referral samples, not one of these chains crossed an intersection with a nontertiary street, and thus none left their tcommunity. The tertiary street network clearly constrained residents’ interaction patterns and the evolving networks of neighbors. Despite the fact that many starting households were located near t-community end points, the link of respondents never crossed these t-community end points but instead traversed the t-community interior. Thus, tcommunity end points were phenomenological realities. Furthermore, in 75 of the 88 sequences, the links of interviewees “doubled back” on themselves before the terminal interviewee (again, all of the households identified by the current respondent as containing people he knew personally were in the spatial direction of the previous respondent). Neighborhood community networks did not, however, prove to be geographically coincident with neighborhoods defined only by their boundaries. While t-communities perfectly delimited the concatenating network of neighbors in these data collections, census tracts did not. Of the 88 interview chains, 70 of them traversed multiple census tracts 115

CHAPTER NINE

TABLE 9.4 Adaptive Link-Tracing Sequences, by Same Census Tract or T-community, by Data Collection 1st Los Angeles sample

2nd Los Angeles sample

68

20

Sequences that “doubled back” on themselves

56 (82%)

19 (95%)

Sequences that crossed into a different census tract

51 (75%)

19 (95%)

Sequences that crossed into three separate census tracts

6 (9%)

4 (20%)

0

0

Data Collection Sequences of interviewees

Sequences that crossed an intersection with a nontertiary street (i.e., crossed into a different t-community)

during their course (10 of them actually crossed into three separate census tracts) while none of them traversed multiple t-communities. Finally, it is said that a picture is worth a thousand words, and nothing could be truer of figure 9.3. It represents not the Los Angeles adaptive link-tracing samples but rather the complete census of the college town. In this figure, the 998 black dots represent households that were interviewed. The location of each dot on the figure represents the household’s geographic location to scale. These dots are connected by 10,883 black lines representing identified neighborly relations. There are no arrowheads in this diagram, as they would have added unnecessary clutter. What is fascinating about this structure is that it did not divaricate into a treelike structure branching out endlessly, nor is it artificially truncated. These 10,883 neighborly relations were followed wherever they led. They just always led back to the cluster. These neighborly relations interconnected 1,129 (this number is greater than the number of respondents because respondent households often identified nonre116

S TAG E 3 N E I G H B O R S

FIGURE 9.3. Complete census of neighborhood network Households (n = 998) who know each other (10,883 relations). Spatial representation of households is to scale.

spondent households) of the 1,214 households into a single connected component. Residents could have identified anyone, but this network is remarkably delimited. If this sample had been the entirety of a single town, this result would be expected, but the sample was only a few percent of the households. Clearly, some sociological force constrained the relations to fold back on themselves. Figure 9.4 is similar to figure 9.3. The white lines dissecting it are the new features. These white lines dissect the figure into three census tracts. Clearly, these census tracts do not identify the sociological force delimiting neighbor networks. Figure 9.5 is also similar to figure 9.3 except that white lines dissect it as well, although these lines are not identical to those in figure 9.4. In this figure, the white lines delineate three elementary school catchments. Similar to census tracts, elementary school catchments do not identify the powerful sociological force constraining community in this neighborhood. Finally, Figure 9.6 is 117

CHAPTER NINE

FIGURE 9.4. Complete census of neighborhood network mapped onto census tracts Black dots represent Households (n = 998). Black lines indicate one household knows another (10,883 relations). Spatial representation of households is to scale. White lines dissect figure into census tracts.

similar to figure 9.3 except that a thick black line outlines the network. As the reader may have already guessed, this line outlines the perimeter of the t-community. In this setting, the t-community proves to be the overwhelming sociological force, acting far more powerfully than either school catchments or census tracts, and providing a striking example of the power of concatenated subtleties. Table 9.1 expresses this picture in numbers: 46 of the 998 respondents identified 138 neighbors they knew personally who lived in a different census tract. These are the lines crossing the polygons. No one, however, identified anyone who lived in another t-community. Not one of the 10,883 identified neighbors lived in a different t-community. Thus, all of the lines fall within the yellow polygon. 118

S TAG E 3 N E I G H B O R S

FIGURE 9.5. Complete census of neighborhood network mapped onto elementary school catchment areas Black dots represent Households (n = 998). Black lines indicate one household knows another (10,883 relations). Spatial representation of households is to scale. White lines dissect figure into elementary school catchment areas.

MORE THAN PROXIMITY This finding is so startling that it may make some readers uncomfortable. Is some unidentified social force involved? As I have stated earlier, nothing but residential neighborhoods surrounded this tertiary street network in the college town. But perhaps this effect results from most residents living very close nearby. Perhaps this effect is merely a spurious confound of geographic distance. Almost a century ago Robert Park declared: “It is because geography . . . and all the other factors which determine the distribution of population determine so irresistibly and fatally . . . the group, and the associates with whom each one of us is bound to live, that spatial relations 119

CHAPTER NINE

FIGURE 9.6. Complete census of neighborhood network mapped onto t-community Black dots represent Households (n = 998). Black lines indicate one household knows another (10,883 relations). Spatial representation of households is to scale. Thick black line around exterior indicates the perimeter of the single t-community.

come to have . . . the importance that they do [and] physical distances so frequently are . . . the indices of social distances.”1 Generations later, but still more than a generation ago, Tobler declared the first law of geography to be, “Everything is related to everything else, but near things are more related to each other than distant things.”2 This observation is embedded in the gravity model of trip distribution and the law of demand, in that interactions between places are inversely proportional to the cost of travel between them, much as the probability of purchasing a good is inversely proportional to the cost. Distance has a long and proud history in social theory. Given its ubiquitous importance, we might also expect raw distance to equate to 120

S TAG E 3 N E I G H B O R S

functional distance and thus to play a factor in the emergence of neighbor networks. On a global scale of distances, neighbors live extremely close to the people who know them. It may prove the case that t-communities appear to delimit neighborly relations merely because they extend beyond the range of most people’s neighbors. For example, perhaps the most likely neighbor one might know lives next door, and this person clearly lives in the same t-community, but this is not really an effect of tertiary streets constraining the evolution of neighbor networks. If, however, people know some others who live at relatively large distances and if most of the other residences that far away are in a different tcommunity but those whom they know are restricted to the same tcommunity, this seems to indicate that t-communities are constraining potential neighborly relations. To account for this possibility, I need to find a way to estimate the numbers of neighbors we would expect to be identified in the same tcommunity. For comparison purposes, I also estimate the numbers of neighbors we would expect to be identified in the same census tract, both the same census tract and the same t-community, or neither. To do this, I treat the distance each neighbor is from the respondent who identified the person as a given, and for each respondent and each neighbor identified, I create an imaginary circle, centered on the respondent with a radius equal to the distance to the identified neighbor. I can then dissect these imaginary circles into four types of arcs: those that are within both the same tract and the same t-community as the respondent, those that are in the same tract but not the same tcommunity as the respondent, those that are in the same t-community but not the same tract as the respondent, and those that are neither in the same t-community nor the same tract as the respondent. Given that each identified neighbor’s circle is defined as the distance the person is from the respondent who identified him, treating distance as a given, the probability is 1 that an identified neighbor will be somewhere along the perimeter of the circle. If simple geography was all that mattered, then the probability that an identified neighbor would be found on any particular segment of one of these circles should be equal to that segment’s proportion of the entire circle. Thus, if one half of a 121

CHAPTER NINE

TABLE 9.5 Probability of Those Identified Living in Same T-community or Census Tract, by Distance from Respondent, College Town Census, All Neighbors Number of circles Circles have some segment that lies in

10,883 Same t-community but not same tract as respondent

1,628 (14.96%)

Same tract but not same t-community as respondent

1,218 (11.19%)

Neither same tract nor same t-community as respondent

1,885 (17.32%)

circle’s perimeter was in the same t-community as the respondent, and if simple geographic distance was all that mattered, there would be a probability of .0.5 that the neighbor would be identified along that particular half of the circle and thus in the same t-community and a probability of 0.5 that the neighbor would be identified along the other half of the circle and thus not in the same t-community. Similarly, if half of 100 circles’ perimeters were in the same t-community as their respective respondents, and if simple geographic distance was all that mattered, I would expect half (50) of 100 of those neighbors to be identified along the t-community halves of these circles. For there to be a “t-community effect” I would need to find both substantively and significantly more of those 100 neighbors along the t-community halves of their circles. To evaluate how what we observe relates to what we expect, I can sum both neighbors and circle fragments to create a density measure. Since there is one circle for each identified neighbor, overall the density of neighbors per circle is 1. Similarly, for each type of arc segment (tcommunity, tract, both, neither), I can sum the number of neighbors found on that type of arc segment and I can sum the proportion of its circle each fragment constitutes to calculate a density measure of neighbors per circle. If the density for a particular type of circle arc segment is greater than 1, neighbors are identified along those arcs more frequently than expected merely on the basis of simple geographic 122

S TAG E 3 N E I G H B O R S

TABLE 9.6 Expected and Actual Number of Neighbors, by Same T-community or Census Tract, College Town Census, All Neighbors

Average proportion of circles

Expected number of identified neighbors

Actual number of identified neighbors

Density

Both the same t-community and tract

0.798

8,685

10,745

1.237

Same t-community but not same tract

0.016

174

138

0.793

Same tract but not same t-community

0.121

1,317

0

0

Neither same t-community nor same tract

0.065

707

0

0

p  .0001 for all.

distance. Conversely, if the density for a particular type of circle arc segment is less than 1, neighbors are identified along those arcs less frequently than expected merely on the basis of simple geographic distance. I can then use a binomial test to determine if these densities differ significantly from expected values. There are 10,883 such circles for the first college town data collection, the exhaustive census, each centered on a respondent with a radius equal to the distance between the respondent and the neighbor they identified. Table 9.6 presents the results. The clear finding is that only the portion of the circles in the same t-community had any density, with some preference toward those portions also in the same census tract. Most respondents identified multiple neighbors, and thus these identifications are not independent. It might be that a few respondents identified many others, who typically lived in the same t-community. To deal with this possibility, I focused on one neighbor for each respondent, the one who lived farthest away since this circle was most likely to transcend their t-community or census tract and thus maximize available variability in arc segment types. If we confine our focus thus 123

CHAPTER NINE

TABLE 9.7 Probability of Those Identified Living in Same T-community or Census Tract, by Distance from Respondent, College Town Census, Farthest Neighbors Only Number of circles Circles have some segment that lies in

979 Same t-community but not the same tract as the respondent

852 (87.0%)

Same tract but not the same t-community as the respondent

534 (54.5%)

Neither same tract nor the same t-community as the respondent

979 (100%)

TABLE 9.8 Expected and Actual Number of Neighbors, by Same T-community or Census Tract, College Town Census, Farthest Neighbors Only

Average proportion of circles

Expected number of identified neighbors

Actual number of identified neighbors

Density

Both the same t-community and tract

0.670

656

933

1.422****

Same t-community but not the same tract

0.041

40

46

1.15*

Same tract but not the same t-community

0.105

103

0

0****

Neither same t-community nor the same tract

0.184

180

0

0****

* p  .05**** p  .0001.

to the neighbor identified by each respondent who lived farthest away, we have only 979 circles.3 Table 9.8 presents the results. Again, only the portion of the circles in the same t-community had any density. This data collection suffered from a potential bias due to the cognitive mapping exercise. This was not a factor for the follow-up data 124

S TAG E 3 N E I G H B O R S

TABLE 9.9 Probability of Those Identified Living in Same T-community or Census Tract, by Distance from Respondent, College Town Follow-up, All Neighbors Number of circles Circles have some segment that lies in

3,578 Same t-community but not the same tract as the respondent

533 (14.89%)

Same tract but not the same t-community as the respondent

421 (11.77%)

Neither same tract nor the same t-community as the respondent

611 (17.08%)

collection, however. Therefore, I also report results from the second college town data collection. This data collection has the added benefit that it not only allowed neighbors to be identified across t-communities and census tracts but also included multiple t-communities and multiple census tracts in the sampling frame.4 In this second college town data collection, there are 3,578 identified neighbors, thus 3,578 circles. While the second college town data collection was not biased by the introduction of cognitive maps, it also did not ascertain the exact locations of neighbors, so we do not know exactly how far away they are. Thus, to examine this data set, I assign distances between respondents and identified neighbors as follows. For each respondent, I identify all of the respondents from the first college town data collection, the exhaustive census collected three years earlier in the exact same neighborhood, who identified exactly the same number of neighbors (i.e., if a respondent in the third data collection identified seven neighbors, I find all of the respondents in the second data collection who identified exactly seven neighbors). I then average the distances between the respondents and the neighbor identified who lived nearest to them for each of these respondents from the second data collection and assign that value to the nearest identified neighbor for the respondent in the third data collection. I repeat this process for each of the second nearest, third nearest, and so on, to the second farthest, and farthest identified neighbors. Thus I can create these imaginary circles, centered on respondents with radii equal to the distances to identified neighbors, for the third data set as well. 125

CHAPTER NINE

TABLE 9.10 Expected and Actual Number of Neighbors, by Same T-community or Census Tract, College Town Follow-up, All Neighbors

Average proportion of circles Same t-community but not same tract Both the same t-community and tract Same tract but not same t-community Neither same t-community nor same tract

Expected number of identified neighbors

Actual number of identified neighbors

Density

0.110

392

455

1.19

0.775

2,771

3,079

1.11

0.075

271

32

0.13

0.040

144

12

0.08

p  .0001 for all.

Table 9.10 presents the results. While not the completely dichotomized split noted above, the results are still striking. The portions of the circles in the same t-community have densities greater than 1 while the portions of the circles not in the same t-community have densities near zero. Whether or not the arc segments are in the same tract is trivial. After accounting for the potential cognitive mapping bias, it is clear that neighbors will almost exclusively be identified in the same tcommunity, whether or not it is part of the same census tract. Finally, if, for the second college town data collection, we confine our focus to the neighbor identified by each respondent who lived farthest away, as we did for the first data collection, we have only 582 circles. Table 9.12 presents the results. Again, it is clear that neighbors will almost exclusively be identified in the same t-community, while the effect of census tracts is virtually nonexistent. In sum, no matter how far away they live, neighbors will almost exclusively be identified in the same t-community, whether or not it is part of the same census tract. The effects t-communities appear to have in constraining neighborly relations are not merely spurious confounds of geographic distance. At any distance, neighborly relations appear to 126

S TAG E 3 N E I G H B O R S

TABLE 9.11 Probability of Those Identified Living in Same T-community or Census Tract, by Distance from Respondent, College Town Follow-up, Farthest Neighbors Only Number of circles Circles have some segment that lies in

582 Same t-community but not the same tract as the respondent

524 (90.0%)

Same tract but not the same t-community as the respondent

323 (55.5%)

Neither same tract nor the same t-community as the respondent

582 (100%)

TABLE 9.12 Expected and Actual Number of Neighbors, by Same T-community or Census Tract, College Town Follow-up, Farthest Neighbors Only

Average proportion of circles

Expected number of identified neighbors

Actual number of identified neighbors

Density

Both the same t-community and tract

0.630

369

473

1.29

Same t-community but not the same tract

0.130

75

97

1.29

Same tract but not the same t-community

0.100

58

5

0.10

Neither the same t-community nor the same tract

0.140

81

7

0.07

p  .0001 for all.

be restricted to the shared tertiary street network but not to shared administrative geography.

MAIN POINTS IN REVIEW In the previous chapter, I showed that no one’s cognitive understanding of her neighborhood escaped tertiary street networks. In this chapter, 127

CHAPTER NINE

I show that the actualized stage 3 neighbor networks, which emerged from stage 2 neighbor networks, did not escape tertiary street networks either. Stage 1 neighboring relations, when measured in terms of housesteps along a tertiary street, powerfully relate to higher stages of neighboring. They do not, however, relate well to raw distance as the crow flies. The tertiary street network not only constrained individual residents’ interaction patterns but also the networks of neighbors they concatenated into; however, neighborhoods defined by shared boundaries such as census geography or elementary school catchments did not. Furthermore, the effects t-communities have in constraining neighboring relations are not merely spurious confounds of geographic distance. At any distance, neighboring relations are restricted to the shared tertiary street network but not to shared administrative geography.

128

C H A P T E R

T E N

The Importance of Neighbor Networks

THREE DEGREES OF NEIGHBORING A neighborhood community network is not typically synonymous with any individual resident’s neighbor network. Instead, it is the aggregation of many individual neighbor networks; a true social entity, beyond any individual. Each resident’s neighbor network connects with the neighbor networks of other residents, who connect to still other residents, concatenating and aggregating, neighbor to neighbor to neighbor, and especially child to child to child, to form a network that extends farther geographically and socially than any one resident’s neighbor network yet maintains relatively short path lengths among them all. I illustrate this neighbor network concatenation with a sequence of 10 actual interview respondents. Between them, these 10 respondents identified a total of 36 other neighbors they knew personally (although this only represented 26 distinct households, as some were identified by multiple respondents). The 26 households that were identified included all 10 interviewees as well as 16 other households. Figure 10.1 is a diagram of the neighbor networks of these residents. In the diagram, numbered nodes represent residents, and a line from node 1 to node 2 indicates that resident 1 reported knowing resident 2 personally. The nodes are placed in the diagram using a spring embedding algorithm that places the nodes relative to each other so that the distances between nodes roughly approximate the length of the shortest path between them, so that edge lengths are roughly equal, and so that nodes do not sit on top of each other. Numbers 1 through 10 represent the 10 respondents (in the order in which they were interviewed), while numbers 11 through 26 represent the 16 households identified who were not interviewed. Some lines have arrows going one way, indicating that one household identified the household to whom the arrow points but 129

CHAPTER TEN

FIGURE 10.1 Sample adaptive link-tracing network

this identification was not reciprocated. Those who were not interviewed, of course, could not report knowing others. Some lines have arrows going both ways, indicating that each household identified the other as being known. Again, mutual identification could occur only if both households were interviewed. Some households knew a lot of others and were known by a lot of others. Other households were known by only a few others or by one person, and some knew no one else. The in-degrees (how many times each household was chosen as a known neighbor) range from one to four (respondent 6). The out-degrees (how many others each respondent chose) range from zero to 11 (again, respondent 6). These 26 individuals are connected within a single, relatively small neighborhood community network. Relations concatenate, neighbor to neighbor to neighbor, to form a network that reaches farther than any one resident’s neighbor network. For example, even in this tiny neighborhood community network, respondent 6’s personal network includes less than half of the entire structure. 130

T H E I M P O RTA N C E O F N E I G H B O R N E T WO R K S

The resultant network is also as fragile as its weakest link. For example, the single relation between respondents 3 and 5, among others, holds the network together. Anything, no matter how trivial, that disconnected respondents 3 and 5 would also disconnect the larger network. This is a sample that can only suggest, rather than verify, the fragility of these networks; but, if it were a complete census, anything that inhibited a relation between those two individuals would have disconnected the two halves of the network. Ignoring directionality, the mean path length between all actors is 2.97 steps, the median path length is three steps, and the maximum path length is six steps between both persons 1 and 16 and person 25. I ignore directionality because, as I have mentioned, 89 percent of the relations that could have been reciprocated (i.e., both households were interviewed) actually were reciprocated. In general, unidirectionality is an artifact of our having interviewed fewer than half of the households in the network. The network in figure 10.1 was one neighborhood community network discovered in one interview sequence in one t-community. In the 88 neighborhood community networks discovered through the 88 interview sequences in 88 t-communities, relations concatenated to form chains that included between four and 151 individuals, with the median being 45. The individual households in these neighborhood community networks were separated by an average of 3.65 steps in the first set of 68 Los Angeles samples and by an average of 3.22 steps in the second set of 20 Los Angeles samples (an average of 3.55 steps over all 88 samples). The shorter path lengths from the longer sample chains, 20 rather than 10, resulted from a greater degree of relinking. In both sets of samples, the median path length was three steps. Thus, in the Los Angeles samples, most residents knew someone personally who knew someone personally who knew any typical neighbor in their neighborhood community network. In the exhaustive census, most residents also knew someone personally who knew someone personally who knew any typical neighbor in their neighborhood community network. On average, 159 other neighborhood members were within two steps of each other (13 percent of the entire neighborhood and 16 percent of the responding neighbor131

CHAPTER TEN

FIGURE 10.2 Distribution of path lengths in samples of size 10

hood). In fact, the exhaustive census had a noticeably shorter mean path length than the other data collections, 2.67 steps as compared to 3.65 and 3.22 for the first and second Los Angeles data collections and 3.28 for the follow-up college town data collection. It is likely that this was because it was an exhaustive census, and thus all neighborly relations that could have shortened path lengths were included. For similar reasons, it is likely that an exhaustive census of all data collection sites would have produced shorter path lengths. I will discuss this likelihood further subsequently. This near-complete census also provided data that allowed me to simulate the process of sampling and determine how likely it was that samples, such as I conducted among my other data sets, would provide relatively accurate estimations of important neighborhood phenomena. I define accurate not by reference to an imagined, hypothetical reality but rather by what a census would have found (or more precisely by what my census of the college town did find), which is, after all, the best we can hope for. How accurate are samples? I ran 50,000 simulations. In each simulation, I randomly sampled a starting resident. I then noted the residents whom this first resident was connected to. Of these, I randomly sampled one. I then noted the residents that person was connected to. I 132

T H E I M P O RTA N C E O F N E I G H B O R N E T WO R K S

FIGURE 10.3 Sample adaptive link-tracing network, respondents only

continued this process until I had randomly sampled 10 neighbors and noted the residents whom they were connected to. This, of course, mirrors what I did when conducting my chain referral samples in Los Angeles, except that I did not attempt to find the geographically most extreme actor. The results are displayed in figure 10.2. Path lengths in the 50,000 samples ranged from 2.56 to 5.64 with over 95 percent of them occurring between 3.4 and 4.4. The mean of these path lengths was 3.64. Overall, thus, sample path lengths were almost a full step longer than the path lengths found internal to the census. This generally resulted from trailing edges. To see what I mean by trailing edges, consider the network in figure 10.3, which is a modified version of figure 10.2. This figure shows only the respondents in one of our chain referral samples. In the figure, as drawn, the average path length is 2.51 steps and the median path length is two steps. Now consider what happens to path length when we add those who were nominated by these respondents but not interviewed (figure 10.1). Recall that when we conducted our chain referral samples, we pursued only one nominated alter for each interviewee. Therefore, other nominated neighbors were often left unattached. Their distance from other network members was often the distance from that network member 133

CHAPTER TEN

to the person who nominated them plus one more step. Thus, the complete sample in figure 10.1 has an average path length of 2.96 steps and a median of three steps, an average of 0.45 steps longer and a median of a step longer. So what does all of this mean? It indicates that my sample path lengths were almost certainly longer than actual paths would have been, primarily because of the effect of trailing edges. Overall, the sampling conducted in Los Angeles is likely to have produced a relatively robust upper bound for the real values. This most likely also accounts for the fact that the Los Angeles samples also had longer path lengths than those in the census of the college town. Thus, the actual neighbor path length we might expect to find in neighborhood communities emerging amid tertiary street networks would be shorter than those found in the Los Angeles samples and closer to those found in the complete census.

A NOTE ABOUT THE EXHAUSTIVE CENSUS I discuss briefly one interesting fact arising in the college town census. In the exhaustive census, the average household knew 10.9 others. The average household’s neighbors, however, knew, on average, 14.6 other households. Thus, a “typical” household had neighborly relations with 11 others. Each of these 11 others has neighborly relations with 13 or 14 other households besides the original household (the original neighbor is the fourteenth or fifteenth neighbor to each). This odd finding, that the neighbors whom the average resident knows, on average, know more neighbors than the resident who identified them results from the fact that the neighbor networks are somewhat centralized and thus the larger our sample (i.e., there are more first neighbors than the focal actor), the greater the likelihood we would sample a central neighbor. To understand why this might happen, consider an admittedly extreme hypothetical network with 20 members. Nineteen of them are connected to the twentieth member. This situation is shown in figure 10.4. For convenience, let us label the highly connected network member as A, and the other 19 members as Bs. Each B is directly connected to only other network member (in all 19 cases, this is A). Network member A, however, is directly connected to all other members. Thus, the total 134

T H E I M P O RTA N C E O F N E I G H B O R N E T WO R K S

FIGURE 10.4 Why your neighbors know more neighbors than you do

degree is (19  1) + 19 or 38. The average degree is 38 / 20 or 1.9. The median (and modal) degrees are, of course, both 1. If we consider who can be reached in two steps, however, the situation changes. Each B can reach the other entire 18 Bs in two steps. A can reach no one new in two steps. Thus, the total two-step degree is (19  18) + 0 or 342. The average two-step degree is 342 / 20 or 17.1. The median and modal degree are 18. Thus, we have a situation where the one-step degree averages 1.9 and the two-step degree averages 17.1 (and where the median one-step degree is 1 and the median two-step degree is 18).1

NEIGHBORING IS A FAMILY RELATION Having shown that residents of a t-community are typically connected to each other by short paths, I now focus on a subset of residents within a t-community. I have argued that children and their families are the quintessence of neighborhood life. How much of neighborhood life 135

CHAPTER TEN

FIGURE 10.5 Household neighbor network sizes, by presence of children

involves households with children? While households with children constitute about half of the population of American neighborhoods, their contribution to neighbor networks is much greater than that figure suggests. Households with children are far more involved in neighborhood life than households without children. They know more neighbors and are known by more neighbors than households without children. In our interviews, households with children were more gregarious, identifying a mean of 13.141 other households compared to 4.759 for households without children, almost three times as many and significantly larger (p  .001). Previously I discussed some of the mathematical oddities resulting from concatenation. When considering households with children, not 136

T H E I M P O RTA N C E O F N E I G H B O R N E T WO R K S

TABLE 10.1 Distribution of Neighboring Relations, by Presence of Children in Either Household Los Angeles Data collection

College town

All

Original

Resample

Census

Resample

Between two households with children

19,231 (85.543%)

4,355 (85.3%)

2,506 (85.59%)

9,308 (85.53%)

3,062 (85.57%)

Between a household with children and a household without children

1,962 (8.727%)

462 (9.07%)

253 (8.64%)

944 (8.67%)

303 (8.47%)

Between two households without children

1,288 (5.729%)

275 (5.40%)

169 (5.77%)

631 (5.80%)

213 (5.95%)

Total

22,481

5,092

2,928

10,883

3,578

only do they know three times as many neighbors as households without children, but when these relations concatenate, they dominate neighborhood community networks. In the Los Angeles chain referral samples, the disproportionate identification of households with children has a multiplier effect. In these data collections, 22 percent more of the households identified by respondents had children than those households that were in the initial sample of starting households. Furthermore, when compared to the proportion of households with children in a t-community’s population, households with children were significantly overrepresented (p  .05) as “known personally” by respondents in a little more than half (45) of the t-communities surveyed. They were not significantly underrepresented as “known personally” by respondents in any t-communities. In the college town complete census, the vast majority (85 percent) of all neighboring relations were between two households with children. The household reporting the relationship had children living in it and noted that the household identified as a neighbor did so as well. Again, as noted earlier, this probably results from children’s limited mobility, which encumbers their parents as well. An additional 9 percent of all neighboring relationships included a household with children as at least one of its participants. This included both respondent households with children identifying households without children 137

CHAPTER TEN

FIGURE 10.6 Neighboring relations, by presence of children in either household

and respondent households without children identifying neighboring households that had children. Only 6 percent of all neighboring relationships involved two households neither of which had children, neither the respondent household nor the household they identified. This seems consistent with the fact that most of the neighborhood effects researchers concern themselves with involve households with children. This result again highlights the pitfalls of adult social researchers theorizing about neighborhoods, observing neighborhoods, and talking to adult respondents about neighborhoods. If one conceptualized neighborhoods in terms of adult neighbors, less than 6 percent of 138

T H E I M P O RTA N C E O F N E I G H B O R N E T WO R K S

neighboring relationships would remain. To put it differently, if one sampled neighboring relationships randomly, one would have only a 6 percent chance of sampling a relationship between two households without children. I do not want to belabor this point, but this number gives one pause when one considers that a probability of less than 5 percent is a typical cut-off point for rejecting a null hypothesis. Neighboring relationships occurring between two households without children are an insignificant proportion of the total picture. Neighborhood communities are about children and their families. Neighbor networks are essentially networks among households with children. The rest of neighborhood life is a peripheral addition. Furthermore, in contrast to the essential uniformity of path lengths among racial groups, as shown in chapter 7, paths among households with children were clearly shorter than paths among households without children. In the college town complete census, among the 669 households with children, 9,309 relations existed in which one household with children “knew personally” members of the other. Conversely, among the 329 households without children, 623 relations existed in which one household without children “knew personally” members of the other. Overall, these households with children had a much shorter path length than did households without children. Any particular household with children could, on average, reach any other particular household with children in 2.38 steps. In contrast, households without children typically were 4.41 steps apart. Two more steps separated typical households with children than did typical households without children. In fact, the shortest path between most households without children led through a household with children.

THE IMPORTANCE OF CONVENIENT AVAILABILITY I have shown that neighborhood communities emerging in tertiary street networks are integrated by relatively short paths among residents and that households with children are the focal point of these communities. How much do these neighboring relations and neighbor networks matter? In subsequent chapters, I will discuss the importance of neighboring relations for neighborhood communities and their effects. 139

CHAPTER TEN

FIGURE 10.7 Percentage of households reachable at each neighbor network step, by presence of children

In the remainder of this chapter, however, I will discuss their importance for individual residents. While neighbors may not be one’s best friends or the primary source of emotional support, most residents attributed great value to their neighboring relations. Neighbors performed important services that would have been quite expensive to purchase and that family and friends, for a variety of reasons, were not asked to do. Not surprisingly, the most important of these services related to children. Neighboring parents routinely monitored their own children and those of their neighbors in spontaneous playgroups. They also acted as parental substitutes, advising, scolding, and disciplining the children of others. In many cases, respondents expressed no concern about neighboring 140

T H E I M P O RTA N C E O F N E I G H B O R N E T WO R K S

parents scolding or disciplining their children as well. Neighboring parents would often drop off (or pick up) each other’s children at school, soccer matches, and a variety of other activities. They would often accept responsibility for watching each other’s children on a regular basis. Numerous times parents reported having dropped off their children with a neighbor in an emergency or having asked their babysitter to do so if they would be late returning from work. Many neighbors were listed as the emergency contact person for each other’s children at school. Furthermore, for households with children that did monitor each other’s children in spontaneous playgroups, many did so multiple times every week. Neighbors also provided each other with substantial services involving local safety and home improvement. About half of all respondents had asked a neighbor to watch their house when they were away. Displaying a substantial amount of trust, about half had exchanged keys with each other so that they could let in service people or children if they were locked out. About half of all respondents said that, if they felt unsafe, they were as likely as not to call a neighbor before the police. About one in six neighbors reported assisting each other with very large and expensive projects such as reroofing a house, paving a driveway, or building a fence. Neighbors also provided the more conventional small, mundane favors such as loaning tools or small amounts of money, providing small services and favors, and exchanging information about local goods and services. Finally, while it is true that neighbors are not generally a large source of emotional support and most residents had not engaged in intentional social activities with their neighbors, a significant minority had done so quite routinely. One might imagine that passively generated contacts, upon which I have focused so directly, signify lots of trivial neighboring but not the more important type of neighboring leading to the generation of neighborhood effects. In contrast to that supposition, however, passively generated contacts proved even more likely than nonpassively generated ones to result in substantively important neighborly relations. For example, if a resident had exchanged keys with a neighbor or asked a 141

CHAPTER TEN

TABLE 10.2 Proportion of Neighboring Relations Providing Specific Services, Households with Children Only (N = 1,321)

Relational type

Percentage of all identified neighborly relations

Monitored other neighbors’ children in spontaneous playgroups Have dropped off their children with a neighbor in an emergency or have asked babysitter to do so if respondent would be late returning from work

85.01% 57.00%

Dropped off (and/or picked up) neighbor’s children at school, soccer matches, and a variety of other activities

53.14%

Expressed no concern about neighboring parents scolding or disciplining their children

43.98%

Listed a neighbor as the emergency contact for their children at school

33.69%

Listed as the emergency contact person for neighbor’s children at school

30.36%

Listed a neighbor as the emergency contact for children at school and was listed by the same neighbor as the emergency contact person for neighbor’s children at school

25.66%

neighbor to watch his house while he was away, over 90 percent of the time these relations had started as passive contacts. Furthermore, if someone had helped a neighbor with a very large, expensive project such as reroofing a house, paving a driveway, or building a fence, almost 90 percent of the time this relation began as a passive contact. For households with children, 95 percent of those who reported dropping off and picking up neighboring children said that their relationship began as a passive contact, and 98 percent of those who reported monitoring other neighbors’ children in spontaneous playgroups said that their relationship with that neighbor began as a passive contact. Thus, 142

T H E I M P O RTA N C E O F N E I G H B O R N E T WO R K S

FIGURE 10.8 Frequency of monitoring neighbors’ children in spontaneous playgroups

not only does neighboring produce important social capital and not only is neighboring the result of geographically dependent passive contacts, it is geographically dependent passively generated relations that produce the important social capital. Overall, many neighbors claimed to engage in most of these activities simply out of convenience. This, however, is a testimony to the importance, rather than the weakness, of the neighboring relation. The choice to live in a neighborhood is to some extent a choice about who we would like to be conveniently available to us and to whom we would like to be conveniently available. 143

CHAPTER TEN

TABLE 10.3 Proportion of Neighboring Relations Providing Specific Services, All Households (N = 2,577)

Relational type

Percentage of all identified neighborly relations who had engaged in this type of relation

Neighbors exchanged keys with each other so that they could let in service people or residents’ children, if they were locked out

51.9%

Neighbors said that, if they felt unsafe, they were as likely as not to call a neighbor first before the police

48.0%

Asked a neighbor to watch house while respondent was away

50.1%

Was asked by a neighbor to watch house while neighbor was away

34.8%

Asked a neighbor to watch house while respondent was away and was asked by same neighbor to watch house while neighbor was away Neighbors reported assisting each other with very large and expensive projects such as reroofing another’s house, paving another’s driveway, or helping each other build fences

29.2%

16.0%

MAIN POINTS IN REVIEW In this chapter, I have shown that a neighborhood community network is not typically identical to any individual resident’s neighbor network; it is a true social entity, beyond any individual. Each resident’s neighbor network connects with the neighbor networks of other residents, who connect to still other residents, concatenating and aggregating, neighbor to neighbor, and especially child to child, to form a network that extends farther geographically and socially than any one resident’s neighbor network. Significantly, however, these aggregated neighbor144

T H E I M P O RTA N C E O F N E I G H B O R N E T WO R K S

FIGURE 10.9 Frequency of socialization with neighbors

hood community networks maintain relatively short internal path lengths, typically three degrees of separation, among residents. In this chapter, I have also shown that households with children are far more involved in neighborhood life than households without children. They know almost three times as many neighbors and are known by more neighbors than households without children. These differences compound so that the vast majority (85 percent) of all neighboring relations are between two households with children and only 6 percent of all neighboring relationships involve two households neither of which has children. Furthermore, neighbor-to-neighbor paths among households with children are half as long as those among households without children. 145

CHAPTER TEN

TABLE 10.4 Proportion of Neighboring Relations Providing Specific Services Who Were Met through Passive Contact, All Households

Relational type

Percentage of relations that began as a result of passive contacts

Neighbors exchanged keys with each other so that they could let in service people or residents’ children, if they were locked out. N = 1,337. Neighbors said that, if they felt unsafe, they were as likely as not to call a neighbor first before the police. N = 1,237 Asked a neighbor to watch house while respondent was away. N = 1,291. Was asked by a neighbor to watch house while neighbor was away. N = 897. Asked a neighbor to watch house while respondent was away and was asked by same neighbor to watch house while neighbor was away. N = 752. Neighbors reported assisting each other with very large and expensive projects such as reroofing another’s house, paving another’s driveway, or helping each other build fences. N = 412. Monitored other neighbors’ children in spontaneous playgroups. N = 1,123.

91%

68%

91% 91% 91%

87%

98%

Finally, in this chapter I have shown that most residents attributed great value to their neighboring relations. Neighbors performed important services for each other. Not surprisingly, the most important of these services related to children. Furthermore, passively generated contacts proved even more likely than nonpassively generated ones to result in substantively important neighboring relations. The choice to live in a neighborhood is to some extent a choice about who we would like to be conveniently available to us and to whom we would like to be conveniently available.

146

T H E I M P O RTA N C E O F N E I G H B O R N E T WO R K S

TABLE 10.5 Proportion of Neighboring Relations Providing Specific Services Met through Passive Contact, Households with Children Only

Relational type

Percentage of relations that began as a result of passive contacts

Have dropped off their children with a neighbor in an emergency or have asked babysitter to do so if respondent would be late returning from work. N = 753.

94%

Dropped off (and/or picked up) neighbor’s children at school, soccer matches, and a variety of other activities. N = 702.

95%

Expressed no concern about neighboring parents scolding or disciplining their children. N = 581.

91%

Listed a neighbor as the emergency contact for their children at school. N = 445.

91%

Listed as the emergency contact person for neighbor’s children at school. N = 401.

91%

Listed a neighbor as the emergency contact for their children at school and was listed by the same neighbor as the emergency contact person for neighbor’s children at school. N = 339.

92%

147

C H A P T E R

E L E V E N

Network Influence Theory

SOCIAL INFLUENCE NETWORK THEORY In chapter 5, I argued that the flow and exchange of norms, values, beliefs, and influences among neighbors along their stage 4 influence networks can generate social capital and collective efficacy and other important neighborhood community effects. Norms and values flow and are exchanged, but precisely who influences whom and how much is a function of the structure of the neighbor networks. Residents’ efforts to integrate discrepant values and to demonstrate socially validated norms occur within these networks. Neighborhood community, social capital, and collective efficacy, and all of the other products of the flow and exchange of norms, values, and expectations along the conduit provided by neighbor networks, are embedded functions of the neighbor networks that convey them. The structure of these networks, therefore, has profound influence on the evolution of norms and values, the content of equilibrium positions concerning norms and values that residents settle on, and the speed and efficiency with which this content is produced. The structure of the neighbor networks guiding the flow and exchange of norms and values also has further implications for the emergence of community, continuously creating and re-creating social collectivities. We can understand emergent neighborhood effects better if we more precisely identify the neighborhood from which the effect could emerge. An important fact distinguishes neighbor networks from some other types of social networks. While residents may attempt to restructure their personal neighbor networks to control this influential flow of norms and values, the geographically constrained nature of neighborly contact makes this control difficult. Unlike other less constrained relations, for those who are unwilling or unable to divest themselves of 148

NETWORK INFLUENCE THEORY

neighborly contact, allowing oneself to be influenced, even in undesirable ways, is often the only feasible alternative. Instead, it is just as likely that, as a result of the interchange of norms and values, shifts in values and norms occur when individual residents alter their opinions to more closely approximate the apparent consensus. Persons integrate conflicting norms and values as if they were utilizing a “cognitive algebra” of weighted averaging. This integration is initiated by the display of differences in norms and values and the susceptibility of person to interpersonal influences, what Friedkin (2003) refers to as “a social comparison trigger.” As a result of this triggering, people attempt to integrate discrepant influences and to form socially validated norms and values. These attempts occur, however, within the preexisting network of interpersonal influences. This network has profound effects on the course of the change process and the revised norms and values that persons may settle on. The specific norms and values held by specific individuals once equilibrium is achieved, the efficiency with which this equilibrium is produced, and the relative net influence of each group member on others depend on the structure of the influence network. Social influence network theory1 is a mathematical formalization of the process of interpersonal influence that occurs in groups, affects norms and values, and produces interpersonal agreements, including group consensus, where it did not previously exist. The theory allows one to predict the outcome of interpersonal influence as well as the consequences of particular modifications of social structures on the expected outcomes of the influence system. Social influence network theory describes processes in which individuals’ norms and values change as they revise their positions by taking weighted averages of the influences of others members. The basic influence model (which I adapt here from Friedkin 1991, 1998, 1999, 2003; Friedkin and Johnsen 1990, 1997, 1999) is as follows. The opinion of individual i at time t is the product of the individual’s own opinion at the previous time period, yit1, and the opinions of others at the previous time period, xit1. The susceptibility of individual i to the influence of others, or the relative impact of others, is given by the weighting factor, ai, with ai = 1 indicating that one’s opinion at time t was entirely subject to the opinions of others and with ai = 0 indicating 149

CHAPTER ELEVEN

that one entirely disregarded, both consciously and subconsciously, the influences of others. Both extremes, of course, are highly unlikely scenarios, and the reality for almost everyone is somewhere in between. The following equation formalizes this. yit = aixit1 + (1  ai)yi1 0  ai  1 t = 1, 2, . . . In this equation, the influences of others are treated as an ambiguous external factor; however, individuals respond to specific influences from specific others. Therefore, the opinions of others at the previous time period, xit1, which have the potential to influence individual i, can be conceptualized as the opinions of all other individuals j at the previous time period, yjt1, weighted by their relative influence on individual i, wij. Thus, xit1 = jwij yjt1 0  wij  1 jwij = 1 wii = 1  ai j = 1, 2, . . . N persons. Inserting this into the previous equation we have yit = aijwijyjt1 + (1 ai)yi1. Thus, social influence network theory models individuals’ opinions at time t as a function of everyone’s initial opinions, the social structure creating the set of interpersonal influences, and individuals’ susceptibilities to influence. Over time, if this model reaches equilibrium it will be given by Y = VY1 V = [I + AW + (AW)2 + (AW)3 + . . .](I  A). Thus, the total interpersonal effect of individual j on the norms and values that individual i holds when equilibrium is achieved arises as a cumulative summation of the direct and indirect flows of interpersonal influence in the system. More precisely, the total interpersonal effect of one individual on another is related to the number and length of the various sequences that join them in the network of interpersonal influences. For example, considering the term (AW)k, in the infinite series, the specific value, awijk, indicates the amount of the flow of influence 150

NETWORK INFLUENCE THEORY

from individual i to individual j that reaches individual j in precisely k steps, having flowed through precisely k  1 intermediaries. For computational purposes, the infinite series [I + AW + (AW)2 + (AW)3 + . . .] reduces to the matrix inverse of (I  AW), and thus V reduces to V = (I  AW)1 (I  A). In actual practice, however, we rarely have complete (or even nearcomplete) information about who influences whom, not to mention how susceptible individuals are likely to be. In such cases, we are forced to use more general measures to proxy the influence process and the structures that guide it. I discuss some of the more important measures below.

BEYOND DENSITY It has been argued that dense neighbor networks allow norms, values, symbols, ideas, and other social and cultural goods and resources to flow extensively among residents, diffusing broadly and being exchanged freely through social interaction. In a dense neighbor network, there might be efficient information spread about how residents are behaving and how neighborhood residents are reacting to these behaviors. Thus, residents would be provided with numerous exposures to other residents’ norms and values, expressed both through their behaviors and their reactions to the behaviors of others, and this would allow for greater consistency in normative behavior, greater consensus in values, and ultimately to emergent community. Furthermore, social disorganization theory has been interpreted to state that neighborhoods with low densities of social ties between neighbors2 or low frequencies of social interaction among neighbors3 are less able to realize common values and maintain the social controls that foster safety.4 All of these images implicitly assume a random or somewhat uniform organization of neighborly relations. In fact, in his classic work, Friedkin (1998, 90) states, “In natural settings . . . the structure of a network is likely to form in a manner that produces a close correspondence between the bases of interpersonal power and the degree of 151

CHAPTER ELEVEN

FIGURE 11.1a FIGURE 11.1 Example networks with identical order (number of vertices) and size (number of edges) but disparate quantity and length of paths

actors.” While this assumption may be appropriate for friendship networks, or for many other commonly studied networks, the situation is quite different for neighbor networks, which are heavily dependent upon the geographic location of a set of individuals and the duration of time that they have been so located. The stage 1 tertiary street networks that constrain higher stages of neighboring relations are clearly not a “natural setting” in the sense implied by Friedkin. I argue, therefore, that density alone is not sufficient to understand how norms, values, and expectations flow along the neighbor network. Neighbor networks with equal density might provide highly unequal opportunities and disparate constraints for the diffusion and blending of norms, values, and expectations that occur as information is transmitted about how residents are behaving and how neighborhood residents are reacting to those behaviors. Figure 11.1 illustrates nonrandom organizations of neighborly relations. All four networks have nine actors and 11 relations, thus perfectly equal density. In the first network shown (fig. 11.1a), there is an average of 1.7 steps between each pair of neighbors. While a little interaction occurs between others, norms and values would generally be interpreted through the central neighbor. Thus, of the 72 (9  8) possible influences occurring between pairs of neighbors (each of the nine 152

NETWORK INFLUENCE THEORY

FIGURE 11.1b

FIGURE 11.1c

FIGURE 11.1d 153

CHAPTER ELEVEN

neighbors could influence each of the eight others), only six acts of influence can happen without involving the central actor.5 This situation might occur in a neighborhood if, for example, relations revolved around a day care center operating out of a residence: Each person might have ties to the household operating the day care center, and be connected only through this household to every other household in the neighborhood. In the second network (fig. 11.1b), there is an average of 2.5 steps between each pair of neighbors, almost a full step longer. Four of the neighbors are intermediaries in the flow of norms, values, and influence between two or three others and the larger group. In the third network (fig. 11.1c), there are actually two distinct groups. Therefore, the distances between neighbors in different groups are infinite. No norms, values, expectations, or influence of any kind will flow among these two groups via the relations under consideration. Considered as a set, these residents have the same density of ties as all of the others, but there is no possibility for influence to flow between the two distinct groups. Within each subgroup, however, path lengths are short, averaging 1.5 steps for the group on the left and one step for the group on the right. Furthermore, both subgroups are robust. Neither subgroup has a neighbor who necessarily acts as an intermediary in the flow of norms, values, and influence between people. In the fourth network (fig. 11.1d), there is an average of 1.92 steps between each pair of neighbors. In this network the neighborhood appears relatively cohesive. There are two or three independent paths between all pairs of neighbors, which means that not only is it impossible for a single neighbor to color the flow of norms and values between any set of neighbors, but often it is equally impossible for two distinct neighbors to do so. It would take no less than two or three neighbors (and they would have to be the right two or three) to unduly influence the flow of norms and values. A shortcoming of many previous approaches to social influence has been that they limited their focus to density, as if all of the information contained within neighbor networks could be found in the individual worlds of individual neighbors. These studies may have summed these worlds, or multiplied these worlds, but they all perceived neighborhood effects to be a product only of these worlds. 154

NETWORK INFLUENCE THEORY

If this is all that I concerned myself with, there would be no need to discuss neighborhoods and their effects in terms of networks. Neighborhoods would just be agglomerations of individuals satisfying similar interactional and locational wants and needs. I argue, however, that neighborhoods are more than just clusters of demographically similar households residing nearby each other and interacting with each other. They are true social entities, or at least they have the potential to be. They emerge from the lives of their constituent residents but are a linear function of those lives. They cannot be understood by simple addition or multiplication. They are something different. This is a fundamental concept. They are the product of the lives and interactions of their residents, but not in a linear fashion. Neighborhood community and its effects emerge from the concatenations of interactions among households,6 interactions that often concatenate in odd, nonlinear ways, interactions whose concatenations are mediated, directed, and constrained by geographic considerations. Neighborhood effects are a function of neighborhood community networks, although they are not a linear function of the density of these networks.

THE HORIZON OF OBSERVABILITY If density is an insufficient proxy of the neighbor influence networks with which we are interested, what else might be? For a neighborhood community to produce an efficacious environment, it must be able to produce actionable social control. What is necessary for this to occur? Social control consists of two distinct processes: (1) monitoring and evaluating performance, and (2) influencing the monitored and evaluated performance.7 Observability of role performance is a prerequisite condition of control, in that reactions to an individual’s behavior cannot occur unless the behavior is first observed. Furthermore, the person whose behavior has been observed must in turn be able to observe the reactions to and evaluations of that behavior. These two conditions must be met for influence or control to occur. 155

CHAPTER ELEVEN

Granted, observability does not guarantee control. It is quite possible to monitor the behavior of another, to evaluate that behavior, and to transmit one’s evaluations to the other person, without any influence or control occurring. Lack of observability, however, guarantees lack of control since the preconditions of the control process are absent. When we think about the role of neighborly relations, what we are really interested in is the following. I behave in a particular normative way. You evaluate my behavior. You behave in a particular way. I evaluate your behavior. We may or may not discuss our evaluations with each other or with various others. By your observing my behavior and by my reacting to my observation of your behavior, I have potentially influenced you. For example, if a resident of a face block observes a group of children other than his own who live on the same face block engaging in delinquent behavior, he has the option to confront the children or not. If he chooses to confront the children, their parents may find out and either reaffirm or denounce this behavior both to the resident who confronted the children and to other neighbors. If all of the residents of the face block know each other and interact somewhat routinely, news of this behavior and of how the parents of the confronted children and other residents of the face block have evaluated it will soon become common knowledge. The behavior will have been monitored and evaluated. Neighbors will have shared norms and values and either subtly or overtly pressured each other toward a common evaluation. The neighbors in this example all knew each other personally. None required an intermediary. They were all one step away from each other (zero steps being yourself). What if they had been two steps away from each other? What if there is a group of residents who do not all know each other but all know at least someone (or several others) who knows everyone else (i.e., if they did not all know each other directly but knew common intermediaries), can observability occur? Can individuals become aware of the behavior of others through retelling by one or more intermediaries, evaluate it, and, again through intermediaries, transmit their reactions, with any force or impact, back to the individual whose behavior was evaluated? Is it possible for any influence or social control to occur under such a condition? 156

NETWORK INFLUENCE THEORY

In his classic study “The Horizon of Observability,” Friedkin (1983) studied this question by asking to what extent one academic is able to observe the current (potentially unpublished) research of another academic through a network of face-to-face communication about current research. Friedkin (1983) found that, with reference to this relation, the horizon of observability appears to generally extend to persons who are two steps removed in a network but not much beyond that. There was a dramatic decline in the likelihood of observability associated with increasing network distance. Thus, persons who discuss their research with a common intermediary may have a useful awareness of each other’s research, even though they do not directly communicate, and may be able to influence the general direction through third-party discussions, but beyond two steps such influence or informal control is rare. To illustrate what the horizon of observability might look like, I revisit figure 11.1. If these networks represented academics discussing their research (i.e., the nodes would be academics and a line would connect them if they discussed their research), then in the first figure, everyone would be within the horizon of observability of everyone else. In the second figure, no one in this network has the entire network within her horizon of observability, and, for the two neighbors at the distant ends of each triangle, the majority of the network is outside their horizon of observability. In the third figure, where there are actually two distinct groups, within each group everyone is within everyone else’s horizon, and between groups, of course, no one is. In the fourth figure, the average neighbor is within the horizon of most, but not all, of the other residents. Again, all figures had perfectly equal density but dramatically different potentials for social control based on their ability to observe and influence each other. How is this relevant to neighbor networks and neighborhood communities? If academics separated by two steps (i.e., they discussed their current research in face-to-face communication with the same third alter) are more likely than those who did not to be aware of each other’s research, it seems reasonable to suppose that any neighborly relation at least as intensive as academics discussing current research in face-toface communication might be expected to create awareness at least two 157

CHAPTER ELEVEN

steps out.8 I will extend this discussion in the applications presented in the two subsequent chapters.

STRUCTURAL COHESION Friedkin (1983) also noted that, for two-step connections, how many there were mattered. For example, if A discusses his research with both C and D and both C and D discuss their research (C’s and D’s, respectively) with B, then B is more likely to be aware of A’s current research than if A had only discussed it with C. Furthermore, Friedkin found that, if A and B were separated by two steps, but there were at least six such paths between them, then they were about as likely to be aware of each other’s research as if they were directly connected, or one step apart.9 This importance of this statistic can be lost if you do not stop to think about what it means. Assume you, the reader, and I, the author, both discuss our current research with six other people, but not with each other. I discuss my research with these six people and you discuss your research with the same six people, but neither you nor I discuss our research with each other. The startling thing is that in the end we will be as aware of each other’s research as if we had discussed it directly. The incidental knowledge we would obtain through those six other people, each of whom has had discussions with both of us, would give us as clear an awareness of one another’s research as if we discussed it face to face. Secondhand knowledge may be less valuable than firsthand knowledge when it comes from one source only, but what it lacks in immediate value, it makes up for in volume. Similarly, when norms and values diffuse broadly and are exchanged freely, “there is more efficient information spread about what members . . . are doing, and thus better ability to shape behavior.”10 This tends to lead to increased consistency in normative behavior,11 increased consensus in beliefs,12 and increased sense of community.13 Moody and White (2003) noted that, while the number of paths interconnecting others is important, the number of independent paths is even more important. For example, in the first network (fig. 11.1a), 92 percent of all possible influences occurring between pairs of neigh158

NETWORK INFLUENCE THEORY

bors somehow involved the central actor. In contrast, in the fourth network (fig. 11.1d), there are two or three independent paths between all pairs of neighbors. As noted above, this means that not only is it impossible for a single neighbor to color the flow of norms and values between any set of neighbors, but often it is equally impossible for two distinct neighbors to do so. Furthermore, it would take two or three neighbors (and they would have to be the right two or three) to unduly influence the flow of norms and values. This last concept is what Moody and White (2003) referred to as the k-connectivity of a network. A network’s k-connectivity simultaneously defines both the minimum number, k, of network members that, if removed, would disconnect the network and the minimum number, k, of node-independent paths14 connecting each pair of network members (the two ks are necessarily mathematically identical).15 What do these definitions mean in terms of the flow of norms and values? The k-connectivity of a set of neighborhood residents defines how many resident-independent paths there are, how many completely distinct paths, using completely distinct residents, will transmit observations and evaluations of behavior and how it was evaluated to and fro between each and every pair of residents in the set. It also gives us a way of defining how free the set of residents is from disruption or undue influence by the evaluations of one, or a small set, of residents. K defines the minimum number of neighbors required to completely color the flow of norms and values to and fro between any pair of neighbors in the set.

MERELY A MECHANISM? While density might be too simple a measure, one could argue that this focus on the complex structure of neighbor networks and the mathematical foundations of network influence is unnecessary and overly complicated. One might further argue that neighborhood communities must necessarily, at some level, be the product of the psychology of individual residents and that neighbor networks are merely a mechanism translating the characteristics of residents into the characteristics of a neighborhood. 159

CHAPTER ELEVEN

There is a sense in which this is true; but, there is also a sense in which most, or perhaps all, things are mechanisms. For example, in his somewhat controversial book The Selfish Gene, Richard Dawkins16 implies that people are merely mechanisms that genes use in the struggle of natural selection. While accurate in its context, most social scientists would be uncomfortable with that viewpoint. One person’s agent is another’s mechanism. Even if neighbor networks are only mechanisms for residents to act through, just as space is a mechanism, it is as important to understand the network that individuals, or households, act through as it is to understand the environmental and metabolic processes through which a genotype produces a phenotype. Values and norms, exchanged through peer influence and modeling, may be the motivations that create neighborhood communities, but they do it through the neighbor networks that are available to them. I could reside in a neighborhood with my norms and values and my predispositions to trust my neighbors and to work together with them to create a mutually beneficial environment. My neighbors could reside adjacent to me with their similar predispositions and values and norms. We could all be rather friendly. Until we actually observe each other and interact and become mutually aware of our similar norms and values and predispositions and until we actually begin to trust each other, we cannot effectively cooperate to create a more mutually beneficial environment. Not everyone has opportunities to observe and interact with everyone else. While residents may choose not to observe or to interact with some or all of those they could interact with, or their personal characteristics or lifestyles may constrain them from doing so, they are severely geographically constrained with respect to whom they can observe and interact with. Furthermore, they are completely powerless to determine who is within their horizon of observability or with whom they share many, or few, or no distinct paths transmitting norms and values. The patterning of stage 1 neighbor networks, of geographic availability, can neither be designed nor altered by the residents, except by moving, yet it completely controls the evolution of all neighborhood outcomes. 160

NETWORK INFLUENCE THEORY

Even the most activist resident, who intentionally attempts to generate a sense of neighborhood community, to generate trust among residents, to actively influence the norms and values of fellow residents, to produce social capital and collective efficacy in the neighborhood setting, is insufficient. The activist may certainly serve as a catalyst; but, if her fellow residents are not readily available to each other, either directly or through short paths, the most ardent activist is doomed to failure. Furthermore, and perhaps more importantly, besides being insufficient, activists are not necessary. Neighborhood communities often produce desirable environments without the aid of activist residents. Neighborhoods are often, perhaps typically, the unintended consequences of neighborly interactions. Most residents do not set out to create an efficacious community. Instead, they concern themselves with their family and the values and norms it is subjected to. They act in trustworthy manners amid their neighbors, affirm those who do also, sanction those who do not, and ultimately relocate if they are unable to create a desirable environment for their family. Community, social capital, collective efficacy, and the rest are side effects of individual households looking out for themselves.

MAIN POINTS IN REVIEW In this chapter, I formally revisited the discussion of how efficacious neighborhood communities emerge from the flow and exchange of norms, values, and beliefs among stage 4 neighbors. Social influence network theory mathematically models the evolution of community norms and values. A simple focus on neighbor network density, treating neighbor networks as if all of the information about them was contained within the relations of individual neighbors, ignores the information content captured in their larger networks. For influence to occur, residents must be within each other’s horizon of observability. The number of distinct paths transmitting norms and values between residents also affects the degree of influence; secondhand knowledge may be less valuable than firsthand knowledge, but what it lacks in immediate value, it can make up for in volume. 161

C H A P T E R

T W E L V E

Influence Networks in a College Town

T-COMMUNITIES, CHILDREN, AND THE HORIZON OF OBSERVABILITY How do stage 4 neighbor networks and the neighborhood community norms and values and social control they produce relate to tertiary street networks? In the previous chapter, I discussed Friedkin’s (1983) work, which found that academics separated by two steps (i.e., they discussed their current research in face-to-face communication with the same third alter) were more likely than those who didn’t to be aware of each other’s research. I argue that Friedkin’ s research provides a useful baseline to study neighbor networks and collective efficacy. It seems reasonable to suppose that any neighborly relation at least as intensive as discussing current research in face-to-face communication might be expected to create awareness at least two steps out. Recall that, in the college town, any two typical households with children knew a common other household with children (with one more intermediary household with children required a minority of times), being separated by an average of 2.38 steps. Thus, this relation of knowing each other is near the edge of Friedkin’s “horizon of observability” for households with children. In contrast, the additional two steps that were required to connect the typical households without children in this community (4.41 steps separated average households of this sort) suggest that they are far beyond any “horizon of observability,” more than doubling it. Furthermore, for the three of four intermediary households typically required to connect the average pair of households without children in this community, at least one typically had children living in it. Thus, while the short path lengths among households with children, near the horizon of observability, may have made them aware of each other as a community, the long path lengths 162

INFLUENCE NETWORKS IN A COLLEGE TOWN

among households without children would have made them very unlikely to be aware of each other as a community, and, when they were aware of a community, this community would have involved households with children. Two households with children who merely know each other may not have a very intense relation. However, one relation we interviewed about, neighbors’ monitoring of each other’s children and their evaluations of other neighbors’ monitoring each other’s children in spontaneous playgroups, seems likely to be at least as important to neighborhood families as academic’s discussions of their current research is to academics. For the exhaustive census, 6,198 such child-monitoring relations were reported among 669 households with children. This relation forms a giant connected component1 with 518 members. Seventy-seven percent of all households with children are in this giant component. The remaining households with children are either isolates (66 households that have no child-monitoring relation with another household) or members of trivially sized components. The trivial components include 13 dyads, five triads, four four-household components, two fivehousehold components, and three six-household components. Within this giant component encompassing over three-fourths of all households with children in the neighborhood, the median path length between households is two and the mean is 2.35. Thus, 77 percent of the households with children in this neighborhood existed in a community where the typical household watched over the children of someone who watched over the children of other typical households (with one more step of “watched over the children of” in some households). Furthermore, these networks were several times as dense as those Friedkin (1983) studied, being far more connected by two-step paths than were Friedkin’s academics. On average, Friedkin’s academics were connected by two distinct two-step paths of sharing research, while in contrast the neighborhood average was 13.72. Thus, a typical household with children had watched over the children of 14 other households, each of whom who watched over the children of the same other typical household with children in the neighborhood.2 Thus, it seems likely that a very large portion of the neighborhood is mutually observable through this relation. 163

C H A P T E R T W E LV E

In sum, the length of this relation, trusting each other to watch over children in spontaneous playgroups, is near the edge of Friedkin’s horizon of observability, and its density is much greater than that of the networks Friedkin studied. In addition, this relation is arguably at least as sociologically substantive as two academics discussing their research. If parents are at least as interested in the values of those who monitor their children as academics are interested in the research of their colleagues, then it seems reasonable to assume that they may have a good sense of the values, at least relating to children, of households monitoring children throughout this neighborhood. Clearly, in this case, the behavior of neighborhood children and those who monitor them is observable to most of the other households with children. The network of relations also appeared to create a community value. Every single person who was part of the neighborhood-sized, but closed, community of households watching over each other’s children in spontaneous playgroups felt it was a good neighborhood for children. While others who were not part of this network also believed it was a good neighborhood for children, 65 of the 151 households with children not included in this giant component did not assess the neighborhood as good for children (44 of these were isolates). The likelihood that 44 of the 65 households who believed it was not a good neighborhood for children might be isolates was statistically insignificant (p  .05). Furthermore, the likelihood that none of the 65 would be in the giant component was statistically infinitesimal. It simply could not have resulted from chance. For those in the giant component, the neighbor network clearly created the social closure necessary to affirm this belief for everyone. Thus, to be part of this community of neighbors watching over each other’s children was equivalent to believing it was a good neighborhood for children.

T-COMMUNITIES AND SOCIAL CONTROL An interesting, although unhappy, example of this observability in action presented itself for us quite fortuitously. During the Halloween immediately following the exhaustive census an incident occurred in the t-community we had been intensively studying. Several parents 164

INFLUENCE NETWORKS IN A COLLEGE TOWN

claimed to have seen a teenager exposing and sexually gratifying himself while watching children “trick or treat.” The parents did not recognize the teenager and believed he did not live in the neighborhood. A police report was filed that provided details of the incident, as well as which parents claimed to have observed the incident, and a date. This incident was intentionally not reported in any local papers, even in the police beats. It was not announced at any of the local schools. In fact, school officials were unaware that it had taken place. This news spread through the t-community, however. The details were modified somewhat as the event was described by one neighbor to another. In one version or another, the event became part of the collective memory of at least some of the neighborhood’s residents. It became such a powerful part of the neighborhood’s informal collective memory that, during the following Halloween, parents organized meetings a few weeks in advance to actively monitor their children while they were going house to house. While this semiorganized response did not occur the subsequent years, parents testified to increased vigilance. Since we sampled this neighborhood, the t-community and surrounding areas, three years later, I had an opportunity to follow up on the story. I could map the degree to which the news had spread through the neighborhood. I could verify the accuracy of the details. I could, more importantly, determine who was and who was not aware of the story. Had it spread primarily among parents, or was it more widely known? Had families who had moved into the neighborhood after that Halloween become aware of it? What were the geographic parameters that guided and constrained its spread? When interviewing residents during the data collection three years later, we asked if they had heard about an incident occurring at Halloween “some years earlier.” While details of their descriptions varied tremendously, of the 424 families with children we surveyed, 151 mentioned the incident in some form. Only nine of the households without children mentioned the incident. Of the families with children, 83 had moved into the neighborhood after the incident had occurred three years earlier. Eighteen of these new families with children mentioned it. Altogether, the families interviewed lived in three elementary school catchments, with 196 families in one catchment, 133 families in another, and 95 families in a third. Of these, 83 families in the first school 165

C H A P T E R T W E LV E

catchment, 24 families in the second school catchment, and 44 families in the third referred to the incident. The most telling fact, however, was that only 222 of the interviewed families (about half) lived in the tcommunity in which it had occurred, but all 151 who referred to it lived in that t-community. Thus, the story—the observed behavior and the observation of parental response, the valuation of behavior by the neighborhood community—had persisted, had been introduced to new residents, had spread almost exclusively among parents, had spread throughout all three school catchments, but had remained entirely within the t-community.

NEIGHBOR INFLUENCE AND T-COMMUNITY CULTURE Having noted the remarkable closure and short range of these neighborly relations, I now explore their utility. What does it matter whether one lives in a neighborhood that is closed and dense with short paths? I argue that it matters because one’s behavior is observable throughout the neighborhood, as are one’s reactions to the behaviors of others. Thus, one’s norms and values are readily transmitted, and the norms and values of others are readily accessible. I have argued that neighbor networks are the mechanism translating individuals and their social norms and values into neighborhood communities. In this section, I explore this mechanism by focusing on four particular neighborhood values: whether neighbors share one’s values, whether people are the best part of the neighborhood one lives in, whether one lives in a safe neighborhood, and whether the neighborhood one lives in is “good for children.” I explore whether one’s values are similar to those of neighbors. Table 12.1 shows that there are correlations of .96, .86, .76, and .75, respectively, between one’s own opinions and those of one’s immediate neighbors about whether the neighborhood is safe, whether people are the best part of the neighborhood, whether neighbors share one’s values, or whether the neighborhood is a good one for families with children. Thus, in one of the typical 10-step interview chains conducted in the first set of 68 Los Angeles samples, these would be the correlations of the fifth interviewee with the fourth and the sixth interviewees, who 166

INFLUENCE NETWORKS IN A COLLEGE TOWN

TABLE 12.1 Similarity in Residents’ Beliefs about Neighborhood, by Neighbor Network Steps Separating Them 1 step 2 steps 3 steps away in the away in the away in the interview interview interview chain chain chain

4 or more steps away in the interview chain

Safe neighborhood

.96***

.92***

.88***

.84***

People are best part of this neighborhood

.86***

.74***

.64***

.55***

Neighbors share my values

.76***

.58***

.45***

.34**

Good neighborhood for families with children

.75***

.57***

.42***

.32**

** p  .01 *** p  .001.

were each one step away. These correlations drop to .92, .74, .58, and .56 between one’s opinions and one’s neighbors’ opinions who are two steps in the interview chain away. Again, in one of the typical 10-step interview chains, these would be the correlations of the fifth interviewee with the third and the seventh interviewees, who were each two steps away. From the perspective of the fifth interviewee in the typical 10step interview chains, three steps away would be the second and the eighth interviewees, while four or more steps away would be the first, ninth, and tenth interviewees. Clearly, one’s perception of one’s neighborhood is very similar to one’s immediate neighbors’ perception, and this similarity of perception concatenates, with only somewhat diminished impact, across neighbor networks, step by step. Especially noteworthy is that, at any distance in the neighbor network, chains of respondents shared a correlation of at least .84 about whether or not it was a safe neighborhood. Not only are one’s perceptions of one’s neighborhood values similar to those of one’s neighbors, but they relate directly to one’s interactions with neighbors. Table 12.2 shows the high correlation between individual-level neighborly relations and one’s general perception of one’s neighborhood. For example, the majority of the variation in whether or not a person believed it was a safe neighborhood could be 167

C H A P T E R T W E LV E

TABLE 12.2 Neighborly Interactions and Perception of Neighborhood, All Households

Safe neighborhood

People are best Neighbors part of this share my neighborhood values

Good neighborhood for families with children

If I felt unsafe, there is a neighbor I would call before I called the police

.90***

.36***

.26***

.49***

There is a neighbor I have given my house keys to

.78***

.69***

.51***

.36***

*** p  .001.

accounted for by whether she had at least one neighbor whom she would call if she felt unsafe before she called the police, or if there was a neighbor someone had given keys to so that he could let in a service person or one’s children if they were locked out. These neighbor-level activities related powerfully to impressions of the neighborhood, and the converse is true as well. If we focus on households with children, similar powerful correlations appear. The majority of the variation in whether or not one believes it is a good neighborhood for children or believes neighbors share one’s values can be accounted for by whether or not there is a neighbor who has monitored one’s children in a spontaneous playgroup and whether or not there is a neighbor one would not be concerned about disciplining one’s child. Again, these neighbor-level activities related powerfully to impressions of the neighborhood, and the converse is true as well. In short, neighbors influence each other’s beliefs both by their actions and by their interactions. The beliefs and values foundational to neighborhood effects, such as the working trust necessary for the development of collective efficacy, emerge from these networked interactions. This suggests that neighbor networks also affect how successful 168

INFLUENCE NETWORKS IN A COLLEGE TOWN

TABLE 12.3 Neighborly Interactions and Perception of Neighborhood, Households with Children Only Good People neighborhood Neighbors are best for families share my part of this Safe with children values neighborhood neighborhood There is a neighbor whom I would not be concerned about disciplining my child

.84***

.82***

.79***

.76***

There is a neighbor who has monitored my child in a spontaneous play group

.80***

.70***

.28***

.54***

*** p  .001.

neighborhood-oriented individuals may be in generating efficacious neighborhood communities. The college town follow-up survey provided an opportunity to directly test the role of influence and to apply Friedkin’s influence model both to determine which geographic feature best proxies the actual influence structure in the neighborhood and to show that consensus in values follows as a result. I was able to do this because I had self-reports of various beliefs about the neighborhood from the same individuals at two different time periods. Thus, I had data on the residents’ change, if any, in their beliefs and values. While there were numerous beliefs and values I could model, I illustrate with two dichotomous beliefs. The first was whether or not the respondent believed that “People are the best part of this neighborhood.” This belief speaks directly to the perceived importance of neighbors. Furthermore, it proved closer to equivalent numbers of affirmations and repudiations than any other statement, thus offering maximal variance to explore. Of the 213 respondents who were interviewed at both time periods, 93 clearly agreed with this statement at time 1. This 169

C H A P T E R T W E LV E

number increased to 106 at time 2. While many more respondents agreed with this statement in part, issuing replies such as “They’re one of the best parts” or “There are many good parts in this neighborhood,” for this test I only count those who unequivocally agreed. The other belief I tested was whether or not the respondent believed that “My neighbors share my values.” Of the 213 individuals interviewed in both samples, 153 agreed with this statement at time 1. An almost identical number of people (not necessarily the same people), 154, agreed with this statement at time 2. My question was what these same respondents would say at time 2 and, if they altered their response, what network structure guided the influence at work upon them. Recall that, in Friedkin’s model, one’s beliefs at a particular time, yit, are a function of one’s beliefs at an earlier time, yit1, one’s susceptibility (i.e., one’s weighting of one’s own beliefs relative to the beliefs of others), ai, and the social structure guiding the flow of interpersonal influences, w. yit = ai j wij yjt1 + (1  ai) yi1. To proxy the influence structure, I used two different weight matrices. The first weight matrix, w1, I created by the following method. fbij = 1 if and only if person i shares a face block with individual j 0 otherwise w1ij = 1/j fbij Thus, the value in the ijth cell was 0 if individuals i and j did not share a face block, which was of course the most common situation, and, if they did share a face block, it was 1 divided by the number of other individuals i shared the face block with. Thus, this model assumed that all those who shared a face block with individual i had equal influence upon individual i and that no one else did. This, of course, is quite an oversimplification, but it was a convenient one to measure. Since the first weight matrix might be expected to form pockets of consensus the size of a face block, the second weight matrix I used expanded the horizon somewhat. It was similar to the first weight matrix except that in this case I assumed that all those who shared a face block with individual i or those who lived on a face block separated from individual i’s face block by one and only one tertiary intersection 170

INFLUENCE NETWORKS IN A COLLEGE TOWN

had equal influence upon individual i and again that no one else did. Thus, tiij = 1 if person i shares a face block with individual j 1 if individual j lives on a face block separated from individual i’s face block by exactly one tertiary intersection 0 otherwise w2ij = 1/jtiij While still a dramatic oversimplification, this weighting allows for overlapping spheres of influence. Thus, in this model, influence is less likely to form distinct pockets. Perhaps the most difficult aspect in using this model was estimating how susceptible a respondent would be, or how much a respondent would value her own opinion relative to her neighbors’ opinions. Given the noted relationship between neighborhood effects and residential longevity and given the apparently reasonable intuition that the longer one has lived in a neighborhood the more established one’s opinion of it will be, I used the number of years someone had lived in the neighborhood. I tried to scale the time in which residents had lived in the neighborhood many ways. The most successful in producing results proved to be the natural logarithm of 1 more than the number of years the person had lived in their current residence (I added 1 to avoid the undefined natural logarithm of zero). Having done this, I divided the resulting scores by the maximum value to scale susceptibility between 0 and 1. This scale, however, placed the highest value on those who had lived in the neighborhood the longest and the lowest value on those who had lived there the shortest. My intuitions were, however, that those who had lived there the longest would be least susceptible and that those who had lived there the shortest would be most susceptible. To fix my scale, therefore, I subtracted this result from 1 (so that 1 becomes 0, 0 becomes 1, etc.) to reverse the direction of the scale. Using this method, I produced estimates for each individual’s response at time 2. I was thus able to determine how accurate my estimate was using either method. Of course, perhaps the most widely argued for estimate of how each individual would respond at time 2 was how they responded at time 1. This is equivalent to setting respondents’ susceptibility to 0 (i.e., 171

C H A P T E R T W E LV E

FIGURE 12.1 “My neighbors share my values” at times 1 and 2

they cannot be influenced but will maintain their opinions). I use this as a baseline. One hundred and fifty of the 213 respondents gave the same response at time 2 to the query whether neighbors shared their values.3 This evidences a remarkable stability in belief structures within the neighborhood. The influence models, however, made better predictions. Using shared face blocks as a weight matrix estimated the correct response for 175 of the 213 respondents, and using separated by no more than one tertiary intersection produced the correct response for 183 of the 213 respondents. With reference to the query whether neighbors were the best part of the neighborhood, 186 of the 213 respondents gave the same response at time 2 (86 of the 93 individuals who said yes the first time said yes 172

INFLUENCE NETWORKS IN A COLLEGE TOWN

FIGURE 12.2 Comparison of models’ predictive power for “My neighbors share my values”

again, and 100 of the 120 individuals who said no the first time said no the second time as well). As before, this evidences a remarkable stability in belief structures within the neighborhood. The influence models, however, provided better estimates. Using shared face blocks as a weight matrix in the influence model estimated the correct response for 204 of the 213 individuals at time 2. Finally, using separated by no more than one tertiary intersection as the weight matrix in an influence model produced a near perfect result. It estimated the correct response for all but two of the 213 individuals at time 2. 173

C H A P T E R T W E LV E

FIGURE 12.3 “People are the best part of this neighborhood” at times 1 and 2

What is remarkable here is that, while it is well known that people tend to hold persistent beliefs across time, the influence models were able to make not only better, but nearly perfect predictions for their beliefs. Thus, in this case, a better approximation of how a respondent would reply at time 2 than merely what they said at time 1 would take into account the influence network structure, other’s beliefs, and their susceptibility to those beliefs. Clearly, in this case at least, structured influence networks matter in the determination of residents’ beliefs about their fellow neighbors’ values and utility. This could not bear more directly on the concepts of social capital and collective efficacy. One might be concerned that these findings result not so much from influence as from self-selection out of the sample. While only 213 re174

INFLUENCE NETWORKS IN A COLLEGE TOWN

FIGURE 12.4 Comparison of models’ predictive power for “People are the best part of this neighborhood”

spondents were interviewed at both time periods, this was primarily due to budgetary limits; we simply did not attempt to interview everyone the second time. Furthermore, we have no clear information on how many residents who had been present at time 1 were around to be interviewed at time 2. We do know, however, that those who were interviewed at both time periods did not differ in a statistically significant way from those who were not, either in number of neighbors identified, or by the presence of children in their household, or by race, or by their responses to the prompts “People are the best part of this 175

C H A P T E R T W E LV E

neighborhood” or “My neighbors share my values.” If selection occurred, it happened independently of these factors. One final point is worth highlighting. While residents’ beliefs and values conformed to those who shared their tertiary street network, not all neighbors in this study conformed to each other. Recall, from the discussion of the data in chapter 6, that there were three distinct tcommunities in the second college sample (and three distinct elementary school catchments as well). Residents’ beliefs and values conformed to those who shared their particular tertiary street network, but the three t-communities were converging to different values. Even in the short period of three years, it was clear that this process could lead to distinct neighborhood cultures. In the next chapter, I discuss an example where residential stability does in fact lead to extremely consistent norms and values shared among residents and a very distinctive neighborhood culture.

MAIN POINTS IN REVIEW In this chapter, I explored the relationship between influence networks and neighborhood-level outcomes in an insular setting, a college town. (I will explore the relationship between influence networks and neighborhood-level outcomes in a distinct insular setting, a gang barrio, in the next chapter.) Exploring the college town in this chapter, I begin by analyzing one particular neighboring relation, trusting each other to watch over children in spontaneous playgroups, and show that it is both dense and short enough to be within the horizon of observability, allowing the behavior of neighborhood children and those who monitor them to be observable to most of the other households with children in the t-community. I provided an example of a particular criminal incident, where the observation of the illicit behavior, the parental response, and the valuation of these behaviors by the neighborhood community was “observed” through influence networks by the residents throughout the t-community but nowhere else. Shared tertiary streets, but not shared elementary school catchments, circumscribed neighborhood collective memory and produced collective efficacy for children. 176

INFLUENCE NETWORKS IN A COLLEGE TOWN

I then showed that neighbors influence each other’s beliefs both by their actions and by their interactions. One’s perceptions of one’s neighborhood’s values are both similar to those of one’s neighbors and directly related to one’s interactions with one’s neighbors. The beliefs and values foundational to neighborhood effects, such as the working trust necessary for the development of collective efficacy, emerge from these networked interactions. The structure of influence networks, which was heavily determined by the structure of the tertiary street network, powerfully affected residents’ beliefs about their fellow neighbors’ values and utility. The norms and values that emerged within one t-community, while internally consistent, differed from those that emerged in neighboring t-communities.

177

C H A P T E R

T H I R T E E N

Influence Networks in a Gang Barrio

GEOGRAPHIC NEIGHBORHOOD AND SOCIOLOGICAL NEIGHBORHOOD In this chapter, I ask again: How does the nature of stage 4 influence networks and the identity, social capital, and efficacy they potentially produce relate to stage 1 tertiary street networks? I ask this question in a quite different way than in the last chapter, however. There I identified a geographic area and asked to what extent it related to a reasonable facsimile of community, trying to measure which type of geographic neighborhood equivalent typically evidences the most community. In this chapter, I demonstrate the relationship between geography and community from the opposite end. I identify a well-established community that provided identification, social capital, and efficacy for its members and attempt to understand why it was associated with a particular geography. The social group I focused on is a large territorial gang in Southern California.1 In this neighborhood, geography and community were tightly bound together. The association between the gang and the geographic neighborhood it claimed was so powerful that almost everyone within the geographic neighborhood identified with this gang, and virtually everyone who identified with the gang was found within the geographic neighborhood. Virtually all younger respondents, whether gang members or not, identified their neighborhood by the gang’s moniker. Furthermore, the majority of respondents of any age labeled their neighborhood with the gang’s name as well. Gang members typically claimed that almost everyone in the neighborhood was part of the gang. Local law enforcement took a more restricted view, claiming that it was a “substantial minority.”2 I conducted a rough statistical estimate using census data. According to the 1990 census, 1,155 18- to 21-year-old males resided 178

INFLUENCE NETWORKS IN A GANG BARRIO

in the census tracts claimed by the gang in 1990. In 1990, local law enforcement had identified 982 of them as members of the gang. At that time, the local police did not identify females as members of the gang; this policy has since been changed. Thus, somewhere around 85 percent of the 18- to 21-year-old males living in the territory claimed by the gang in 1990 had actually been identified by local law enforcement as a member of it. Records dating back to 1986 and forward to 1996, at the conclusion of the study, showed that the number remained relatively stable, despite the fact that very few of the individuals identified in 1986 were the same individuals identified in 1996. Thus, the vast majority of males in the locale identified with the gang had at some point been personally associated with the gang. The gang and local law enforcement agreed that most members of the local community of any age were sympathetic to the gang. My ethnographic experience confirmed this assertion. This powerful identification with the gang, however, was not limited to youth. Most residents of any age or gender identified with the gang. Furthermore, just about everyone was the cousin, sibling, or parent of someone the police identified with the gang. Two years of ethnographic study confirmed that most of the neighborhood’s residents saw the world in two parts: those who lived in their neighborhood, which identified with the gang, and those who did not. An interesting anecdote illustrates this identification well. After a church meeting one Wednesday night, I was talking with a couple dozen church members about their experience of life in this gang community. An Anglo youth who claimed never to have been a gang member (and no one present disputed this claim) recounted a time when he had cashed his paycheck at a check cashing service and emerged with several hundred dollars in small bills. On his way home he was approached by three gang members, who demanded that he turn over his money. He protested that he lived in the neighborhood. They claimed never to have seen him, but followed him home. When he arrived home, at a residence in the neighborhood, they apologized for accosting him and left without the money. The response from the others present, most of whom were adults, was that this how a neighborhood was supposed to be. One middle-aged man even claimed this was the way “God designed community.” 179

CHAPTER THIRTEEN

FIGURE 13.1 Gang neighborhood tertiary street network

NEIGHBORHOOD COMMUNITY AND TERTIARY STREET NETWORKS The foundation of this community identity was not racial or economic, however. Most of the residents of this neighborhood were people of Mexican ancestry who spoke Spanish as their first language and aspired to be working class at best, but so were most of the residents of the city of which the community was a part, and they belonged to different neighborhoods not associated with this particular gang. The only apparent sociological force that engendered the association of the gang with its geographic neighborhood was the stable discontinuities in the tertiary street network. The boundaries that the gang claimed had remained consistent for as long as the local police gang unit had existed. The boundaries that the gang claimed overlapped portions of several middle school catchments, and split across two high school catchments and two Catholic parishes (most of the residents of the neighborhood identified as Catholic). The boundaries that the gang claimed, however, perfectly coincided with a tertiary street island (pictured in figure 13.1), which had remained consistent for as far back as I was able to find street records. While school catchments had changed and parish boundaries had changed, the tertiary street network had not. 180

INFLUENCE NETWORKS IN A GANG BARRIO

What brought this connection between the gang’s domain and the tertiary street island to my attention was that the gang did not expand into two unclaimed neighborhoods immediately adjacent to their territory (most of that part of their city was claimed by one gang or another). Extensive discussion of this phenomenon with gang members made me realize that I was thinking in terms of maps, but they were thinking in terms of who they could walk down the street to hang out with. The communities I was interested in, which were quite close as the crow flies, seemed far away to them since they would have had to drive to get there or to travel down high-traffic streets, something they did not often do, and certainly not en masse. While most gang members were old enough to drive, the choice about whether to join or otherwise associate with the gang was made long before, in late elementary school and finalized by middle school. Children of this age are much more geographically dependent than adults, or even youth. As children and teenagers who grew up in the neighborhood played and went about their daily routines, they shared lives with each other. When they went outside to skateboard, to ride a bicycle, or just to hang out, the individuals they interacted with, especially the older role models they learned to look up to, were those who were geographically available. When they played games in the street, they played with those with whom they shared tertiary streets. While they could certainly interact with others at school, those relationships did not persist unless they were reinforced at home in the neighborhood. While youth visited other youth, and more distant gang members visited nearby gang members, these networks reflected the tertiary street system precisely because they had been formed in alignment with that system. The social network of each generation of youth mapped onto the social network they had formed as children. In sum, in this neighborhood, geography and community were tightly bound together; most residents of any age or gender identified with the local gang to some extent. The only apparent sociological force that engendered the association of this gang with its geographic neighborhood was the stable discontinuities in the tertiary street network. 181

CHAPTER THIRTEEN

AN EFFICACIOUS NEIGHBORHOOD How real was this neighborhood community that associated with this geographic context? Did the neighborhood community offer its residents anything more than identity? My ethnographic experience revealed a friendly, vibrant community, economically poor but culturally and spiritually rich. Putnam (2000, 315) in his well-known book Bowling Alone interpreted some gangs as an “attempt at neighborhoodbased social capital building in areas where constructive institutions are sadly lacking . . . closer to mutual aid societies based on horizontal bonds of interpersonal trust, reciprocity, and friendship that is defended to the death. In many cases gang members are tolerated and well integrated into the mainstream community.”3 The gang we are concerned with fits this description well. The neighbor network that identified with the gang generated an enormous amount of social capital. Numerous semiformal organizations existed in the neighborhood that were not typically composed of gang members and were nearly always composed only of adults. Some of these groups were handyman teams that went around the neighborhood fixing things, especially for single mothers. Other groups, primarily composed of older women, held parenting classes and conducted child abuse programs. Still others pooled books to form lending libraries, tutored students in English, and helped them with their homework. Groups organized to inform people about sexually transmitted diseases and AIDS or to provide career counseling and vocational training. Drug detoxification and prevention programs emerged in the neighborhood. Local businessmen banded together to raise scholarship funds for Catholic schools. Youth organized to paint murals and street art, most of which the local community enjoyed, but all of which the city officially disapproved of. Informal groups, teams, and organizations formed to do basically anything that “makes our barrio a better place to live,” as one resident put it. It is key to understand that all these groups, teams, and organizations were informal, that all of them were emergent, and that all of them associated with the neighborhood defined only by this gang. Virtually all did this by making the gang part of the name of the organization, which was not the formal real-estate name of the neigh182

INFLUENCE NETWORKS IN A GANG BARRIO

borhood or any of its parts (real-estate agents identified this gang’s territory as three distinct neighborhoods). The gang exemplified how the flow of neighborhood norms and values created social pressure, social control, and sanctioned behavior. While one may immediately think that the gang controlled the nongang community, perhaps the greatest control was exercised not by the gang on the community but by the community on the gang’s members, especially senior veteranos. The community pressured gang members, especially veteranos, to redistribute some of their surpluses back to the neighborhood rather than to consume them. Burglary and automobile theft of nearby neighborhoods brought resources into the community under the control of veteranos and other gang leaders. These gang leaders then gave some useful items, such as computers, televisions, Nintendos, and money to community churches and charities and sold the rest back into the formal economy for money, some of which they also donated. Churches and charities used some of this money to buy blankets or food and distributed these items along with the goods they received to community members. Sometimes veteranos loaned money to community members to buy houses or granted scholarships to promising young community members. Veteranos also helped sponsor community religious festivals. In 1995, two of the gang’s veteranos were the primary sponsors for the local Our Lady of Guadalupe festival. Most of the churches adopted a “Don’t ask, don’t investigate” policy with respect to the origin of the donated money and goods, although one monsignor mentioned liberation theology when discussing the donations made by the gang. The pressure exerted by the community on the gang was mentioned repeatedly in interviews with community members, gang members, and the veteranos themselves. The veteranos commented they were just “doing their part” and that if they failed to do so, they would be putos.4 By extensively exchanging trust and sharing beliefs and expectations in each other’s capability and willingness to act for the collective good, the flow of norms, values, expectations, and trust gave rise to working trust, or collective efficacy, although this application, a gang attempting to act on behalf of the neighborhood, is somewhat different than the typical applications of this concept. Most residents, however, did in fact perceive the gang to generally act on behalf of the neighborhood. 183

CHAPTER THIRTEEN

Finally, the most obvious example of this community efficacy was that the gang prevented the sale of drugs within the neighborhood. Around 1988, the gang had committed itself to keeping its community drug free. While they were not opposed to their members selling drugs to residents of surrounding neighborhoods, it was widely and loudly proclaimed that they would kill anyone who tried to sell drugs within the gang’s neighborhood. The local police claimed that, while no one was actually killed, several individuals who had tried to sell drugs in the gang’s territory were beaten savagely. Over time, this mandate became increasingly effective; by 1994, the gang had created the lowest rate of drug usage of any of the surrounding areas. While the police denied that the gang had eradicated drugs from their territory, it had been successful enough that the drug unit no longer paid attention to the area, considering whatever trafficking that did occur there to be trivial. In early 1995, the head of the local antigang unit told me that the gang’s actions had been so successful that it essentially allowed his unit to “split the city” with the antidrug unit.

NEIGHBORHOOD EFFICACY AS A FUNCTION OF INFLUENCE NETWORKS What was the basis of this efficacious (at least to its residents) community? In my ethnographic experience, I discovered numerous strong relations of trust and loyalty among many residents in this neighborhood. While there were several ways trust and loyalty was expressed, one phrase that captured the poignancy of the sentiment was that someone was “for you.”5 Everyone I encountered in the neighborhood recognized the phrase and, as far as I could determine, understood its meaning in the same way. Family members were the most common persons identified as being “for you.” Teachers, clergy, and other mentors were also commonly identified as being “for you.” Gang members, too often identified other gang members as being “for you.” The relation of being “for” someone was strong and not readily attributed. While the 184 interviewed respondents identified 2,880 individuals they knew personally, they identified 764 others they believed 184

INFLUENCE NETWORKS IN A GANG BARRIO

FIGURE 13.2 Number of alters respondents believed to be “for” them

were “for” them. The pie chart in figure 13.2 shows the distribution of those respondents’ beliefs about how many others were “for” them. Despite not being readily attributed, this “for you” relation formed a single component with a mean path length of 2.91. In the neighborhood associated with the gang, most residents were a little less than three steps away from each other in the “for you” network. The typical other respondent was “for” someone who was “for” someone who was “for” him. A very powerful relation connects that person with someone who is connected by a very strong relation to someone who is connected to the typical other respondent. Going beyond simple connectivity, we return to the concept of kconnectivity discussed earlier. The reader will recall that this concept focuses not merely on whether he community was a single unit in terms 185

CHAPTER THIRTEEN

of the flow of norms and trust but also the robustness and intensity of that flow. Everyone in the neighborhood was connected in the “for you” network (while some believed no one was “for” them, they themselves were “for” others), being a single entity in terms of the intense bonds of trust and loyalty, this community is even more intensely connected. A full 80 percent of the residents are involved in multiple paths of trust with other neighborhood residents, either giving or receiving norms, values, expectations, and influence. Almost half (47 percent) are involved in five or more such paths; a fifth (20 percent) are involved in 10 or more such paths; and a tenth (10 percent) are involved in 15 or more such paths. The average person was involved in 6.76 paths. In sum, many short paths of trust and loyalty interconnected residents. Most residents were a little less than three steps away from each other in the “for you” network. This flow of trust and loyalty along short, dense paths formed the lifeblood of this vibrant community. I certainly do not intend to argue that all neighborhood communities emerging from tertiary street networks will provide as much utility for their residents as this one did. However, as in my discussion about the importance of emergence in chapter 11, I argue that the tertiary street networks provided the necessary mechanism for the community to evolve. One more factor seemed to be important. The effects of the stage 1 neighboring relations provided by the tertiary street network became even more pronounced because of how long residents and their families had lived close to each other and thus how many opportunities to interact they had been given. While few adults had been born into that neighborhood (although some had), some had lived there for decades, and most of those who had moved there had done so to be with family members who already lived in the neighborhood. Similarly, and perhaps more importantly, even if they did not desire to interact with the neighborhood, every passing day they lived close to the same others, every day they encountered each other and had passive contacts with each other, made this decision not to interact increasingly difficult. Residents either had to become recluses in their own homes or choose to interact with members of the established neighborhood community. Finally, when they did interact, they had to choose to interact prudently 186

INFLUENCE NETWORKS IN A GANG BARRIO

as well, given the extreme residential stability. Neighbors of today were likely to be the neighbors of tomorrow as well. While the association between length of residency and geographic community is not surprising, what is important for my argument is which geography this long residency used as a catalyst to form community. Residents of many neighboring geographic areas had also lived in their area for just as long but were not part of this community. Many residents had attended the same middle school or high school as residents in areas claimed by competing gangs (and, in some cases, their parents had as well), but they were not part of this community. Residential longevity highlighted the effects of tertiary street proximity, but not mere spatial proximity or even residing in the same school catchment. Residing in the same school catchment or being spatially proximal did not prove to be sufficient stage 1 neighboring relations to be translated into higher stages of neighboring over time. Sharing tertiary streets, however, did.

INFLUENCE NETWORKS AS A FUNCTION OF TERTIARY STREET NETWORKS I have shown that, in this context, neighborhood community perfectly coincided with the tertiary street network. Residential stability worked together with the tertiary street network to engender many short paths of trust and loyalty, interconnecting residents. From these short paths of trust and loyalty emerged a community that provided identification, social capital, and efficacy for its members. A question remains: why was it a gang neighborhood? While this vibrant social entity sanctioned behavior, including the behavior of some of the most powerful members of the gang, and while it prevented the sale of drugs within their neighborhood, something that the police had been unable to do, it was still centered on a gang, definitely a criminal enterprise, an organization generally perceived by the larger society to be antisocial. Its members certainly victimized surrounding neighborhoods. They robbed and stole from, and sometimes sold drugs to, other parts of the city with impunity. 187

CHAPTER THIRTEEN

While any complete answer to the question of why the neighborhood community identified as a gang would undoubtedly be complex, I suggest one small factor that may have contributed. I argue that the neighborhood identified itself with the gang because those who were most intimately involved in the “for you” network identified with the gang. How does this correspond to my arguments about the relationship between neighborhood geography and community? Those who were most intimately involved in the “for you” network could be geographically determined by their location in the tertiary street network. First, I show that the neighborhood identified itself with the gang because those who were most intimately involved in the “for you” network identified with the gang. During the course of the interviews, the respondents were asked to which neighborhood entity (if any) they ascribed primary identification. Of these 184 respondents, 58 claimed to identify primarily with the gang; 35 with the local Catholic parish; and 91 identified either with both (which was not one of their original options) or another neighborhood entity (typically one’s family) or none at all. For the other neighborhood entities, none was identified with by more than three residents (again, these entities were typically respondents’ families). Even though the interviews showed more respondents to identify with the gang than any other single entity, the largest group of respondents refused to primarily identify with it. I argue that the factor that determined what this neighborhood community was about, which norms and values were accepted by the community and which were not, was not the sheer number of those who identified with the gang but rather the location of various neighborhood members within the structure of the stage 4 neighborhood community networks that guided the flow of influence. I argue that it was a “gang” community because those identifying with the gang were deeply embedded in the cycles of connectivity created by these trust networks, the neighborhood’s cohesive core. Residents who primarily identified with the gang were involved in an average of 5.07 cycles of k-connectivity in the “for you” network, while those who did not were involved in an average of 1.75 cycles of k-connectivity. Furthermore, everyone who primarily identified with the gang was involved in at least one cycle of k-connectivity while almost half of those who did not were not involved in any. Neigh188

INFLUENCE NETWORKS IN A GANG BARRIO

FIGURE 13.3 Alters identified as “for” respondents by distance (in face blocks) between alters’ and respondents’ residences

borhood community members had different primary identifications, but the heart of the neighborhood influence network, the neighborhood’s cohesive core, primarily involved those who identified with the gang. As a result, the neighborhood identified itself as a gang. The neighborhood identified itself with the gang because those who were most intimately involved in the “for you” network identified with the gang. This is, in itself, an interesting sociological finding. It becomes even more interesting when one considers the fact that those who were most intimately involved in the “for you” network could be geographically determined. Above I noted how the tertiary street network constrained interaction. Beyond that, however, it also tightly constrained networks of trust. Figure 13.3 illustrates. For each identified “for you” relation, I was able to calculate the distance in tertiary face blocks from the respondent to the person they believed was “for her.” Clearly, most of the people one believed were “for them” lived within a few tertiary face blocks. Thus, the simple fact of overlapping propinquity made this relationship possible. Furthermore, no one who lived beyond the tertiary street 189

CHAPTER THIRTEEN

network was identified by a respondent as being “for” them. Even for those respondents who lived at the edge of the tertiary street networks, all of those whom they identified as being “for” them lived along tertiary face blocks. Thus, those residents who occupied the center of the “for you” network also resided in the geographic center of the tertiary street network. By constraining the pedestrian interactions of local children and youth, the tertiary street network constrained neighborly interactions, including the powerful trust and loyalty relations. Furthermore, these neighbor influence networks embedded within the street networks helped determine what this community would be about. Who residents thought they were reflected whom they were for, which further reflected the street networks, and these identifications ultimately guided the neighborhood’s identification and cultural norms. This, of course, brings up an interesting chicken-and-egg question, which I will pose but am not prepared to answer (or at least to answer with any confidence). Did those who influenced the neighborhood community’s norms and values the most choose to live in a central location so that they could have influence, or did the neighborhood community merely reflect the norms and values of those who fortuitously lived there (and thus, if others had lived there, would the neighborhood community have reflected others’ norms and values)? From my ethnographic experience, the latter answer seems more plausible. Given the policy implications, this would be an interesting question for future study.

MAIN POINTS IN REVIEW In this chapter, I explored neighborhood community processes from a different direction. Instead of identifying a geographic area and asking to what extent it related to some reasonable facsimile of community, I identified a well-established community that provided identification, social capital, and efficacy for its members and attempted to understand why it was associated with a particular geography. I showed that the geography identified by residents of this community perfectly coincided with a tertiary street network but not with school catchments or parish 190

INFLUENCE NETWORKS IN A GANG BARRIO

boundaries or other potentially competing neighborhood foci. Within this neighborhood community, the neighbor influence network generated an enormous amount of social capital and collective efficacy, including actively preventing the sale of drugs within the neighborhood amid a city rife with the drug trade. More importantly, I showed that the neighborhood community took its powerful norms and values from those most intimately involved in the network of trust and loyalty, but who was most intimately involved in the trust and loyalty network was determined by where they lived in the tertiary street network.

191

C H A P T E R

F O U R T E E N

Implications

SUMMARY In this book, I have argued that neighborhood communities are not mere confounds of their geographic context but rather emerge from networks of neighbors interacting within them. I have further argued that neighborhood communities are geographically constrained because the interactions that produce them are geographically constrained. More importantly, I have argued that neighborhood communities are geographically identifiable because the networks of interactions that produce them, that translate neighbor-level interactions into neighborhood communities, are constrained by predictable urban geographic substrates. Administrative units are not those substrates. I began by arguing that the neighboring relationship is a multiplestage relationship in which each stage is superimposed on the previous stage. The initial stage of neighboring occurs between two individuals if they are geographically available to each other. While this proximity is typically conceptualized in terms of the absence of neighborhoodsized boundaries, I reconceptualized stage 1 neighboring in terms of neighbor-sized distances and the absence of neighbor-sized boundaries, specifically tertiary face blocks and intersections. Because each subsequent stage of neighboring is embedded within the previous stages, an accurate definition of stage 1 neighborly relations is crucial. Most sociological studies of neighborhoods, however, use administrative geography that implicitly defines two households to be stage 1 neighbors if they live in the same administratively defined area. It is not clear, however, whether residents of these spatially defined analytic units are geographically available to each other. In contrast, I defined new neighborhood equivalents in terms of the concatenated network of walking arenas as represented by tertiary face blocks. These new neighborhood 192

I M P L I C AT I O N S

equivalents focused on geographic availability, the potential for passive contacts, and thus the interactional aspect of neighborhoods. I showed that households relocated so that their stage 1 neighbors would be similar to themselves, when those stage 1 neighbors are defined in terms of the tertiary street network. They sometimes decided to move from their previous home if those who shared their tertiary streets were sufficiently different from themselves; they considered with whom they would share tertiary streets in potential future residences; and, if their attempts at homophily proved unsuccessful, they desired to move once again. To the extent to which neighbors have the same goals for homophily, these relocations concatenate within the tertiary street network, and thus segregation patterns reflect it. I used administrative data to show that discontinuities in the distribution of racial demographics do indeed map onto discontinuities in the tertiary street system, especially for the racial distribution of households with children. “Invisible” discontinuities in the network of tertiary streets are just as disruptive to population distributions as natural barriers or freeways. Finally, I showed that it does not require many tertiary streets to interconnect communities, and for racial disparities to dissipate. The most substantial distinction occurs between those that are in the same tertiary street network and those that are not. Thus, connecting two tertiary street networks by a single tertiary street may connect two otherwise disconnected communities and thus may dramatically affect their demographic composition. The next stage of neighboring relation, stage 2, exists between two residents when their lifestyles cause them to unintentionally encounter each other and thus have the opportunity to learn about each other through observation and to acknowledge each other’s presence. I showed that this stage is indeed embedded within the previous stage; the effects of proximity on passive contacts (and thus all higher stages of neighboring) are quite pronounced and quite short-distance, best counted in house-steps not “as the crow flies” but as the tertiary street runs. I showed that these passive contacts are sociologically real phenomena, not merely a theoretical construct. Respondents had no difficulty identifying whether or not an activity was an unintentional meeting resulting from the mere fact of being neighbors. 193

CHAPTER FOURTEEN

Most passive contacts began when children casually played together, and, in general, meetings involving children were identified as passive and meetings not involving them were not. Furthermore, these geographically dependent children and their families engaged much more in neighboring and neighborhood life than did households without children. The vast majority of all neighboring relations were between households with children. This corresponds to the fact that most of the effects of neighborhood communities with which researchers concern themselves involve children or adolescents either exclusively or primarily. Children and their families proved to be the quintessence of neighboring and neighborhood communities. A stage 3 neighboring relation exists between two residents if they have intentionally initiated contact. I showed that individual residents’ neighbor networks concatenated to form networks that extend farther geographically and socially than any individual resident’s neighbor network, but which nonetheless maintain relatively short path lengths among them all. Stage 3 neighbor networks proved to be a subset of stage 1 neighbor networks when those stage 1 neighbor networks were defined by shared tertiary face blocks and shared tertiary intersections but not when they were defined by the shared boundaries of census geography or elementary school catchments. The effects t-communities appeared to have in constraining neighborly relations were not spurious confounds of geographic distance. The effects of shared administrative geography were such spurious confounds. Finally, a stage 4 neighboring relation exists between two residents who engage in a substantial activity that indicates mutual trust or a realization of shared norms and values, when they share a belief in each other’s willingness and ability to act together to achieve a common goal, when they influence each other, either actively or passively. Neither stage 3 nor stage 4 implies that the involved parties understand their relationship to be intimate or strong in the sense of friendship or an affective bond. Efficacious neighborhood communities emerge from the flow and exchange of norms, values, and beliefs along these stage 4 neighbor networks. I demonstrated some of the potential emergent effects of neighborhood communities in two insular settings, a college town and a gang barrio. In the college town, I showed that the behavior of 194

I M P L I C AT I O N S

most neighborhood children and those who monitored them was observable to most of the other households with children in the neighborhood and created a community value. Every single person who was part of the neighborhood-sized, but closed, community of households watching over each other’s children in spontaneous playgroups felt it was a good neighborhood for children. Furthermore, I showed that this belief may have been accurate, as these neighbor networks produced at least some collective efficacy. Finally, residents’ beliefs about their neighborhood conformed to each other over time, tending to resemble those of their neighbors more than those respondents held in the past. The structure of influence networks, which reflected the tertiary street network, powerfully affected residents’ beliefs about their fellow neighbors’ values and utility and their expectations and actions in response to these beliefs. In the gang barrio, I identified a well-established community that provided identification, social capital, and efficacy for its members. I attempted to understand why it was associated with a particular geography. I showed that this intense community perfectly coincided with a tertiary street network but not with school catchment areas or parish boundaries or other potentially competing neighborhood foci. Within this neighborhood community, many short paths of trust and loyalty interconnected residents. I argued that, as a result of this intense interaction, the neighbor influence network generated an enormous amount of social capital and collective efficacy. Finally, I showed that the neighborhood community took its identity and cultural norms from those most intimately involved in the network of trust and loyalty, but that who was most intimately involved in the trust and loyalty network was to a large extent determined by where they lived in the tertiary street network. Locational-based neighborhood effects such as segregation correspond to the influence-based neighborhood effects such as social capital and collective efficacy because each acts on different stages of neighboring. Relocation determines stage 1 neighbors and thus, of necessity, the higher stages that influence works upon. In resettling, households segregate their phase 1 neighbor networks. To the extent that they have settled randomly with respect to the other components of phase 2 neighbor networks, these will segregate as well. If their household char195

CHAPTER FOURTEEN

acteristics that influence the transition of their phase 2 neighbor networks into phase 3 neighbor networks and that influence the transition of their phase 3 neighbor networks into phase 4 neighbor networks do not also covary with phase 1 neighbor networks, which is unlikely, their phase 2 neighbor networks will translate into phase 3 and phase 4 neighbor networks in a straightforward way. Thus, segregation patterns will correspond to neighborhood effects because the same concatenated, multistage processes are guiding both. Finally, I return to a discussion of my neighborhood equivalents. I did define two distinct new neighborhood equivalents. While they share a focus on the concatenated network of walking arenas as represented by tertiary face blocks, they differed with reference to the intersections they allowed to connect face blocks with each other. T-communities used only tertiary intersections. Islands used both tertiary and nontertiary intersections. Together, they allow me to test not only for the influence of tertiary face blocks upon neighboring but for the constraining power of tertiary intersections. How constraining were tertiary intersections? With reference to racial segregation, for the black and Asian populations (but not for the Hispanic or non-Hispanic white populations), settlement patterns clearly favored t-communities over islands, indicating that relocating households were not troubled by differences in the black and Asian demographics of nearby face blocks as long as a nontertiary intersection separated them. Thus, when neighbors desire a homophilous neighborhood, they appear to desire only that their face block and all face blocks that can access it via only tertiary intersections be homophilous. Overlapping networks of passive contacts, which emerged from the lived experience of these tertiary street networks, resembled tcommunities much more than islands. Ninety-nine percent of all intersections identified by respondents on their cognitive maps were tertiary in nature, implying that residents’ lived experience of their neighborhoods rarely crossed nontertiary intersections. A tiny fraction (1 percent) of the cognitive awareness of residents did transgress nontertiary intersections. They did not, however, ever traverse nontertiary face blocks. Of the 22,481 actualized neighboring relations that emerged from passive contacts constrained by the tertiary street networks, only 44 196

I M P L I C AT I O N S

crossed nontertiary intersections, and, again, none traversed a nontertiary face block. Thus, while 1 percent of residents’ cognitive neighborhoods may have expanded beyond t-communities into islands, only one-fifth of 1 percent (0.2 percent) of their actualized relations did. Again, none escaped islands. The influence networks at work in the college town and the gang barrio behaved differently from each other. In the college town, influence networks were completely constrained within t-communities. Not only were none of the child-monitoring or value-influencing relations able to traverse nontertiary face blocks and thus exit the tertiary street network, none was even able to cross nontertiary intersections. In this context, t-communities perfectly inhibited all stage 4 influence relations. The gang barrio, however, was a different story. Whatever pedestrian interactions were at work influencing neighboring children and youth with reference to the gang, these were not inhibited by nontertiary intersections. They were, however, inhibited by nontertiary face blocks. They could cross a nontertiary intersection but not go down a nontertiary street. This ability to cross nontertiary intersections probably corresponded to the long-term settlement patterns of the area. Even if only a small number of contacts crossed nontertiary intersections on a regular basis, as noted above, these may have been able to aggregate over time. In sum, clearly neighboring relations at any stage were not able to extend along nontertiary face blocks. A fraction, albeit a tiny fraction, were able to cross nontertiary intersections. Thus, when one conceptualizes the circumscribing effects of t-communities and islands upon neighborhood dynamics ranging from homophilous neighbor selection to unintentional encounters to actualized interaction to influencing each other’s norms and values, islands behave like sure containers, while t-communities behave like bags with small leaks in them.

WHAT IT ALL MEANS I wish that I had been able to gather data from neighborhoods around the world that varied along many meaningful parameters and that rep197

CHAPTER FOURTEEN

resented meaningful points along a continuum of efficacy. Unfortunately, that was not possible. The data collections reported here were difficult enough to undertake. As a result, I cannot definitively link the stages of neighboring relations with specific neighborhood outcomes. I believe I have, however, been able to reach several conclusions. First, neighborhood communities are more than residents with correlated attributes sharing similar conditions; they are neighborly interactions concatenating and consolidating to form emerging networks. By precisely identifying the latent social ties of passive contacts, tertiary street networks allow us to better understand why some translate into meaningful communities with various degrees of social capital, and some do not. In the neighborhood context, tertiary street networks provide us with a lens to focus more closely on agentive social capital. I am not arguing that I have identified the only possible type of neighborhood equivalent or the only useful one. Numerous neighborhood equivalents have been proposed, and each may be relevant for various purposes. What I am arguing is that the neighborhood equivalents I have identified, t-communities and islands, map particular underlying social processes more closely than do alternatives. They map some, but not all, of the forces that mediate neighborhood effects and should therefore be relevant to researchers interested in distinguishing such things. Furthermore, distinguishing which neighborhood effects, or other neighborhood-related processes, map onto t-communities and islands more than they do to other equivalents and which ones do not will help researchers understand which forces are at work in creating them. T-communities use an entirely different metric than census geography, one based on close-at-hand interactions. Any importance of census geography may result, in part, from its confounding with tcommunities. The importance of census geography may also result, in part, from its identification of distinct types of communities with distinct neighborhood effects. It may be the case that, similar to human society writ large, neighborhoods are “constituted of multiple overlapping and intersecting sociospatial networks.”1 Perhaps two or more different mechanisms are at work in neighborhoods leading to two or more different communities. One of these mechanisms is probably the concatenation of neighboring relations into neighborhood community 198

I M P L I C AT I O N S

networks. A second may be service areas, provided by schools, marketplaces, police, and so on, uniting residents around similar needs and opportunities. T-communities precisely measure the first. Neighborhood equivalents, defined solely by their boundaries, measure the second to the extent their boundaries coincide with the service areas. A careful use of both t-communities and neighborhood equivalents defined by their boundaries could tease apart the different mechanisms at work. I have previously performed this type of study,2 disentangling the effects of t-communities and census geography and showing that tcommunities better account for the racial distribution of households with children than do either census or school geography, while the results were more mixed when one examined the population in general. This suggests that households with children are primarily influenced by the concatenation of residential-street-based neighboring relations into neighborhood community networks, while service areas, provided by marketplaces, police, and so on, may play a larger role in the lives of households without children. This is consistent with the findings presented earlier concerning the special role of households with children. More importantly, studying neighborhoods precisely as networks rather than vaguely as diffuse entities highlights their nonlinear response to apparently similar conditions. Relatively minor modifications in the urban ecological environment that mediates individual-level interactions can result in disproportionate sociological outcomes. Thus, we can influence the evolution of neighborhood communities by modifying the patterning of the tertiary street network. For example, the most substantial difference in the demographic composition of adjacent block groups occurs when they are connected by their first tertiary street. Simply adding a single tertiary street, often no longer than a city block, between a pair of t-communities that are spatially adjacent would allow neighborly relations to flow along channels not previously available. A few relatively trivial, but well-considered, modifications made to the tertiary street system could have dramatic impacts on the health of local neighborhood communities, the development of social capital, the emergence of collective efficacy, and other neighborhood effects. By modifying tertiary streets, we can build better neighborhoods for the future. 199

CHAPTER FOURTEEN

Finally, I believe that this research speaks to fields of study besides neighborhood research, by refocusing on the need for the precise identification and measurement of the latent substrate that governs the probability social networks will form. When we cannot directly measure these social networks, and often we cannot, it is essential that we find appropriate proxies.

200

Appendix SURVEY INSTRUMENT

THE SURVEYS consisted of at least seven distinct components. The surveys began with three portions asking some basic questions about the respondents, their neighborhood, and their previous neighborhood. The fourth portion of the survey was a “name generator” that asked questions designed to elicit the names of any neighbors the respondent knew. The fifth portion of the survey asked specific questions about each neighbor they had identified, one at a time. Much information, however, had already been gathered through the name generators. The sixth portion of the survey asked about the respondent’s neighbors as a set. The seventh, and final, portion of the survey asked about the respondent’s neighborhood geography and included the cognitive mapping exercise.

ABOUT THE RESPONDENT How old are you? Who else lives in the household with you? How old are they? What is their relationship to you? What race would you identify yourself as? Would anyone living in this household identify as being of a different race? What ethnicity would you identify yourself as? Would anyone living in this household identify as being of a different ethnicity? What language do you most often speak at home? Is there a religion you identify or affiliate with? Is this the same for other members of your household? 201

APPENDIX

If not, how do you believe they would identify their religious affiliation? How long have you lived here? Do you rent this home or do you own it?

ABOUT THE RESPONDENT’S NEIGHBORHOOD There are many things people like about a neighborhood. What would you say is the best part of this neighborhood? People choose where they live for a variety of reasons. What was the most important reason for you to choose to live in this area? What other reasons were important? If you were able to, would you like to live in this area for many more years? When you think about the values that are most important to you, would you say that your neighbors share your values? Is this a good neighborhood for families with children? Is this a safe neighborhood? On a scale from 0 to 10, with “10” being virtually identical and “0” being completely different, how much are your neighbors in your current neighborhood like you? Were the people who live here an important reason for you to move into this area?

ABOUT THE RESPONDENT’S PREVIOUS NEIGHBORHOOD Now, thinking about the neighborhood you lived in immediately before you came to live in this neighborhood Why did you move from your previous home? On a scale from 0 to 10, with “10” being virtually identical and “0” being completely different, how much were your neighbors in your previous neighborhood like you? Did your previous neighbors influence your decision to move from that area? 202

SURVEY INSTRUMENT

NEIGHBORS THE RESPONDENT KNOWS Can you identify any neighbors that you know personally? Is there a neighbor you have socialized with? How often would you say you have done this? Is there a neighbor you have done a small favor for, such as lending them a tool or some food? Is there a neighbor who has done a small favor for you, such as lending you a tool or some food? Is there a neighbor you have given your keys so that they could let in service people or your children, if they were locked out, or for any other reason? Is there a neighbor who has given you their keys so that you could let in service people or their children, if they were locked out, or for any other reason? Is there a neighbor who has asked you to watch their house while they were away? What tasks did this involve? Is there a neighbor you have asked to watch your house while you were away? What tasks did this involve? Is there a neighbor who has assisted you with a large, expensive project such as reroofing a house, paving a driveway, or building a fence? Is there a neighbor you have assisted with a large, expensive project such as reroofing a house, paving a driveway, or building a fence? Is there a neighbor you might call before you called the police if you felt unsafe? Is there a neighbor who has ever called you when they felt unsafe? Is there a neighbor who you think might, even though they haven’t yet done so? Is there a neighbor whose children you have watched over in casual or spontaneous playgroups, without the other neighbor being present? How often would you say that you do this? Is there a neighbor who has watched over your children, without you being present, in casual or spontaneous playgroups? How often would you say that they do this? 203

APPENDIX

Is there a neighbor whose children you have dropped off (and/or picked up) at school, soccer matches, or any other activity? What types of activities or events do you typically drop them off or pick them up for? How often do you do this? Is there a neighbor who has dropped off (and/or picked up) your children at school, soccer matches, or any other activity? For what types of activities or events do they typically drop off your children or pick them up? How often do they do this? Is there a neighbor who has dropped off their children with you in an emergency or asked a babysitter to do so if they would be late returning from work? Is there a neighbor you have dropped off your children with in an emergency or asked a babysitter to do so if you would be late returning from work? Is there a neighbor you have listed as an emergency contact for your children at school? Is there a neighbor who has listed you as an emergency contact for their children at school? Is there a neighboring child you have scolded or disciplined? Is there a neighbor who has scolded or disciplined your child? Is there a neighbor you wouldn’t be concerned about scolding or disciplining your child? Do you have any relatives who are neighbors but who don’t live in your household? Are there any other neighbors you know personally?

ABOUT THE NEIGHBORS WHOM THE RESPONDENT KNOWS One by one, for each and every neighbor identified by the previous set of questions, we asked the following questions: How often do you see them, even if you don’t greet them? How often do you greet them? Do you talk to them on a regular basis? What kinds of activities do you do with this neighbor? 204

SURVEY INSTRUMENT

Does their household include any children similar in age to the children in your household? Do you think that neighbor is older than you, younger than you, or about your age? Is this neighbor the same race as you? If not, what race are they? Is this neighbor the same ethnicity as you? If not, what ethnicity are they? What language do you believe they most often speak at home? Does this neighbor share your religious identity or affiliation? If not, what religion, if any, do they identify or affiliate with? Did they live in their residence before you moved into the neighborhood? If you know, how long have they lived in their residence? Did you know this person before you moved into your current residence? If not, how soon after moving in did you meet? How did you meet this neighbor? Would you say that this meeting was basically unintentional and resulted primarily from the mere fact of being neighbors?

ABOUT THE RESPONDENT’S NEIGHBORS GENERALLY Think of all the people you know at least as well as those you have identified. What fraction of these people are your neighbors? What activities, if any, do you engage in with the people you just identified as neighbors which you do not engage in with anyone else? In what way, if any, are the people you just identified unique among the network of people you know? What fraction of your children’s friends are neighbors?

ABOUT THE RESPONDENT’S NEIGHBORHOOD GEOGRAPHY One by one, for each child identified as living in the household, we asked: 205

APPENDIX

Are there any boundaries beyond which this child is not supposed to go? Respondents were then given a blank piece of paper and a pencil and asked to “Draw the streets in your neighborhood and identify on the map you have just created your residence as well as the residences of any neighbors you know.” If they failed to identify the location of one of the neighbors listed, the interviewer reminded them of this person.

206

Notes

CHAPTER ONE NEIGHBORHOODS AND NEIGHBORING 1. Wellman and Wortley 1990; Coleman 1988; Leventhal and Brooks-Gunn 2000; Warner and Rountree 1997; Rountree and Warner 1999; Bellair 1997, 2000. 2. In later chapters I will explore how individual households’ self-interests translate into social capital. 3. Sampson, Raudenbush, and Earls 1997. Sampson et al.’s argument was similar to Putnam’s (1993) discussion of norms and trust. 4. Morenoff, Sampson, and Raudenbush 2001. 5. Shaw and McKay 1969. 6. Brantingham and Brantingham 1993. 7. Massey and Mullen 1984. 8. Massey and Denton 1993. 9. For an overview, see Brooks-Gunn, Duncan, and Aber 1997. 10. Jencks and Mayer 1990. 11. Brooks-Gunn et al. 1993. 12. Catsambis and Beveridge 2001; Ginther, Haveman, and Wolfe 2000. 13. Boardman et al. 2001; Jencks and Meyer 1990. 14. Sampson 1997. 15. Tita, Cohen, and Engberg 2006. 16. Browning et al. 2004. 17. South and Baumer 2001. 18. Ginther, Haveman, and Wolfe 2000; Brooks-Gunn et al. 1993. 19. Pinderhughes et al. 2001. 20. Ainsworth 2002; Blau 1957; Connell and Halpern-Felsher 1997; Davis, Spacth, and Huson 1961; Duncan 1967, 1994; Entwisle, Alexander, and Olson 1994. 21. Bellair 1997, 2000; Elliott et al. 1996; Hirschfield and Bowers 1997; LaGrange 1999; Morenoff, Sampson, and Raudenbush 2001; Peterson, Krivo, and Harris 2000; Rountree and Warner 1999; Sampson 1986; Sampson and Groves 1989; Sampson, Raudenbush, and Earls 1997; Smith, Frazee, and Davison 2000; South and Baumer 2000; Veysey and Messner 1999; Warner and Rountree 1997; Wilson 1987. 207

NOTES TO CHAPTER ONE

22. Sampson and Morenoff 1997. 23. Cohen et al. 2000; Markowitz et al. 2001; Perkins and Taylor 1996; Ross 2000; Rountree and Land 1996; Sampson and Raudenbush 1999; Wilson and Kelling 1982. 24. For black respondents, exposure to whites during formative years was a more important factor in the number of white acquaintances and friends they had than where they lived, worked, or worshipped as adults (Sigelman et al. 1996, 1326). 25. See Sampson, Morenoff, and Gannon-Rowley’s (2002) review of the literature on neighborhood effects. 26. Shaw and McKay 1969. 27. Galster 2001, 2111. 28. Studies that focus on people have clear analytic units. What is a neighborhood, however, is ambiguous. The modifiable areal unit problem refers to the fact that different relationships are found among spatial variables (varying not only in magnitude but sometimes even in direction) depending upon the neighborhood equivalent one uses. 29. Openshaw 1984. 30. Openshaw 1977, 1996; You, Nedovic-Budic, and Kim 1997; Ding 1998; Guo 2000; and Alvanides, Openshaw, and Macgill 2001. 31. Fotheringham and Wong 1991. 32. Sampson, Morenoff, and Gannon-Rowley 2002, 445. 33. A geographic information system (GIS) is a computer program that one can use to store, edit, analyze, and display data that refer to or are linked to geographic location. 34. The precise figure is 51.0 percent; while only 33.2 percent of all households have children living in them, these households are more populous than households in general. 35. The precise figure is 43.1 percent. 36. Lindzey and Byrne 1968; Nahemow and Lawton 1975; Grannis 2003. 37. Appleyard 1981, 4. 38. Appleyard 1981, 71. 39. Brower 1977. 40. Alexander 1972; Michelson and Roberts 1979; Ward 1978; Appleyard 1981; Brower 1977. 41. Fischer 1982; Hillier 1996; Hillier and Hanson 1984; Gehl 1987; Cochran 1994. 42. Grannis 2003. 43. Alexander 1972; Michelson and Roberts 1979; Ward 1978; Appleyard 1981; Brower 1977. 208

NOTES TO CHAPTER TWO

44. Some of this may be lost on academics because neighborhoods are less significant for children in well-to-do families whose outside activities take place in many different locations and whose parents may ferry them to school, music lessons, swimming classes, sporting games, and organized events with specialized facilities (Zeiher 2003). 45. Fischer 1982; Hillier 1996; Hillier and Hanson 1984; Gehl 1987; Cochran 1994. 46. Massey and Mullen 1984. 47. Hunter 1979. 48. Sampson, Morenoff, and Earls 1999. 49. Sampson, Raudenbush, and Earls 1997; Bandura 1997. 50. Coleman 1990; Furstenberg et al. 1999. 51. Sandefur and Laumann 1998. 52. Hunter 1979. 53. Zito 1974; Michelson 1977; Cochran 1994; Abu-Gazzeh 1999; Lindzey and Byrne 1968; Nahemow and Lawton 1975; Grannis 2001. 54. Massey and Mullen 1984; Bogue 1984; Hatt 1946; Jacobs 1961; Park 1984; Perry 1933; Zorbaugh 1929. 55. Massey and Mullen 1984; Hayes and Taylor 1996. 56. Using a hedonic price analysis, Black (1999, 595) found that parents are willing to pay more for better “school peers and other unmeasured components of school quality.” 57. Bogart and Cromwell 1997; Kain and Quigley 1975; Li and Brown 1980. 58. Keep it simple, stupid. 59. I will formally define these geographically based opportunity structures, tertiary face blocks and tertiary intersections, when I discuss them in chapter 3. 60. I will define t-communities and islands in detail when I discuss them in chapter 4.

CHAPTER TWO THE STAGES OF NEIGHBORING 1. This is, of course, unless later processes disturb this arrangement. 2. Neighboring, however, is not the only type of geographically constrained relation (e.g., office-mates, roommates, etc.). 3. Haythornthwaite 2002. 4. Burt 1992; Granovetter 1973. 5. Haythornthwaite 2002. 209

NOTES TO CHAPTER THREE

6. Cochran 1994; Festinger, Schachter, and Back 1950; Fischer 1982; Gehl 1987; Hillier 1996; Hillier and Hanson 1984; Kuper 1953. 7. Jacobs 1961, 56. 8. Michelson 1977; Silverman 1983; Skjaeveland and Garling 1997. 9. American Society of Civil Engineers 1996. 10. Hillier 1996. 11. Kalmijn 1998. 12. McPherson and Ranger-Moore 1991; Blokland 2003. 13. Greenbaum 1982; Greenbaum and Greenbaum 1985. 14. Sampson, Morenoff, and Gannon-Rowley 2002, 446. 15. Putnam 1993. 16. Axelrod 1984. 17. Campbell 1990; DiPasquale and Glaeser 1999. 18. This argument was first outlined by Gans (1961); see also Gutman 1966. 19. La Gory 1982. 20. Greenbaum and Greenbaum 1985; Athanasiou and Yoshioka 1973; Caplow and Forman 1950; Ebbesen, Kjos, and Konecni 1976; Kuper 1953; Priest and Sawyer 1967; Whyte 1957; Deutsch and Collins 1951; Wilner et al. 1955; as well as Nahemow and Lawton 1975. 21. Grannis 2003. 22. Bellair 1997.

CHAPTER THREE RECONCEPTUALIZING STAGE 1 NEIGHBORING 1. In perhaps the first such studies, Festinger, Schachter, and Back (1950) and Caplow and Foreman (1950) examined socializing among the residents of post–World War II married student housing at MIT and found that the most important factor in the actualization of neighborly relations to be the joint proximity of dwellings and that centrally located apartments were more involved in social contacts. 2. Festinger, Schachter, and Back 1950, 39. 3. Whyte 1957; Bossard 1932; Kennedy 1978; Koller 1948; Catton and Smircich 1964, Hanson, Marble, and Pitts 1972; Peach 1981. 4. Grannis 2003. 5. Grannis 2003. 6. Greenbaum and Greenbaum 1985. 7. Park 1984; Zorbaugh 1929. 210

NOTES TO CHAPTER THREE

8. Suttles 1972; Bogue 1984; Hatt 1946. 9. Lynch 1971. 10. Albrecht and Cunningham 1979. 11. Rabin 1987, 217. 12. Anderson 1992, 154. 13. Drake and Cayton 1993, 190. 14. Ryan and McNally 1995; Appleyard and Lintell 1986; Schoenberg and Rosenbaum 1981; Calthorpe 1991. 15. Tienda 1991. 16. Park 1984; Zorbaugh 1929. 17. Suttles 1972; Bogue 1984; Hatt 1946. 18. Tienda 1991. 19. Baldassare 1975a, 1975b; Fowler 1987; Fox, Fox, and Marans 1980; Franck 1983; Kennedy 1978; Lansing and Marans 1969; Popenoe 1977; Skjaeveland and Garling 1997. 20. Suttles 1972. 21. Greenbaum 1982. 22. For example, Appleyard’s (1981) and Appleyard and Lintell’s (1986) studies of San Francisco. 23. Appleyard and Lintell 1986, 98. 24. Appleyard and Lintell 1986, 106. 25. Appleyard 1981, 22. 26. Appleyard and Lintell 1986, 109. 27. The Census Bureau, in conjunction with the United States Geological Survey (USGS), classifies all roads for the Topologically Integrated Geographic Encoding and Referencing (TIGER) digital line graph (DLG) files. 28. Suttles 1972, 56–57. 29. Rabin 1987. 30. de Jong 1986. 31. Appleyard 1981, 71. 32. Greenbaum 1982. 33. Appleyard and Lintell 1986, 109. 34. Wellman 1979, 1214. 35. Storm 1985. 36. One could logically argue that one would want to distinguish a nontertiary street intersecting a tertiary street from two nontertiary streets intersection, but I will completely ignore the latter case, as I would argue it is of no sociological relevance to the emergence of neighboring. 37. American Society of Civil Engineers 1996, 20. 211

NOTES TO CHAPTER FIVE

38. Anderson 1992, 46. 39. Lynch 1971. 40. Hillier 1996.

CHAPTER FOUR RECONCEPTUALIZING STAGE 1 NEIGHBOR NETWORKS 1. Grannis 2005. 2. The label t-community is short for tertiary street community, indicating a neighborhood equivalent that is connected by tertiary streets (Grannis 1998). 3. Why the name islands? A pair of neighborhoods would be geographically unreachable if they could only reach each other by crossing water; therefore, I analogously label these discontinuous networks of tertiary face blocks as islands since it is impossible for households in two different islands to access each other using tertiary face blocks, even if they use nontertiary intersections. 4. Grannis 2001, 2003, 2005. 5. Berry and Kasarda 1977; Hawley 1950. 6. Ryan and McNally 1995.

CHAPTER FIVE SELECTION AND INFLUENCE 1. Duncan, Connell, and Klebanov 1997. 2. Galster 2001. 3. Park 1916, 147–54. 4. Nechyba and Strauss 1998; Hunt, McMillan, and Abraham 1994. 5. Ben-Akiva and Bowman 1998; Nechyba and Strauss 1998. 6. Abraham and Hunt 1997; Freedman and Kern 1997; Waddell 1992, 1993; Hunt, McMillan, and Abraham 1994. 7. “[I]n short-run, modal choice decisions, individuals’ income may influence their time/money tradeoffs, but when it comes to longer-term choices such as where to live, all people find themselves within the same twenty-fourhour day. Cycles of work, leisure, and sleep may then become more important than opportunities to save some time by spending some money, or vice versa” (Levine 1998, 148). 8. See, for example, Burgess 1984; Hoyt 1939; Harris and Ullman 1945; Berry and Kasarda 1977; and Flanagan 1993. 9. Waddell 1992; Sermons and Koppelman 1998. 212

NOTES TO CHAPTER SIX

10. Berry and Kasarda 1977, 90. 11. Brantingham and Brantingham 1993; Quillian and Pager 2002. 12. Massey and Mullen 1984. 13. Massey and Denton 1993. 14. Farley and Frey 1994; Farley et al. 1994. 15. Sampson 2001, 9. 16. Coleman 1990. 17. Hogan and Kitagawa 1985. 18. Connell et al. 1995; Sewell and Armer 1966; Sewell, Haller, and Portes 1969; South and Baumer 2000. 19. Yabiku 2004. 20. Successful role models establish normative environments that emphasize personal success and mainstream behavior (Brooks-Gunn 1993; Halpern-Flesher et al. 1997). The discussion in this paragraph draws heavily upon Jencks and Mayer 1990; Montgomery and Casterline 1996; and Crane 1991. 21. Coleman 1988; Friedkin 1998. 22. See Burt 1987 for a review. 23. Bearman 1997; see also Alba 1973; Fershtman 1997; Frank 1995; Richards 1995. 24. Brooks-Gunn et al. 1993; Bursik and Grasonick 1993; Jencks and Mayer 1990; Sampson and Groves 1989; Sampson and Morenoff 1997; Sampson, Raudenbush, and Earls 1997.

CHAPTER SIX RESPONDENTS, INTERVIEWS, AND OTHER DATA 1. A snowball sample is one in which existing study subjects recruit future subjects; thus the sample group appears to grow like a rolling snowball. This sampling technique is often used in hidden populations, such as gangs, which are difficult for researchers to access. 2. See, respectively, Casciaro 1998; Calloway, Morrissey, and Paulson 1993; and Brewer and Yang 1994 3. Freeman and Romney 1987; Freeman, Romney, and Freeman 1987. 4. Romney and Faust 1982. 5. For discussions of the utility of cognitive mapping exercises, see Appleyard 1981; Appleyard and Lintell 1986; Lloyd and Hooper 1991; and Lee and Campbell 1997. 213

NOTES TO CHAPTER EIGHT

6. While this phrase is no longer widely used among juveniles in the neighborhood, it has been replaced by similarly pithy expressions. 7. I identified t-communities using a program I wrote to run in ESRI’s ArcView geographic information system in conjunction with street data from the 1990 Census Topologically Integrated Geographic Encoding and Referencing (TIGER) digital line graph (DLG) files. I used data distributed by Geographic Data Technology. 8. For a discussion of the small-world effect, see Milgram 1969; Killworth and Bernard 1979; and Watts and Strogatz 1998. 9. Elsewhere (Grannis 1998, 2003, 2005), I have explored those subjects more extensively with nationally representative samples. CHAPTER SEVEN SELECTING STAGE 1 NEIGHBORS 1. Given that we did not place a time limit on how long ago one had moved, everyone reported having moved into their current household at some point. 2. Informally, a path is a sequence of nodes connected by edges. The geodesic distance between two nodes is the length of the shortest path between them, where distance is measured by the number of lines traversed. The characteristic path length of a network is the average geodesic distance between all pairs of nodes in the network. Note that the network must be connected for the characteristic path length to be defined. CHAPTER EIGHT UNINTENTIONAL ENCOUNTERS 1. I will discuss apartment complexes in more detail subsequently. 2. Festinger, Schachter, and Back (1950) and Caplow and Foreman (1950) found that within apartment buildings the specific arrangement of walkways, staircases, and other points of contact between residents had an important effect. For example, apartments near stairways were more often involved in social contacts. CHAPTER NINE STAGE 3 NEIGHBORS AND TERTIARY STREETS 1. Park 1984, 21. 2. Tobler 1970. 214

NOTES TO CHAPTER ELEVEN

3. This number is less than 998 because 19 individuals identified no one whom they knew personally. 4. Results from the Los Angeles data collections, while more complicated to report, differ only in more powerfully verifying my argument for stage 3 neighboring relations being constrained by tertiary street networks.

CHAPTER TEN THE IMPORTANCE OF NEIGHBOR NETWORKS 1. Scott Feld (1991) derived a similar result in his article “Why Your Friends Have More Friends Than You Do.” He showed that, in general, the number of second-order relations was ( xi2) / ( xi). Thus, in this case, we would determine the average number of second neighbors to be 10. Our example differs because the actors with various degrees are distributed nonrandomly.

CHAPTER ELEVEN NETWORK INFLUENCE THEORY 1. Friedkin 1991, 1998, 2003; Friedkin and Johnsen 1990, 1997, 1999. 2. Rountree and Warner 1999; Elliott et al. 1996; Veysey and Messner 1999; Morenoff, Sampson, and Raudenbush 2001. 3. Bellair 1997. 4. Sampson’s (2004) paraphrase of Putnam’s model of the relationship between social capital theory and social disorganization theory. 5. Sixteen of the 72 interneighbor influences directly involve the central neighbor, eight as a recipient of influence and eight as a giver of influence. The other 50 interneighbor influences involve situations where the central neighbor is necessarily an intermediary of information about how residents are behaving and how neighborhood residents are reacting to those behaviors, and thus the norms, values, and expectations flowing between these neighbors are necessarily transmitted through and interpreted by the central neighbor. 6. The literature on neighborhood organizing, for example, explicitly conceptualizes a neighborhood as a network of networks (Cohen 1979; Warren and Warren 1977), neighborhoods being composed of block-level networks linked horizontally to adjoining block-level networks. 7. Friedkin 1983. 8. If Figure 11.1 represented academics discussing their research, then in figure 11.1a, everyone would be within the horizon of observability of every215

N O T E S T O C H A P T E R T W E LV E

one else. In figure 11.1b, no one in this network has the entire network within their horizon the horizon of observability, and, for the two neighbors at the distant ends of each triangle, the majority of the network is outside their horizon of observability. In figure 11.1c, where there are actually two distinct groups, within each group everyone is within everyone else’s horizon, and between groups, of course, no one is. In figure 11.1d, the average neighbor is within the horizon of most, but not all, of the other residents. Again, all figures had perfectly equal density but dramatically different potentials for social control based on their ability to observe and influence each other. 9. Specifically, 321 out of 1,324 (24.2 percent) for six or more two-step paths and 244 out of 848 (28.8 percent) for seven or more two-step paths. 10. Granovetter 1992, 35. 11. Coleman 1988; Friedkin 1998. 12. See Burt 1987 for a review. 13. Bearman 1997 . 14. Two paths from i to j are node-independent if they have only nodes i and j in common. Thus, paths 1 and 2 between nodes i and j are node-independent if none of the intermediaries on path 1 are also intermediaries on path 2. 15. Menger proved that a graph in which k is the minimum number of nodes whose removal would disconnect the graph also has at least k nodeindependent paths connecting every pair of nodes, and vice versa (see Harary 1969 for Menger’s proof).While the fourth network in the example has a kconnectivity of three, the first and second networks had k-connectivities of one, and the third network had a k-connectivity of zero (it is already disconnected). These values, however, are for the entire network, and, in the first three networks; certain subgroups had higher k-connectivities. For example, in the first network, while most of the network depends upon the central neighbor, the three triangles all have internal k-connectivities of two; similarly, in the second network, the triangles have a k-connectivity of two. In the third network, the subgroup on the left has a k-connectivity of two, and the subgroup on the right has a k-connectivity of three. 16. Dawkins 1976.

CHAPTER TWELVE INFLUENCE NETWORKS IN A COLLEGE TOWN 1. A connected component is a maximal subset of vertices with paths connecting them. In this example, it is the largest set of households such that each household in the set is connected to every other household by a sequence of 216

NOTES TO CHAPTER FOURTEEN

child-monitoring relations (i.e., they monitored the child of a household who monitored the child of a household, etc.). 2. Again, this involved a significant minority, more than a few but not a majority, of households watching over large numbers of others’ children. 3. The remainder split almost perfectly between a “Yes” becoming a “No” and a “No” becoming a “Yes.”

CHAPTER THIRTEEN INFLUENCE NETWORKS IN A GANG BARRIO 1. At the request of several respondents, I have chosen not to identify the gang by name. 2. This phrase was repeated by everyone in the gang unit on numerous occasions. leading me to believe it was the “official” description. 3. Such interactions between local gangs and their neighborhood communities have been extensively studied by Moore, Sullivan, Sanchez-Jankowski, Patillo-McCoy, Vankatesh, and others. I do not attempt to contribute to the wealth of literature that exists on gang communities, only to establish that, in this context, they provide an excellent example of stage 4 neighborly relations. 4. Puto is clearly a term of derision or contempt. According to Wikipedia’s article on Spanish profanity (as of November 12, 2006) puto is “probably the strongest profanity . . . it is highly offensive.” 5. While this particular phrase is no longer widely used among juveniles in the neighborhood, it has been replaced by similarly pithy expressions.

CHAPTER FOURTEEN IMPLICATIONS 1. Mann 1986, 1. 2. Grannis 2005.

217

References

Abraham, J. E., and J. D. Hunt. 1997. Specification and estimation of a nested logit model of home, workplace and commuter mode choice by multiple worker households. Transportation Research Record 1606:17–24. Abu-Gazzeh, Tawfiq M. 1999. Housing layout, social interaction, and the place of contact in Abu-Nuseir, Jordan. Journal of Environmental Psychology 19:41–73. Ainsworth, J. W. 2002. Why does it take a village? The mediation of neighborhood effects on educational achievement. Social Forces 81:117–52. Alba, Richard D. 1973. A graph-theoretic definition of a sociometric clique. Journal of Mathematical Sociology 3:113–26. Albrecht, Roger S., and James V. Cunningham. 1979. A New Public Policy for Neighborhood Preservation. New York: Praeger. Alexander, Christopher. 1972. The city as a mechanism for sustaining human contact. Pp. 406–34 in People and Buildings, ed. Robert Gutman. New York: Basic Books. Almgren, G., A. Guest, G. Imerwahr, and M. Spittel. 1998. Joblessness, family disruption, and violent death in Chicago, 1970–1990. Social Forces 76:1465–93. Alvanides, A., S. Openshaw, and J. Macgill. 2001. Zone design as a spatial analysis tool. Pp. 141–57 in Modelling Scale in Geographical Information Science, ed. Nicholas J. Tate and Peter M. Atkinson. Chicester, UK: John Wiley and Sons. American Society of Civil Engineers, National Association of Homebuilders of the United States, and the Urban Land Institute. 1996. Residential Streets. 2nd ed. Anderson, Elijah. 1992. Streetwise: Race, Class, and Change in an Urban Community. Chicago: University of Chicago Press. Appleyard, Donald. 1981. Livable Streets. Berkeley and Los Angeles: University of California Press. Appleyard, Donald, and Mark Lintell. 1986. The environmental quality of city streets: The residents’ viewpoint. Pp. 93–120 in Transport Sociology: Social Aspects of Transport Planning, ed. Enne de Boer. Oxford: Pergamon. Athanasiou, R., and G. A. Yoshioka. 1973. The spatial character of friendship formation. Environment and Behavior 5:43–65. 219

REFERENCES

Axelrod, Robert. 1984. The Evolution of Cooperation. New York: Basic Books. Baldassare, Mark. 1975a. The effects of density on social behavior and attitudes. American Behavioral Scientist 18:815–25. ———. 1975b. Residential density, local ties, and neighborhood attitudes. Sociological Symposium 14:93–102. Bandura, A. 1997. Self-Efficacy: The Exercise of Control. New York: Freeman. Bearman, Peter. 1997. Generalized exchange. American Journal of Sociology 102:1383–1415. Bell, W., and M. Boat. 1957. Urban neighborhoods and informal social relations. American Journal of Sociology 62:391–98. Bellair, P. E. 1997. Social interaction and community crime: Examining the importance of neighbor networks. Criminology 35:677–703. ———. 2000. Informal surveillance and street crime: A complex relationship. Criminology 38:137–67. Ben-Akiva, M., and J. L. Bowman. 1998. Integration of an activity-based model system and a residential location model. Urban Studies 35:1131–53. Berry, Brian J., and John D. Kasarda. 1977. Contemporary Urban Ecology. New York: Macmillan. Black, Sandra E. 1999. Do better schools matter? Parental valuation of elementary education. Quarterly Journal of Economics 114:577–99. Blau, Peter M. 1957. Formal organizations: Dimensions of analysis. American Journal of Sociology 63:58–69. Blockland, T. 2003. Urban Bonds. Cambridge: Polity. Boardman, J. D., B. K. Finch, and C. G. Ellison. 2001. Neighborhood disadvantage, stress, and drug use among adults. Journal of Health and Social Behavior 42:151–65. Bogart, William T., and Brian A. Cromwell. 1997. How much more is a good school district worth? National Tax Journal 50:215–32. Bogue, David J. 1984. Ecological community areas. Pps. 1–25 in Essays in Human Ecology II, ed. Donald J. Bogue and Michael J. White. Chicago: Community and Family Study Center. Bossard, James H. S. 1932. Residential propinquity as a factor in marriage selection. American Journal of Sociology 38:219–24. Brantingham, Patricia L., and Paul J. Brantingham. 1993. Nodes, paths, and edges: Considerations on the complexity of crime and the physical environment. Journal of Environmental Psychology 13:3–28. Brewer, Devon D., and Bihchii Laura Yang. 1994. Patterns in the recall of persons in a religious community. Social Networks 16:347–79. Brooks-Gunn, J., G. J. Duncan, and J. L. Aber, eds. 1997. Neighborhood Poverty. 2 vols. New York: Russell Sage Foundation. 220

REFERENCES

Brooks-Gunn, J., G. J. Duncan, P. K. Klebanov, and N. Sealand. 1993. Do neighborhoods influence child and adolescent development? American Journal of Sociology 99:353–95. Brower, Sidney N. 1977a. The Design of Neighborhood Parks. Baltimore: Department of Planning. ———. 1977b. Streetfronts and backyards: Two ways of looking at neighborhood open spaces. Baltimore: Department of Planning. Browning, Christopher R., Tama Leventhal, and Jeanne Brooks-Gunn. 2004. Neighborhood context and racial differences in early sexual activity. Demography 41:697–720. Burgess, Ernest W. 1984. The growth of a city: An introduction to a research project. Pp. 47–62 in The City, ed. Robert E. Park, Ernest W. Burgess, and Roderick McKenzie. The Heritage of Sociology. Chicago: University of Chicago Press. Bursik, R. J., and H. G. Grasonick. 1993. Neighborhoods and Crime: The Dimensions of Social Control. New York: Lexington. Burt, Ronald S. 1987. Social contagion and innovation: Cohesion versus structural equivalence. American Journal of Sociology 92:1287–1335. ———. 1992. Structural Holes: The Social Structure of Competition. Cambridge: Harvard University Press. Calloway, M., J. P. Morrissey, and R. I. Paulson. 1993. Accuracy and reliability of self-reported data in interorganizational networks. Social Networks 15:377–98. Calthorpe, P. 1991. The post-suburban metropolis. Whole Earth Review 78: 44–51. Campbell, K. E. 1990. Network past: A 1939 Bloomington neighborhood. Social Forces 69:139–55. Caplow, T., and R. Forman. 1950. Neighborhood interaction in a homogenous community. American Sociological Review 15:357–66. Casciaro, Tiziana. 1998. Seeing things clearly: Social structure, personality, and accuracy in social network perception. Social Networks 20:331–51. Catsambis, S., and A. A. Beveridge. 2001. Does neighborhood matter? Family, neighborhood, and school influences on eighth grade mathematics achievement. Sociological Focus 34:435–57. Catton, William R., and R. J. Smircich. 1964. A comparison of mathematical models for the effects of residential propinquity on mate selection. American Sociological Review 29:522–29. Cochran, S. 1994. Understanding and enhancing neighbourhood sense of community. Journal of Planning Literature 9(1): 92–99. 221

REFERENCES

Cohen, Albert K. 1955. Delinquent Boys: The Culture of the Gang. New York: Free Press. Cohen, D., S. Spear, R. Scribner, P. Kissinger, K. Mason, and J. Widgen. 2000. “Broken windows” and the risk of gonorrhea. American Journal of Public Health 90:230–36. Cohen, Rick. 1979. Neighborhood planning and political capacity. Urban Affairs Quarterly 14(3): 337–62. Coleman, James S. 1988. Social capital in the creation of human capital. American Journal of Sociology 94 (supp.): S95–S120. ———. 1990. Foundations of Social Theory. Cambridge: Harvard University Press. Connell, J. P., and B. L. Halpern-Felsher. 1997. How neighborhoods affect educational outcomes in middle childhood and adolescence: Conceptual issues and an empirical example. Pp. 174–99 in Neighborhood Poverty, vol. 1, Context and Consequences for Children, ed. J. Brooks-Gunn, G. J. Duncan, and J. L. Aber. New York: Russell Sage Foundation. Connell, J. P., B. Halpern-Felsher, E. Clifford, W. Crichlow, and P. Usinger. 1995. Hanging in there: Behavioral, psychological, and contextual factors affecting whether African American adolescents stay in high school. Journal of Adolescent Research 10:41–63. Crane, J. 1991. The epidemic of ghettos and neighborhood effects on dropping out and teenage childbearing. American Journal of Sociology 96:1226–59. Craven, Paul, and Barry Wellman. 1973. The network city. Sociological Inquiry 43:57–88. Davis, J. A., J.L.B. Spaeth, and C. Huson. 1961. A technique for analyzing the effects of group composition. American Sociological Review 26:215–25. Dawkins, Richard. 1976. The Selfish Gene. Oxford: Oxford University Press. de Jong, Rudolf. 1986. The recapture of the street? Pp. 77–91 in Transport Sociology: Social Aspects of Transport Planning, ed. Enne de Boer. Oxford: Pergamon. Dennis, N. 1968. The popularity of the neighborhood community idea. Pp. 74–92 in Readings in Urban Sociology, ed. R. E. Pahl. London: Pergamon. Deutsch, Morton, and M. E. Collins. 1951. Interactial Housing: A Psychological Evaluation of a Social Experiment. New York: Russell and Russell. Ding, C. 1998. The GIS-based human-interactive TAZ design algorithm: Examining the impacts of data aggregation on transportation-planning analysis. Environment and Planning B 25:601–16. DiPasquale, Denise, and Edward L. Glaeser. 1999. Incentives and social capital: Are homeowners better citizens? Journal of Urban Economics 45:354–84. 222

REFERENCES

Drake, St. Clair, and Horace R. Cayton. 1993. Black Metropolis: A Study of Negro Life in a Northern City. Chicago: University of Chicago Press. Dubin, R. A. 1992. Spatial autocorrelation and neighborhood quality. Regional Science and Urban Economics 22:433–52. Duncan, G. J. 1994. Families and neighbors as sources of disadvantage in the schooling decisions of white and black adolescents. American Journal of Education 103:20–53. Duncan, G. J., J. P. Connell, and P. K. Klebanov. 1997. Conceptual and methodological issues in estimating causal effects of neighborhoods and family conditions on individual development. Pp. 219–50 in Neighborhood Poverty, vol. 1, Context and Consequences for Children, ed. J. Brooks-Gunn, G. J. Duncan, and J. L. Aber. New York: Russell Sage Foundation. Ebbesen, E. B., G. L. Kjos, and V. Konecni. 1976. Spatial ecology: Its effects on the choice of friends and enemies. Journal of Experimental Social Psychology 12:505–18. Elliott, D. S., W. J. Wilson, D. Huizinga, R. J. Sampson, A. Elliott, and B. Rankin. 1996. The effects of neighborhood disadvantage on adolescent development. Journal of Research in Crime and Delinquency 33:389–426. Entwisle D. R., K. L. Alexander, and L. S. Olson. 1994. The gender gap in math: Its possible origins in neighborhood effects. American Sociological Review 59:822–38. Farley, Reynolds, and William H. Frey. 1994. Changes in the segregation of whites from blacks during the 1980s: Small steps toward a more integrated society. American Sociological Review 59:23–45. Farley, Reynolds, Charlotte Steeh, Tara Jackson, Maria Krysan, and Keith Reeves. 1994. The causes of continued racial residential segregation: Chocolate city, vanilla suburbs revisited. Journal of Housing Research 4:1–38. Feld, Scott L. 1991. Why your friends have more friends than you do. American Journal of Sociology 96:1464–77. Fellin, P., and E. Litwak. 1963. Neighborhood cohesion under conditions of mobility. American Sociological Review 28:364–76. Fershtman, Meir. 1997. Cohesive group detection in a social network by the segregation matrix index. Social Networks 19:193–207. Festinger, L., S. Schachter, and K. Back. 1950. Social Pressures in Informal Groups: A Study of Human Factors in Housing. New York: Harper and Row. Fischer, Claude. 1976. The Urban Experience. New York: Harcourt Brace Jovanovich. ———. 1982. To Dwell among Friends: Personal Networks in Town and City. Chicago: University of Chicago Press. 223

REFERENCES

Flanagan, William G. 1993. Contemporary Urban Sociology. Cambridge: Cambridge University. Fotheringham, A. S., and D.W.S. Wong. 1991. The modifiable areal unit problem in multivariate statistical analysis. Environment and Planning A 23:1025–44. Fowler, E. P. 1987. Street management and city design. Social Forces 66: 365–69. Fox, Bonnie J., John Fox, and Robert Marans. 1980. Residential density and neighbor interaction. Sociological Quarterly 21:349–59. Franck, Karen A. 1983. Community by design. Sociological Inquiry 53:289– 313. Frank, Kenneth A. 1995. Identifying cohesive subgroups. Social Networks 17:27–56. Freedman, Ora, and Clifford R. Kern. 1997. A model of workplace and residence choice in two-worker households. Regional Science and Urban Economics 27:241–60. Freeman, Linton C., and A. Kimball Romney. 1987. Words, deeds and social structure. Human Organization 46:330–34. Freeman, Linton C., A. Kimball Romney, and Sue C. Freeman. 1987. Cognitive structure and informant accuracy. American Anthropologist 89:310–25. Friedkin, Noah. 1983. Horizons of observability and limits of informal control in organizations. Social Forces 62:54–77. ———. 1991. Theoretical foundations for centrality measures. American Journal of Sociology 96:1478–1504. ———. 1998. A Structural Theory of Social Influence. Cambridge: Cambridge University Press. ———. 1999. Choice shift and group polarization. American Sociological Review 64:856–75. ———. 2003. Social influence network theory: Towards a strategic modification of interpersonal influence systems. Pp. 89–100 in Dynamic Social Network Modeling and Analysis, ed. Ronald Breiger, Kathleen Carley, and Philippa Pattison. Washington, DC: National Academies Press. Friedkin, N. E., and E. C. Johnsen. 1990. Social influence and opinions. Journal of Mathematical Sociology 15:193–205. ———. 1997. Social positions in influence networks. Social Networks 19:209–22. ———. 1999. Social influence networks and opinion change. Advances in Group Processes 16:1–29. 224

REFERENCES

Furstenberg, F. F., Jr., T. D. Cook, J. Eccles, G. H. Elder, Jr., and A. Sameroff. 1999. Managing to Make It: Urban Families and Adolescent Success. Chicago: University of Chicago Press. Galster, G. C. 2001. On the nature of neighborhood. Urban Studies 38: 2111–24. Gans, Herbert. 1961. Planning and social life: An evaluation of friendship and neighbor relations in suburban communities. Journal of the American Institute of Planners 27(2): 134–40. ———. 1967. Levittowners: Ways of Life and Politics in a New Suburban Community. New York: Pantheon. Gehl, J. 1987. Life between Buildings: Using Public Space. New York: Van Nostrand Reinhold. Ginther, Donna, Robert Haveman, and Barbara Wolfe. 2000. Neighborhood attributes as determinants of children’s outcomes: How robust are the relationships? Journal of Human Resources 35:603–42. Goldstein, Harvey. 1995. Multilevel Statistical Models. 2nd ed. London: E. Arnold; New York: Halstead Press. Goldstein, Jeffrey. 1999. Emergence as a construct: History and issues. Emergence: Complexity and Organization 1:49–72. Grannis, Rick. 1998. The importance of trivial streets: Residential streets and residential segregation. American Journal of Sociology 103:1530–64. ———. 2001. From neighbors to neighborhoods: Social networks and street networks. Paper presented at the Annual Meetings of the American Sociological Association, Anaheim, CA. ———. 2003. Islands in the city: Social networks and street networks. Working paper, Department of Sociology, University of California, Los Angeles. ———. 2005. T-Communities: Pedestrian street networks and residential segregation in Chicago, Los Angeles, and New York. City and Community 4:295–321. Granovetter, Mark S. 1973. The strength of weak ties. American Journal of Sociology 78:1360–80. ———. 1985. Economic action and social structure: The problem of embeddedness. American Journal of Sociology 91:481–510. ———. 1992. Economic action and social structure: The problem of embeddedness. Pp. 304–33 in Decision Making, ed. M. Zey. Newbury Park, CA: Sage. Greenbaum, P. E., and S. D. Greenbaum. 1985. The ecology of social networks in four neighborhoods. Social Networks 7:47–76. 225

REFERENCES

Greenbaum, Susan. 1982. Bridging ties at the neighborhood level. Social Networks 4:367–84. Guest, A. M., and B. A. Lee. 1984. How urbanites define their neighborhoods. Population and Environment 7(1): 32–56. Guo, J. Y. 2000. A graph partitioning approach to transport analysis zone design in a geographical information system environment. Master’s thesis, Transport Research Center, Royal Melbourne Institute of Technology. Guo, Jessica. 2004. Addressing spatial complexities in residential location choice models. Ph.D. diss., University of Texas at Austin. Gutman, Robert. 1966. Site planning and social behavior. Journal of Social Issues 22:103–15. Halpern-Flesher, B. L., J. P. Connell, M. B. Spencer, J. L. Aber, G. J. Duncan, E. Clifford, W. E. Crichlow, P. A. Usinger, S. P. Cole, L. Allen, and E. Seidman. 1997. Neighborhood and family factors predicting educational risk and attainment in African American and white children and adolescents. Pp. 146–73 in Neighborhood Poverty, vol. 1, Context and Consequences for Children, ed. J. Brooks-Gunn, G. J. Duncan, and J. L. Aber. New York: Russell Sage Foundation. Hanson, P. O., D. F. Marble, and F. Pitts. 1972. Individual movement and communication fields. Regional Science Perspectives 2:80–94. Harary, Frank. 1969. Graph Theory. Reading, MA: Addison-Wesley. Harris, Chauncy D., and Edward L. Ullman. 1945. The nature of cities. Annals of the American Academy of Political and Social Sciences 242:7–17. Hatt, Paul. 1946. The concept of natural area. American Sociological Review 11:423–27. Hawley, Amos H. 1950. Human Ecology: A Theory of Community Structure. New York: Ronald. Hayek, F. A. von. 1942. Scientism and the Study of Society. Economica 9: 267–91. Hayes, Kathy, and Lori Taylor. 1996. Neighborhood school characteristics: What signals quality to homebuyers? Federal Reserve Bank of Dallas. Economic Review. Fourth Quarter: 2–9. Haythornthwaite, Caroline. 2002. Strong, weak, and latent ties and the impact of new media. Information Society 18:385–401. Hillier, Bill. 1996. Space Is the Machine. Cambridge: Cambridge University Press. Hillier, Bill, and J. Hanson. 1984. The Social Logic of Space. Cambridge: Cambridge University Press. Hirschfield, A., and K. J. Bowers. 1997. The effect of social cohesion on levels of recorded crime in disadvantaged areas. Urban Studies 34:1275–95. 226

REFERENCES

Hogan, D. P., and E. M. Kitagawa. 1985. The impact of social status, family structure, and neighborhood on the fertility of black adolescents. American Journal of Sociology 90:825–52. Horton, F. E., and D. R. Reynolds. 1971. Effects of urban spatial structure on individual behavior. Economic Geography 47(1): 36–48. Hoyt, Homer. 1939. The Structure and Growth of Residential Neighborhoods in American Cities. Washington, DC: Federal Housing Administration. Hunt, J. D., J.D.P. McMillan, and J. E. Abraham. 1994. Stated preference investigation of influences on attractiveness of residential locations. Transportation Research Record 1466:79–87. Hunter, Albert. 1979. The urban neighborhood: Its analytical and social contexts. Urban Affairs Quarterly 14:267–88. Irving, H. W. 1979. Distance, intensity, kinship: Key dimensions in social interaction. Sociology and Social Research 60(1): 77–86. Jacobs, Jane. 1961. The Death and Life of Great American Cities. New York: Random House. Jencks, Christopher, and Susan E. Mayer. 1990. The social consequences of growing up in a poor neighborhood. Pp. 111–86 In Inner-City Poverty in the United States, ed. Laurence E. Lynn Jr. and Michael G. H. McGeary. Washington, DC: National Academy. Kain, John F., and John M. Quigley. 1975. Housing Markets and Racial Discrimination: A Microeconomic Analysis. New York: National Bureau of Economic Research. Kalmijn, Matthijs. 1998. Intermarriage and homogamy: Causes, patterns, trends. Annual Review of Sociology 24:395–421. Keller, Suzanne. 1968. The Urban Neighborhood: A Sociological Perspective. New York: Random House. Kennedy, Leslie. 1978. Environmental opportunity and social contact: A true or spurious relationship? Pacific Sociological Review 21:173–86. Killworth, P. D., and H. R. Bernard. 1979. The reversal small-world experiment. Social Networks 1:159–92. Kirby, J., and T. Kaneda. 2005. Neighborhood socioeconomic disadvantage and access to health care. Journal of Health and Social Behavior 46:15–31. Knoke, David, and James H. Kuklinski. 1982. Network Analysis. Beverly Hills, CA: Sage. Koller, Marvin R. 1948. Residential propinquity of white males at marriage in relation to age and occupation of males, Columbus, Ohio, 1938 and 1946. American Sociological Review 13:613–16. Kuper, Leo, ed. 1953. Living in Towns. London: Cresset Press. 227

REFERENCES

La Gory, Mark. 1982. Toward a sociology of space: The constrained choice model. Symbolic Interaction 5(1): 65–78. LaGrange, T. C. 1999. The impact of neighborhoods, schools, and malls on the spatial distribution of property damage. Journal of Research in Crime and Delinquency 36:393–422. Lansing, John, and Robert Marans. 1969. Evaluation of neighborhood quality. Journal of the American Institute of Planners 35(3): 195–99. Laumann, Edward O. 1979. Network analysis in some large social systems: Some theoretical and methodological problems. Pp. 379–401 in Perspectives on Social Network Research, ed. Paul W. Holland and Samuel Leinhard. New York: Academic Press. LeClere, Felicia B., Richard G. Rogers, and Kimberley Peters. 1998. Neighborhood social context and racial differences in women’s heart disease mortality. Journal of Health and Social Behavior 39:91–107. Lee, Barrett A., and Karen E. Campbell. 1997. Common ground? Urban neighborhoods as survey respondents see them. Social Science Quarterly 78:922–36. Lee, Barrett A., Karen E. Campbell, and O. Miller. 1991. Racial differences in urban neighboring. Sociological Forum 6:525–50. Leventhal, T., and J. Brooks-Gunn. 2000. The neighborhoods they live in: The effects of neighborhood residence on child and adolescent outcomes. Psychological Bulletin 126(2): 309–37. Levine, J. 1998. Rethinking accessibility and jobs-housing balance. Journal of American Planning Association 64(2): 133–49. Li, Mingche M., and H. James Brown. 1980. Micro-neighborhood externalities and hedonic housing prices. Land Economics 56:125–41. Lindzey, G., and D. Byrne. 1968. Measurement of social choice and interpersonal attractiveness. Pps. 452–525 in Handbook of Social Psychology, ed. G. Lindzey and E. Aronson. Reading, MA: Addison-Wesley. Lloyd, Robert, and Hartwell Hooper. 1991. Urban cognitive maps: Computation and structure. Professional Geographer 43:15–28. Lynch, Kevin. 1971. The Image of the City. Cambridge: MIT Press. Mann, Michael. 1986. The Sources of Social Power. Vol. 1, A History of Power from the Beginning to AD 1760. Cambridge University Press. Markowitz, F. E., P. E. Bellair, A. E. Liska, and J. Liu. 2001. Extending social disorganization theory: Modeling the relationships between cohesion, disorder, and fear. Criminology 39:293–319. Massey, Douglas S., and Nancy A. Denton. 1993. American Apartheid: Segregation and the Making of the Underclass. Cambridge: Harvard University Press. 228

REFERENCES

Massey, Douglas S., and Brendan Mullen. 1984. Processes of Hispanic and black spatial assimilation. American Journal of Sociology 89:836–73. Mayer, Susan E., and Christopher C. Jencks. 1989. Growing up in poor neighborhoods: How much does it matter? Science 243:1441–46. McGahan, Peter. 1972. The neighbor role and neighboring in a highly urban area. Sociological Quarterly 13:397–408. McPherson, J. Miller, and James Ranger-Moore. 1991. Evolution on a dancing landscape: Organizations and networks in dynamic Blau space. Social Forces 70:19–42. Michelson, William H. 1970. Man and His Urban Environment: A Sociological Approach. Reading, MA: Addison-Wesley. ———. 1977. Environmental Choice, Human Behavior, and Residential Satisfaction. Oxford: Oxford University Press. Michelson, William H., and Ellis Roberts. 1979. Children and the urban physical environment. Pp. 410–77 in The Child in the City: Changes and Challenges, ed. William Michelson et al. Toronto: University of Toronto Press. Miles-Doan, Rebecca. 1998. Violence between spouses and intimates: Does neighborhood context matter? Social Forces 77:623–45. Milgram, Stanley. 1969. The small world problem. Psychology Today 22:61–67. Miller, H. 1999. Potential contributions of spatial analysis to geographic information systems for transportation (GIS-T). Geographical Analysis 31:373–99. Montgomery, Mark R., and John B. Casterline. 1996. Social learning, social influence, and new models of fertility. Population and Development Review 22:151–75. Moody, James, and Douglas R. White. 2003. Structural cohesion and embeddedness: A hierarchical concept of social groups. American Sociological Review 68:103–27. Morenoff, Jeffrey D., Robert J. Sampson, and Stephen W. Raudenbush. 2001. Neighborhood inequality, collective efficacy, and the spatial dynamics of urban violence. Criminology 39:517–60. Nahemow, L., and M. P. Lawton. 1975. Similarity and propinquity in friendship formation. Journal of Personality and Social Psychology 32(2): 205–13. Nechyba, T. J., and R. P. Strauss. 1998. Community choice and local public services: A discrete choice approach. Regional Science and Urban Economics 28:51–73. O’Campo, P. 2003. Invited commentary: Advancing theory and methods for multilevel models of residential neighborhoods and health. American Journal of Epidemiology 157:9–13. 229

REFERENCES

Openshaw, Stan. 1977. Optimal zoning systems for spatial interaction models. Environment and Planning A 9:169–84. ———. 1984. The Modifiable Areal Unit Problem. Norwick: Geo Books. ———. 1996. Developing GIS-relevant zone-based spatial analysis methods. Pp. 55–73 in Spatial Analysis: Modelling in a GIS Environment, ed. P. Longley and M. Batty. Cambridge: Geoinformation International. Park, Robert E. 1916. The city: Suggestions for the investigations of human behavior in the urban environment. American Journal of Sociology 20: 577–612. ———. 1984. The urban community as a spatial pattern and a moral order. Pp. 21–31 in The City, ed. Robert E. Park, Ernest W. Burgess, and Roderick McKenzie. The Heritage of Sociology. Chicago: University of Chicago Press. Peach, Ceri. 1981. Conflicting interpretations of segregation. Pp. 19–33 in Social Interaction and Ethnic Segregation, ed. Peter Jackson and Susan J. Smith. New York: Academic Press. Perkins, D. D., and R. B. Taylor. 1996. Ecological assessments of community disorder: Their relationship to fear of crime and theoretical implications. American Journal of Community Psychology 24:63–107. Perry, Clarence A. 1933. The Rebuilding of Blighted Areas. New York: Russell Sage. Peterson, R. D., L. J. Krivo, and M. A. Harris. 2000. Disadvantage and neighborhood violent crime: Do local institutions matter? Journal of Research in Crime and Delinquency 37:31–63. Pinderhughes, Ellen E., Robert Nix, E. Michael Foster, and Damon Jones. 2001. Parenting in context: Impact of neighborhood poverty, residential stability, public services, social networks, and danger on parental behaviors. Journal of Marriage and Family 63:941–53. Popenoe, David. 1977. The Suburban Environment. Chicago: University of Chicago Press. Priest, R. F., and J. Sawyer. 1967. Proximity and peership: Bases of balance in interpersonal attraction. American Journal of Sociology 72:633–49. Putnam, Robert. 1993. The prosperous community: Social capital and community life. American Prospect 13:3542. ———. 2000. Bowling Alone: The Collapse and Revival of American Community. New York: Simon and Schuster. Quillian, Lincoln, and Devah Pager. 2002. Black neighbors, higher crime? The role of racial stereotypes in evaluations of neighborhood crime. American Journal of Sociology 107:717–67. 230

REFERENCES

Rabin, Yale. 1987. The roots of segregation in the eighties: The role of local government actions. Pp. 208–26 in Divided Neighborhoods: Changing Patterns of Racial Segregation, ed. Gary A. Tobin. Newbury Park, CA: Sage. Raudenbush, Stephen W., and Anthony S. Bryk. 1992. Hierarchical Linear Models: Applications and Data Analysis Methods. Thousand Oaks, CA: Sage. Richards, William D. 1995. NEGOPY. Version 4.30. Brunaby, British Columbia: Simon Fraser University. Romney, A. Kimball, and Katherine Faust. 1982. Predicting the structure of a communications network from recalled data. Social Networks 4:285–304. Ross, Catherine E. 2000. Neighborhood disadvantage and adult depression. Journal of Health and Social Behavior 41:177–87. Ross, Catherine E., and John Mirowsky. 2001. Neighborhood disadvantage, disorder, and health. Journal of Health and Social Behavior 42:258–76. Rountree, P. W., and K. C. Land. 1996. Perceived risk versus fear of crime: Empirical evidence of conceptually distinct reactions in survey data. Social Forces 74:1353–76. Rountree, P. W., and B. D. Warner. 1999. Social ties and crime: Is the relationship gendered? Criminology 37:789–813. Ryan, S., and M. G. McNally. 1995. Accessibility of neotraditional neighborhoods: A review of design concepts, policies, and recent literature. Transportation Research A 29:87–105. Sampson, Robert J. 1986. The effects of urbanization and neighborhood characteristics on criminal victimization. Pp. 3–26 in Metropolitan Crime Patterns, ed. Robert Figlio, Simon Hakim, and George Rengert. Monsey, New York: Criminal Justice Press. ———. 1997. Collective regulation of adolescent misbehavior: Validation results from Chicago neighborhoods. Journal of Adolescent Research 12: 227–44. ———. 2001. How do communities undergird or undermine human development? Relevant contexts and social mechanisms. Pp. 3–30 in Does It Take a Village? Community Effects on Children, Adolescents, and Families, ed. A. Booth and N. Crouter. Mahwah, NJ: Lawrence Erlbaum. ———. 2004. Networks and neighbourhoods: The implications of connectivity for thinking about crime in the modern city. Pp. 157–66 in Network Logic: Who Governs in an Interconnected World? ed. Helen McCarthy, Paul Miller, and Paul Skidmore. London: Demos. Sampson, Robert J., and W. B. Groves. 1989. Community structure and crime: Testing social-disorganization theory. American Journal of Sociology 94: 774–802. 231

REFERENCES

Sampson, Robert J., and J. Morenoff. 1997. Ecological perspectives on the neighborhood context of urban poverty: Past and present. Pp. 1–22 in Neighborhood Poverty, vol. 2, Policy Implications in Studying Neighborhoods, ed. J. Brooks-Gunn, G. J. Duncan, and J. L. Aber. New York: Russell Sage Foundation. Sampson, Robert J., Jeffrey D. Morenoff, and Felton Earls. 1999. Beyond social capital: Spatial dynamics of collective efficacy for children. American Sociological Review 64:633–60. Sampson, Robert J., Jeffrey D. Morenoff, and T. Gannon-Rowley. 2002. Assessing “neighborhood effects”: Social processes and new directions in research. Annual Review of Sociology 28:443–78. Sampson, Robert J., and S. W. Raudenbush. 1999. Systematic social observation of public spaces: A new look at disorder in urban neighborhoods. American Journal of Sociology 105:603–51. Sampson, Robert J., S. W. Raudenbush, and F. Earls. 1997. Neighborhoods and violent crime: A multilevel study of collective efficacy. Science 277:918–24. Sandefur, R. L., and E. O. Laumann. 1998. A paradigm for social capital. Rationality and Society 10:481–501. Schelling, T. 1971. Dynamic models of segregation. Journal of Mathematical Sociology 1:143–86. Schoenberg, Sandra, and Patricia L. Rosenbaum. 1981. Neighborhoods That Work. New Brunswick, NJ: Rutgers University Press. Sermons, M. William, and Frank S. Koppelman. 1998. A factor analytic approach to incorporating systematic taste variation into models of residential location choice. Transportation Research Record 1617. Sewell, William H., and J. Michael Armer. 1966. Neighborhood context and college plans. American Sociological Review 31:159–68. Sewell, William H., Archibald O. Haller, and Alejandro Portes. 1969. The educational and early occupational attainment process. American Sociological Review 34:82–92. Shaw, Clifford R., and Henry D. McKay. 1969. Juvenile Delinquency and Urban Areas. Rev. ed. Chicago: University of Chicago Press. Sigelman, Lee, Timothy Bledsue, Susan Welch, and Michael W. Combs. 1996. Making contact? Black-white social interaction in an urban setting. American Journal of Sociology 101:1306–32. Silverman, C. J. 1983. Neighbors and neighborhoods. Ph.D. diss., University of California, Berkeley. Skjaeveland, Oddvar, and Tommy Garling. 1997. Effects of interactional space on neighbouring. Journal of Environmental Psychology 17:181–98. 232

REFERENCES

Small, Mario L., and Katherine Newman. 2001. Urban poverty after the truly disadvantaged: The rediscovery of the family, the neighborhood, and culture. Annual Review of Sociology 27:23–45. Smith, Ken R., and Norman J. Waitzman. 1997. Effects of marital status on the risk of mortality in poor and non-poor neighborhoods. Annals of Epidemiology 7:343–49. Smith, W. R., S. G. Frazee, and E. L. Davison. 2000. Furthering the integration of routine activity and social disorganization theories: Small units of analysis and the study of street robbery as a diffusion process. Criminology 38: 489–523. Sorensen, A. B. 1998. Theoretical mechanisms and the empirical study of social processes. Pp. 238–66 in Social Mechanisms: An Analytical Approach to Social Theory, ed. Peter Hedstro¨m and Richard Swedberg. Cambridge: Cambridge University Press. South, Scott J., and Eric P. Baumer. 2000. Deciphering community and race effects on adolescent pre-marital childbearing. Social Forces 78:1379–1407. ———. 2001. Community effects on the resolution of adolescent premarital pregnancy. Journal of Family Issues 22:1025–43. South, Scott J., and Kyle Crowder. 1997. Escaping distressed neighborhoods: Individual, community, and metropolitan influences. American Journal of Sociology 102:1040–84. Stein, Maurice. 1964. The Eclipse of Community. New York: Harper and Row. Storm, Murray. 1985. Master Plan of Arterial Highways. Orange County, CA: Orange County Environmental Management Agency. Stutz, F. P. 1973. Intra-urban social visiting and leisure behavior. Journal of Leisure Research 5:6–15. Suttles, Gerald. 1968. The Social Order of the Slum: Ethnicity and Territory in the Inner City. Chicago: University of Chicago Press. ———. 1972. The Social Construction of Communities. Chicago: University of Chicago Press. Tienda, Marta. 1991. Poor people and poor places: Deciphering neighborhood effects on poverty outcomes. Pp. 245–62 in Macro-Micro Linkages in Sociology, ed. J. Haber. Newbury Park, CA: Sage. Tita, George, Jacqueline Cohen, and John Engberg. 2005. An ecological study of the location of gang set space. Social Problems 52:272–99. Tobler, Waldo R. 1970. A computer model simulation of urban growth in the Detroit region. Economic Geography 46(2): 234–40. Veysey, B. M., and S. F. Messner. 1999. Further testing of social disorganization theory: An elaboration of Sampson and Groves’s “Community Structure and Crime.” Journal of Research in Crime and Delinquency 36:156–74. 233

REFERENCES

Waddell, P. 1992. A multinomial logit model of race and urban structure. Urban Geography 13(2): 127–41. ———. 1993. Exogenous workplace choice in residential location models: Is the assumption valid? Geographical Analysis 25:65–82. Ward, Colin. 1978. The Child in the City. London: Architectural Press. Warner, B. D., and P. W. Rountree. 1997. Local social ties in a community and crime model: Questioning the systemic nature of informal social control. Social Problems 44:520–36. Warren, R., and D. Warren. 1977. The Neighborhood Organizer’s Handbook. Notre Dame, IN: University of Notre Dame Press. Watts, Duncan, and Steven Strogatz. 1998. Collective dynamics of small world networks. Nature 393:440–42. Webber, Melvin. 1963. Order in diversity: Community without propinquity. Pp. 63–93 in Cities and Space: The Future Use of Urban Land, ed. L. Wingo Jr. Baltimore: Johns Hopkins Press. Wellman, Barry. 1979. The community question: The intimate networks of East Yorkers. American Journal of Sociology 84:1201–31. ———. 1999. The network community: An introduction. Pp. 1–47 in Networks in the Global Village: Life in Contemporary Communities, ed. B. Wellman. Boulder, CO: Westview Press. Wellman, Barry, Peter Carrington, and Alan Hall. 1983. Networks as personal communities. Pp. 130–84 in Structural Sociology, ed. S. D. Berkowitz and Barry Wellman. Cambridge: Cambridge University Press. Wellman, Barry, and Scott Wortley. 1990. Different strokes from different folks: Community ties and social support. American Journal of Sociology 96:558–88. White, Harrison C., Scott A. Boorman, and Ronald R. Breiger. 1976. Social structure from multiple networks. I. Blockmodels of Roles and Positions. American Journal of Sociology 81:730–80. Whyte, William H., Jr. 1957. The Organization Man. Garden City, NJ: Anchor. Wilner, D. M., R. Walkley, and S. W. Cook. 1955. Human Relations in Interracial Housing: A Study of the Contact Hypothesis. Minneapolis: University of Minnesota Press. Wilson, J. Q., and G. Kelling. 1982. The police and neighborhood safety: Broken windows. Atlantic Monthly 127:29–38. Wilson, William Julius. 1987. The Truly Disadvantaged: The Inner City, the Underclass, and Public Policy. Chicago: University of Chicago Press. Wolfgang, Marvin E., and Franco Ferracuti. 1967. The Subculture of Violence: Towards an Integrated Theory in Criminology. New York: Tavistock. 234

REFERENCES

Yabiku, S. T. 2004. Marriage timing in Nepal: Organizational effects and individual mechanisms. Social Forces 83:559–86. Yen, Irene H., and George A. Kaplan. 1999. Neighborhood social environment and risk of death: Multilevel evidence from the Alameda County study. American Journal of Epidemiology 149:898–907. You, Jinsoo, Zorica Nedovic-Budic, and Tschangho John Kim. 1997. A GISbased traffic analysis zone design: Implementation and evaluation. Transportation Planning and Technology 21:69–91. Zeiher, Helga. 2003. Shaping Daily Life in Urban Environments. Pp. 66–81 in Children in the City: Home, Neighbourhood and Community, ed. P. Christensen and M. O’Brien. London: Routledge; New York: Falmer. Zito, J. 1974. Anonymity and neighboring in an urban high rise complex. Urban Life and Culture 3:243–63. Zorbaugh, Harvey Warren. 1929. The Gold Coast and the Slum: A Sociological Study of Chicago’s Near North Side. Chicago: University of Chicago Press.

235

Index

adaptive link tracing, 11, 59, 66–67, 72, 75, 78, 114–16, 130, 133 administrative data, 4, 11–12, 51, 70, 72, 78, 92, 193 administrative geography, 3–4, 8, 10, 13, 30, 47, 79, 127–28, 192, 194. See also school (attendance zone, catchment, district, geography), census (block group, geography, tract) African American, 29, 51, 79, 82–84, 87–88, 196 alters, 133, 157, 162, 185, 189 Anglos. See non-Hispanic whites apartment complexes, 65–66, 101,105–6 arterial streets, 29, 32 Asians, 76–77, 82–84, 86, 88, 196 babysitters, 95, 141–42, 147 barriers, 12, 29, 31, 33, 82, 88–89, 92, 193; natural, 12, 88–89, 92, 193 barrio, 11, 14, 59–60, 176, 178, 182, 194–95, 197 block parties, 95–96, 111 border block groups, 12, 84–85, 88–90, 92 boundaries, 3, 9, 13, 15, 20, 27, 29–31, 33–34, 36, 43–44, 52, 66, 88, 94, 115, 128, 180, 191–92, 194–95, 199 Bowling Alone, 182 building fences, 141–42, 144, 146 built environment, 31 Burgess, Ernest W., 49 calling neighbor before police, 141, 144, 146, 168 Carrington, Peter, 24 Catholic, 60, 180, 182, 188

census block groups, 3, 70–71, 79, 81–91, 199 Census Bureau, 3–5, 30, 78 census data, 69–70, 178 census geography, 3–4, 13, 30, 51, 81, 111, 128, 194, 198–99 census tracts, 3, 28, 30, 51, 64, 110–13, 115–118, 121–27, 179 census feature classification code (CFCC), 33 child abuse programs, 55, 182 child monitoring relations, 6, 14, 140– 43, 146, 163–65, 168–69, 176, 195, 197 children, 1–2, 4–8, 12–14, 19, 23, 53–56, 59, 62, 72, 83–84, 90, 92–95, 97, 108, 111, 129, 135–47, 156, 162–69, 175– 76, 181–82, 190, 193–95, 197, 199; disciplining, 140–42, 147, 168–69; households with (see households with children); presence of, 84, 95, 136–38, 140, 175; school-age, 6–7; seeking work, 95, 111 choice, 7, 10, 14, 18, 24–25, 40, 49, 52, 57–58, 77, 143, 146 churches, 30, 35, 49, 95–97, 100, 179, 183 cognitive algebra, 149 cognitive maps, 62–64, 68, 98, 100–107, 109, 111, 115, 124–26, 196 cognitive neighborhoods, 12, 108, 197 cohesion, 158 collective efficacy, 1, 5, 8, 11, 14–16, 26, 38, 48, 54–59, 148, 161–62, 168, 174, 176–77, 183, 191, 195, 199 collective socialization, 53 college town, 11, 14, 59, 61–64, 68–70, 72, 76–78, 100–102, 105–6, 110–13,

237

INDEX

college town (cont’d) 116, 119, 122–27, 132, 134, 137, 139, 162, 169, 176, 194, 197 college town census. See exhaustive census college town resample, 61–64, 68, 110–13, 137 commercial zone, 65 community, 1, 4, 11, 13–15, 37–39, 41, 55–56, 58, 60, 68, 76–77, 114–15, 129–31, 137, 144, 148, 155, 161–62, 176, 180, 182, 186–88, 190–91, 195, 198–99; sense of, 38, 54, 158 commuting time, 49 concatenation, 9–10, 13, 15, 30, 37–39, 42–44, 46–48, 52, 57–58, 72, 75, 78, 89, 102–5, 113, 115, 118, 129–31, 136– 37, 144, 155, 167, 192–94, 196, 198–99 concentrated disadvantage, 2 consensus, 54–55, 149, 151, 158, 169–70 constraint, 8, 11, 13, 19, 22, 25, 39, 42– 43, 48, 59, 61, 66, 72, 115, 117, 121, 126, 128, 148, 152, 155, 160, 165, 189– 90, 192, 194, 196–97 convenient availability, 18, 139 crime, 2, 49–50, 55, 65, 67, 78 cul-de-sac, 33 data collection events, 61–65, 67–70, 75, 77, 81, 95, 102, 110–13, 115–16, 123–26, 132, 137, 165, 198 Dawkins, Richard, 160 delinquency, 2, 4 demographics, 2, 4, 12, 48–53, 70, 80– 85, 89–92, 155, 193, 196, 199 density, 14, 65, 67, 122–24, 126–27, 151– 52, 154–55, 157, 159, 161, 164 Detroit area study, 51 disciplining children, 140–42, 147, 168–69 distance, 9, 11, 13, 20–23, 28–29, 31, 34– 36, 45, 50, 52, 59, 67–68, 109–10, 119–29, 133, 154, 157, 167, 189, 192– 94; in face blocks, 189; functional, 21, 23, 34, 45, 121; geographic, 13, 119,

238

122–23, 126, 128, 194; in house-steps, 13, 109–10, 128, 193; in neighbor network steps, 140, 167; physical, 20, 31, 45, 120; tertiary street, 109, 187 dogs, 95, 97, 111 doubled back, 115–16 early childbearing, 2, 55 ecological units, 29–30 ecology, 24 emergency contact person, 141–42, 147 ethnography, 11–12, 59–61, 70, 92, 179, 182, 184, 190 exchanged keys, 141, 144, 146 exhaustive census, 11, 59, 61–63, 68–70, 72, 76–78, 100, 105, 112–13, 116–25, 131–34, 137, 139, 163–64 face block, 9–10, 12, 20, 22, 31–36, 42–45, 47, 51–52, 98–105, 108, 111, 114, 156, 170–73, 189–90, 192, 194, 196–97 family relation, 135 for you relation, 184–86, 188–90 freeways, 29, 31, 33, 88–89, 193 Friedkin, Noah, 149, 151–52, 157–58, 162–64, 169, 170 functional distance, 21, 23, 34, 45, 121 gang, 2, 4, 11, 14, 59–61, 72, 176, 178– 85, 187–89, 194–95, 197 gang barrio, 11, 14, 59–60, 176, 178, 194–195, 197. See also gang neighborhood gang neighborhood 60, 72, 180, 187. See also gang barrio Gans, Herbert, 24 Geographic Data Technology, 71 geographic information system (GIS), 4, 78 geography, 1, 4, 6, 10, 13, 15, 20, 22, 30– 31, 47–49, 51, 63–64, 81, 88, 95, 111– 12, 119–21, 127–28, 178, 181, 187–88,

INDEX

190, 192, 194–95, 198–99; administrative, 3–4, 8, 10, 13, 30, 47, 79, 127–28, 192, 194; census, 3–4, 13, 30, 51, 81, 111, 128, 194, 198–99; school, 199 ghettos, 79 Girl Scout cookies, 95 Granovetter, Mark, 25 Guttman scale, 19 Halloween, 164–65 handyman teams, 182 highways, 29, 31, 33, 69, 88 hills, 88 Hispanic, 81–84, 87–88, 196 homophily, 11, 48, 50–52, 56, 73, 78, 92– 93, 193. See also neighbor similarity horizon of observability, 14, 155, 157, 160–62, 164, 176 households with children, 1, 5–7, 12–14, 54, 56, 72, 83–84, 90, 92, 136–37, 139, 141–42, 145, 147, 162–64, 168–69, 176, 193–95, 199 ideas, 53–54, 151 influence networks, 14–15, 38, 43, 58, 148–50, 155, 161–62, 174, 176–78, 184, 187, 189–91, 195, 197 informal control, 157 internal adjacency, 85, 88–90 intersection, 9–10, 20, 31–32, 34–36, 43– 45, 47, 52, 80, 82, 88–89, 96, 98, 100– 103, 107, 111, 114–16, 170–73, 192, 194, 196–97; non-tertiary, 43; tertiary, 9–10, 20, 35–36, 43–45, 47, 80, 82, 89, 100, 102–3, 114, 170–73, 194, 196–97 interviews, 11–12, 59–70, 72–78, 92–93, 100–101, 103, 107, 109, 115–16, 129– 31, 133, 136, 165–67, 169, 175, 183– 84, 188; structured, 11, 59–61 island, 10, 12, 42, 44–47, 81–83, 92, 102– 7, 114–15, 180–81, 196–98 isolates, 163–64

Jenks, Christopher, 53 Johnsen, E. C., 149 Keller, Suzanne, 24 Korean, 76–77 landmarks, 29 latent ties, 19, 21 lending libraries, 182 linguistic isolation, 12, 22, 77, 99 link tracing, 11, 59, 66–67, 72, 75, 78, 114–16, 130, 133 lived experience, 96, 108, 196 locational choice, 10, 50, 52, 57–58, 75, 78, 195 Los Angeles, 11, 45, 59, 61–62, 64–69, 72, 75, 77–79, 100, 102–4, 106, 111– 16, 131–34, 137, 166 map, 4, 45–46, 63–64, 78–79, 98–100, 102–7 market-based models, 49 marketplaces, 199 markets, 49 Mayer, Susan, 53 mechanism, 30, 53, 159–60, 166, 186, 198–99 micro-macro transition, 37 Miller, H., 3 modifiable areal unit problem (MAUP), 3 Moody, James, 158–59 multi-unit complexes, 65–66, 101,105–6 natural areas, 29–30, 49 natural barriers, 12, 88–89, 92, 193 natural movement, 35 natural settings, 151–52 neighbor networks, 1, 9–14, 18, 34, 37– 43, 46–48, 52, 54, 56–57, 59, 61, 68– 69, 72–73, 76–77, 90, 92, 96, 103–4, 111, 113–15, 117, 121, 128–30, 134, 136, 139–40, 144, 148, 151–52, 154,

239

INDEX

neighbor networks (cont’d) 157, 159–62, 164, 166–68, 182, 194– 96; fragility of, 37, 131; higher stages of, 9–11, 13, 27, 44, 57–58, 73, 76–77, 92, 110, 128, 152, 187, 193, 195; layers of, 1, 17–18, 37; sizes of, 136; stage 1 (see stage 1 neighbor network); stage 2 (see stage 2 neighbor network); stage 3 (see stage 3 neighbor network); stage 4 (see stage 4 neighbor network) neighbor network step, 140, 167 neighbor similarity, 11–12, 74–75, 90, 92, 167; in beliefs 167 neighborhood: beliefs about, 10, 11, 14– 15, 54, 57–58, 72, 148, 158, 161, 167– 70, 174, 176–77, 183, 185, 194–95; cognitive, 12, 108, 197; effect (see neighborhood effect); efficacious, 14, 41– 42, 54–55, 155, 161, 169, 182, 184, 194; efficacy of, 184; equivalent (see neighborhood equivalent); gang, 60, 72, 180, 187; geographic, 72, 178, 180–81; and shops, 35; high crime, 65, 67; people are the best part of, 166– 67, 169–70, 172, 174–75; perception of, 14, 32, 102, 167–69, 177; sociological, 178 neighborhood community, 1, 4, 11, 13– 15, 37–39, 41, 55–56, 58, 60, 68, 76– 77, 114–15, 129–31, 137, 144, 148, 155, 161–62, 176, 180, 182, 186–88, 190–91, 195, 198–99 neighborhood community networks, 13, 37–39, 68, 114–15, 117–20, 129–31, 137, 144, 155, 188, 198–99 neighborhood effects, 3–5, 15, 18, 23, 43, 48–49, 53, 55–57, 138, 141, 148, 154– 55, 168, 171, 177, 195–96, 198–99 neighborhood equivalent, 3–4, 8, 10, 12, 42–44, 47, 52, 72, 79, 108, 178, 192, 196, 198–99 neighborhood watch, 96, 111

240

neighbors; accepting, 76; immediate, 10, 48, 51–52, 55–57, 75, 166–67; and influence, 15, 155, 166, 190–91, 195; known personally, 69, 110, 112, 137; selecting, 48, 73; socialization with, 145; tertiary street, 11, 48, 73, 76–78, 90, 92 network structure, 37, 72, 170, 174; complex, 37, 159 non-Hispanic whites, 81–82, 179, 196 norms, 1, 6, 8–10, 14–15, 17–20, 25–27, 38–39, 42, 48, 53–55, 57–58, 72, 89, 148–52, 154, 156, 158–62, 166, 176– 77, 183, 186, 188, 190–91, 194–95, 197 parent, 2, 5–7, 53, 94–95, 97, 137, 140–42, 147, 156, 164–66, 179, 187 parenting, 2, 182 parish, 15, 60, 100, 180, 188, 190, 195 Park, Robert, 20, 49, 119 parking lot, 101, 106 parks, 6, 29, 35 Pasadena, 79–81 passive contact, 6, 9–10, 12, 19, 21–25, 27, 36, 38–40, 42–45, 47, 52, 54, 73– 73, 93–98, 100, 102–4, 107–9, 111–13, 142–43, 146–47, 186, 183–84, 196, 198 path lengths, 13, 76–77, 129, 131–34, 139, 145, 154, 162–63, 185, 194 paving another’s driveway, 141–42, 144, 146 pedestrian, 9, 22, 32–36, 79, 190, 197; arena, 6–7, 9–10, 22, 36, 42–44, 47–48, 59, 71, 192, 196 peer influence, 53, 160 permeable, 33–34 police district, 4, 30 poverty, 2, 55 premarital sex, 55 primary metropolitan statistical area (PMSA), 65, 67, 79 propinquity, 28, 189. See also distance

INDEX

proximity, 5, 19–21, 23, 28, 32, 49, 79, 105, 109, 119, 187, 192–93. See also distance Putnam, Robert, 55, 182

sidewalk, 6, 33, 35 small world, 66 social capital, 1, 8, 11, 15–16, 26, 38, 48, 53–59, 72, 143, 148, 161, 174, 178, 182, 187, 190–91, 195, 198–99 race, 11, 22, 50, 52, 75–78, 83, 92, 175 social comparison trigger, 149 raffle tickets, 131 social control, 6, 54, 151, 155–57, 162, railroads, 29, 33–34 164, 183 reachable, 140 social disorganization theory, 54, 151 real estate agents, 1, 183 social distance, 20, 22, 50, 52, 120 relations, 1, 6, 8–9, 14, 17–20, 22–28, 37, social distance theory, 50 131, 135, 142–44, 157, 162–64, 176, social influence network theory, 14, 184–85, 189, 193–94 148–50, 161 relinking, 131 socioeconomic status, 2, 49–50, 52, 77–78 relocation, 10, 50, 52, 57–58, 75, 78, 195 spatially adjacent block groups, 85–88, re-roofing another’s house, 141–42, 90–91, 199. See also border block 144, 146 groups, internal adjacencies residential stability, 23, 176, 187 spontaneous playgroups, 14, 140–43, rivers, 29, 88 146, 163–64, 168–69, 176, 195 routine, 19, 60, 140–41, 156, 181 stage 1 neighbor(ing), 8–12, 19–20, 27–28, 35–37, 41, 47, 51, 57–58, 73, Sampson, Robert, 55 75, 77, 90, 92, 96, 103, 109, 111, 114, Schelling, Thomas, 51 128, 160, 186–87, 192–195; and school, 2, 4–7, 12–15, 30, 49, 55, 60, selection, 73 64, 69–71, 94–95, 97, 100, 108, 111, stage 1 neighbor network, 10–12, 37, 41, 117–19, 128, 142, 147, 165–66, 176, 90, 92, 96, 103, 111, 114, 160, 194 180–82, 190, 194–95, 199; dropping stage 2 neighbor(ing), 8–10, 12, 19–20, out of, 55 22, 25, 36–37, 39–41, 46–47, 96, 98, school attendance zone, 71 100, 102–4, 107–8, 113, 128 school catchments, 13–15, 64, 69–71, stage 2 neighbor network, 9–10, 12, 37, 117–19, 128, 166, 176, 180, 187, 190, 39–41, 46–47, 96, 103, 113, 128 194–95 stage 3 neighbor(ing), 9–10, 12, 19, 23, school district, 4, 7, 12, 30, 100, 108, 165 25, 27, 37, 39–40, 46, 92, 96, 103, 109, school fundraisers, 95, 182 111–13, 128, 194 school geography, 199 stage 3 Neighbor Networks, 9–10, 12, 37, secondary streets, 31 39–40, 46, 103, 111, 113, 128, 194 segregation, 2, 5, 10–12, 29, 50–52, 56–58, stage 4 neighbor(ing), 9, 14, 19, 25–27, 69, 72–73, 76–79, 90, 92, 193, 195–96; 37–39, 46, 48, 161–62, 188, 194 and tertiary street networks, 79 stage 4 neighbor network, 9, 37–39, 46, selection, 10, 48, 57–58, 109, 174, 176, 162, 194 197; and stage 1 neighbors, 73 stages of neighboring, 9–11, 13, 17–18, The Selfish Gene, 160 27, 37–38, 51, 56–58, 72–73, 92, 110, selling, 97, 184 152, 187, 193, 198

241

INDEX

Star Trek, 21, 93 Steno, Nicolas, 17 street network, 10, 12–13, 15, 57–58, 72, 78–79, 84, 89–90, 92, 95–96, 104–5, 111, 113, 115, 119, 127–28, 134, 139, 162, 176–78, 180–81, 186–91, 193, 195–99; residential, 65–66 substrate, 8, 21, 39, 72, 192, 200 superimposition, 8, 27, 103, 192 superposition, 17–18, 37 symbols, 53, 151 t-communities, 10, 12–15, 42–47, 52, 61, 64–72, 77, 79–85, 88–90, 92, 100, 102– 4, 110–16, 118, 120–28, 131, 135, 137, 162, 164–66, 176–77, 194, 196–199; sizes of, 71 t-community culture, 166 tertiary face block, 9–10, 12, 20, 32–36, 42–45, 47, 51–52, 99–100, 102– 5, 108, 114, 189–90, 192, 194, 196–97 tertiary street, 10–15, 33–35, 43–46, 48, 52, 57–59, 69–73, 75–81, 83–84, 88– 92, 95–96, 98, 103–11, 113–15, 119, 121, 127–28, 134, 139, 152, 162, 176– 78, 180–81, 186–191, 193, 195–99; island, (see island); neighbors, (see tertiary street neighbors) tertiary street neighbors, 11, 48, 73, 76–78, 90, 92 tertiary street proximity, 109, 187 tertiary street network, 10, 12–13, 15, 57–58, 72, 78–79, 84, 89–90, 92, 95– 96, 104–5, 111, 113, 115, 119, 127–28, 134, 139, 162, 176–78, 180–81, 186–

242

91, 193, 195–99; borders, 12, 84–85, 88–90, 92; segregating, 79; shared, 13–14, 34, 91, 111, 114, 127–28, 176, 181, 194; and single tertiary street connection, 89–90, 193, 199 three degrees of neighboring, 129, 145 Tobler, Waldo, 20, 120 tool loaning, 141 tutoring, 182 unintentional encounters, 9–10, 21–22, 25, 36, 42, 47, 59, 73, 93, 107, 111, 197 values, 1–2, 6, 8–10, 13–15, 17–20, 25– 27, 38–39, 42, 48, 53–58, 72, 76, 89, 123,125, 134, 140, 146, 148–52, 154, 156, 158–62, 164, 166–74, 176–77, 183, 186, 188, 190–91, 194–95, 197; shared, 54 violence, 2, 50, 55 walking arena, 6–7, 9–10, 22, 36, 42–44, 47–48, 59, 71, 192, 196 watching neighbors’ houses, 141 weak ties, 25–26 Wellman, Barry, 24 White, Douglas, 158–59 workplaces, 49 Wortley, Scott, 24 yard work, 95 zoning, 65, 71