 

“gain.

2‘

1.,
‘ P
«NH

1 ~
8.: .
‘1.

s
v

—. r

 

«a
n.l..
5. .1.- Cu
tn . , , . ., . “Rm.

a

.st5? u
.l.
a) .‘

, ma".

Fﬁﬁu

L}: .
‘ 233.7}

. an? .
.L.

«m .
é. .
, .r. > .
. .1 . (i
..
.«r. 1.1., «a.
)Av‘LE

.. ....
«mum»:
fir. ..

.,
. .V.
hawk“ .

.3...
)4
L

_ ... .

 

 

’J'Héfkés"

1 (g,

g:

This is to certify that the .

thesis entitled
THE IMPACT OF THE NEIGHBORHOOD MICROSYSTEM ON

ADOLESCENT OUTCOMES: A BLOCK‘BY-BLOCK
ANALYSIS OF SENSE OF COMMUNITY

presented by

DANIEL M. CANTILLON, JR.

has been accepted towards fulﬁllment
of the requirements for

M.A. . Psychology
degree in

 

 

 

Major professor \

Date 11/22/00

 

0.7639 MS U is an Afﬁrmative Action/Equal Opportunity Institution

 

 

LIBRARY

M'Chigan State
University

 

 

PLACE IN RETURN BOX to remove this checkout from your record.
TO AVOID FINES return on or before date due.
MAY BE RECALLED with earlier due date if requested.

 

DATE DUE

DATE DUE

DATE DUE

 

J18! 1 $2903

 

 

 

 

 

 

 

 

 

 

 

 

 

6/01 cJCIRC/DateDue.p65—p. 15

 

THE IMPACT OF THE NEIGHBORHOOD MICROSYSTEM ON ADOLESCENT
OUTCOMES: A BLOCK-BY-BLOCK ANALYSIS OF SENSE OF COMMUNITY

By

Daniel M. Cantillon, Jr.

A THESIS
Submitted to
Michigan State University
in partial fulﬁllment of the requirements
for the degree of
MASTER OF ARTS
Department of Psychology — Urban Affairs

2000

ABSTRACT

THE IMPACT OF THE NEIGHBORHOOD MICROSYSTEM ON ADOLESCENT
OUTCOMES: A BLOCK-BY-BLOCK ANALYSIS OF SENSE OF COMMUNITY

By

Daniel M. Cantillon, Jr.

This study utilized an updated systemic model of social disorganization to assess
community effects on adolescent outcomes. The research on neighborhood or
community effects has substantially increased in recent years yet it is argued that current
conceptualizations of the important social characteristics of communities is still lacking.
This study utilized Sense of Community (SOC) as a possible mediating variable of social
disorganization theory in an attempt to introduce a comprehensive and systemic measure
to the ﬁeld. The results of this study indicate that SOC was able to measure and
discriminate among differential levels of social organization within a community and that
these levels are related to both positive and negative youth outcomes. The results suggest
two possible modiﬁcations for ﬁrture investigations of the neighborhood microsystem.
One, there are varying conceptualizations of neighborhood and research needs to begin
the assessment of neighborhood effects from the block or face-block level. It was found
that there are indeed important differences between face-blocks in the same census tract
and these differences must be accounted for in any truly contextual model of
neighborhood effects. Two, the sense of community construct is recommended for
inclusion in future neighborhood based studies. It is a more systemic and applicable
measure than current conceptualizations and may provide more insights into how

neighborhoods affect youth outcomes.

ACKNOWLEDGEMENTS

First and foremost, I would like to thank the Lansing residents who took their
time to speak with project staff about their community. I was continuously and happily
surprised by the number of residents that were willing to let us into their homes to
interview them about the possible impact of community conditions on their children.
Second, I would like to thank Dr. John Schweitzer for the opportunity to conduct this
research project. Third, I would like to thank all of the interviewers on the project who
spent an incredible amount of time driving throughout Lansing to conduct the interviews:
Heather Van Dyke, Hamza Bunnya, Anne Hussey, Michelle Busch, Julie Winnick, Becky
Gardner, Debbie Johnson, Jim Clark, Maunda Burk, Jennifer Poat, Jenny McMann, and
Diana Menefee. I am truly indebted to their hard work and persistence. I would also like
to thank the Sense of Community project staff members who helped train many of the
aforementioned interviewers. Finally, I would like to thank my committee, in particular
Dr. Bill Davidson, for the considerable time he spent guiding me through the Master’s

process.

iii

TABLE OF CONTENTS

LIST OF TABLES
LIST OF FIGURES

I. INTRODUCTION
A. Social Theories of Delinquency
B. History of the Sense of Community (SOC)Construct
C. The Re—Emergence of Social Disorganization
D. Reformulated Systemic Theories of Social Disorganization

E. The Current Study

II. METHODOLOGY
A. Procedures
B. Interviewer Training & Supervision
C. Interview Procedure
D. Face-Block Measures
E. Individual Youth Measures

11. RESULTS
A. Data Analytic Strategy
B. Research Objective 1: Dimensionality of the SOC & SRD Constructs
C. Research Objective 2: Evaluating the Validity of the Face-Block
as a Neighborhood Unit
D. Research Objective 3: Sense of Community (SOC) as 3 Mediator
of Neighborhood Advantage (NA) & Youth Outcomes

III. DISCUSSION
A. Research Objective 1: Dimensionality of the SOC & SRD Constructs
B. Research Objective 2: Evaluating the Validity of the Face-Block
as a Neighborhood Unit
C. Research Objective 3: Sense of Community (SOC) as a Mediator
of Neighborhood Advantage (NA) & Youth Outcomes
IV. IMPLICATIONS FOR FUTURE RESEARCH
V. APPENDIX A: Initial Letter to Parents
VI. APPENDIX B: Follow-Up Letter to Parents
VII. APPENDIX C: Youth Assent Form

VIII. APPENDIX D: Parent Introduction

vi

vii

13
16
22
29
32
32
35
36
36
4o
45
45
46
50
52

66
66

68

71

76

80

82

84

86

IX.

.><

XII.

XIII.

APPENDIX E: SOC Measure with Six Components

APPENDIX F: Youth Survey

APPENDD( G: Principal Component Analysis of the SOC Construct
APPENDD( H: SOC’s Three Components

REFERENCES

88

90

94

96

99

LIST OF TABLES

Table 1 Race/Ethnic Background of Lansing High School Students &

Research Participants 42
Table 2 Residential Stability Items, Corrected Item-Total Correlations,

& Reliability Coefﬁcients for Parent & Neighbor 44
Table 3 Emotion Items, Corrected Item-Total Correlations, &

Reliability Coefﬁcients for Parent & Neighbor 56
Table 4 Action Items, Corrected Item-Total Correlations, & Reliability

Coefﬁcients for Parent & Neighbor 57
Table 5 Safety Items, Corrected Item-Total Correlations, & Reliability

Coefﬁcients for Parent & Neighbor 58
Table 6 Principal Component Analysis of the Self-Reported Delinquency

(SRD) Construct 59
Table 7 Face-Block Descriptive Statistics 60
Table 8 Uncorrected & Corrected Correlations of the SOC construct

Between Parent & Neighbor 61
Table 9 Correlation Matrix of Independent (IVs), Mediating (MVs),

& Dependent Variables (DVs) 62
Table 10 Uncorrected & Corrected Correlation Matrix of Independent (IVs)

& Mediating Variables (MVs) 64
Table 11 WLS Regression Estimates of Conventional Activity (CA) 65

Table 12 Principal Components Analysis of the Sense of Community
(SOC) Construct 95

LIST OF FIGURES

Figure 1 Signiﬁcant Findings of Mediating Variables From Previous
Neighborhood Disadvantage Studies 31

Figure 2 Research Participant Recruitment Flowchart 43

Figure 3 Measurement Model 63

vii

INTRODUCTION

“ I urge you to consider this: As you demand tougher penalties for those who
choose violence, let us also remember how we came to this sad point We have seen a
stunning and simultaneous breakdown of community, family, and work. This has created
a vast vacuum which has been filled by violence and drugs and gangs. So I ask you to
remember that even as we say no to crime, we must give people, especially our young
people something to say yes to. ”

President Clinton, State of the Union Address, January 24, I 994

America’s juvenile delinquency problem has garnered signiﬁcant public attention
in recent decades and is one of the deﬁning political issues of the day. A primary reason
for the public’s current concern with delinquency was due to the dramatic increase which
occurred in the late eighties and early nineties. For instance, from 1985 to 1994, arrests
of juveniles for rape increased by 25%, robbery by 53%, aggravated assault by 34% and,
even more alarming, arrests for murder increased by 144% (Butts, 1996). While these
statistics are quite disconcerting, more recent data displays an overall decline in juvenile
crime over the past several years. Juvenile arrests for violent crimes dropped 19% ﬁ'om
1994 to 1998 and juvenile arrests for violence in 1998 were the lowest they had been in a
decade (OJJDP, 1998). Despite the current statistics, however, there is a perception

among the general public and policymakers that juvenile crime is forever on the rise and

that each generation of youth are simply more violent than the previous generation (Eddy
& Gribskov, 1998).

This misperception has translated into changes in juvenile justice policy. Over
the last decade, the majority of states have reduced the age at which juveniles can be tried
as adults and have also provided stiffer penalties for convicted juveniles (Yee, 1998). For
instance, in 1996, Michigan passed legislation that reduced the age limit for transfer to
adult court from 15 to 14 and also doubled the number of offenses for this possible
transfer from nine to 18 (Clark, 1996). Another substantial change enacted by this
legislation was that prosecutors could now try juveniles in the traditional juvenile court as
adults for any speciﬁed juvenile offense without age restrictions.

Michigan’s implementation of “get tough” juvenile offender laws recently
brought the state national attention when Nathaniel Abraham was charged as an adult for
murder in the ﬁrst degree (in juvenile court) for a crime he committed when he was only
11 years-old. Fortunately for Nathaniel and youth advocates, Judge Eugene Moore
refused to sentence him as an adult or impose the politically palatable “blended
sentence.” Instead, Judge Moore sentenced Nathaniel to a maximum eight year sentence
at one of the state’s training schools, derided the 1996 law, and summed up the case by
stating, “He is a boy who has been neglected by his home, our community, and our
juvenile justice system. He represents our collective failings” (Knott & Brand-Williams,
2000).

As the above quote and President Clinton’s 1994 State of the Union address
indicate, much of the dialogue regarding the etiology of juvenile crime has focused on the

breakdown of the neighborhood as a nurturing and caring environment. The intent of the

current study is to focus on how neighborhoods inﬂuence the children who are reared
within them. While there may be agreement that it takes a village to raise a child, one
vital and unanswered question remains: exactly what type of village? Which
neighborhood characteristics lead to positive youth outcomes and which characteristics
lead to deleterious youth outcomes?

The majority of studies that have addressed the impact of neighborhood
characteristics on juvenile crime have limited their analyses to the socioeconomic status
(SES) of the neighborhood (Coulton, Korbin, & Su, 1996; Komhauser, 1978). While low
SES has long been associated with higher incidence of crime, many past and
contemporary analyses have not explicated the true nature of this relationship. What is it
about growing up in poor environments that leads to increased contact with the juvenile
justice system? Some researchers disagree with these prior statements altogether and
suggest that the increased contact with the police and justice systems are simply
reﬂections of bias rather than differential offending rates (Sampson, 1986). Also, there
are examples of low-income neighborhoods across the United States that have not had
delinquency problems or high rates of other negative youth outcomes (Maccoby,
Johnson, & Church, 1968). How have these areas avoided the pervasive negative impact
of poverty on child developmental outcomes like delinquency? So, while there is
agreement that neighborhoods matter, the question that remains unanswered is - how?

This question of how neighborhoods matter has recently been re-addressed from a
social-ecological framework. This framework includes an emphasis on how SES affects
the structural or normative characteristics of a neighborhood, its social dimensions, and

ultimately delinquency rates (Brooks-Gunn, Duncan, & Aber, 1997; Bursik, 1988;

Sampson, 1993; Simcha-Fagan & Schwartz, 1986). The current study continues this
exploration of the nature of the relationship between neighborhood disadvantage and
delinquency. The current study offers both theoretical and methodological improvements
over previous work. Speciﬁcally, neighborhood is conceptualized in a much smaller,
objective, and relevant unit for urban residents - the face-block. Face-blocks are houses
on a street that face each other and are bounded by cross-streets. Other research has
indicated the usefulness of the face-block in understanding residents’ perception of
informal social control, crime, and fear of crime (Perkins & Taylor, 1996; Schweitzer,
Kim, & Mackin, 1999; Taylor, 1997; Taylor, Gottfredson, & Brower, 1984). At the
methodological level, it is argued that sense of community (SOC) provides a more
comprehensive and systemic measure of the intervening variables in social
disorganization theory. The vast majority of recent neighborhood studies have lacked
this relevant proximal unit of the face-block and have not fully measured or captured the
intervening variables proposed by social disorganization theory.

Before discussing in detail the need for the current study, a review of social
theories of delinquency is warranted. There will be an emphasis on social
disorganization theory in this review since it is the grounding theory for the study. Next,
an historical overview of the sense of community construct will be presented. Third, an
argument will be made regarding the re-emergence of social disorganization theory in
delinquency research and the need for improved methodology in the ﬁeld. Fourth, recent
literature on neighborhood-based, social disorganization studies will be examined.

Finally, the author will describe how the current study will improve the ﬁeld’s

understanding of the impact of the neighborhood microsystem on multiple youth
outcomes, with a particular emphasis on juvenile delinquency.
Social Theories of 1361th92

Subcultural or Cultu_r_al Devia_ng:e Th_eories

Walter Miller (1958) is one of the prominent theoreticians of delinquency in
urban communities which were often labeled as poor, lower-class, or poverty-stricken.
Miller was largely concerned with gang or group delinquency that occurred in poor
communities across the United States during the 19503. One of the main assumptions of
Miller’s subcultural theories was that juveniles were perceived as goal-oriented and their
delinquent behavior was a direct attempt to achieve status by following lower-class
guidelines for appropriate behavior. Miller contended that lower-class values were
different ﬁom the middle-class and those values were the main reason for the increased
delinquency rates that occurred in these areas. In a related vein, Miller also emphasized
that the lack of intact families in these areas was another major cause of the increased
delinquency rates.

There was a strong sense of victim-blaming throughout Miller’s analysis of poor
communities and their ‘deviant’ value system:

In the case of gang delinquency, the cultural system which exerts the most direct

inﬂuence on behavior is that of the lower-class community itself - a long-

established, distinctively patterned tradition with an integrity of its own - a rather

so-called “delinquent subculture” which has arisen through conﬂict with middle-

class culture and is oriented to the deliberate violation of middle-class norms (p.

5).

In his analysis, Miller was not concerned with conditions of oppression and factors like
prejudice, racism, and inadequate housing and public schools which isolated individuals
in these circumstances and created this “delinquent subculture.” He simply attributed the
effects of poverty and other negative social conditions to those who happened to be born
in these chaotic communities. Furthermore, while there may have existed a delinquent or
criminal subculture, it was certainly true and continues to be true that the majority of
individuals in poor communities have the same exact values as those in the middle and
upper-classes.

Despite this and other criticisms, the importance of Miller’s work resided in his
attempt to understand a particular type of delinquency - gang delinquency in poor urban
areas. It was primarily due to his efforts, and other pioneers concerned with gang and
lower-class delinquency (e.g., strain theorists), that social scientists paid more attention to
the inﬂuence of poor neighborhoods, delinquent peers, and the existence of illegal
economies.

W

Strain theories also attempted to explain delinquency that occmred in lower SES
neighborhoods. In contrast to subcultural theory, strain theory asserted that the
motivation toward delinquency was not to attain status or reputation by following a
different (i.e., lower-class) value orientation. Rather, delinquent behaviors were the
product of the desire to attain ‘normal’ societal aspirations, while lacking the same
opporttmity as those who resided in the middle and upper-social classes (Merton, 1938;
Cloward & Ohlin, 1960). This disadvantage brought about frustration at the gap between

socially desired goals and socially approved means of achieving them, and illegitimate or

delinquent paths were more likely to be considered and taken, especially in a group
context (Shoemaker, 1990). So, although there was a similarity between the two theories
in that both viewed individuals as rational and goal-oriented, their explanations as to the
path of delinquency differed substantially.
Social Learning Theories

Social learning was an all-encompassing theory of behavior which did not seek to
solely explain deviance or delinquency. This theory postulated that all human behaviors
are predicated on the anticipated reinforcements for behavior. The decision to perform a
behavior was the result of a rational assessment of the rewards and possible negative
sanctions for that behavior compared to the rewards and possible negative sanctions for
alternative behaviors (Akers, 1990; Akers, 1977; Elliot, Huizinga, & Ageton, 1985). A
particular strength of this theory was that it attempted to explain behavior across all SES
groups. Also, this theory did not assume any constant strain or motivation for delinquent
behavior. Rather, social learning theory postulated that delinquency was the result of
direct socialization with a deviant peer group and conformity was the result of direct
socialization with a conventional peer group. In this manner, social learning theory was
similar to social control theory, as will be presented, in that both viewed delinquency as a
result of variations in a youth’s socialization experience. Meanwhile, this theory matched
adequately with subcultural theories if one assumed there was a greater presence of
deviant peer groups in disadvantaged communities.
52W

This perspective assumed that there must be some form of internal or external

constraint if criminal or delinquent behavior was to be held to a minimum. In other

words, delinquent or criminal behavior was universal and society needed appropriate
control mechanisms to keep it to a minimum. It was hypothesized that there were both
internal or personal and external or social control mechanisms. Personal social control
theories focused on individual psychological constructs like self-esteem and internalized
norms. Social control theories focused on individuals’ attachments to social institutions
like the family, school, church, and neighborhood (Shoemaker, 1990). As Hirschi (1969)
stated, social control extends beyond mere attachment to also address one’s commitment,
involvement and belief in the social system. In sum, social control theory postulated that
delinquency was the result of a weakened attachment and commitment to society and its
key social institutions. There has been extensive research on social control theory;
however, the primary focus has been on proximal units like the individual and family
without incorporating more macrolevel variables. (Brooks-Gunn, Duncan, Klebanov, &
Sealand, 1993).

Recently however, there has been considerable research on the effects of the
neighborhood and school microsystems on delinquency participation and other youth
developmental outcomes (Brooks-Gum, Duncan, & Aber, 1997; Pretty, Conroy, Dugay,
Fowler, & Williams, 1996). In this aspect, social control and social disorganization
theories, as will be discussed, drew the similar conclusion that a reduction in or lack of a
bond to our basic societal institutions resulted in increased delinquency (Agnew, 1993;
Elliot et al., 1985; Shoemaker, 1990). The current study follows this line of inquiry by
investigating how the sense of community in neighborhoods is related to “control” over
youth that reside there. While social control theory and social disorganization postulated

the inﬂuence of our basic societal institutions (family, school, church, neighborhood, etc.)

on adolescent development, their relative importance has not been adequately explored or
empirically demonstrated. Therefore, this study will add to our knowledge of one of the
important social control mechanisms, the neighborhood or local community, and allow
one to draw inferences on its relative importance.
Social Disorganization Theories

Old Chicago-school. One of the earliest and most inﬂuential social theories of
delinquency was developed at what is now referred to as the old Chicago-school. The
sociology department at University of Chicago was the ﬁrst institution in the United
States which assessed the impact of America’s urbanization in the early to mid-19003. In
The Gig, Park, Burgess, and McKenzie (1967) presented an ecological theory, borrowed
from biology’s attempt to understand how plant life adapted to changing environments,
and theorized how humans adapted to their rapidly changing environment. Speciﬁcally,
the authors were interested in how people adapted to the rise of urban centers. In this
case, the entire city of Chicago had become a research laboratory. The old Chicago-
school was concerned with measuring and assessing the impact of the expansion of our
cities and the concomitant changes in social life for its residents. In the second chapter of
TM, Burgess proposed his inﬂuential concentric zone theory in which each he
argued that each successive concentric circle represented the extension or growth of a
city from the original circle, which represented the central business district, the Loop in
this case. From these zones, one could predict the type of physical structures, businesses,
and demographic characteristics of the individuals who had chosen to locate there.
Basically, the poorest and least attractive housing and neighborhoods were found directly

outside the central business district, the Loop, followed by more and more attractive

neighborhoods characterized by higher standards of living and considerably less social
problems (e.g., crime, prostitution, drugs). As each social group gained more ﬁnancial
resources and social status there was an effort to move outward to a better community.
This resulted in expansion of the city, the tendency for each inner zone to extend its area
by invading the adjacent outer zone. Thus, cities were in a constant state of change and
expansion.

As previously mentioned, the old Chicago-school was mainly concerned with
how this growth process affected the social organization of these communities which
were in a constant state of transformation. So, as one zone was invaded by another, how
were all of the individuals redistributed and what was the overall effect on the involved
communities? It was postulated that as communities expanded and incorporated outer
zones these areas would go through massive changes due to population increase and
turnover (invasion) and the community would be in a state of disorganization until, after
some time, it stabilized into a new outer zone (succession). Therefore, the present level
of invasion and succession in a particular neighborhood could predict the social
organization of communities. According to this ecological theory, social problems could
be found in their most acute forms in communities undergoing these rapid population
changes - in disorganized communities.

W. Shaw and McKay (1942) extended this ecological
approach in the speciﬁc area of delinquency. These two sociologists are largely credited
for popularizing a social ecological theory of delinquency, social disorganization theory,
which explained how poor urban environments affected adolescent development and

delinquency. However, the general premise that environmental factors, especially aﬁer

lO

the industrial revolution, had an impact on delinquency and other social problems had a
long history dating back to nineteenth-century European scholars (Shoemaker, 1990;
Krisberg & Austin, 1978).

Shaw and McKay argued that there was not a direct relationship between the
econorrric composition of a community and delinquency rates. Rather, there were
additional variables that explained the relationship between low economic status and high
rates of delinquency. These authors, in the ecological tradition of previous University of
Chicago scholars, proposed that urban neighborhoods also had higher rates of population
turnover (invasion and succession) and a greater mix of racial and ethnic groups. The
crux of their argument was that it is all of these variables -- low SES, population
turnover, and cultural heterogeneity -- which led to poorly organized communities which
in turn led to higher delinquency and crime rates. Thus, social disorganization was
deﬁned as the inability of local communities to realize the common values of their
residents or solve commonly experienced problems (Komhauser, 1978).

Shaw and McKay argued that there was great difﬁculty in maintaining a strong
institutional base when a community underwent rapid population turnover. Most
members of the community wanted to live elsewhere and were intent on moving as soon
as it was economically feasible. Therefore, little effort was exerted in maintaining or
improving the basic social institutions of the community — schools, churches, social
service organizations, etc. This resulted in the breakdown of the formal institutions that
helped communities come together and solve their problems (Shoemaker, 1990).
Population turnover and cultural heterogeneity also impeded the natural process of

communication and ﬁiendships that develop over time within the community. This

11

communication process was painstakingly slow and unfeasible in a corrrrnurrity with a
constant inﬂux of new residents, especially those who may speak little to no English.
Thus, high rates of population turnover and large amounts of heterogeneity, coupled with
low SES and the corresponding lack of informal and institutional resources, reduced the
development of strong local ties, norms, and informal social control which served to
regulate inappropriate public behavior like delinquency (Bursik and Grasmick, 1993;
Greenberg & Rohe, 1986; Sampson, 1997).

To summarize, this theory began with certain structural or compositional
characteristics of urban communities and explained how certain combinations of these
components led to a decrease in the formal and informal mechanisms of a community
that controlled individual behavior. Social disorganization posited that it is the level of
social organization, or, more accurately, high levels of social disorganization that
mediated the relationship between a neighborhood’s SES and crime and delinquency
rates.

Given the negative and value-laden term of social disorganization, this author
preferred to gauge communities in terms of the relative levels of social organization.
Therefore, the term neighborhood disadvantage is utilized to refer to what the old
Chicago-school labeled social disorganization In accordance with Shaw and McKay, the
major postulate of the current study is that communities characterized by neighborhood
disadvantage (low levels of social organization) will have lower rates of sense of
community and greater incidence of delinquency and other negative youth outcomes.

The previous description of social theories of delinquency was not all-inclusive

and the goal of this study is not to test one social theory against another. Rather, the

12

discussion was intended to provide some theoretical background on the development of
social theories of delinquency and show that there was a great deal of overlap even
though their basic assumptions often dramatically differed. For instance, social
disorganization contained elements of social control, strain, and subcultural theories
(Komhauser, 1978). However, for theoretical grounding, social disorganization’s basis
on understanding rapid ecological changes in urban areas is best approached from a
control-theoretic perspective (Bursik, 1988; Bursik & Grasmick, 1993).

The current study follows the guidelines and principles of social disorganization
theory and is also consistent with social control theory’s emphasis on external
institutional and informal social controls. The belief that SOC is related to delinquency
rates and other youth outcomes also corresponds to social learning theory and subcultural
theory in that youths who are socialized in stable communities are more likely to bond
with conventional peer groups and follow the formal and informal guidelines of the
particular community. Sense of community was not conceptualized as a mediating
construct speciﬁcally for social disorganization theory, yet this variable epitomizes what
Shaw and McKay meant by organized communities, those that were able to meet the
diverse needs of its members. Since this construct is borrowed from community
psychology and has not been previously proposed as a mediating construct for social
disorganization theory, the next section discusses the history of the construct and reasons
for its inclusion in the neighborhood disadvantage framework.

WWW
SOC is deﬁned as “a feeling [italics added] that members have of belonging and

being important to each other, and a shared faith that members’ needs will be met by

13

their commitment to be together” (McMillan, 1976). According to McMillan, there are
four essential elements to this construct: membership, inﬂuence, sharing of values with
an integration and fulﬁllment of needs, and a shared emotional connection. The most
commonly used scale for SOC has been the shortened version of the sense of community
index that was originally proposed and empirically evaluated in Chavis, Hogge,
McMillan, & Wandersman’s (1986) landmark study. However, there have been multiple
scales utilized over the years (Buckner, 1988; Davidson & Cotter, 1986; Doolittle &
MacDonald, 1978; Glynn, 1981). The scale used in the present study was created by the
Sense of Community in Lansing Neighborhood Project team in 1995 and contains many
items from these aforementioned scales.

Since this variable’s initial conception, it has been heralded by community
psychologists as one of the most important theoretical propositions since the birth of
community psychology. For instance, in 1974 Sarason referred to SOC as one of the
overarching values of community psychology. In a recent special edition in the Journal
of Community Psychology, Lorion and Newbrough (1996) stated,

For both of us, that concept (PSOC), above all others, represents the essence of

the ﬁeld’s unique perspective on understanding the human condition. We also

viewed this concept as most reﬂective of the ﬁeld’s initial raison d’etre, i.e., to
reconnect the social sciences generally and psychology speciﬁcally to life as
experienced by people under everyday circumstances. To us, the PSOC is as

central to the view of community psychology as the psychology of everyday life

(p. 311).

14

To date there have been a large number of publications and four special issues of
the Journal of Community Psychology devoted entirely to the topic psychological sense
of community (Chavis & Pretty, 1999). Despite the vast amount of attention paid by
researchers and journals to the topic, a recent literature review in the 1996 special issue
indicated that there were not many studies that actually measured the variable. Further,
those that did presented contradictory ﬁndings as to what correlated with the construct
and even its basic factor structure (Hill, 1996). One of the reasons for the inability of the
ﬁeld to advance this theory was simply because it is extremely difﬁcult to assess
community level constructs or contextual effects (Coulton, Korbin, & Su, 1996;
Sampson, 1991; Shinn, 1990), and, as a ﬁeld, we were still conﬁned to methodological
individualism at the same time that we eschewed it: “Where is the community in
community psychology?” (Heller, 1989, p.1).

However, this is not to say that there was no valid or productive research
regarding sense of community, especially in the years since the 1996 special issue. An
individual’s level of SOC has been found to be a prime determinant in providing both
instrumental and emotional support to one’s neighbors (U nger & Wandersman, 1982) and
may act as a catalyst for both community participation (Chavis & Wandersman, 1990)
and political participation (Davidson & Cotter, 1989). The SOC construct has also been
studied in a number of diverse settings such as the workplace (Burroughs & Eby, 1998),
high schools (Royal & Rossi, 1996), community organizations (Hughey, Speer, &
Peterson, 1999), and housing for the elderly (Zaff & Devlin, 1999). In the residential-
based neighborhood setting, SOC has been documented as a quantiﬁable neighborhood

level construct which can be targeted in designing intervention and rehabilitation

15

programs for disadvantaged urban neighborhoods (Buckner, 1988; Chavis et al., 1986;
Glynn, 1981; Kingston, Mitchell, Florin, & Stevenson, 1999).

Thus, it would seem more than plausible that sense of community also taps into
the informal social processes that are necessary for a community to self-regulate and
inhibit behaviors harmful to the collective, such as crime and delinquency. Therefore, it
meets all the deﬁning criteria of social organization and should be considered as a
possible key mediating construct. However, before one can assert SOC as a possible
mediating measure for SD theory, a review of the burgeoning literature on neighborhood
effects is necessary. First, is the general model of social disorganization still valid? Is it
receiving empirical support in current studies? If so, how is the level of social
organization/disorganization being measured? The following review will trace the
development of social disorganization theory over the past two decades and answer these
pertinent questions.

The Re-Emergence of Social Disorganization

Social disorganization theory was prominent in the middle part of this century and
has once again re-emerged as the prominent social theory in explaining the role
neighborhood characteristics play in explaining variations in crime and delinquency rates
(Bursik, 1988; Bursik and Grasmick, 1993). In fact, there have been a number of recent
neighborhood studies that utilized components of this theory to understand the
importance of contextual effects on delinquency (Elliot et al., 1996; Sampson & Groves,
1989; Sampson, Raudenbush, & Earls, 1997; Simcha-Fagan & Schwartz, 1986; Simons,

Johnson, Beaman, Conger, & Whitbeck, 1996).

16

One of the main criticisms of the social disorganization model, which was
recently addressed, was the lack of empirical support for the proposed intervening
relationship between neighborhood disadvantage and crime. This was due, in large part,
to the aforementioned problems in measuring community level constructs and on the
over-reliance on census data. While census data provides a large amount of demographic
information, it does not provide measures for the proposed mediating variables of
neighborhood composition. SES, population turnover and heterogeneity are easily
quantiﬁed through census data, but measurements of the social organization of a
neighborhood required costly and time-consuming survey and interview collection
(Bursik, 1988; Heitgard & Bursik, 1987; Sampson, 1993).

Simcha-Fagan & Schwartz’s (1986) study marked the return of social
disorganization theory by answering the long-standing criticism that, although
theoretically popular, it had never been empirically demonstrated in a large-scale study.
The main ﬁnding of this study was that there were two neighborhood level constructs,
proposed by social theories of delinquency, which mediated the relationship between
neighborhood disadvantage and crime. The authors operationalized this disorganization
or neighborhood disadvantage by measuring the individual level of organizational
participation in neighborhood activities and programs. The second neighborhood level
construct, often labeled subcultural theory, was operationalized by measuring the extent
of disorder-criminal subculture. Speciﬁcally, the authors found that a community’s level
of organizational participation and extent of disorder-criminal subculture accounted for
52% of self-reported and 80% of ofﬁcially recorded delinquency. In other words, the

community effects on delinquency were largely mediated by socialization experiences.

17

In addition, Simcha-Fagan & Schwartz found that the level of organizational
participation was better at predicting self-reported delinquency while extent of disorder-
criminal subculture was better at predicting ofﬁcially recorded delinquency. This ﬁnding
reﬂected the prevalent criticism that ofﬁcially recorded delinquency is biased in that
police often concentrate their efforts at apprehending individuals in high poverty areas
(Sampson, 1986). This often cited study was the ﬁrst to empirically document support
for the intervening variables of social disorganization theory.

Sampson and Groves (1989) provided an even more direct test on the mediating
variables in social disorganization theory. Due to the inherent problems of testing
macrosocial intervening variables via census data, these authors utilized national survey
data from Great Britain and Wales. The unique design of the British Crime Survey
allowed the authors to create measures of both social disorganization and crime and
victimization rates for more than 200 local communities. They hypothesized that
neighborhoods characterized by low SES, ethnic heterogeneity, residential mobility, and
family disruption led to community social disorganization which, in turn, led to increased
crime and delinquency. They measured a community’s level of social organization in
terms of local ﬁiendship networks, control of street-corner teenage peer groups, and
prevalence of organizational participation.

The results supported the hypothesis that variations in the social organization of
local communities mediated much of the effects of demographic community
characteristics (SES, residential mobility, etc.) on crime and victimization. Speciﬁcally,
the level of unsupervised teenage peer groups, a proxy measure for informal control, had

the largest mediating effect on rates of personal violence and victimization. One

18

interesting ﬁnding was that organizational participation proved to be the weakest
mediating variable in contrast to the ﬁndings from Simcha-Fagan and Schwartz’s (1986)
landmark study.

While providing signiﬁcant support for social disorganization theory, there were
some limitations to this study. First, each of the proposed mediating variables was
assessed by single items and, thus, suffered from a lack of reliability. Further, the
systerrric concept of social organization is difﬁcult to measure and one of the many
reasons for the lack of empirical support of social disorganization theory. It is not likely
that social organization can be accurately measured through the use of single survey
items. Therefore, given this limitation, signiﬁcant ﬁndings from a culture outside the
United States palpably displayed both the validity and generalizability of the theory. In
fact, the authors were well aware of this limitation and in their conclusion suggested that
better measures of social disorganization needed to be created to thoroughly and
accurately measure this systemic concept:

And most important, we believe that the ability to measure dimensions of social

disorganization at the community level represents an essential ﬁrst step in directly

testing macrosocial control theory... We therefore hope that future research will
improve on the present effort by directing attention toward more precise measures
of the salient dimensions of community social disorganization (Sampson &

Groves, 1989, p.800).

A recent multi-site study answered this plea for a more comprehensive and
systemic measure of the social organization in a community (Elliott et al., 1996). The

study’s neighborhood contextual model not only allowed for further empirical assessment

19

of social organization as a mediating variable, but also assessed the impact of this
neighborhood level construct on individual behavior and development. Speciﬁcally, the
authors hypothesized that the relationship between neighborhood disadvantage and crime
is mediated by neighborhood organization and culture as proposed by social
disorganization theory. Neighborhood organization and culture was measured in terms of
the level of social integration, informal control, and prevalence of informal networks.

The dependent variable consisted of three individual-level outcomes — prosocial
competence, conventional friends, and problem behavior [delinquency].

Once again, results indicated support for an intervening relationship and the
authors utilized hierarchical linear modeling to display which component of
neighborhood organization was the strongest predictor of the three outcomes at the
individual level. Informal control was the only component of neighborhood organization
that was a signiﬁcant predictor for all three developmental outcomes at both sites
(Chicago and Denver). Social integration signiﬁcantly predicted the outcome of
prosocial competence, but only in Chicago. In a related manner, the mediating construct
of informal networks signiﬁcantly predicted the level of conventional friends only in the
Denver sample.

This study continued the trend of more sophisticated neighborhood level analyses
on the mediating variables pr0posed by social disorganization theory. This study was
unique in many ways. Unlike other criminological endeavors, the goal of present-day
neighborhood studies is to identify protective or resiliency factors, as well as impeding or
risk factors that can be targeted through prevention and intervention programs.

Therefore, positive behaviors like conventional ﬁiends and prosocial competence were

20

measured in addition to problem behavior [delinquency]. Further, this was the ﬁrst study
that utilized a mediating model of social disorganization which assessed the inﬂuence of
neighborhood level variables on individual development and behavior through
hierarchical linear modeling (I-ILM).

This study also addressed, albeit incompletely, the criticism that the majority of
social disorganization or ND studies viewed, or at least analyzed, neighborhoods as if
they were static structures (Bursik, 1988; Sampson, 1993). This valid critique has
disturbed many in the ﬁeld because social disorganization theory attempted to explain the
rapidly changing demographic characteristics typical of urban areas that resulted in
increased crime and delinquency rates. Obviously, there are serious ﬁnancial and time
constraints which preclude longitudinal analyses which could assess such an impact of
change in a neighborhood. Therefore, every attempt should be made to at least include
multi-site studies so comparisons can be made across communities. For example, the
results of this study indicated that there were indeed differences between Chicago and
Denver and that these differences need to be taken into consideration before
implementing any type of intervention program.

The largest and most recent neighborhood study included over 8,000 participants
representing 343 neighborhoods in Chicago (Sampson, Raudenbush, & Earls, 1997).
This study’s main focus was to explicate the variables responsible for mediating the
relationship between neighborhood disadvantage and violence rates. Once again, the
mediating variable proposed was derived from social disorganization theory: “Our basic
premise is that social and organizational characteristics of neighborhoods explain

variations in crime rates that are not solely attributable to the aggregated demographic

21

characteristics of individuals” (Sampson, Raudenbush, & Earls, 1997, p. 918). These
authors measured social organization via informal control (5 items) and social
cohesiveness and trust (5 items). Since these two subscales were signiﬁcantly correlated,
they were combined to create a summary construct the authors labeled collective efficacy
(CE). There were three different measures of violence in this study. Perceived violence
and personal victimization were measured through survey items, and, for a comparison
statistic, ofﬁcially recorded measures of horrricide were also included. The authors’
speciﬁc model was that the relationship between neighborhood disadvantage and
violence rates would be mediated by collective efﬁcacy. Neighborhood disadvantage
was measured by concentrated disadvantage (low SES), immigrant concentration
(population heterogeneity), and residential stability (population turnover).

These demographic or compositional characteristics signiﬁcantly explained over
70% of the variability in perceived violence rates across neighborhoods. However, CE
proved to mediate this relationship as both population turnover and population
heterogeneity reduced to nonsigniﬁcant levels and low SES’s effects were substantially
reduced. CE also demonstrated discrirrrinant validity by proving to be a stronger
predictor than other social processes such as ﬁiendship and kinship ties, neighborhood
services, and organizational participation.

Reformulated Svgemicﬂmmwm

In summary, updated social disorganization models have received recent
empirical support and ﬁndings from the two most recent studies indicated that one of the
more important components of this mediating construct was the level of informal control

in the neighborhood (see Figure 1). It would seem logical that the informal social control

22

variable would tap into the ability of the community to realize its common values and self
regulate behavior that would be harmful to the collective. In fact, there has been
extensive discussion on the important role of informal social control in controlling crime
and delinquency (Greenberg & Robe, 1985; Sampson, 1997). Adding momentum to
these empirical ﬁndings, Bursik & Grasnrick (1993) published a seminal book that
utilized social disorganization theory to focus on the importance of informal relations and
social control across multiple neighborhood levels.

This book presented an updated social disorganization model which also
addressed one of the strongest critiques of social disorganization or neighborhood
disadvantage in that it often neglected important external factors that inﬂuenced
neighborhood social organization. Before discussing the speciﬁcs of the book, it is
necessary to ﬁrst illustrate the strength of extralocal community factors and how they can
disrupt the natural process of community ties and cohesion. The following example from
Chicago will display both the strength and importance of accounting for all factors
outside the community which have an impact on a neighborhood’s level of social
organization. After the example from Chicago, a discussion of this updated social
disorganization model will be presented with a focus on how it utilizes multiple levels of
neighborhood units to account for extralocal community forces and to explain the process
of informal social control in the local community.

Wm

Logan and Molotch (1987) are perhaps the most well known of old Chicago-

school critics who felt ecologists grossly underestimated the undemocratic and unnatural

processes that occur in the distribution of people and resources within neighborhoods

23

across cities in the United States. As they discuss in Ueran Fortunes, there are many
political, economic, and social forces which inﬂuenced the growth or demise of certain
areas in a metropolis. For instance, on the South Side of Chicago the African-American
population was isolated by discrirrrinatory housing policies in a small area known as the
Black Belt until 1948 when the Supreme Court struck down race-restrictive covenants in
housing policies. As African-Americans slowly gained greater mobility and choice in
housing, there was tremendous opposition by the surrounding ethnic white
neighborhoods which resulted in numerous racial conﬂicts, real estate speculation, and
white ﬂight to the suburbs (Bursik and Webb, 1982; Hirsch, 1998). Obviously, Burgess’s
concentric zone theory and invasion/succession principles captured in social
disorganization theory were unable to explain the inﬂuence of these ‘unnatural’ forces in
neighborhood change. The result was that Shaw and McKay’s (1962) important
theoretical ﬁnding that delinquency rates remained relatively stable despite continuous
changes in the racial and ethnic composition of the community was no longer valid.

However, as Bursik and Webb demonstrated (1982) in a re-analysis of the
original data, social disorganization still maintained validity if one looked at the nature
and pace of change in the racial and ethnic composition which now dramatically differed
from the past. Community change in racial composition in this era was characterized as
foothold change, turnover change, or entrenchment change. Foothold change represented
an increase of families of color in a neighborhood by at least 10% over ten years, but
where they remained in the minority, while turnover change represented an increase of at
least 10% but where families of color moved from minority to majority status.

Entrenchment change occurred when families of color increased their composition by

24

10% and were in the majority in the community in all ten years. The main ﬁnding of this
re-analysis was that delinquency rates signiﬁcantly increased for foothold and turnover
change but remained stable for entrenchment change: “The most established nonwhite,
changing communities [entrenchment areas] had delinquency rates not much different
than would have been expected from their previous patterns” (Bursik and Webb, 1982,
p.39).

So, the variation in delinquency patterns reﬂected the nature of racial change
rather than the speciﬁc groups involved. The incredible patterns of population turnover
during this era excluded the development of an institutional and social network base that
was responsible for keeping delinquency rates stable among Chicago neighborhoods up
to 1950. Therefore, social disorganization theory still retained validity if one analyzed
the external social forces ( e.g., discriminatory housing practices, racism, white ﬂight)
that dramatically altered the gradual process of invasion/succession for African
Americans since 1950. Updated social disorganization models need to be able to explain
these important political, social, and econorrric forces both inside and outside the
neighborhood that affect neighborhood composition and development as evidenced by
the example of Chicago. Some of these recent external forces which have altered the
growth or decline of our cities include: loss of traditional employment opportunities in
industrial cities and increasing concentration of the extremely poor in the central city
(Jargowsky, 1997; Jargowsky & Bane, 1991;Wilson, 1996); erosion of the local tax base
which has decreased the ability of cities to provide essential public services (Bursik &
Grasmick, 1993); and the impact of pro-business incentives and discriminatory housing

policies (Hirsch, 1998; Molotch, 1987).

25

An U ted and S stemic Model

Bursik and Grasrrrick’s systemic theory of social disorganization utilizes Hunter’s
(1985) three-level approach of community control to connect the neighborhood resident
to these external forces. This approach begins with the most immediate geographical
area of the resident and extends outward to include an increasingly larger and larger area.
The ﬁrst level of control is termed the private level of community control. It represents
the intimate yet informal relations that develop over time directly surrounding the
resident’s dwelling. These are the type of relationships that develop among neighbors
who provide instrumental support, such as the lending of tools or direct assistance, and
the emotional attachment and support that comes from day-to-day contact. In this
conceptualization of the neighborhood, shared norms exist about the appropriateness of
public behavior and, if broken, can result in direct criticism, ridicule or ostracism from
your fellow neighbors (Black, 1989).

The second level of community control is called the parochial level and extends
from the geographical area immediately surrounding the resident’s dwelling outward to
include other nearby areas and the institutions and resources located there such as
schools, social service agencies, churches, parks, and libraries. The parochial level of
community captures the social attachments among close neighbors as well as the
relationships that are not so well established or relationships with those individuals whose
geography results in less daily or weekly contact. This level also explains the more
formal aspect of relationships dictated by community institutions such as the local school

or market exchanges such as those that occur at the local grocery or dry cleaners.

26

The public level of control is the third tier and extends from the private level
outward to explain relationships to other outside communities and the corresponding
political, economic, social, and even global forces. These include local, state, and
national governmental policies, cultural forces, land-controlling elite, and other external
actors who affect the organization of a community at both the parochial and private
levels. The Chicago example displayed how strong an impact these extralocal forces can
exert in local community dynamics, and unfortunately, these external forces have
dramatically altered the urban landscape over the last several decades.

Although the traditional focus of social disorganization theories has been to
explain delinquency and crime rates, reformulated models have widened this focus to
include a host of other individual outcomes such as child maltreatment rates (Coulton,
Korbin, Chow, & Su, 1995; Coulton, Korbin, & Su, 1996), academic achievement
(Gonzales, Cauce, Friedman, & Mason, 1996), interpersonal aggression (Grifﬁn, Scheier,
Botvin, Diaz, & Miller, 1999), psychological distress (Simons, Johnson, Beaman,
Conger, & Whitbeck, 1996), and fear of crime (Perkins & Taylor, 1996; Taylor, 1984).
The importance of Taylor’s research resides in his attempt to reign in the overutilization
of census data and national samples and focus on a much smaller yet more valid unit, the
block. It must be acknowledged that the current study is, in many ways, a response to
Taylor’s (1997) article which called for a further exploration and investigation into the
private and parochial levels of control. The current study follows this prescription to
further assess the importance of smaller neighborhood units and their inﬂuence on both

positive and negative youth outcomes.

27

Before the details of the current study are discussed, it is important to reiterate
that updated social disorganization models found signiﬁcant effects for measures of
social organization other than informal control such as: organizational participation
(Simcha-Fagan & Schwartz, 1986; Sampson & Groves, 1989), and informal networks
and social integration (Elliot et al., 1996). Neighborhood social organization is a
systemic concept that involves many transactional relationships with direct and indirect
effects. Thus, even the private and parochial levels of community control, the more basic
levels, cannot be characterized by just one dimension like informal control. Moreover,
the private level of community control is usually enforced through unmeasurable actions
such as possible withdrawal of support and respect, direct criticism and ridicule, and
ostracism (Black, 1989; Hunter, 1985).

The results of previous studies on the importance of informal control may simply
be due to the fact that the ﬁeld has been constrained to measuring the behavioral
components or outcomes of social organization. Yet, this criticism seems minor and
perhaps unwarranted since it was not long ago (i.e., 1986) that studies ﬁrst demonstrated
empirical support for the basic mediating variables of social disorganization. When you
consider that the theory dates back to the middle and earlier parts of this century, it is not
surprising that the model has not been fully delineated. The re-emergence of social
disorganization and its reformulated systemic models are still in an embryonic state.
While conceptual and methodological improvements (HLM) have advanced theory and
empirical support, the ﬁeld is still struggling for a complete and adequate measure of

neighborhood social organization.

28

The CuIrrent Studv

There were three main research objectives for the current study. The ﬁrst
research issue was to evaluate the dimensionality of the sense of community (SOC) and
self-reported delinquency (SRD) constructs. As previously mentioned, there is still
debate in the literature on whether SOC is a unidirnensional or multifaceted construct
(Hill, 1996). Also, the SOC construct for this study was originally intended to tap into
six components although results from pilot studies have only supported a three-factor
solution. Thus, examination of the factor structure of SOC construct will add to the
growing literature and also direct possible modiﬁcations of the measure for future
projects by the SOC in Lansing Neighborhoods research team. Since the self-reported
delinquency scale utilized for this study tapped into minor, medium, and major offenses,
a principal components analysis will verify if youth evidenced differential grouping by
these severity levels.

The second goal of the current study was to evaluate the validity of a block-level
conceptualization of neighborhood. For instance, are urban blocks large enough to
represent a neighborhood? Do blocks within the same administrative unit such as a
census tract display differential levels of SOC? While comparisons across neighborhood
units can be made, it must be noted that the main focus of this study is the face-block
conceptualization of neighborhood and how it corresponds to the private level of
community control as postulated by Bursik and Grasmick (1993).

The third research goal of the current study is to ascertain if SOC mediated the
relationship between neighborhood disadvantage and youth outcome variables such as

delinquency, conventional activities, and average grade. It was hypothesized high levels

29

of neighborhood disadvantage will lead to lower levels of SOC and consequently higher
levels of negative youth outcomes. Conversely, low levels of neighborhood disadvantage
(neighborhood advantage) will lead to higher levels of SOC and result in higher levels of
positive youth outcomes. Related to this research objective, the systemic nature of this
construct and its potential for application in future studies and neighborhood

developmental efforts will be discussed.

30

Figurel

Signiﬁcant Findings of Mediating Variables From Previou_s_ Neighborhood Disadvantage
Studies

  
 
  

   
  
   
  

  
    
 
 

  
   
     
 

Measures of Delinquency,
Neighborhood Neighborhood Violence Rates,
Disadvantage Social & Related

Organization Outcomes

 

1) Organizational Participation (1986 & 1989)

2) Community Disorder-Criminal Subculture (1986)
3) Local Friendship Networks (1989)

4) Control of Street Corner Youth (1989)

5) Social Integration (1996)

6) Informal Control (1996)

7) Collective Efficacy (CE) (1997)
CE = Social cohesiveness and informal control

 

 

 

31

Methodology
Procedures

me

This research study was conducted in the city of Lansing, Michigan. Lansing is
the state capital and an industrial city with a population of approximately 127,000. The
Sense of Community in Lansing Neighborhoods Project ofﬁce is located at Michigan
State University in East Lansing. All interviews were conducted in either the homes or
on the front porches of the respondents.
Participants

Participants were referred from the local school district. There were 103 tenth
grade males who participated in the study along with one of their parents and one of their
neighbors. Therefore, for each student there was a total of three interviews (student,
parent, and neighbor) for a combined total of 309. Of the 103 youth, 41% were White,
40% were Aﬁican American, 11% were Hispanic, 6% Asian, and 3% mixed race. As can
be seen in Table 1, student respondents closely matched the racial/ethnic percentages of
Lansing High School students.

Primary caretakers and neighbors had to be over the age of 18. Only one adult
was surveyed at each household to avoid the potential bias or inﬂuence of others.
Of the 206 participating adults, 67% were female and 67% owned their own homes. The
average length of time on the block was approximately 10 years and covered a range
from three months to 39 years. An interpreter was utilized to increase Asian respondents
since a large percentage were of Hmong descent and many of these youths’ parents spoke

little or no English.

32

Biruitment Procedggs

This research project was conducted with assistance from the Lansing school
district ofﬁce of research and evaluation services. There was a three-step consent
procedure employed to assure that no parent, youth, or neighbor felt coerced to
participate. First, the school district provided the research team with 300 mailing labels
which were sent out from the school district’s own ofﬁce. The SOC research team
mailed an initial letter to these 300 parents which explained the rationale and goals of the
study and contained a request for participants (see Appendix A). A self-addressed,
stamped postcard was included and parents were instructed to return this postcard if they
did not want their youth to be included in a pool of possible research participants. The
phone number to the SOC project office was also included and parents were encouraged
to call if they had any questions, comments, or concerns. Once an adequate time had
passed from this initial mailing (3 months), the Lansing school district ofﬁce of research
and evaluation services provided the SOC project team with the names and addresses of
235 parents for an initial 22% refusal rate. This complied with local school district
policy, and again, was only the ﬁrst of a three-step procedure.

For the second component of the recruitment procedure, trained interviewers
traveled to the youth’s home to explain the study to the youth’s primary caretaker and ask
if they would like to participate. A follow-up letter with the interviewer’s name was left
if the primary caregiver was not home or too busy to discuss the study at that time (see
Appendix B). If a primary caregiver was home and agreed to participate, the interviewer
reviewed the consent form, obtained a signature, and interviewed the parent either at that

time or set up an appointment for an interview in the future. At this point of the consent

33

procedure, an additional 41 students (17%) were dropped because their parents refused to
participate. Also, 49 students (21%) were dropped for various reasons including
incorrect addresses, if a farrrily moved, or if they did not live on a street where a face-
block could be approximated.

In the third component of the consent procedure, each youth was contacted during
the initial parent visit, or at a follow-up visit, and asked if they would like to participate.
If so, the interviewer reviewed the project assent form (see Appendix C), obtained a
signature, and either interviewed the youth then or set up a time for a future interview.
So, even if parents signed written consent, only those youths that also assented on their
own were included in the study. Of all youth that were approached, only one refused to
participate and was dropped ﬁom the study. Neighbors were approached via a random
sampling procedure and in the same manner as the primary caretaker although only
verbal consent was needed for their participation. There were three requirements for a
neighbor’s participation. They had to be at least 18, agree to participate, and live on the
same face-block as the youth.

This three-step consent procedure required a substantial amount of travel time,
particularly with hard-to-reach families, but increased the pool of participants by not
assuming all families had a working phone. Also, the project’s refusal rate was
substantially lower than if a telephone interview format had been utilized. One hundred
and three families participated out of a possible 186, yielding a 55% response rate. See

ﬁgure 2 for a graphical summary of the three-step participant recruitment strategy.

34

Interviewer Training and Sumrvision

Upper-level, undergraduate psychology students received college credits for
participating as part of an independent study course at the University. Initial interviewers
were trained for one and a half months prior to the data collection process in a class
format which was similar in nature to that of a graduate level seminar. First, interviewers
were introduced to seminal works in the area of Sense of Community and delinquency.
Second, interviewers were responsible for reading and discussing the recent research on
neighborhood studies that incorporated social disorganization theory. Third, interviewers
were exposed to methodological issues in community research with an emphasis on
interview techniques. Finally, and most importantly, interviewers reviewed the various
survey instruments and practiced role-playing interviews until there was adequate
consistency in their administration of the survey. Particular attention was paid to the
youth interviews and how to build rapport, trust, and explain, in terms that the youths
could understand, the meaning of conﬁdentiality. Interviewers that were recruited after
this training period received an abbreviated version of this training course. All
interviewers were accompanied by a project staff member for their ﬁrst two interviews.

Also, the project team developed an extensive interviewer training manual which
detailed the multiple issues that interviewers may confront and strategies to overcome
these potential pitfalls. Approximately 6% of the interviews were coded by two
interviewers to assess interrater reliability. Of a total of 868 items, only six were
inconsistently marked for an overall agreement rate greater than 99%.

During the data collection phase, interviewers attempted to conduct eight hours of

interviews per week. Interviewers were required to go out on different days and times

35

but were instructed to focus their time during the early evening hours and on weekends.
Interviewers also kept a logbook that consisted of the dates, days, and times of attempted
contacts in order to facilitate face-to-face contact of all households. More importantly,
no household was excluded because it was difﬁcult to establish contact, regardless of the
number of times an interviewer had to return. Households with an invalid address (e.g.,
abandoned, family moved) were returned to the Lansing school district for the correct or
updated address. Interviewers also met one time per week for a two-hour supervisory
group meeting during the ﬁrst month and a half of data collection. After this,
interviewers met individually for 30-minute sessions with the SOC project coordinator.
Project staff were also always available to interviewers for help or feedback as often as
needed.
Interview Procedure

Trained interviewers traveled to the youth’s homes to conduct the interviews for
the study. First, interviewers introduced themselves to the primary caretaker as a project
member on the Sense of Community in Lansing Neighborhoods Project team. Each
interviewer had a Michigan State University ID card displayed along with the SOC
project logo. Interviewers provided a brief scripted introduction and description of the
study (see Appendix D) prior to obtaining consent. Again, written consent was obtained
from the primary caretaker, written assent from the youth, and verbal consent was
obtained from the neighbor prior to interviews.

gage-Block Measures
This study differed from all previous neighborhood studies based on social

disorganization theory by directly measuring the proposed independent variables of

36

income, residential stability, cultural homogeneity, and family composition. Previous
studies have all utilized some combination of census indicators or other administrative
unit data to derive these measures. In general, income or SES has been measured in
numerous ways that are representative of social science research. Most prior studies have
summed and standardized a number of measures within a census tract such as: percentage
of families below the poverty line, percentage of residents employed in professional or
managerial positions, percentage of college educated residents, average income or
housing value, etc. Residential stability has most often been calculated as the proportion
of families that have moved in the last ﬁve years in a census tract. Cultural homogeneity
has usually been measured by the presence and number of race/ethnic groups greater than
10% in a census tract. Finally, family composition has been measured in the past by the
proportion of single parent families within a census tract.

Another major difference in the current study was the way in which the
independent variables were operationalized. The independent variables were
conceptualized and measured on a level of advantage versus the traditional method of
focusing on levels of disadvantage. It is argued that this modiﬁcation was more than just
a matter of semantics. Researchers need to be more sensitive to the communities under
study as what is being investigated surround issues of safety, quality of life, and possible
life outcomes of youth in these communities. By conceptualizing and measuring these
neighborhoods in a positive manner, the dialogue can change from one focused on
deﬁcits to one focused on investigating and explaining the various strengths of these

neighborhoods and the people who live within them.

37

Mum

Income was individually calculated by dividing monthly income by the number of
residents in the house. To obtain a block income measure, the parent and neighbor’s
income were simply averaged.
Residential Stability

Individual scale scores (e. g., parent, neighbor) were calculated by summing four
likert items such as: (1) People move in and out of this block a lot —- reverse coded. Block
scores were obtained by averaging the parent and neighbor’s total scale score.
Respondents were asked to rate from 5 (strongly agree) to 1 (strongly disagree) their
endorsement of the four items. Reliability was .79 for parents, .70 for the neighbor, and
.84 for the block. See Table 2 for all residential stability items and the corrected item-
total correlations at the individual level.
Homogeneig

Individual scale scores were calculated by summing two likert items on a ﬁve-
point scale ﬁom 5 (strongly agree) to 1 (strongly disagree). Parent and neighbor’s scores
were again averaged for the face-block score. Items included: (1) The residents on this
block are from my racial/ethnic group, and (2) This block is racially/ethnically diverse —
reverse coded.
WG—11

Block composition was simply calculated as the percentage of two ‘parent’
households based on a family structure item. The item asked the parent and neighbor if
they had a spouse or live-in partner. Thus, the variable of interest in this study was if

there were two adults living in the house, regardless of whether these individuals were

38

married or if they were the birthparents. Each block could fall into one of three
categories: 100% two-parent households, 50% two-parent households, or 0% two-parent
households.

Neighborhood advantage (NA)

This composite measure was created by standardizing and summing the four
independent variables as proposed by social disorganization theory: income, stability,
homogeneity, and composition. Each score on one of these independent variables
indicated the degree of advantage shared by residents on that face-block compared to
residents on other blocks for that particular variable or category (i.e., block income or
block stability). All negatively worded items were reverse coded prior to analyses.
Therefore, this composite measure summarized the overall level of advantage for the
face-block. Thus, a high block score indicated a greater degree of structural advantage
and a low score was indicative of a disadvantaged neighborhood.

Sense of community (SOCtotal)

This scale was created to measure the sense of community that exists on
residential face-blocks within an urban community. The SOC in Lansing
Neighborhood’s Project team originally constructed a six component measure that
incorporated 24 items, many of which were from previously published SOC measures.
The hypothesized six components include items related to: connection, belonging,
support, participation, empowerment, and safety. See Appendix E for a copy of this
measure.

Parents and neighbors of the targeted youth were asked to respond to this measure

and their scores were averaged to construct the SOCtotal measure for the face-block of

39

each youth. Respondents were asked to rate from 5 (strongly agree) to 1 (strongly
disagree) their endorsement of the SOC items. Examples include: (1) People on this
block socialize with each other, (2) If faced with a problem on the block, residents would
be unable to create a solution - reverse coded, and (3) People on this block feel they
belong here. All negatively worded items were reverse coded prior to analyses. Alpha
was .93 for the parent survey, .92 for the neighbor survey, and .86 for the block.
Qdividual Youth Mm

There are ﬁve elements to the youth survey measure: (1) self-reported
delinquency, (2) conventional friends, (3) delinquent peers, (4) conventional activity, and
(5) average grade. The vast majority of the items from the youth survey were utilized in
Elliot et al.’s (1996) study on neighborhood effects although there were some
modiﬁcations to the delinquency and drug use sections. See Appendix F for a copy of
the youth survey.

Self-Report Delinquency (SRDtotal)
This is a basic self-report delinquency scale that incorporated 17 items. The

 

questions survey a range of delinquent behaviors from minor offenses (e. g., stolen or
tried to steal something worth $5 or less?) to major infractions of the juvenile code (e. g.,
attacked someone with the idea of seriously hurting him or her?). Youth were asked to
respond by the number of times in the last year they had performed the speciﬁc
behaviors. Alpha was .66.
Conventional Friends_(_(_3£)

This scale consisted of three questions on prosocial ﬁiends. The introductory

preface before each question was “How many of your ﬁiends ” and youth responded

40

using a ﬁve-point scale from 5 (all of them) to 1 (none of them). An example includes:
(1) How many of your ﬁiends get good grades in school? Alpha was .60.
Delinquent Peers (DP)

This scale combined ﬁve questions on delinquent friends and utilized the same
introductory sentence and response scale as the conventional ﬁiend’s measure. An
example includes: (1) How many of your ﬁiends use hard drugs? Alpha was .66.
Conventional Activity (CA)

This outcome variable was measured with one item which directly asked the
youth for the total number (if any) of school activities, clubs, sport teams, etc. of which
he was a member in the past year.

Average Grade
This outcome variable was measured by a question which directly asked for the

youth’s current grade point average.

41

Table 1

Race/Ethnic Backgtound of Lansing High School Students & Research Participants

 

 

Race/Ethnic Lansing H.S. Research
Category Students Participants
White 45% 41%
Aﬁican-American 36% 40%
Hispanic 12% 11%
Asian 7% 6%
Mixed Race NA." 3%
Native American 1% 0%

 

Note. Race/Ethnic background of Lansing H.S. students was obtained from the Lansing
school district office of research and evaluation services. Percentages are rounded so can

sum to over 100%.
8NA. = Not available. Mixed race/ethnicity is not an option on school demographic

fonrrs.

42

Figure 2

Research Participant Recruitment Flowchart

 

Initial refusal rate

 

 

 

 

Miscellaneous drops:

 

 

22% m=65) <—
i
p=235
Conngzgﬁjs-‘talgmte l "
p=f86
- p=103
55% Response rate

43

’ wrong address, age,
gender, etc.
21% (r_r=49)

Table 2

Residential Stabilitv Item_s, Corrected Item-Total Correlations & Reliability Coefficients
for Parent & Nei bor

 

The majority of the residents on this block rent their houses. (R)°
When people come to live on this block, they tend pg to stay here long. (R)
People move in and out of this block a lot. (R)

I would like to live on this block for at least another ﬁve years.

 

 

Residential Stability Corrected Item- Corrected Item-
Items Total Correlation Total Correlation
Parent Neighbor

Rent their houses .60 .57

Not to stay .64 .47

Move in and out .69 .58
Another ﬁve years .50 .35

Alpha .79 .70

 

Note. Reliability at the block-level was .84.
’(R) = Reverse coded.

Results
Data Analmc Strategy

There were three main research objectives of this study. The ﬁrst was to evaluate
the dimensionality of the sense of community construct as well as the self-reported
delinquency scale. The former was accomplished by conducting a series of analyses.
First, a principal component analysis was conducted on the SOC construct but its
inconclusive results warranted further empirical evaluation of its factor structure. Thus,
reliability analyses were conducted on three theoretically-driven components of SOC,
which were also partially supported by the principal component analysis, to assess if this
construct could measure distinct community processes. A principal component analysis
was also performed on the SRD construct to ascertain if the overall scale could be broken
down into more meaningful components since the questions differed along severity
levels.

The second research goal was to evaluate the validity of a block-level
conceptualization of neighborhood. The validity of this conceptualization was assessed
in three different ways. First, descriptive statistics were run to evaluate whether or not
blocks substantially differed in their levels of neighborhood advantage (NA) and sense of
community (SOC). Second, correlations between parent and neighbor scores on the SOC
construct were run to ascertain if there was a block-level consensus. Third, discrirrrinant
validity was assessed by evaluating the pattern of relationships among NA, SOC, and the
various outcome variables.

The third research aim was to evaluate the hypothesis that SOC mediated the

relationship between neighborhood characteristics and youth outcomes. Speciﬁcally, it

'45

was hypothesized that neighborhood advantage leads to higher levels of sense of
community which, in turn, leads to more positive youth outcomes. In a related manner,
low levels of block advantage (neighborhood disadvantage) should lead to lower levels of
SOC and a greater degree of negative outcomes like delinquency, lower grade point
average, etc. Correlation and regression analyses were utilized to assess this contention.
By ﬁrst running correlations, variables and relationships that were not signiﬁcant were
dropped. This helped reduce the number of regression equations required to test for the
possible mediating role of SOC in the relationship between neighborhood advantage and
positive and negative youth outcomes. At this point in the analyses, the systemic nature
of the SOC construct was also investigated.
Mh Obiective 1: Dirnens_ir&alitv of SOC & SRD Con_structs

Principal Comppnent Analysis of the SOC Con_struct

The dimensionality of the SOC construct was ﬁrst assessed by conducting a
principal component analysis with promax rotation. Although there were six components
with eigenvalues greater than one, the scree plot indicated a three or four factor solution.
Taking into consideration prior factor analysis from a pilot study, substantive theory, and
a desire for parsimony, a three factor solution was chosen. These three factors accounted
for a total of 52% of the variance and are labeled emotion, action, and safety (see
Appendix G). However, while the PCA helped verify that SOC was indeed
multidimensional, the results also presented as many questions as it answered.

For instance, the emotion component accounted for the majority of the variance in
the construct as a whole and some of its items that were included fell below the

traditional cut-off value of .5. The safety component also contained multiple items from

46

the other two components. Therefore, it was decided to further examine the factor
structure of the instrument by conducting Cronbach’s coefﬁcient alpha on the three SOC
components for the parent, neighbor, and block-level. Block-level alpha coefﬁcients
represent reliability of SOC across the face-block, and thus, incorporated both the parent
and neighbor responses to the SOC survey. Next, a brief description of the items that
comprise these components and their reliabilities are presented. In addition to reliability
coefﬁcients, tables with item-total correlations for each component will be presented
since SOC is the major construct in the current study. See Appendix H for a copy of the
SOC measure broken down by its three components of emotion, action, and safety.

gnLtiop. The emotion component of the SOC construct contained 14 items
related to connection, belonging, and support among residents on the face-block. This
construct measured the emotional connection between residents on the block and was
representative of the affective component of SOC. Reliability was .90 for the parent, .88
for the neighbor, and .89 for the block. Table 3 lists all emotion items and corrected
item-total correlations for both the parent and neighbor surveys.

Aptio_n. The action component of the SOC construct contained eight items
pertaining to residents’ feelings of empowerment and actual participation in block-related
activities. Reliability was .80 for both the parent and neighbor and .84 for the face-block.
See Table 4 for a listing of all action items and their corrected item-total correlations at
the individual level.

§_af_ety. The safety component contained two items regarding safety issues on the
block. While the third component from the PCA included at least two items from the

other two components, only the safety items were utilized for theoretical purposes since it

47

made little sense to create a catch-all component. This is especially true with principal
component analysis since this procedure capitalizes on analyzing all of the variance of
the items, essentially equating measurement error with shared variance. Reliability was
.60 for the parent, .65 for the neighbor, and .63 for the block. Table 5 lists the two safety
items and corrected item-total correlations for both the parent and neighbor surveys.

Since the results from the principal components analysis presented some
confusion as to the exact factor structure and item composition of each factor for the SOC
construct, the results from Cronbach’s coefﬁcient alpha were utilized to determine how
the construct would be operationalized for the current study. The results from the
reliability analyses suggested that it was indeed correct and reliable to interpret the sense
of community construct as being composed of these three distinct components of emotion
action, and safety. Therefore, all subsequent analyses in the current study utilized a
SOCtotal score as well as scores for its three components.

Principal Comppnent Analysis of th_e SRD Construct

 

A principal component analysis was conducted on the 17 items from the SRDtotal
scale to investigate if there were distinct groupings of delinquent actions within this scale.
The scree plot method of deterrrrining the number of component loadings with varimax
rotation was utilized to determine the factor structure for this construct. While Kaiser’s
criterion would have resulted in seven separate components, the scree plot indicated a
deﬁnite four factor solution which captured 57% of the variance. Since some of these
components were thought to be highly correlated with each other, a PCA with promax
rotation was also conducted. The results were almost identical to the PCA with varimax

rotation, minus the order of loadings within the third component, steal and ﬁght

48

delinquency. It was decided to present and interpret the PCA with varimax rotation since
this is the most common method of rotating simple structures in factor analysis. See
Table 6 for the component loadings from the PCA.

Of the 17 items entered in the PCA, four were thrown out and were not presented
in Table 6 as they did not signiﬁcantly load on any of the four components. These items
were: (Pb 1) Gone onto someone’s land or into someone’s house or building when you
weren’t suppose to be there?, (Pb8) Been stopped and questioned by police but not
arrested?, (Pb14) Carried some sort of hidden weapon (gun, knife, razor blade, etc.) in
school?, and (Pb20) Been involved in gang ﬁghts? All subsequent analyses utilized a
SRDtotal score as well as scores for each of the four components. The following section
provides a quick description of the items that comprise each construct along with their
reliabilities.

Stial & deal delinquency(SDDel.). This component contained ﬁve items. Four
of these items related to stealing and one asked if the participant had sold drugs. Alpha
was .69.

School delinquency (SchmlDeD. This component contained three items. One
item asked about skipping classes while in school and another item asked about skipping
a ﬁrll day of school. The ﬁnal item in this component asked about carrying a weapon
outside of school. The ﬁrst two items are status offenses and indicative of “minor”
delinquency while the third would be a crime if committed by an adult. Alpha was .87.

W. This component was comprised of three
items. Two of these items related to stealing and the ﬁnal item asked if the respondent

had fought someone physically in the past year. Alpha was .48.

49

Severe delinquency (Seve_reDel.). Two items constituted this component of self-
reported delinquency. These two items were on the severe end of the scale and asked the
participant if they had attacked someone in the past year, while the other item asked if
they used force (may have included a weapon) to take money or things from someone
else. Alpha was .52.

Research Obiective 2: Evaluating the Validig of the Face-Block as a Neighborhood Unit

Descriptive analyses were run to evaluate if there was adequate variance between

 

 

blocks on both the independent (NA) and mediating (SOC) variables of social
disorganization. This was the ﬁrst of a three-step procedure to establish the validity of
the conceptualization of the urban face-block as a neighborhood entity. As Table 7
illustrates, results indicated that income, stability, homogeneity, composition, and the
composite measure of neighborhood advantage vary across the different face-blocks
utilized for the study. For instance, the mean of residential stability was 14 with a
standard deviation of 3.48. The rrrinimum rating given by a block was 6 while the
maximum block rating was 20, which was the highest rating possible. The mediating
variable of SOC and its three components also displayed adequate ranges. Emotion had a
mean of 48, a standard deviation of 7.27, and covered a range from 33 to 63. SOCtotal
and its other two components displayed similar variability across the face-blocks.

Thus, the next step of the validation procedure entailed correlating parent and
neighbor responses to SOCtotal and its three components to assess if there was a block-
level consensus on the degree of SOC within the neighborhood (see Table 8). Both the
uncorrected and corrected correlation coefﬁcients are presented to display the effect of

measurement error on block-level correlations for the SOC construct. Since the measures

50

utilized in the study were reliable, coefﬁcients do not change dramatically when
corrected for attenuation. As Table 8 also displays, correlation coefﬁcients were all
signiﬁcant and ranged from small to moderate, with .21 at the lower end and .37 at the
upper end. The least amount of agreement was with the emotion component while the
highest level of agreement was with the action component.

For the third step, correlations among NA, SOC, and the outcome variables were
conducted to assess discriminant validity (see Table 9). Overall, patterns among these
constructs conformed to the prediction that high levels of NA and SOC would be related,
in direction and magnitude, to positive youth outcomes while low levels of these
variables would be related to negative outcomes. For instance, in the case of grade point
average, NA, SOC, and all of their components were positively correlated with this
variable and coefﬁcients ranged ﬁ'om .02 to .23. However, there were a couple of
instances where relationships did not conform to expectations. In the case of one of the
delinquency components, steal and deal, both NA and SOC were positively related to this
outcome variable. The other case where there was a discrepancy also came from the
delinquency construct. For steal and ﬁght, most of the NA components were negatively
related while the SOC components were all positively related. Aside from these two
deviations, discrirrrinant validity was conﬁrmed in the case of the other two delinquency
components, delinquent peers, conventional ﬁiends, conventional activity, and grade

point average.

51

Research Obiective 3: Sense of Community (SOC) as a Mediator of Neighborhood
Advantage (NA) & Youth Outcomes

Correlation analyses of the independent, mediating, and dependent variables were

 

conducted to reduce the total number of regression equations needed in order to establish
a mediating model. The main ﬁnding from this initial correlation matrix was that there
was no relationship between neighborhood advantage and the outcome variables (see
Table 9). Therefore, it was impossible to test the proposed hypothesis that sense of
community mediated this relationship. However, since there were multiple indicators for
each of the main constructs, it was possible to look at their components and evaluate if
there were any mediating relationships in the implied submodel (see Figure 3).

According to Baron and Kenny (1986), three regression equations are required to
test for mediating relationships. In the current study, neighborhood advantage was the
independent variable and sense of community was the mediating variable. The
dependent variable consisted of positive and negative youth outcomes such as
delinquency and conventional activity. For the ﬁrst regression equation, the IV must
signiﬁcantly predict the MV. Second, the MV must be a signiﬁcant predictor of the DV.
Finally, when the DV is regressed on both the IV and MV, the relationship between the
DV and MV (SOC) must remain signiﬁcant while the relationship between the DV and
IV (NA) must reduce to a nonsigniﬁcant level.

Therefore for organizational purposes, results for the third research objective will
be explained in a step-wise fashion following Baron and Kenny’s method. First, the
relationship between the independent (IV) and mediating variables (MV) and their

components were evaluated via correlation analyses. Again, if there was an insigniﬁcant

52

correlation at this stage of analysis, the variables can be dropped or ignored, as there is no
possibility of a mediating relationship. For variables that exhibit signiﬁcant correlations,
the second step consisted of adding the dependent variables (DV) and their components
to the model. If signiﬁcant correlations continued to emerge at this stage, the third step
required running the combined or multivariate regression equation.

The multivariate regression equation assessed both the independent and mediating
variables’ ability to predict the dependent variable. For a signiﬁcant intervening
relationship to exist, the mediator (SOC) must have retained a signiﬁcant effect while the

independent variable’s (NA) effect must have reduced to a nonsigniﬁcant level.

Stgp 1: Relationship between the Independent (IV) and Mediating Variables (M!) and

Their Comppnents

The independent variable, neighborhood advantage (NA), was comprised of four
components: block income, block homogeneity, block stability, and block composition.
The originally hypothesized six indicators of the mediating variable of sense of
community (SOC) were broken down to three components as suggested by the reliability
analyses and labeled emotion, action, and safety. See Table 10 for the correlation matrix
of these variables. Again, both the uncorrected and corrected correlation coefﬁcients are
presented.

As can be seen from the correlation matrix, NA and SOCtotal are signiﬁcantly
correlated (.48). In fact, NA highly correlates with all three components of SOC as
indicated by the correlation range of .44 to .63. The components of neighborhood
advantage, however, display differential relationships with the components of SOC.

Block stability and block homogeneity signiﬁcantly correlate with SOC and all of its

53

respective components while block income only signiﬁcantly correlates with the safety
component (.35). Block composition only correlates with the action component of the

SOC construct (.20). Overall though, this correlation matrix suggests that there may be
mediating paths among the SOC construct and its components with block homogeneity,

block stability, and block income.

Step 2: Relationship between the Independent (IVs), Mediating (MVs), & Demndent
Variables (DVs) and Their Commnents

The outcome or dependent variable of self-reported delinquency was also broken
down into four components as suggested by the principal components analysis and
labeled steal and deal delinquency (SDDel.), school delinquency (SchoolDel.), steal and
ﬁght delinquency (SF Del.), and severe delinquency (SevereDel.). Additional outcome
variables included in these correlation analyses were conventional activity (CA),
conventional friends (CF), delinquent peers (DP), and average grade (see Table 9). The
highlighted correlation coefﬁcients in Table 9 indicate where possible mediating
relationships may have existed. As can be seen in this table, there was only one outcome
variable, conventional activity, that corresponded to the directional hypothesis that SOC
mediated the effect of neighborhood advantage on youth outcomes. Thus, only for CA
was it necessary to proceed to the third and ﬁnal step and run a combined or multivariate
regression equation to test for a mediating model.

Step 3: Regtession Analyses to Test the Hyppthesis that SOC Mediates the Relationship
Between Niand Youth Outcomes
SOCtotal and its three components signiﬁcantly correlated with conventional

activity (CA) and so did the independent variable of block stability. Thus, there were

54

four possible mediating paths to test for an intervening relationship between block
stability and CA. The independent variable of block income also signiﬁcantly correlated
with CA, but there was only one path to test since block income only signiﬁcantly
correlated with the safety component of the SOC construct. In sum, there were a total of
ﬁve multivariate regression equations run to evaluate the mediating hypothesis (see Table
11).

For the independent variable of block stability, none of the four SOC paths proved
to be a signiﬁcant mediator (Equations 1 — 4). However, there was a deﬁnite pattern
toward an intervening relationship as both SOCtotal (B = .22, p= .053) and emotion (B =
.21, p = .059) closely approached signiﬁcance while the effect of block stability
substantially decreased (Equations 1 -— 2).

Again, there was only one possible mediating path to test for an intervening
relationship between the SOC construct and block income and this was through the safety
component. Results displayed that safety did not mediate the effects of block income on

conventional activity (Equation 5).

55

Table 3

Emotion Items Corrected Item-Total Correlations, & Reliability Coefﬁcients for Parent
& Neighbor

People on this block know each other.

People on this block socialize with each other.

People on this block .ch1n_’t trust each other. (R)‘

People on this block feel connected to each other.

On this block people talk to each other about community problems.
People on this block feel isolated from each other. (R)

People who live on this block think of themselves as a community.

A feeling of community spirit exists among the residents of this block.
Residents don’t care about the block’s future. (R)

People on this block feel they belong here.

People on this block take care of each others’ plants, pets, kids.

When someone on this block has a problem, it’s hard to get help ﬁ'om neighbors. (R)
People on this block would give rides to each other if needed.

People on this block watch out for each other.

 

 

 

 

Emotion Component Corrected Item- Corrected Item-
Items Total Total
Correlation Correlation
Parent Neighbor

Know each other .48 .55
Socialize with each .60 .54
Don’t trust each .56 .42
Feel connected to .74 .66
Talk to each .65 .43
Feel isolated from .64 .70
Think of themselves .67 .63
Feeling of community .69 .67
Resident’s don’t care .56 .56
Feel they belong .35 .29
Take care of .69 .51
Has a problem .55 .60
Would give rides .54 .44
Watch out for .78 .61
Alpha .91 .88

 

Note. Reliability at the block-level was .89.
3‘(R) = Reverse coded.

56

Table 4

Action Item_s_. Corrected Item-Total Correlations & Reliability Coefﬁcients for Parent &

Neighbor

People on this block participate in social activities (e. g., pot lucks, group garage
sales, etc).

People on this block participate in community improvement activities (e.g., community
clean-ups, ﬂower plantings, etc.).

People on this block participate in the curbside recycling program.

People on this block participate in neighborhood organizations like block groups,
neighborhood watches, neighborhood associations, etc.

People on this block hair do things together to improve the block. (R)ll

If faced with a problem on the block, residents would be unable to create a solution. (R)

People on this block have a voice regarding important community issues.

Together, people on this block can persuade the city to respond to their needs and
concerns.

 

 

 

Action Component Corrected Item- Corrected Item-
Items Total Correlation Total Correlation
Parent Neighbor

Participate in social .58 .47
Participate in community .50 .59
Participate in curbside .37 .33
Participate in neighborhood .70 .71
Never do things .54 .67
Faced with problem .32 .56

Have a voice .70 .56

Can persuade city .54 .52
Alpha .80 .80

 

Note. Reliability at the block-level was .84.
a(R) = Reverse coded.

57

Table 5

S_afety Items, Corrected Item-Total: Correlations. & Reliability Coefﬁcients for Parent &

Neighbor

It is fairly safe to walk on this block at night.

People on this block make it a safer place to live.

 

 

Safety Component Corrected Item- Corrected Item-

Items Total Correlation Total Correlation
Parent Neighbor

Safe to walk .43 .49

Make it safer .43 .49

Alpha .60 .65

 

Note. Reliability at the block-level was .63.

58

Table 6

Principal Comppnent Analysis of the Self-Reported Delinquency (SRD) Construct

 

 

SRD Items SDDel. SchoolDel. SFDel. SevereDel. h2
Stolen > $50 but < $100 .84 -.04 .17 .05 .74
Taken some part of a car .81 .17 .13 -.02 .70
Sold or dealt drugs .79 .07 .09 .23 .69
Stolen > $100 .60 .06 -.16 -.09 .39
Stolen something 85 or less .60 -.07 .34 .09 .49
Skipped a full day of school .05 .93 -.02 -.03 .87
Skipped class while in school -.02 .90 .14 -.02 .83
Carried weapon outside of school -.04 .78 -.05 .03 .61
Fought someone physically -.O4 .09 .86 .17 .7 8
Bought, sold, or held stolen goods .49 .07 .78 .01 .85
Gone into building to steal -.03 -. 10 .71 -.07 .52
Used force to take 3 or things -.04 -.02 -.02 .92 .85
Attacked someone .08 .15 .05 .90 .84
Eignevalues 3.56 2.55 1.81 1.71

 

Note. Principal component analysis with varimax rotation. h2 equals the sum of squared component
loadings for each item. SDDel. = steal & deal delinquency; SchoolDel. = school delinquency; SFDel. =
steal & ﬁght delinquency; SevereDel. = severe delinquency.

59

Table 7

Face-Block Descriptive Statistics

 

 

Variables Items M SD Min-Max
Independent
Income 2 918.48 575.47 122 - 3,750
Stability 4 14.02 3.48 6 - 20
Homogeneity 2 5.65 1.37 3 — 9
Composition 1 1.64 .35 l —- 2
Neighborhood 4 .02 .59 -1.07 - 1.51
Advantage
Mediating
Safety 2 7.63 1.43 4 — 10
Emotion 14 47.97 7.27 33 - 63
Action 8 26.10 4.79 14 - 36
SOCtotal 24 81.69 12.40 55 — 107
Dependent
SDDel. 5 1.14 2.84 0 - 23
SchoolDel. 3 l 1.82 32.03 0 — 284
SFDel. 3 2.22 6.63 0 — 64
SevereDel. 2 .25 .85 0 - 7
SRDtotal 17 18.76 36.90 0 - 292
CF 3 10.99 2.01 5 — 15
DP 5 7.58 2.34 5 — 18
Grade 1 2.57 .83 0 — 4
CA 1 1.49 1.45 0 - 5

 

Note. The Neighborhood Advantage variable is standardized. SDDel. = steal & deal
delinquency; SchoolDel. = school delinquency; SFDel. = steal & ﬁght delinquency;
SevereDel. = severe delinquency; CF = conventional friends; DP = delinquent peers; CA

= conventional activity.

Table 8

Uncorrected & Corrected Correlations of the SOC Con_struct Between Parrent & Neighbor

 

 

Sense of Community Uncorrected Corrected

Construct Correlations Correlations
Emotion .20“ .21 *
Action .33" .37“
Safety .29" .37"

SOCtotal .29” .30"

 

Note. To obtain the corrected correlation, the bivariate correlation is divided by the
square root of the reliability of the measure.
*p < .05. "p < .01.

61

Table 9

Correlation Matrix of Independent (IVs). Mediating (MVs), & Demndent Variables (QVs)

 

 

 

91s
Variables SDD. SchD. SFD. SevD. SRDt. DP CF CA Grade
IVs
Income .01 -. 12 -.05 . 13 -. 14 .01 .19 .22* .12
Stability .04 -.04 -.05 -.29" -.08 -.18 .09 .19* .08
Homog. .22* .05 .04 -.10 .08 -.09 -.1 1 -.O7 -.02
Comp. .00 .l l -.17 .03 .06 .19 -.13 .01 .10
NA .1 1 .00 -.10 -.10 -.04 -.03 .02 . 15 . 13
W5
Emotion .21' -. 14 .13 -. 18 -.07 -.03 .05 .26“ .06
Action .25' " -.09 . 10 -.06 -.03 .00 . 14 .23* .23 *
Safety .13 .05 .07 -.07 .06 -.or .03 .2o* .02
SOCtotal .24“ -.1 l . 12 -.14 -.05 -.01 . 10 .26" . 12

 

Note. Bolded correlation coefﬁcients indicate potential mediating models which require further regression
analyses. SDD. = stealing & dealing; SchD. = school delinquency; SFD. = stealing & ﬁghting; SevD. =
severe delinquency; DP = delinquent peers; CF = conventional friends; CA = conventional

activity; NA = neighborhood advantage; Grade = average grade for the past school year.

.2 < .05. ”p < .01.

62

 

3.8
.w><

 

 

mn—

 

 

m0

 

 

.8663...”

 

 

39.5

 

 

_oQ
Eon—om

 

 

 

_oQQm

 

 

 

 

 

 

$5
58> 8.82362
60%
.03—.8800
mo 85m
303 532‘ cote—um

 

 

 

 

 

 

 

 

 

 

8538800
goo—m

 

 

 

36598.5
x35

 

 

 

eaaam
x85

 

 

0883
385

 

 

 

m 3:5

63

Table 10

Uncorrected & Corrected CorreLatipn My} of Independent (IVs) & Mediating Variables (MES)

 

 

 

I_V_§
MVS NA NA' 8 3' BH. BH.‘| BS. BS.‘ BC. BC.II
Emotion .44" .47" . 15 . 16 .28” .30" .49" .60" .l l .12
Action .39" .44" . 15 .17 .21"I .23"I .37“ .48” .18 .20"
Safety .50“ .63" .28" .35" .26" .33“ .57" .72" .09 .l l

SOCtotal .46" .48'"I .18 .19 .28” .29"l .49“l .59"”'I .15 .16

 

Me. Corrected correlations were obtained by dividing the bivariate correlation by the square root of the
reliability of the measure. NA = neighborhood advantage; S = block income; BH. = block homogeneity;
BS. = block stability; BC. = block composition.

'Indicates correlations that were corrected for attenuation due to measurement error.

‘p < .05. "p < .01.

Table 11

OLS Regtession Estimates of Conventional Activiﬂ (CA)

 

 

 

 

Regression Equations Convention_al Activitv (CA)
i S_E B B t-ratio
Equation 1
Block Stability .13 .16 .09 .790
SOCtotal .03 .01 .22 l .96
Equation 2
Block Stability .13 .16 .09 .830
Emotion .04 .02 .21 1.91
Equation 3
Block Stability .19 .15 .13 1.23
Action .05 .03 .18 1.72
Equation 4
Block Stability .17 .17 .12 1.02
Safety .13 .12 .13 1.09
Equation 5
Block Income .26 .15 .18 1.81
Safety .15 .10 15 1.47

 

Note. Equations 1 - 5 represent all possible mediating paths for
conventional activity.

65

Discussion
Resparch Obiective 1: Dimensionaligz of the SOC 85 SRD Conptructs

The ﬁrst goal of the current study was to examine the factor structure of the SOC
construct since previous publications have called attention to the lack of consistent
ﬁndings for either a unidimensional or multidimensional solution (Hill, 1996). The
results from the present study certainly do not resolve the confusion surrounding the
dimensionality and composition of the SOC construct. The results from the principal
component analysis tentatively indicated that the SOC construct contained three
components that accounted for 52% of the variance and loaded according to whether the
item measured an emotion, action, or safety component of the neighborhood. However,
there were several results from the PCA which contradicted this multidimensional
interpretation, or at the least, the exact item composition of each of the three components.
For instance, the emotion component accounted for a considerable degree of the overall
variance and the safety component included items from the other two constructs and
really created a catch-all component that made little theoretical sense. Therefore,
reliability analyses were conducted on all of the items hypothesized to fall into the
emotion, action, and safety components. As these results indicated, reliability
coefﬁcients were all in the adequate range and it was decided to follow these empirical
results and continue analyses with these three factors. In sum, similar to the sense of
community index and most theoretical writings on the topic, the SOC measure in this
study proved to be multidimensional and could quantify distinct aspects of community

life.

66

As Chavis and Pretty (1999) recommend, the ﬁeld could beneﬁt from a
collaborative effort to develop a standardized SOC measure. Presently, the most
common measure employed, the short form of the sense of community index, suffers
from reliability problems (Chipuer & Pretty, 1999). Utilizing four separate sets of data,
these authors found that the overall reliability of the short form of the SCI index ranged
from .64 to .69. The four subscale reliabilities rarely approached acceptable levels and
ranged ﬁ'om .16 to .72. Comparatively, the reliability of the SOC scale was impressive.
For parents, the reliability of the overall scale, SOCtotal, was .93. Parent alpha for the
emotion component was .90, action component was .80, and safety component was .60.
Neighbor and block reliability scores also displayed similar loadings on SOC and its
three components.

There certainly is a need for collaboration in the ﬁeld as the proliferation of new
SOC measures, like the one utilized in the current study, inhibit comparisons across
studies and limits community psychology’s understanding of one of its overarching
values. However, rather than simply returning to the long form of the SCI (Chipuer &
Pretty, 1999), there is enough empirical data on various SOC scales that a thorough
review of their commonalties and strengths is warranted. Such a practical approach can
build upon what has been learned since the sense of community index was ﬁrst presented
(Chavis et al., 1986).

The results of this study also found that the self-reported delinquency scale could
be separated into meaningful components. Speciﬁcally, four components accounted for
57% of the variance. Two of the components mainly consisted of stealing items while

one component dealt overwhelmingly with school issues and the ﬁnal component dealt

67

with severe delinquency. Thus, a strength of the self-report delinquency scale was that it
allowed differentiation between minor delinquency (e. g., stole $5 or less) and severe
delinquency (e.g., attacked someone with the idea of seriously hurting him or her?).
More importantly, it allowed one to assess if some neighborhoods created an
environment where certain types of delinquency proliferated.

Research Obiective; Evaluating the Validity of the Face-Block as a Neighborhood Unit
The face-block or block-level conceptualization of neighborhood, as formulated

 

by Bursik and Grasrrrick (1993), represents the smallest and most immediate geographic
area around one’s home. It is only one of three important neighborhood units postulated
to affect community dynamics but most likely exerts the most inﬂuence and is the focus
of the current study. Descriptive statistics were computed to begin the assessment of the
validity of both NA and SOC at the block-level. Results vividly displayed that urban
blocks differentially possess both the exogenous structural characteristics and the
mediating social characteristics as proposed by social disorganization theory. One
interesting ﬁnding was that many of these face-blocks were nested within the same
census tract yet displayed strikingly different results on both the exogenous and
mediating variables proposed by social disorganization theory. In essence, while it is
reasonable to conceive of neighborhoods at a census tract level, results indicate it may be
more ecologically valid and meaningful to urban residents to conceive of their
neighborhood in a much smaller and objective unit. Thus, for a true contextual analysis
of neighborhoods and how they inﬂuence their inhabitants, the results of this study
indicated that it is necessary to begin from the block-level and build up to larger

neighborhood systems. This ﬁnding supports Bursik and Grasrrrick’s (1993) updated

68

social disorganization model which utilizes three levels of community control (private,
parochial, and public) to explain delinquency rates within urban areas.

Correlations were also run to ascertain whether or not there was a block-level
consensus on these constructs. For the SOC construct, correlations between parent and
neighbor were all signiﬁcant and responses ranged from a low of .21 to a high of .37.
While these correlations were signiﬁcant and indicate a block-level consensus of sense of
community, they were only in the small to moderate ranges. It was hoped that there
would be greater block-level consensus and that coefﬁcients would range from moderate
to high. However, only two respondents were assessed to create a face-block summary
measure due to the time-consuming nature of the neighborhood surveys. Future studies
should improve upon the current endeavor by interviewing more households per block.

Discriminant validity was assessed by evaluating how the NA and SOC constructs
related to each other and the various positive and negative youth outcomes. Results
supported the prediction that high levels of NA would translate into high levels of SOC.
For instance, block homogeneity, block stability, and the composite measure of
neighborhood advantage signiﬁcantly correlated with SOC and all of its components.
When corrected for attenuation, correlations ranged from moderate to large with the
correlation between block stability and safety reaching .72.

An important ﬁnding regarding the independent and mediating variables was that
block income only correlated with the safety component of SOC (.35). While lower
income neighborhoods were not perceived to be as safe as other neighborhoods, it did not
have a signiﬁcant impact on the emotional attachment residents felt toward their

neighborhood or on residents’ participation in neighborhood organizations and activities.

69

Meanwhile, block stability signiﬁcantly related to SOC and all of its components.
Correlations ranged from .48 to .72 and thus dwarfed the effects of income in this study.
This ﬁnding points to the importance of not limiting analyses of neighborhood effects to
just the income level or SES of the neighborhood. In this study, block instability exerts
considerably more deleterious effects than low income. Nevertheless, the main point is
that it is not simply residential stability or income level, or any other independent
variable, but the combination of these variables that usually leads to the extreme disparity
between disadvantaged and advantaged neighborhoods across the United States.

Another interesting ﬁnding from the components of the independent variable
neighborhood advantage was the fact that block composition only signiﬁcantly correlated
with the action component of the SOC construct. The sample for this study had a large
percentage of dual parent families in contrast to some neighborhood studies where single
parent families predominate within a community. Perhaps, as others have hypothesized,
there is a tipping point or threshold that must be surpassed before the collective
supervision capacity of a village suffers (Simons et al., 1996).

Discriminant validity was also assessed by evaluating how both positive and
negative outcome variables related to NA and SOC. In seven of the nine outcome
variables, the majority of NA and SOC components correlated with the outcome
variables in the expected direction. Many of these correlation coefﬁcients were also at
signiﬁcant levels. In one of the two deviating cases, steal and deal, both NA and SOC
correlated in the opposite direction as expected. In this case, higher structural and social
organizational levels resulted in greater negative outcomes and this ﬁnding will be

discussed at length later. For the other variable, steal and ﬁght, NA correlated in the

70

expected direction but SOC positively correlated with the outcome variable. These
contradictory ﬁndings both came from the self-report delinquency measure and involved
items on stealing.

Overall, validity of the face-block as a neighborhood unit was established. This
result validates previous research on blocks, which found that they are indeed valid and
meaningful neighborhood units for urban residents (Perkins & Taylor, 1996; Taylor,
1997; Taylor, Gottﬁ'edson, & Brower, 1984). Unfortunately however, the overwhelming
majority of researchers continue to utilize census tract to deﬁne a neighborhood (Coulton,
Korbin, & Su, 1996; Gonzalez et al., 1996). While there is a tremendous amount of
information collected at the census tract level, any true contextual analysis of
neighborhoods must begin at the block-level and build up to this larger neighborhood
system.

RLearch Objective 3: Sense of Communi SOC as a Mediptor of Neighborhood
Advantage (NA) & Youth Outcome;

The third guiding research goal and main hypothesis of the study was that SOC
mediated the relationship between neighborhood advantage and youth outcomes. This
hypothesis did not hold under empirical investigation. In fact, none of the components of
neighborhood advantage signiﬁcantly correlated with any of the outcome variables.

Also, the proposed mediating variable of SOC only correlated with one of the ﬁve
delinquency outcomes and conventional activity. Although the overall model proved
insigniﬁcant, implied submodels were tested since there were multiple indicators for all
of the main variables in the study. This investigation of the submodels also corresponded

to the stated research issue of exploring the systemic and applicable nature of the SOC

71

construct. Thus, a strength of the proposed mediator is that researchers can test the
various dimensions of the SOC construct for possible differential effects on youth
outcomes. Moreover, it is argued that a comprehensive and systemic construct like SOC
is needed to truly measure the dynamic nature of urban areas. For instance, this
multidimensionality even allows researchers the ability to make distinctions among
communities that may have similar overall SOC scores but substantially different totals
among the three components of emotion, action, and safety.

The results from the submodels displayed that only one outcome variable could be
tested for the directional hypothesis that SOC mediated the effect of NA on youth
outcomes. This outcome variable was conventional activity and was deﬁned as the
number of school activities the youth participated in the past year. Although results
indicated that none of the SOC components were signiﬁcant mediators, both SOCtotal
and emotion approached statistically signiﬁcant levels. Considering the small sample
size, this ﬁnding seems to indicate that youth reared in communities characterized by a
high level of SOC, especially those with a greater degree of emotional attachment, are
more likely to bond to and participate in conventional activities. Further, participation in
school activities was the best predictor of good grades in this study. Thus, youth from
high SOC neighborhoods participated more in school activities and attained more success
in the school environment than youth in low SOC neighborhoods. Therefore, it seems
quite evident that the neighborhood microsystem has an important impact on youth’s
attitudes and behaviors outside of that speciﬁc setting.

While some research has been completed on both an adolescent’s neighborhood

SOC and school SOC (Pretty, Andrews, & Collett, 1994; Pretty, Conroy, Dugay, Fowler,

72

& Williams, 1996), future research should incorporate the school microsystem to a much
greater extent and its relationship with neighborhood SOC. A greater understanding of
the relationship between these two settings and their impact on youth outcomes would
certainly aid in developing a positive and supportive context for all youth. Further, it
could also direct possible interventions for those youth who are experiencing turmoil in
one or both of these important developmental settings.

Although CA was the only dependent variable that met the criteria for the main
hypothesis of the study, steal and deal delinquency also evidenced a signiﬁcant
correlation with both NA and SOC components. However, it was not evaluated because
the current study had a directional hypothesis. If the direction of the hypothesis was
ignored, there would have been three possible paths to test for this outcome variable.
While two of these paths proved to be nonsigniﬁcant, the action component of SOC was
found to mediate the effect of block homogeneity on stealing and dealing. In the
multivariate regression equation, action retains it signiﬁcant effect (B = .22) while block
homogeneity drops to an insigniﬁcant level (B = .18). Thus, neighborhoods characterized
by a greater degree of both NA and SOC led to increased stealing and dealing
delinquency rates.

This result poses serious problems for the social disorganization model in general
and particularly for the main hypothesis of this study. However, this ﬁnding is not
shocking when one considers the perpetual debate in the ﬁeld of criminology as to what
self-report scales and official records of delinquency measure. In general, ofﬁcial records
have been criticized as simply reﬂecting the bias in police practices and record keeping

while self-report measures have been criticized for equating minor delinquency with the

73

more serious infractions that comprise the uniform crime reports (Hagan, Gillis, & Chan,
1978; Sampson, 1986). Since steal and deal is one of the more minor scales of the
overall self-reported delinquency construct, it certainly could follow that structural
advantage and sense of community levels have little to no inﬂuence on this type of
offending pattern compared to its role in more severe delinquency. While recent social
disorganization studies have found support for explaining self-report and victimization
data, it is worthwhile to note that the study that launched social disorganization’s re-
emergence found distinct pattern differences between self-report and ofﬁcial delinquency
(Simcha-Fagan & Schwartz, 1986). Unfortunately, this line of inquiry cannot be tested in
the current study since only self-report measures of delinquency were collected, and, any
conclusion regarding these largely null ﬁndings are unwarranted at this time.

Overall, the lack of signiﬁcant ﬁndings for the proposed mediating model and
submodels may simply be due to the fact that this was a city-based study. In an extensive
review of the literature on neighborhood effects (Leventhal & Brooks-Gunn, 2000), it
was found that national and multisite studies had a much better chance of ﬁnding
signiﬁcant effects than city-based studies. This is due to the greater sampling variability
of neighborhoods and greater sample sizes in national studies. In a city-based study such
as this one, higher intercorrelations among neighborhood dimensions leads to a higher
probability of a null ﬁnding. The current study also had a sample size only slightly
greater than 100 and this limited the power to detect signiﬁcant effects.

Limitations
There were several limitations to this study which warrant attention. First, results

from the principal component analysis of the SOC construct were largely inconclusive

74

and indicative of the controversy in the ﬁeld regarding the dimensionality of the SOC
construct. For the self-reported delinquency construct, there were two components that
suffered from reliability problems. Both steal and ﬁght delinquency (Alpha = .48) and
severe delinquency (Alpha = .52) had inadequate reliability levels. Another major
drawback in the current study was that block income was utilized as one of the major
independent variables versus the traditional method of measuring SES in social
disorganization studies. Therefore, caution must be taken in the interpretation of the role
SES exerts on the level of sense of community within a neighborhood and its inﬂuence

on youth outcomes such as delinquency.

75

Implications for Future Research

This study has shown that the urban face-block is a valid way to conceptualize
and measure both the independent and mediating variables as proposed by social
disorganization theory. This corresponds with previous work from the SOC in Lansing
Neighborhoods Project where it was found that there is constant variation in SOC levels
from block to block within the same census tract. Previous project research also found
that higher SOC levels were positively related to voter turnout in local elections and
participation in recycling, and negatively to fear of crime and actual criminal activity on
the block (Schweitzer, Kim, & Mackin, 1999). Other research at the block-level has also
found validity for this unit of analysis and its relationship to informal control, crime, fear
of crime, and SOC (Chavis & Wandersman, 1990; Perkins & Taylor, 1996; Taylor, 1997;
Taylor, Gottfresdson, & Brower, 1984).

While a strong argument has been made for understanding block-level dynamics
in an urban setting, this study does not mean to neglect the larger communities and forces
which affect both the structural and social composition of urban blocks. Updated,
systemic social disorganization models attempt to explain community effects on
delinquency and other youth outcomes at the local level by incorporating the role and
inﬂuence of these outside econorrric, political, social, and even global forces. Future
research needs to combine various neighborhood levels, such as the private, parochial,
and public, in order to gain a more complete understanding of how these outside forces
inﬂuence block-level dynamics, and, ultimately individual youth outcomes (Bursik &

Grasmick, 1993).

76

This study was, in part, instigated by Taylor’s (1997) call for a greater degree of
scrutiny at the most immediate source of neighborhood social control, the primary level.
Utilizing sense of community as a potential mediating variable for social disorganization
theory, this study extended our understanding of local neighborhood dynamics. As this
study has shown, there is great variability in the structural and social dimensions across
urban blocks, even within the same community. Previous research has also found that
programs based on census tract averages will not be sensitive enough to build upon the
varied strengths and deﬁcits of neighborhoods within the same census tract (Caughy,
O’Campo, & Brodsky, 1999). If the main goal of neighborhood effects research is to
utilize results to direct neighborhood development and intervention programs, variability
across blocks must be captured and understood for successful program implementation.
Thus, it is recommended that future neighborhood studies should incorporate the urban or
face-block as the primary unit of analysis.

Moreover, as recent work on community building has found, successful programs
need to be strengths-based, consistent with the values of the local community, and
devised and implemented by local citizens (McNeely, 1999). Thus, successful
community development programs call for grassroots action that is best approached at the
block-level. Block associations are the most typical of such programs and success at this
level can result in greater participation and empowerment (Perkins, Florin, Rich,
Wandersman, & Chavis, 1990) and possibly lead to participation in larger neighborhood
organizations that form to combat more macrolevel problems like child maltreatment,

delinquency, and high school dropout rates. Again, the ﬁeld needs to start small and

77

understand block-level processes before attempting to understand and solve larger (e.g.,
parochial, census tract) neighborhood processes and problems.

Sense of community is a systemic and applicable construct unlike previous
measures utilized in neighborhood studies. Incorporation of this construct into the
mediating model of neighborhood effects could ultimately lead to successful
neighborhood development efforts and prevention and intervention programs for youth.
Research on the importance of understanding the inﬂuence of the neighborhood
rrricrosystem has evolved considerably in the last two decades. It is now time for this
research to go beyond an explanatory framework and out into the neighborhoods. By
investigating SOC at the face-block level, researchers, neighborhood leaders and
organizers could come together in their efforts to create, foster, and maintain healthy
communities. As neighborhood leaders and organizations implement various policies
and programs, evaluators could work alongside studying their effects: block-by-block.
Research ﬁndings could then be utilized to inform federal and state policymakers of
successful neighborhood development policies and programs.

In addition to directing youth intervention and neighborhood redevelopment
programs, the SOC construct also corresponds to community psychology’s goal of
operating on a prevention or health promotion level. Efforts to increase SOC within a
neighborhood will lead to more positive outcomes for all children and adolescents as well
as greater ratings of quality of life for all community residents. In closing, as Hill (1996)
stated, “We could concentrate on forming healthy communities, and rely on the
communities to form healthy individuals. Then we could truly become community

psychologists” (p. 437).

78

APPENDICES

79

Appendix A

Initial Letter to Parents

80

Sense of Community in Lansing neighborhoods Project
Michigan State University Urban Affairs Program
W-29 Owen Graduate Hall
East Lansing, Michigan 48823-1109

Dear Parent:

You may have heard the expression that "it takes a village to raise a child." As part of the
Sense of Community in Lansing Neighborhoods Project, we are studying the impact of
neighborhoods on public school children in Lansing. Your child was selected by lottery
through cooperation with the Lansing School District to be part of this study.

Some time within the next several months, one of the interviewers ﬁom Michigan State
University's Sense of Community team will visit you at your home to seek your consent
to include your child in the study. If you agree, the interviewer will survey you, your
child, and one or two of your neighbors. The survey should take no more than 30
rrrinutes to complete, and participation will be completely voluntary.

Any information you, your child, or neighbor provides will remain strictly conﬁdential.
Within these restrictions, results will be made available to you at the completion of the
study in the Fall of 2000.

Because the goal of this study will help us to understand the lives and development of all
the children in Lansing, I hope that you will agree to participate. If you have any
questions, comments, or concerns, please feel ﬂee to contact me at 353-9145 or 353-9144
(voice mail). If you would rather not participate in the study, please return the self-
addressed, stamped postcard.

Sincerely yours,

John Schweitzer
Professor, Urban Affairs Programs

81

Appendix B

Follow-up Letter to Parents

82

Sense of Community in Lansing Neighborhoods Project
Michigan State University Urban Affairs Program
W-29 Owen Graduate Hall
East Lansing, Michigan 48823-1109

Dear Parent:

This is a follow-up to the initial letter we mailed to you in late May regarding MSU’s
Sense of Community Study (SOC) study. Some time within in the next couple of days, I
will be stopping by again to see if you would like to participate in the study. Again, we
are trying to evaluate how neighborhood cohesion may affect a child or teenager’s
development. We hope that this research can be utilized in Lansing to foster community
development and cohesion to the beneﬁt of everyone in the community, particularly our
youth.

If you agree to participate, I will interview you and one of your neighbors to get an idea
of the Sense of Community on your block. I will also talk to your 10th grade son about
his school, activities, and friends. I hope that you can participate and look forward to
talking to you about your neighborhood. If you have any questions, comments, or
concerns, please feel free to contact the Sense of Community project ofﬁce at 353-9145
or 353-9144 (voice mail). Thanks for your time and consideration.

Sincerely,

SOC Project Interviewer

83

Appendix C

Youth Assent Form

84

Youth Assent Form

Hi. My name is from Michigan State University. We’re working on
a study that looks at how living in certain neighborhoods affects youth, like you, who live
there. I have spoken to one of your parents who said that it is OK. if I talk to you.
Before we begin, I would like to make sure that it is also OK. with you. I will ask you
questions about your school, friends, and outside activities. Everything that you tell me
will remain completely conﬁdential. This means that only members of this study, like
myself, will have access to this information and I will not share it with anyone else like
your parents, ﬁiends, school, police, etc.

Why is this (conﬁdentiality) necessary?

0 Some of the questions deal with legal and illegal activities often committed by youth.
Conﬁdentiality assures you that what you say is between you and me and will not be
reported to any outside party.

0 Conﬁdentiality assures you that I will not tell anyone anything that you tell me. This
is always a necessary component when you talk to individuals about their behaviors,
perceptions, etc.

It is important for you to know that it is your choice if you want to answer our questions.
Also, if you do choose to answer our questions, you can still refuse to answer any
question that bothers you. By signing your name below, you are showing me that the
study has been fully explained to you and that you are choosing to participate. So, if it is
OK. to speak with you, can you please sign your name on the line below.

Participant signature Date

 

 

0 Make sure to clarify any questions the student may have.
0 Emphasize and explain conﬁdentiality in language they understand.

0 If unable to complete interview then, ask if another day/time would be better.

85

Appendix D

Parent Introduction

86

Sense of Community Introduction

Hello. My name is from Michigan State University. We’re
working on a study that evaluates the Sense of Community (SOC) in Lansing
neighborhoods. By SOC, we want to know your feelings about this M — how people
get along, if they feel safe, and if they enjoy living in this neighborhood, etc. It is our
goal to understand how a strong community can aid in the positive development of youth
that reside there.

We recently mailed a letter to you regarding this study and I wanted to know if you could
take some time to answer a survey about your perceptions of SOC on this immediate
hlpc_k along with some questions about your child/teen? (Provide them with a copy of the
letter and explain the exact dimensions of their face-block).

YES — OK. then, ﬁrst I need to review this consent form with you in order to conduct the
survey. Provide respondent with a copy and review the form with them highlighting the
numbered items. Have them sign your copy and let them keep their copy of the consent
form.

m — OK, thanks for your time and have a good day.

0 If the person is hesitant, ﬁnd out if a different time or day would be better.
0 If the person wants to know how long it takes, make sure to let them know that it
depends on how much they have to say but that it will probably take around 20 - 30

minutes.

0 Refer to John Schweitzer if the person has concerns or questions about the survey
(353-9144).

0 Respondent needs to be at least 18 years old. Ask if you are unsure.

87

Appendix E

SOC Measure with Six Components

88

Sense of Community Measure

Connection

Cl. People on this block know each other.

C2. People on this block socialize with each other.

C3. People on this block ﬁrm trust each other. (R)

C4. People on this block feel connected to each other.

C5. On this block people talk to each other about community problems.
C6. People on this block feel isolated from each other. (R)

Belonging
B 1. People who live on this block think of themselves as a community.

82. A feeling of community spirit exists among the residents on this block.
B3. Residents don’t care about the block’s future. (R)
B4. People on this block feel they belong here.

Supmrt
S 1. People on this block take care of each other’s plants, pets, kids.

82. When someone on this block has a problem, it’s hard to get help from neighbors. (R)
S3. People on this block would give rides to each other if needed.
84. People on this block watch out for each other.

Participation

Pl. People on this block participate in social activities (e.g., potlucks, group garage
sales, etc.).

P2. People on this block participate in community improvement activities (e.g.,
community clean-ups, ﬂower plantings, etc.).

P3. People on this block participate in the curbside recycling program.

P4. PeOple on this block participate in neighborhood organizations like block groups,
neighborhood watches, neighborhood associations, etc.

Emmwerment
E1. Pe0ple on this block never do things together to improve the block. (R)

E2. If faced with a problem on the block, residents would be unable to create a solution. (R)
E3. People on this block have a voice regarding important community issues.
E4. Together, people on this block can persuade the city to respond to their needs and concerns.

Safeg
SAl. It is fairly safe to walk on this block at night.

8A2. The People on this block make it a safer place to live.

89

Appendix F

Youth Survey

90

Youth Survey

Prosocial Competence

Personal Efﬁcacy
The majority of questions that I am going to ask you will require that you answer with the
speciﬁc # of times you have performed a certain behavior. Or, they will require you to
answer with the responses I provide you. For the ﬁrst three questions, I would like you to
use these following response options -— Give card and review categories slowly.
Requnse Categoty — (5) Strongly Agree, (4) Agree, (3) Neutral, (2) Disagree,

(1) Strongly Disagree.

5. Strongly Agree
[You should use this category if you have a strongly held response to the
statement. By this, I mean that this statement is very true of you — your
behavior, perception, belief: etc. ]

4. Agree
[You should use this response if the statement is true about you but not as
descriptive or true as the previous choice]

3. Neutral
[That the statement is neither true nor false about you or does not really
apply to you. ]

2. Disagree

[You should use this response if the statement about you is false]

1. Strongly Disagree
[That the statement is deﬁnitely false and it elicits a very strong response
- it is deﬁnitely not you]

0 Before we begin, I’d also like to inform you that there are no right or wrong answers
to this survey. 1 want you to be completely honest, as everything you tell me will
remain conﬁdential -— in other words, between you and me. Do you have any
questions before we begin?

1. When people I know are having problems, I feel like I should try to help them.
2. There is really no way I can solve some of the problems I have. (R)
3. I can do just about anything I set my mind to.

Educational Expectations
Response Category - # or Year

1. How far do you think you will actually go in school?

Grades
Response Category - # or letter

1. Average school grade for the past year.

91

Commitment to Conventionality
Response Category — Record in # of minutes per day

1. Average time spent on homework and studying each night.

Resmnse Categoty: (5) Very Important, (4) Pretty Important, (3) Somewhat Important,
(2) Not Too Important, (1) Not Important at all.

2. Importance of grades.
3. Importance of education for getting a job later on.

Involvement in Conventional Activity
Response Category - # activities youth is involved in:

l. __ School activities such as student government, clubs, sports, band, choir, etc.
2. __ Community sports teams.
3. _ Community activities such as scouts, service clubs, Boys & Girls’ Club,
YMCA, etc.
4. __ Religious youth groups.
Peer Involvement

Prosocial Friends & Delinqqgnt Peers

How many of your ﬁiends
Resmnse Categoty: (5) All of them, (4) Most of them, (3) Half of them, (2) Few of
them, and (1) None of them.

Get good grades in school?

Rob or bully someone into giving you something?
Break into buildings?

Are interested in school?

Sell drugs?

Use marijuana?

Use hard drugs?

Attend school regularly?

9°39??ow

Problem Behaviors

Delinqupncy

How many times in the past year have you

Response Category - #

I. _ Gone onto someone’s land or into someone’s house or building when you
weren’t suppose to be there?*
Purposely set ﬁre to a house, building, car, or other property or tried to do so?
Gone into or tried to go into a building to steal something?
Fought someone physically?‘
Stolen or tried to steal something worth $5 or less?

9'91”!"

92

6. Stolen or tried to steal something worth more than $50 but less than $100?

7. Stolen or tried to steal something worth $100 or more?

8 Been stopped and questioned by police but not arrested?“

9 Knowingly bought, sold, or held stolen goods or tried to do any of these
things?

10. Snatched someone’s purse, wallet, or picked someone’s pocket?

11. Taken some part of a car (includes radio) without permission from the
owner?‘

12. Stolen or tried to steal a motor vehicle such as a car or motorcycle?“

13. Skipped a full day of school?*

14. Carried some sort of hidden weapon (gun, knife, razor blade, etc.) in school?*

15. Carried a gun or knife for protection - outside of school?

16. Attacked someone with the idea of seriously hurting him or her?

17. Used force (may or may not include a weapon) to take money or things from
people?

18. Sold or dealt drugs?

19. Skipped class when you were in school?‘

20. Been involved in gang ﬁghts?

Drug Use

 

How many times in the past year have you
Response Category - #

h Used marijuana?

2. Used paint, glue, or other things you inhale?
3. Consumed alcohol until intoxicated?

4. Used drugs or pills, other than marijuana?
Arrests

Response Category — Y/N with #

1. Have you been arrested in the last year?

2. If so, how many times (in the last year)?

93

Appendix G

Principal Component Analysis of the SOC Construct

94

Table 12

Principal Component Analysis of the Sense of Communig (SOC) Construct

 

 

SOC Items Emotion Action Safety h2
Know each other .93 .05 -.36 .99
Socialize with each other .78 .09 -. l 8 .65
Think of themselves as a community .71 .09 .03 .51
Feel connected to each other .70 .09 .14 .52
Feel isolated from each other .62 .08 .19 .42
Watch out for each other .44 .08 .37 .34
Would give rides to each other .40 -.05 .23 .22
Take care of each others’ plants, etc. .31 .23 .16 .17
Participate in community improvement .02 .89 -.22 .84
Have a voice ... community issues -.04 .73 .04 .54
Participate in neighborhood organizations .03 .70 .15 .51
Participate in social activities .32 .59 -.29 .53
Never do things to improve the block .09 .54 .27 .37
Feeling of community spirit exists .26 .51 .19 .36
Talk to each other . .. community .15 .48 -.04 .25
Persuade the city to respond -.15 .46 .39 .39
Don’t trust each other .25 -.46 .80 .91
Don’t care about the block’s future -.03 .17 .76 .61
Feel they belong here -.15 -.06 .66 .47
Participate in curbside recycling -.40 .24 .60 .58
Make it a safer place to live .08 .23 .59 .41
Hard to get help from neighbors .30 -.11 .59 .45
Unable to create a solution .02 .33 .46 .32
Fairly safe to walk on block at night .36 -.08 .40 .30
Eigenvalues 8.72 2.19 1 .60

 

Note. Principal component analysis with promax rotation. El equals the sum of squared
component loadings for each item.

95

Appendix H

SOC’s Three Components

96

SOC’s Three Components

Emotion

C1.
C2.
C3.
C4.
C5.
C6.
B1.
B2.
B3.
B4.
81.
82.
S3.
S4.

People on this block know each other.

People on this block socialize with each other.

People on this block _d_qp:t trust each other. (R)

People on this block feel connected to each other.

On this block people talk to each other about community problems.
People on this block feel isolated from each other. (R)

People who live on this block think of themselves as a community.

A feeling of commmrity spirit exists among the residents on this block.
Residents M care about the block’s future. (R)

People on this block feel they belong here.

People on this block take care of each other’s plants, pets, kids.

When someone on this block has a problem, it’s hard to get help from neighbors. (R)
People on this block would give rides to each other if needed.

People on this block watch out for each other.

Action

P1.

P2.

P3.
P4.

E1.
E2.

E3.
E4.

People on this block participate in social activities (e.g., potlucks, group garage
sales, etc.).

People on this block participate in community improvement activities (e.g.,
community clean-ups, ﬂower plantings, etc.).

People on this block participate in the curbside recycling program.

People on this block participate in neighborhood organizations like block groups,
neighborhood watches, neighborhood associations, etc.

People on this block pep/gr do things together to improve the block. (R)

If faced with a problem on the block, residents would be unable to create a solution.
(R)

People on this block have a voice regarding important community issues.
Together, people on this block can persuade the city to respond to their needs and
concerns.

 

Safety
SAl. It is fairly safe to walk on this block at night.

8A2. The people on this block make it a safer place to live.

97

REFERENCES

98

REFERENCES

Agnew, R (1993). Why do they do it? An exarrrination of the intervening
mechanisms between “social control” variables and delinquency. Journal of Research in
Crime and Delinquency, 30 (3), 245-266.

Akers, R. L. (1990). Rational choice, deterrence, and social learning theory in
criminology: The path not taken. The Journal of Criminal Law and Criminology, 81 (3),
653-676.

Akers, R. L. (1977). Deviant behavior: A social learning approach. Belmont,
CA: Wadsworth Publishing Company.

Barker, R. G. (1969). Ecological psychology. Stanford, CA: Stanford University
Press.

Baron, R. & Kenny, D. A. (1986). The moderator-mediator variable distinction in
social psychological research: Conceptual, strategic, and statistical considerations.
Journal of Personality and Social Psychology, 51 (6), 1173-1182.

Black, D. (1989). Toward a general theory of social control: Vol. 4. Orlando,
FL: Academic Press.

Bronfenbrenner, U. (1979). The ecology of human development: Exmriments by
nature and desigp. Cambridge, MA: Harvard University Press.

Brooks-Gum, J., Duncan, G. J ., Aber, J. L. (Eds). (1997). Natghm
mvegy: Vol. 1. Context and consequences for childrep. New York: Russell Sage

Foundation.

Brooks-Gum, J ., Duncan, G. J ., Klebanov, P. K., & Sealand, N. (1993). Do
neighborhoods inﬂuence child and adolescent development? American Journal of
Socioltgy. 99 (2), 353-395.

Buckner, J. C. (1988). The development of an instrument to measure
neighborhood cohesion. American Journal of Communig Psychology, 16 (6), 771 -791.

Buroughs, S. M., & Eby, L. T. (1998). Psychological sense of community at
work: A measurement system and explanatory framework. Journal of Comqu
Psychology, 26 (6), 509-532.

Bursik, R. J ., Jr. (1988). Social disorganization and theories of crime and
delinquency: Problems and prospects. Criminology, 26 (4), 519-551.

Bursik, R. J ., Jr., & Grasrrrick, H. G. (1993). Neighborhoods and crime: The
dimensions of effective communig control. New York: Lexington Books.

Bursik, R. J ., Jr., & Webb, J. (1982). Community change and patterns of
delinquency. American Journal of Sociology, 88, 24-42.

Butts, J. A. (1996). Ofﬁce of Juvenile Justice and Delinquency Prevention. Fact
# 47. October, 1996. Delinquency cases in juvenile courts, 1994. Online. Available:
http://www.ncirs.orgtxtﬁles/delc94.txt.

Caughy, M. 0., O’Campo, P., & Brodsky, A. E. (1999). Neighborhoods,
farrrilies, and children: Implications for policy and practice. WM
Psychology, 27 (5), 615-633.

Chavis, D. M., & Pretty, G. M. H. (1999). Sense of community: Advances in

measurement and application. Journal of Community Psychology, 27 (6), 63 5-642.

100

Chavis, D. M., & Wandersman, A. (1990). Sense of community in the urban

environment: A catalyst for participation and community development. American

Journal of Community Psychology, 18 (1), 55-81.

Chavis, D. M., Hogge, J. H., McMillan, D. W. & Wandersman, A. (1986). Sense
of community through Brunswik’s lens: A ﬁrst look. Journal of Community Psychology,
1_4, 24-40.

Chipuer, H. M., & Pretty, G. M. H. (1999). A review of the sense of community
index: Current uses, factor structure, reliability, and further development. Journal of
Communig Psychology. 27(6), 643-65 8.

Clark, P. (1996). 1996 Juvenile justice reform legislation. Michigan Judicial

Institute 13 (5), 51 -54.

 

Cloward, R. A., & Ohlin, L. E. (1960). Delinquency and Opmrtunity: A theoty
of delinquent gangs. Glencoe, IL: Free Press.

Coulton C. J ., Korbin, J. E., Chow, J ., & Su, M. (1995). Community level factors
and child maltreatment rates. Child Development, 66, 1262-1276.

Coulton, C. J ., Korbin, J. E., Su, M. (1996). Measuring neighborhood context for
young children in an urban area. American Journal of Community Paychology4ﬁ, (1),
5-32.

Davidson, W., & Cotter, P. (1986). Measurement of sense of community within
the sphere of a city. Journal of Applied Social Psychology, 16, 608-619.

Doolittle, R. J ., & MacDonald, D. (1978). Communication and a sense of

community in a metropolitan neighborhood: A factor analytic examination.

Communication Quarterly, 26(3), 2-7.

101

Eddy, J. M., & Gribskov, L. S. (1998). Juvenile justice and delinquency
prevention in the United States: The inﬂuence of theories and traditions on policies and
practices. In T. P. Gullotta, G. R. Adams, & R. Montemayor (Eds), Dﬂment Violent
Youth: Theory and Interventions. Thousand Oaks, CA: Sage.

Elliott, D. S., Huizinga, D., & Ageton, S. S. (1985). Explaining delinquency and
drug abuse. Beverly Hills, CA: Sage.

Elliott, D. S., Wilson, W. J ., Huizinga, D., Sampson, R. J ., Elliott, A., & Rankin,
B. (1996). The effects of neighborhood disadvantage on adolescent development.
Journal of Research in Crime and Delinquency. 33 (4), 389-426.

Glynn, T. J. (1981). Psychological sense of community: Measurement and

application. Human Relations 34 (7), 789-818.

 

Gonzalez, N. A., Cauce, A. M., Friedman, R. J ., & Mason, C. A. (1996). Family,
peer, and neighborhood inﬂuences on academic achievement among Aﬁican-American
adolescents: One-year prospective effects. American Journal of Communig Psychology,
261 (3), 365-387.

Griffin, K. W., Scheier, L.M., Botvin, G. J., Diaz, T., & Miller, N. (1999).
Interpersonal aggression in urban minority youth: Mediators of perceived neighborhood,
peer, and parental inﬂuences. Journal of Community Psychology, 27 (3), 281-298.

Greenberg, S. W., & Rohe, W. M. (1986). Informal social control and crime
prevention in modern urban neighborhoods. In R. B. Taylor, (Ed.), thigh
neighborhoods: Research and mlicy (pp. 79-118). New York: Praeger.

Hagan, J ., Gillis, A. R., & Chan, J. (1978). Explaining ofﬁcial delinquency: A

spatial study of class, conﬂict, and control. The Sociological Merly, 12, 386-398.

102

Heitgerd, J. L., & Bursik, R. J ., Jr. (1987). Extracommunity dynamics and the
ecology of delinquency. American Journal of Sociology, 92, 775-787.

Heller, K. (1989). The return to community. American Journal of Community
Psychology, 17(1), 1-15.

Hill, J. L. (1996). Psychological sense of community: Suggestions for future
research. Journal of Community Paychology,_2_4 (4), 431-438.

Hirsch, A. R. (1998). Making the second ghetto: Race and housing in Chicago
1940 — 1960. Chicago: University of Chicago Press.

Hirschi, T. (1969). Causes of delinquency. Berkley, CA: University of California
Press.

Hughey, J ., Speer, P. W., & Peterson, N. A. (1999). Sense of community in
community organizations: Structure and evidence of validity. Journal of Communig
Psychology, 27 (1), 97-113.

Hunter, A. J. (1985). Private, parochial, and public school orders: The problem of

Crime and Incivility in Urban Communities. In G. D. Suttles & M. N. Zald (Eds), 1113

challenge of social control: Citizenship and institution building in modern socieg (pp.

 

230-242). Norwood, NJ: Ablex Publishing.

Jargowsky, P. A. (1997). Povepty and place: Ghettos, barrios, and the Americap

 

c_ity. New York: Russell Sage Foundation.
Jargowsky, P. A., & Bane, M. J. (1991). Ghetto poverty in the United States,
1970 - 1980. In C. Jencks and P. Peterson (Eds), The urban underclass (pp.235-273).

Washington, DC: Brookings Institution.

103

Knott, L., & Brand-Williams, O. (2000, January 14). Young killer gets juvenile
detention. ﬁe Detroit Newa, p. 2.

Kasarda, J. D., & Janowitz, M. (1974). Community attachment in mass society.
American Sociological Review, 39 (June), 328-339.

Kingston, 8., Mitchell, R., Florin, P., & Stevenson, J. (1999). Sense of
community in neighborhoods as a multi-level construct. Journal of Community
Psychology, 27 (6), 681-694.

Komhauser, R. R. (1978). Social sources of delinquency. Chicago: University of
Chicago Press.

Krisberg, B., & Austin, J. (1978). The childre_n of ishmael: Critical parspectives
on juvenile justice. Palo Alto, CA: Mayﬁeld Publishing Company.

Leventhal, T., & Brooks-Gunn, J. (2000). The neighborhoods they live in: The
effects of neighborhood residence on child and adolescent outcomes. Psychological
Bulletip, 126(2), 309-337.

Logan, J. R., & Molotch, H. L. (1987). Urban Fortunaa: The Political Ecorhrmy
qulaqa. CA: University of California Press.

Lorion, R. P. & Newbrough, J. R. (1996). Psychological sense of community:
Pursuit of a ﬁeld’s spirit. MW (4), 311-314.

Maccoby, E. E., Johnson, J. P., & Church, R. M. (1968). Community integration
and the social control of juvenile delinquency. In J. R. Stratton & Terry, R. M. (Eds),
Prevention of delinquency: Problems and programs. New York: Macmillan.

McNeely, J. (1999). Community building. W81.

2,1 (6), 741-750.

104

Merton, R. K. (193 8). Social structure and anomie. American Sociological

Review 3 672-682.

 

Miller, W. B. (1958). Lower class culture as a generating milieu of gang

delinquency. Journal of Social Issues 14(3), 5-19

 

McMillan, D. (1976). Sense of community: An attempt at deﬁnition.
Unpublished manuscript, George Peabody College for Teachers, Nashville, TN.

Ofﬁce of Juvenile Justice and Delinquency Prevention (OJJDP). Statistical
brieﬁng book (1998). Online. Available: http://ojjdp.ncjrs.otg[ojsstatbb/qa255.htrnl.

Park, R. E., Burgess, E. W., & McKenzie, R. D. (1967). 1he_city. (Rev. Ed.).
Chicago: University of Chicago Press.

Perkins, D. D., Florin, P., Rich, R. C., Wandersman, A., & Chavis, D. M. (1990).
Participation and the social and physical environment of residential blocks: Crime and
community context. American Journal of Community Psychology, 18 (1), 83-115.

Perkins, D. D., & Taylor, R. B. (1996). Ecological assessments of community
disorder: Their relationship to fear of crime and theoretical implications. American
Journal of Communig Psychology, 24 (1), 63-105.

Pretty, G. H., Andrews, L., & Collett, C. (1994). Exploring adolescents’ sense of
community and its relationship to loneliness. WWW,
346-358.

Pretty, G. H., Conroy, C., Dugay, J., Fowler, K., & Williams, D. (1996). Sense of

community and its relevance to adolescents of all ages. Journal of Communig

Psychology, 24(4), 365-379.

105

Royal, M. A., & Rossi, R. J. (1996). Individual correlates of sense of community:
Findings from workplace and school. Journal of Communig Psychology, 24, 395-416.

Sampson, R. J. (1986). The effects of socioeconomic context on ofﬁcial reaction
to juvenile delinquency. American Sociological Review, 51, 876-885.

Sampson, R. J. (1991). Linking the micro- and macrolevel dimensions of
community social organization. Soci_al Forcea. 70(1), 43-64.

Sampson, R. J. (1993). Linking time and place: Dynamic contextualism and the
future of criminological theory. Journal of Research in Crime and Delinquency, 30 (4),
426-444.

Sampson, R. J. (1997). Collective regulation of adolescent misbehavior:
Validation results from eighty Chicago neighborhoods. Journal of Adolescence

Research 12 (2), 227-244.

 

Sampson, R. J., & Groves, W. B. (1989). Community structure and crime:
Testing social disorganization theory. American Journal of Sociology, 94 (4), 774-802.

Sampson, R. J ., Raudenbush, S. W., & Earls, F. (1997). Neighborhoods and
violent crime: A multilevel study of collective efﬁcacy. Scienqe, 277, 918-924.

Sarason, S. B. (1986). Commentary: The emergence of a conceptual center.
Journal of Communig Psychology, 14, 405-407.

Schweitzer, J. H., Kim, J. W., & Mackin, J. R. (1999). The impact of the built
environment on crime and fear of crime in urban neighborhoods. Journal of Urban
Technology, 6(3), 59-73.

Shaw, C. R., & McKay, H. D. (1942). Juvenile delinquency and urban areas.

Chicago: University of Chicago Press.

106

Shaw, C. R., & McKay, H. D. (1969). Juvenile delinquency and urbaLareaa
(Rev. ed.). Chicago: University of Chicago Press.

Shin, M. (1996). Ecological assessment: Introduction to the special issue.
American Journal of Community Psychology, 24 (1), 1-3.

Shoemaker, D. J. (1990). Theories of delinquency: An exmmn of delinquent
hehah/iol (2nd ed.). New York: Oxford University Press.

Simcha-Fagan, 0., & Schwartz, J. E. (1986). Neighborhood and delinquency:
Assessment of contextual effects. Criminology, 24 (4), 667-698.

Simons, R. L., Johnson, C., Beaman, J ., Conger, R. D., & Whitbeck, L. B. (1996).
Parents and peer group as mediators of the effect of community structure on adolescent
problem behavior. American Journal of Community Psychology, 24 (1), 145-172.

Taylor, R. B. (1997). Social order and disorder of street blocks and
neighborhoods: Ecology, rrricroecology, and the systemic model of social
disorganization. Journal of Research in Crime an‘d Delinqpency. 34 (1), 113-155.

Taylor, R. B., Gottfredson, S. D., & Brower, S. (1984). Block crime and fear:
Defensible space, local social ties, and territorial functioning. Journal of Research in
Crime and Delinquency, 21 (4), 303-331.

Unger, D. G., & Wandersman, A. (1982). Neighboring in an urban environment.
American Journal of Community Psychology, 10 (5), 493-509.

Wilson, W. J. (1996). Wars: The world of the new urbah poor.

New York: Vintage Books.

107

Yee, A. (1998). Juvenile crime and justice legislation: 1998. State Legislative
Report: December 1998 (Vol. 23) pp. 1-23. Online. Available:
http://www.ncsl.org/programs/cj/98juv.htm.

Zaff, J. & Devlin, A. S. (1998). Sense of community in housing for the elderly.

Journal of Community Paychology, 26 (4), 381-398.

108

IIIIIIIIIIIIIIIIIIIIIIIIIIIIII

IlllililiillIlllllllllilIlllllll

3 02088 449