Rediscovering the communication engine of neighborhood effects: how the interaction of residents and community institutions impacts health literacy and how it can be leveraged to improve health c...

PDF

Download

Open document

Flip pages

Summarize

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content

REDISCOVERING THE COMMUNICATION ENGINE
OF NEIGHBORHOOD EFFECTS:
HOW THE INTERACTION OF RESIDENTS AND COMMUNITY INSTITUTIONS
IMPACTS HEALTH LITERACY AND HOW IT CAN BE LEVERAGED
TO IMPROVE HEALTH CARE ACCESS

by
Matthew D. Matsaganis
______________________________________________________________

A Dissertation Presented to the
FACULTY OF THE GRADUATE SCHOOL
UNIVERSITY OF SOUTHERN CALIFORNIA
In Partial Fulfillment of the
Requirements for the Degree
DOCTOR OF PHILOSOPHY
(COMMUNICATION)

December 2008

Copyright 2008 Matthew D. Matsaganis

ii

DEDICATION
To my grandfather, William F. Richardson, Jr. (1921-2008), who planted in me the
curiosity about how the world works, and how to make it better.
This is for you Bompin.

iii

ACKNOWLEDGEMENTS
It is hard to imagine this project coming into fruition without having benefited
from the ongoing academic discussion with my advisor and mentor Sandra Ball-
Rokeach; a discussion, among other things, on the complex nature of the role
organizations play in the everyday lives of diverse ethnic communities. While this project
has come to an end – even tentatively – I am thankful that our discussion continues.
While burdened with additional responsibilities at the university during the past year, she
has always made herself available when needed and provided me with her valuable
feedback.
I would also like to thank the other members of my dissertation committee, Sheila
Murphy, Tom Valente, and Michael Cody. Both Sheila Murphy’s and Tom Valente’s
feedback in the early stages of the project was critical, particularly in designing the
survey instruments used to collect the data from organizations in Greater Crenshaw. And
although I should thank Michael Cody for his feedback and comments during the
development of the blueprint to my dissertation (i.e., my dissertation proposal), I feel
more compelled to mention the positive effect of the pressure I felt knowing that, as our
offices are across the hallway from each other, that I would sooner or later run into him
and might have to answer a question like, “so how is the dissertation going?” or “how
many chapters to go?”
Graduate school can be a roller-coaster of a ride sometimes and one of the most
valuable lessons your learn is that you need at least one kindred spirit to help you weather
the storms and celebrate those little or more significant victories that, most likely, no
other friend or family member will ever appreciate. Through these ups and downs, my

iv

friend and colleague Vikki Katz has always been there and has encouraged me to become
a better researcher – often in ways that elude her.
I can’t but recognize the very important role the Metamorphosis Project research
team has played over the years in fine-tuning my ideas and research agenda. I have to
give a special thanks to Meghan Moran for helping me troubleshoot through
methodological and analytical problems, Carmen for her insights into the world of
community-based organizations in South Los Angeles, and Evelyn for helping me co-
ordinate the fieldwork and the administrative aspects of my relationship with my research
assistants, Matt Grindal, LaShawn White, and Christine Sathre. A special thanks goes out
to Matt for his diligence to the project, his persistence in getting more organizations to
participate in the project, and his continued interest in bettering our interview instruments
and data management procedures.
I want to also thank two more (now former) members of the Metamorphosis
Project, Assistant Professors Holley Wilkin and Yong Chan Kim for their input. Over the
past year we have been in frequent communication about a National Institutes of Health
grant we submitted. Through our meetings I got the opportunity to test some of the ideas
that are found in the chapters that lie ahead and refine them. It was an immensely helpful
exercise. Moreover, I am sure that the slight pressure imposed on me by having to
provide short dissertation progress reports during our teleconferences helped, too.
Starting up a fairly large fieldwork project can be challenging and occasionally
daunting. In the early stages of this project I was more than happy to have the support and
friendly ear of Marc Davidson from the First 5 L.A. Commission, with whom we met
through the Metamorphosis Project, and who helped me connect with some of the

v

organizations in the study’s sample. In addition, I would like to thank Nancy Watson of
the Community Health Councils, Reverend Chip Murray, Shannon Whaley of PHFE-
WIC, Heidi Kent and Stephan Baranov of BIOMED-WIC, and Charlene Klink of the Los
Angeles Public Library system for their help in getting in touch with organizations and
people who we needed to interview.
I could say many things about the role my parents played in this project, my
decision to re-enter graduate school to get a Ph.D., how they have influenced the way I
see the world and think about it. But what I am most grateful to them for is their
unbending support and encouragement to pursue my goals and dreams. During this past
year my mother has been, on more than one occasion, my second pair of eyes, reviewing
things I have written and helping me improve my writing style. And, of course, as
always, my father has made sure that I stayed the course I planned by ‘demanding’
frequent progress reports and wanting to be part of troubleshooting problems I
encountered along the way.
I also have to thank my friend Angeliki Kanavou, who I met at USC just as she
was getting ready to leave the university with her Ph.D., for her terrific support over these
past few years and her complete faith in my abilities as a researcher, even at times when I
questioned myself, as well as the phone calls to check on how I was progressing even
when she was half a world away.
During this past year I have also been tremendously grateful for the support of my
personal cheerleading squad, Christo, Effie, and Solero, who encouraged when needed,
pushed when necessary, and insisted on getting away from my computer screen every
now and then to chat about anything but what I was working on.

vi

Last, but definitely not least, I am grateful to Marc for his patience with me and
his ability to keep me sane, even at the expense of his sanity. I could not have arrived at
this point without him.

vii

TABLE OF CONTENTS
DEDICATION ii
ACKNOWLEDGMENTS iii
LIST OF TABLES xi
LIST OF FIGURES xiv
ABSTRACT xv
CHAPTER 1
Introduction: Research problems and goals of the study, and why they matter
The key research problems this study sets out to address 3
The absence of communication from the neighborhood effects literature 4
Understanding the dual role of institutional neighborhood resources 6
Community Communication Capital 7
Extending communication infrastructure theory: the role of meso-level actors 9
Extending CIT: how does ethnicity impact the STN; what are the implications? 10
Contributions to the public health literature 13
Institutional neighborhood actors and health: the potential, broader impact of
this study 15
The impact of neighborhood health effects research on policy and grassroots
activism 16
Organization of the dissertation 21
CHAPTER 2
Theoretical framework I:
Communication as a mechanism of neighborhood effects
Communication in the neighborhood effects literature 25
The role of the residential community in communication research 27
(a) Communication behavior and media studies in social context 27
(b) The role of communication in shaping the social neighborhood environment 28
Re-introducing communication as a social mechanism of neighborhood effects 30
The general model of neighborhood effects 31
Structural neighborhood characteristics 32
(a) The effects of concentrated disadvantage 32
(b) Social-ecological aspects of neighborhood structure 34
Beyond structural characteristics: processes and mechanisms of effects 35
Communication as an antecedent process of neighborhood effects 38
Communication infrastructure theory and neighborhood effects 40
i. The storytelling network of neighborhood actors 40

viii

ii. The neighborhood environment as a communication action context 41
iii. CIT and neighborhood effects mechanisms 42
iv. Limitations of communication infrastructure theory 47
The research site 48
CHAPTER 3
Theoretical framework II & Hypotheses:
The neighborhood communication infrastructure and health
Disparities in access to health care and health literacy 56
Disparities in access to health care 56
Access to health care and health literacy disparities 57
Neighborhood determinants of health disparities 59
Pathways of neighborhood influences on health 60
I. Socioeconomic status 60
II. Mechanisms of neighborhood influence on health 62
III. Socioeconomic status and other neighborhood structural characteristics 63
IV. Neighborhood social environment stressors 64
V. Neighborhood physical environment stressors 65
Neighborhood effects and health: limitations of the existing research 65
Applying communication infrastructure theory to the study of neighborhood health 67
The integrated storytelling network (STN) 67
Neighborhood effects 78
I. Direct effects 78
(1) Health literacy 78
Hypothesis 1 79
(2) Access to health care resources 79
Research Question 1 79
(3) Civic engagement 79
Hypothesis 2 79
II. Indirect effects 80
Hypotheses 3a & 3b 80
III. The impact of the communication action context: direct and indirect effects 80
(1) Density of institutional resources 81
Hypothesis 4 81
Hypotheses 5a & 5b 82
Research Questions 2a & 2b 82
(2) The impact of the ethnic heterogeneity 83
Research Question 3 84
Research Questions 4a & 4b 85
Research Questions 5a & 5b 85
Research Questions 6a & 6b 86
(3) Residential stability 86
Hypotheses 6a & 6b 87

ix

CHAPTER 4
Methodology
The multi-level neighborhood storytelling network: data collection and measures 90
I. Residents 90
a. Sampling & Data Collection 90
b. Measures 91
II. Organizational neighborhood actors 98
a. Sampling & Data Collection 98
b. Measures 105
The neighborhood communication action context: data and measures 114
I. Density of institutional resources 114
II. Population composition-related variables 116
Analysis 117
I. Preliminary analyses 117
II. Analysis plan by hypothesis 120
CHAPTER 5
Results
Descriptive Statistics 127
The role of the storytelling network in civic engagement and health 128
Hypothesis 1 129
Research Question 1 144
Hypothesis 2 157
The storytelling network as a mechanism of neighborhood health effects 168
Hypotheses 3a & 3b 168
The role of the communication action context in neighborhood health effects 178
Hypothesis 4 179
Hypotheses 5a & 5b 182
Research Questions 2a & 2b 183
Research Question 3 186
Research Questions 4a & 4b 186
Research Questions 5a & 5b 189
Research Questions 6a & 6b 189
Hypotheses 6a & 6b 190
Summary of results 194
CHAPTER 6
Discussion and Implications
Theoretical and methodological contributions 202
The five challenges 202
Towards a communication-based theoretical approach to neighborhood effects 205
Understanding the dual role of institutional neighborhood actors 206
The pragmatic benefits of understanding the dual role of institutional resources 207
Diagnosing a fragmented storytelling network at the micro & meso level 207
Methodological contributions: assessing organizations’ integration into the STN 209

x

Community communication capital: an ecometric measure 211
Major findings and implications 211
Residents’ STN integration, a significant predictor of prevention health literacy 212
Individual-level communication capital and health literacy: the power is in the
interaction 213
Organizations matter, but the onus is on the residents: the ideal scenario versus
reality 214
Communities are not monolithic: dealing with a fragmented STN 216
The effects of an ethnic group divide and a bifurcated STN on health literacy 218
Positive signs of change and the challenge of inclusivity: the example of
SCOPE 220
Health care access and the STN 221
Do your homework: the importance of knowing your research site well 222
Negative STN effects and using the STN to turn things around 223
The role of organizations in the STN with respect to health access 223
The integrated multilevel STN & other neighborhood effects mechanisms 223
The role of organizations’ integration into the storytelling network in civic
engagement 224
The impact of individual-level communication capital on health literacy via
civic engagement 225
The role of the communication action context in neighborhood effects 226
Role of organizational density in civic engagement 226
Institutional resource density and health literacy 227
The impact of the CAC on individual-level communication capital in the
neighborhoods of Greater Crenshaw 228
The communication action context and health literacy 228
Limitations 229
Suggestions for future research 231
REFERENCES 235
APPENDICES
Appendix A. Health outcomes: health literacy-related items 252
Appendix B. Classification scheme for organizations mentioned by participant
organizations for the scope of inter-organizational connectedness measure 256
Appendix C. Organizations African Americans and Latinos in Crenshaw cited as
most important 257

xi

LIST OF TABLES
Table 1.1. Top media connections for African Americans and Latinos in Greater
Crenshaw 11
Table 2.1. Population profile of Greater Crenshaw and Los Angeles County 49
Table 2.2. Top-10 most pressing issues in Greater Crenshaw, 2006 (N=607) 50
Table 2.3. Health status, access to health care, and unhealthy behaviors in SPA 6,
in 2007 51
Table 2.4. Health outcomes in SPA 6 (2007) 52
Table 3.1. Scenarios for a bifurcated neighborhood storytelling network along
ethnic lines 85
Table 4.1. Top media choices for African American and Latino residents
of Greater Crenshaw 100
Table 4.2. Summary of institutional resources in Greater Crenshaw
(media not included) 100
Table 4.3. Rules applied for sampling organizations in Greater Crenshaw 102
Table 4.4. Storytelling network-related variables created and used in the
present study 115
Table 4.5. Moran’s I test of spatial autocorrelation for variables used as outcomes 119
Table 5.1. Independent and outcome variable descriptive statistics for Greater
Crenshaw 128
Table 5.2. Hierarchical multiple regression: STN integration & health literacy
(prevention) 131
Table 5.3a. STN & prevention health literacy for Greater Crenshaw African
Americans 133
Table 5.3b. STN & prevention health literacy for Greater Crenshaw Latinos 134
Table 5.4. The role of IIOC in civic engagement & health outcomes
(results of 1-way ANOVAs) 135
Table 5.5. The role of SIOC in civic engagement & health outcomes
(results of 1-way ANOVAs) 136
Table 5.6. The role of MSNC in civic engagement & health outcomes
(results of 1-way ANOVAs) 137

xii

Table 5.7. The role of CSNI
GD
in civic engagement & health outcomes:
1-way ANOVA results 138
Table 5.8. The role of CSNI
CE
in civic engagement & health outcomes:
1-way ANOVA results 138
Table 5.9a-5.9b. The role of OISN in civic engagement & health literacy:
1-way ANOVAs 139
Table 5.10a. Geo-demographic reach & prevention health literacy:
hierarchical regression 141
Table 5.10b. Impact of predictors including CSNIGD on health literacy
(regression coefficients) 142
Table 5.11. Hierarchical multiple regression: STN integration & difficulty getting
health care 146
Table 5.12a. Hierarchical multiple regression: STN integration & neighborhood
belonging 154
Table 5.12b. Hierarchical multiple regression: STN integration & collective
efficacy 155
Table 5.12c. Hierarchical multiple regression: STN integration & civic participation 156
Table 5.13a. Hierarchical regression: STN & belonging for African Americans in
Crenshaw 157
Table 5.13b. Hierarchical regression: STN & belonging for Latinos in Crenshaw 158
Table 5.14. Hierarchical regression: STN & collective efficacy for Latinos in
Crenshaw 160
Table 5.15a. Hierarchical regression: STN & civic participation for African
Americans 161
Table 5.15b. Hierarchical regression: STN & civic participation for Latinos 163
Table 5.16. Correlations among exogenous, endogenous variables in the SEM
analyses 170
Table 5.17. Theoretical model: significant direct and indirect effects 173
Table 5.18. Summary of LISREL model revisions 174
Table 5.19. Revised model: significant direct and indirect effects 176

xiii

Table 5.20. Density of institutional resources and collective efficacy
(hierarchical regression) 181
Table 5.21. Interaction of organizational density & collective efficacy
(post-hoc analysis) 185
Table 5.22. Summary of results 194

xiv

LIST OF FIGURES
Figure 1.1. Citations for the study of neighborhood effects and public health 2
Figure 1.2. Hypothetical scenario for a neighborhood with more than two
storytelling networks 12
Figure 2.1. Articles with ‘Neighborhood’ or ‘Social Capital’ in title
(Social Citation Index) 24
Figure 2.2. How the neighborhood impacts the lives of residents 38
Figure 2.3. Components of the neighborhood communication infrastructure 42
Figure 3.1. Geo-ethnic media and their relationship to other neighborhood
storytellers in the storytelling network 68
Figure 3.2. Community-based organizations and their relationship to other
neighborhood actors in the storytelling network 69
Figure 3.3. The health storytelling network of the neighborhood communication
infrastructure 70
Figure 3.4. Health service providers and their relationship to other neighborhood
actors in the storytelling network 72
Figure 3.5. Residents and their relationship to other neighborhood actors in the
storytelling network 73
Figure 3.6. The communication action context and storytelling network relationship 81
Figure 3.7. The theoretical model and hypotheses guiding the study 87
Figure 4.1. Distribution of population of institutional resources in Greater Crenshaw
(N=333) 101
Figure 4.2. Organizations interviewed in Greater Crenshaw by type (N=80) 104
Figure 5.1. Theoretical model of neighborhood health effects from
a CIT perspective 172
Figure 5.2. Revised model of neighborhood health effects from a CIT perspective 174
Figure 5.3a. Final model of neighborhood health effects for African Americans 177
Figure 5.3b. Final model of neighborhood health effects for Latinos 177
Figure 5.4. Ethnic reach of organizations interviewed in Greater Crenshaw (N=80) 187

xv

ABSTRACT
While the impetus for research that will help us understand what it is about the
places we live in that matters when it comes to our health seems to be growing
exponentially in the public health literature, our knowledge about the social mechanisms
through which place or context-related effects manifest is still limited. The cross-
pollination of public health research on the social determinants of health, on one hand,
and, on the other, sociological investigation into the influences of various social
processes on different facets of human life and behavior has been fruitful. However,
important gaps remain unaddressed.
The primary objectives of this dissertation are to re-introduce communication as
an elementary social process through which individuals and communities organize their
lives, and to develop a communication-based model of neighborhood health effects that
can be applied by researchers, communication campaign professionals, and policy-
makers to improve health care access and health literacy in diverse ethnic communities.
Employing a multi-methodological and multi-level analytical framework, the
author also addresses the question of how institutional resources available in a residential
community operate as mechanisms of neighborhood effects, and investigates ways
through which to gauge their impact as neighborhood actors in building health literacy
and improving health care access.
The theoretical framework guiding this project extends prior research on
communication infrastructure theory and neighborhood effects theoretical models
developed in sociology.

xvi

The findings of the study indicate that the extent to which individuals are
connected to other neighbors, local and ethnically-targeted media, as well as community-
based organizations, is a critical factor in predicting prevention-oriented health literacy.
Institutional community actors can amplify the positive effects of being part of such a
neighborhood-wide storytelling network, even in circumstances where their independent
influence may be small or negligible. The significance of the interaction effect between
residents and institutional level actors is even larger in the case of predicting health care
access. However, as the results indicate in this case, the influence of an integrated
neighborhood storytelling network (STN) may be strongly and negatively affected by
environmental factors. As an information resource, the storytelling network may be
influenced by troubling developments in the community and breaking news, such as the
closure of a medical facility. This type of ‘bad stories’ can impact residents’ perceptions
of vulnerability and capacity to deal with health problems. The more connected to the
STN residents are, the more susceptible they are to these negative effects.

1

CHAPTER 1
INTRODUCTION: RESEARCH PROBLEMS
AND GOALS OF THE STUDY, AND WHY THEY MATTER
In the preface to their book Neighborhoods and Health, Kawachi and Berkman
(2003) note that,
[t]here are clear signals to indicate that researchers in public health and allied
social sciences are converging on the search for place-based influences on health.
Even a cursory search of the major professional journals in public health reveals
dozens of relevant studies published just in the past few years. Research funding
bodies, including the U.S. National Institutes of Health (NIH), have assigned
priority to the search for neighborhood effects, especially in the context of
explaining social inequalities in health (p. v, italics added).

The most compelling and recent example of the NIH’s commitment to research on
neighborhood effects and health is to be found in the launching of the National Children’s
Study (NCS) with the Children’s Health Act, passed by Congress in 2000. NCS is a
massive research endeavor with a 20-year horizon to examine the effects of
environmental influences on the health and development of more than 100,000 children
across the United States. NCS investigators will follow these children from before birth
and until their 21
st
birthday. In its mission statement, NCS defines the ‘environment’
broadly to include: (a) natural and man-made (or built) environment elements, (b)
biological and chemical factors, (c) social conditions, (d) behavioral influences and
outcomes, (e) genetics, (f) cultural and family influences and differences, as well as (g)
geographic locations. For the NCS to achieve its goals, the National Institute of Child
Health and Human Development, the umbrella organization under which NCS is
sponsored, has recruited researchers from the biomedical, life, physical, and social
sciences.

2

The convergence of multiple lines of scientific inquiry over the question of what
is it about the place people live in that matters when it comes to public health and the
lack of any “comprehensive text that surveys the theoretical and methodological
challenges involved in conducting research in this area” (p. v) created the impetus for
Kawachi and Berkman (2003) to write their book. Five years after that volume was
published, a simple citation search through the search engine of Google Scholar returned
thousands of references to research done on the relationship between neighborhoods and
health from 2000 to 2008 (see Figure 1.1).
Figure 1.1. Citations for the study of neighborhood effects and public health

Figure 1.1. The number of citations on neighborhood effects and public health research produced
by the Google Scholar search engine using four different combinations of keywords (April 2,
2008). The combination of “neighborhood” and “health” is the least restrictive as it allows the
search engine to return references for studies developed by scholars concerned with how place
factors into health, but who may not be accustomed to using the term “neighborhood effects”
(e.g., geographers), nor have a research agenda focused specifically on public health issues.

0
10,000
20,000
30,000
40,000
50,000
60,000
Neighborhood Effects and Public Health
Neighborhood Effects and Health
Neighborhood and Public Health
Neighborhood and Health
1,350
2,880
24,200
59,600
198
647
14,100
44,700
NUMBER OFCITATIONS PERCOMBINATION OFKEYWORDSSEARCHED
Citations (1990‐1999) Citations (2000‐2008)

3

In addition, there was a notable increase in the number of works produced in the
2000-2008 period, compared to the previous decade (1990-1999), a fact that suggests
interest in this area of research continues unabated (again, see Figure 1.1).
The key research problems this study sets out to address
Despite the explosion in the number of publications on neighborhood effects and
health produced in recent years, there are still several gaps in the research that remain to
be addressed. Kawachi and Berkman (2003), among others, have pointed out that
reviewing the quickly burgeoning literature on neighborhood effects and public health
reveals a fascination with exploring the potential of recently developed methodological
tools (e.g., multilevel modeling tools), which enable, for example, the examination of
how characteristics of the contexts in which we live our lives in (e.g., workgroup, school,
neighborhood, city) influence our perceptions, attitudes, behaviors, and activities. Often,
however, testing the applicability of new methodological tools happens at the expense of
furthering the conceptual thinking about how place and health are linked (Subramanian,
Jones, & Duncan, 2003; Raudenbush, 2003).
Arguably, the issue is not just developing new conceptual tools or fine-tuning
ones that have gained popularity recently in the public health literature. The idea of social
capital, for instance, as a pathway of neighborhood influence was first developed in the
work of sociologists, but it has become increasingly popular in public health research in
recent years (e.g., Altschuler, Somkin, & Adler, 2004; Carpiano, 2004; Kawachi, 1999).
As the interest in neighborhood effects spans many disciplines, it seems critically
important to improve scholars’ familiarity with the history of this line of research and
work being done across disciplinary divides. Making headway on both these fronts would

4

help avoid the creation of unnecessary jargon; most importantly though it would help fill
conceptual gaps, some of which are common across fields of inquiry and some that are
not. In the public health literature, for instance, the impact of neighborhood socio-
economic conditions on health have been studied rather extensively, but the influence of
social processes (i.e., other than social capital) on the health of residents has received
limited attention to date (see Chapter 3 for a more extensive review).
The absence of communication from the neighborhood effects literature
From a communication researcher’s point of view, it is curious that, even in the
neighborhood effects research conducted by sociologists, communication as a social
process through which place of residence impacts people’s lives is largely absent – often
masked behind terms such as ‘social interaction’ or implied in the examination of social
mechanisms like ‘social ties.’ It is curious, because communication was featured
prominently in the work of early 20
th
century Chicago School urban sociologists, to
which much of the current neighborhood effects literature is theoretically indebted.
Chicago School theorists like Park, Burgess, and McKenzie, argued that communication
was one of the critical processes of social organization shaping Chicago and other cities
of their time. The absence of communication in the burgeoning literature on
neighborhood effects and health is also rather surprising, given the importance health
researchers and professionals place on designing effective communication campaigns to
help people prevent, detect, and manage diseases. This ‘absence’ represents the first gap
that this study sets out to address.
I argue that communication should be (re-)considered as a social mechanism of
neighborhood effects that involves all neighborhood actors: individual residents,

5

interpersonal networks, and institutional community stakeholders. In this project,
building on previous communication research (e.g., Ball-Rokeach, Kim, & Matei, 2001;
Greer, 1962; Keller, 1977; Kim & Ball-Rokeach, 2006a) I test the idea that
communication is a more elementary process, through which other mechanisms of
neighborhood effects, including neighborhood belonging, collective efficacy, and civic
participation, are activated and influence residents’ lives.
Moreover, building on sociology-based theories of neighborhood effects and
communication infrastructure theory (e.g., Ball-Rokeach, Kim, & Matei, 2001; Kim &
Ball-Rokeach, 2006a), I develop a theoretical and analytical framework that allows for
the study of how communication can impact neighborhood health, and especially access
to health care and levels of health literacy in ethnically diverse, urban neighborhoods.
Communication infrastructure theory reflects an ecological approach to the study of
urban communities rooted in communication research. The communication infrastructure
consists of a neighborhood storytelling network (STN) situated in its residential
communication action context (CAC; see Chapter 2 and Figure 2.3 for more details). The
STN of a neighborhood emerges as result of a storytelling process, in which individual
residents and institutional actors (e.g., community organizations, health service providers,
and media) interact and construct a collective vision and a reality of the neighborhood as
a place where they belong and engage shared concerns. The strength of the STN depends
on the makeup of the CAC, or, in terms more commonly found in the neighborhood
effects and public health research, on the natural, built, and social environment of the
neighborhood. I suggest that the configuration of the communication infrastructure as a

6

whole affects whether or not other social mechanisms of effects are activated and the
nature and strength of the effects observed.
Understanding the dual role of institutional neighborhood resources
The significance of institutional resources available in a community has been
highlighted in many neighborhood effects studies, including several focused on health
(e.g., McKnight, 1995; Horowitz et al., 2004). People living in neighborhoods that have
more and better quality health care providers, more active community organizations,
more grocery stores that carry high quality foods and fresh produce, and more
recreational facilities where residents can exercise and children can play, are considered
to be better off. However, the review of the literature indicates that there is a felt
difficulty in distinguishing between the structural role institutions play in a community
and mediating institutional processes (see also: Sampson, Morenoff, & Gannon-Rowley,
2002). While a variety of studies have developed measures to capture and assess the
impact that density, for example, of particular resources has in a neighborhood, or the
extent to which residents become involved in the activities of community organizations,
they fail to conceptualize and examine the processes through which neighborhood
institutions actually shape the life of the communities they serve. This is the second gap
that this project aims to address.
In this study, I argue that taking a communication infrastructure theory approach
allows us to account for the dual role that institutional resources play in the lives of
residents. With no doubt, institutional resources are part of the infrastructure that helps
keep a community alive and in motion. Schools, churches, WIC centers, a variety of other
community-based organizations, health centers and clinics, all represent places residents

7

and their families visit to take care of certain needs, to receive services, and to meet, talk,
and work with neighbors to resolve personal or common problems. The absence of these
resources makes life in a community more difficult. Children may have to attend
overcrowded classrooms, which is likely to compromise their ability to learn; residents
may have to travel longer distances to see a doctor, which could mean that they will not
seek medical attention when they need it; and low-income, single mothers may not be
able to get the food stamps they need to pay for the food their young children must have,
thereby putting their well-being at risk.
However institutions are also neighborhood actors. To accomplish their mission
they develop partnerships to achieve common goals, they exchange information and learn
from each other’s experiences, successes and failures. One can imagine institutional
actors as knots in a net cast over a neighborhood or nodes in a network that encompasses
an entire residential community. The more tightly woven the net is or the stronger the ties
between the nodes of the network are, the more support institutional actors can provide to
the people they serve. Therefore it is not just the sheer number of organizations that are
present in a community that matters, but also the extent to which organizations actively
reach out to residents in the community and forge ties with other institutional
neighborhood actors.
Community Communication Capital
The ‘engine’ that powers the communication infrastructure is the neighborhood
storytelling network (STN). Prior research suggests that the more integrated the network
is, the more likely it is that residents will feel attached to their neighborhood (i.e.,
neighborhood belonging), the more likely it is that they will feel they can come together

8

to solve common problems (i.e., higher collective efficacy), and the higher the level of
civic participation will be (e.g., Kim & Ball-Rokeach, 2006b).
1
Neighborhood belonging,
collective efficacy, and civic participation can all be considered as mechanisms of
neighborhood effects that are linked to the neighborhood communication infrastructure. I
argue that the integration of the STN can be conceptualized as a measure of community
communication capital (CCC). As a concept, CCC is akin to social capital, as they are
both realized through ties that emerge and connect residents in the process of their
everyday life. However, social capital is usually limited to the individual level (i.e.,
interpersonal networks), and does not account for the integration of the multi-level
network of neighborhood actors (or storytellers). In addition, the storytelling aspect
incorporates the communicative dynamics not considered in social capital, or in most
neighborhood effects studies. Thus, the CCC differs in being both multi-level and
focused upon a communicative process.
The introduction of CCC as a community resource addresses a third gap in the
neighborhood effects literature. Sampson and Raudenbush (1999; also: Raudenbush,
2003; Sampson et al., 2002), among others, have highlighted the need to create measures
that capture neighborhood properties, and that are not derived solely from individual
level measures. They discuss this issue in terms of developing ecometric measures (as
opposed to psychometrics). CCC can be defended as a neighborhood level ecometric
measure, as it accounts for the degree of connectedness of neighborhood stakeholders

1
The strength of the storytelling network (STN), as mentioned earlier and as is discussed in Chapter 2 in
more detail, depends on the conditions that exist in the communication action context (CAC). The
availability of plenty of meeting and greeting places, for instance, to which residents of a neighborhood can
get easily and in which they feel safe, could support the building of communicative ties among residents
and other neighborhood actors.

9

across levels of analysis, within a neighborhood with particular environmental
characteristics.
Extending communication infrastructure theory: assessing the role of meso-level actors
The development of CCC also reflects an innovation made within communication
infrastructure theory. In communication infrastructure research (CIT), thus far, it has
been argued that the strength of the STN depends on the extent to which all neighborhood
actors connect to each other and prompt connections among other actors (i.e., how
integrated actors become into the network). Methodologically-speaking, however, the
impact of organizational or meso-level actors has only been assessed indirectly, utilizing
individual level survey data. Those data speak to the question of what meso-level actors
residents connect to, but do not directly capture the degree and scope of connections that
organizations forge in a community, or their reach across a neighborhood’s, often very
diverse, population. This is the fourth gap that this study seeks to address. It is not a gap
in the neighborhood effects research per se, but it is an important problem that needs to
be addressed so as to render communication research and communication infrastructure
theory, in particular, more useful in the study of neighborhood effects, health-related or
otherwise. For the purposes of this study, I undertook fieldwork that would allow me to
capture the communicative ties of different types of organizational actors in the
neighborhoods of the Greater Crenshaw area in South Los Angeles; that is, ties to other
institutional actors and to residents. These data are used in this project in combination
with data gathered previously from individual residents to diagnose the status of the
neighborhood STN.

10

Extending CIT: how does ethnicity impact the STN and what are the implications?
CIT research to date has been done in a variety of different residential
communities. However, for the most part, investigators have focused on specific ethno-
racial groups living in those neighborhoods (e.g., Latino residents of Southeast Los
Angeles neighborhoods or Chinese-origin residents in Monterey Park). In this study, the
individual level survey data come from two different ethnic groups living in the
neighborhoods of Greater Crenshaw: African Americans and Latinos (see Chapter 2 for a
more detailed description of the study area). Preliminary analyses of survey data and
focus groups with residents from both populations conducted by the Metamorphosis
Project, of which I have been a member since 2003, suggest that African Americans and
Latinos may connect to very different institutional resources to accomplish similar goals.
For instance, they connect to very different media to get information they need to stay on
top of and understand what is going on in their community. Table 1.1 presents the media
mentioned most frequently by African Americans and Latino survey participants in
Greater Crenshaw for this purpose. The media are ranked based on the number of times
each one was mentioned by the respondents. It is worth noting that only when it comes to
newspapers is there overlap between the two groups in terms of (a) the media they
connect to, and with respect (b) to the number of African Americans and Latinos who
mentioned these media.
This type of differential connection patterns along lines of ethnic identification
may also be observed at the meso level. Neighborhood institutional actors may in fact be
targeting particular ethnic groups in a community. This is not an uncommon practice
among some media and community-based organizations.

11

Table 1.1. Top media connections for African Americans and Latinos
in Greater Crenshaw

Television Radio Newspapers
African
Americans
KABC (Channel 7) KJLH 102.3 L.A. Times
KTTV (Fox, Channel 11) KKBT 100.3 (THE BEAT) L.A. Sentinel
KCAL (Channel 9) KFWB‐AM 980 The Wave
KCBS (Channel 2) KNX‐AM 1070 L.A. Watts Times
KNBC (Channel 4)
KTLA (Channel 5)

Latinos KMEX (Univision, Channel 34) KLAX 97.9 (LA RAZA) L.A. Times
Telemundo (Channel 52) KLVE 107.5 La Opinion
KWHY (Channel 22, Spanish) KSCA‐FM 101 (LA NUEVA) The Wave
KTTV (Fox, Channel 11) KTNQ‐AM 1020
KABC (Channel 7)
KCAL (Channel 9)
Legend
Mentioned by both groups
Mentioned by both groups as (or almost as) frequently

Different patterns of connections to institutional resources along ethnic lines
suggest that there is the possibility of observing more than one STN at work in a
particular residential community. In this study, I attempt to test whether or not there are
in fact two different storytelling networks at work in Greater Crenshaw neighborhoods;
and, if this is the case, to examine the implications for understanding neighborhood
effects, particularly with regard to health-related outcomes. Pursuing this line of research
questions adds a building block to CIT research, but also to neighborhood effects work.
Figure 1.2 presents a hypothetical scenario, in which there are two STNs at work
in a neighborhood. The STN for ethnic group 1 is highly integrated and strong. The STN
for ethnic group 2 is also integrated and strong. Looking at the community as a whole,
however, the STN appears weaker than it does for each one of the two groups

12

individually, and therefore the neighborhood’s community communication capital is
lower. The power of the whole neighborhood’s STN ‘engine’ will depend: (a) on the
degree to which there are institutional resources that serve the entire community (and not
one ethnic group), (b) the extent to which all residents, regardless of ethnic background,
connect to a number of the same institutional actors present in the neighborhood, and (c)
on the strength and quality of the ties that exist among institutional resources in the
community.

Figure 1.2. Hypothetical scenario for a neighborhood with
more than two storytelling networks

Figure 1.2. In this scenario, the STN for the neighborhood as a whole is weaker than the STN for
the two ethnic groups that comprise the residential community. Note that there is not an
‘averaging effect.’ The whole, in this case, may not be stronger than its parts.

Individuals and families living in a neighborhood with a completely bifurcated
STN may find it harder to tackle larger problems that affect the entire residential
community (e.g., the building of new schools), even if the different ethnic groups
comprising the neighborhood may hold the potential of effectively pursuing other goals

13

(e.g., effective dissemination of health information pertaining to diseases a particular
ethnic group is at a higher risk of developing).
Contributions to the public health literature

From a public health perspective, the development of a communication
infrastructure approach for the study of neighborhood effects creates an opportunity to
bridge the quickly expanding literature that aims to understand how where we live
matters to our health, with research done on health communication campaigns and health
disparities.
The communication action context (CAC) is one of the two key elements of a
neighborhood’s communication infrastructure. As previously noted, the CAC
encompasses all the characteristics of the physical, built, and social environment that
enables and constrains the neighborhood storytelling network (Ball-Rokeach, Kim, &
Matei, 2001). Various aspects of this multi-dimensional neighborhood environment (e.g.,
air pollution, presence of fast food restaurants, socio-economic status) have been studied
as determinants of health-related outcomes. Knowing how the environment impacts our
health is critical if we want to protect or improve people’s health. However, this
knowledge is not sufficient for health professionals, policymakers, and activists who are
called upon to design, develop, and implement community/place-based solutions. These
stakeholders need: (a) to know what the universe of possible solutions are, (b) to
communicate this information to diverse audiences, and (c) often negotiate available
options with different publics. Combating obesity and preventing Type-2 diabetes in a
community, for instance, may entail launching a health communication campaign to
inform residents about healthier eating habits; but it could also involve making more

14

complicated decisions about the type of new restaurants that are given a license to operate
in a community (e.g., fast food franchises versus restaurants that use only organic
produce). In both cases, the implementation of a solution requires that there is a
mechanism in place to support community-wide problem-solving initiatives. From a CIT
point of view, the storytelling network is such a mechanism. A strong and integrated STN
can minimize the effort, time, and cost necessary for information to reach the entire
community, and for neighborhood stakeholders to communicate with each other and
mobilize to address shared concerns.
In a communication infrastructure-based model of neighborhood effects, the
storytelling network is the ‘motor’ of the community. The STN may produce effects on
its own, as may be the case when it is harnessed to communicate important disease
prevention information to at-risk populations. But this ‘motor’ also activates other social
mechanisms that have been associated with particular neighborhood health-related
effects. For this project, I investigate the relationship of the STN and three mediating
social processes – i.e., neighborhood belonging, collective efficacy, and civic
participation – with respect to their impact on two outcomes: health literacy and health
care access. I selected health care access because it is one of the most critical health-
related problems residents in Greater Crenshaw are confronted with; and I chose health
literacy, (a) because prior research has linked it to health care access (i.e., higher health
literacy has been associated with increased levels of health care access) and individuals’
access to information resources(e.g., Andrulis, 2000), and (b) because of the increased
attention it has received in recent years in the public health literature (e.g., Ratzan, 2001).

15

The power of a residential community’s storytelling network and the
neighborhood environment are inextricably linked. Better understanding this relationship
allows researchers to examine how the environment is impacting residents and to design
intervention strategies tailored to the specific challenges a community faces. A one-size-
fits-all-type of media campaign to combat obesity, for instance, in an ethnically diverse
community may prove ineffectual, if the different populations that make up the
community connect to different media, participate in or contact different community
organizations, and trust different providers for their health care. The communication
infrastructure-based model of neighborhood effects developed in this project can be
applied as an intervention model; that is, a model, which integrates the diagnosis and
prescription elements of public health research.
Institutional neighborhood actors and health:
the potential, broader impact of this study
The impact of this study may be assessed from a research point of view, a policy
perspective, and from the standpoint of professionals in the health service sector, in
community organizations, and local media.
From a research point of view, this study delivers a comprehensive theoretical
model that allows for the study of the influence of both the structural and the agentic
roles of institutional resources in a residential community.
From the point of view of institutional actors, the neighborhood storytelling
network can serve as a valuable tool in planning their activities (e.g., campaigns to raise
awareness about child obesity) in a particular community and a guide in pursuing the
synergistic relationships necessary to achieve their goals. The model of neighborhood

16

effects developed in the pages that follow may also be used to assess the impact of
community-specific contextual factors, such as ethnic heterogeneity, in the
implementation phase of health campaigns. The impact of ethnic heterogeneity is
reflected, for instance, in the different media people of different ethnic origins connect to
most for health-related information (e.g., Wilkin, 2005). Moreover, this approach to the
study of neighborhood effects may be utilized by health service providers and community
organizations to evaluate the impact of their efforts.
Understanding the idiosyncrasies of a particular community is also important for
policy-makers. They need to be able to evaluate the factors that improve and limit the
effectiveness of institutional resources and to determine what changes in legislation and
funding priorities are necessary to achieve desired outcomes. Funding new health care
centers in a neighborhood, for instance, may be less important than ensuring that the ones
that exist are accessible to everyone in the residential community.
The impact of neighborhood health effects research on policy
and grassroots activism
The expansion of the literature on neighborhood effects and health has increased
the awareness of multiple publics particularly around issues of health disparities and
environmental justice. Policymakers have been urged by scientists to take action to limit
health inequalities, while grassroots organizations have been heartened to continue their
efforts to inform local communities about the social and environmental determinants of
health, to mobilize these communities to improve their living conditions, and to continue
to pressure legislators to enact policies that can lead to better health for more people.

17

At the international level, the World Health Organization (WHO) has shown a
keen interest in better understanding the effects of the social environment on people’s
health around the world. In March 2005, the WHO launched a special Commission on the
Social Determinants of Health (CSDH). The CSDH, comprised of 20 members, was
charged to work towards improving health equity globally, by bringing together leading
scientists and practitioners to provide evidence on policies that improve health by
addressing the social conditions, in which people live and work. CSDH members have
been working with countries to support policy change and have been monitoring the
results of such efforts. In the introduction to their Interim Statement in 2007, the
commissioners articulated their mission, expand on the global realities that created the
impetus for the establishment of the Commission, and at least indirectly call for more
research to be done on the interplay of the environment (i.e., social, built, and natural) we
live in and health:
Health is a universal human aspiration and a basic human need. The development
of society, rich or poor, can be judged by the quality of its population’s health,
how fairly health is distributed across the social spectrum, and the degree of
protection provided from disadvantage as a result of ill-health. Health equity is
central to this premise and to the work of the Commission on Social Determinants
of Health.
Strengthening health equity – globally and within countries – means going
beyond contemporary concentration on the immediate causes of disease. More
than any other global health endeavor, the Commission focuses on the “causes of
the causes” – the fundamental structures of social hierarchy and the socially
determined conditions these structures create in which people grow, live, work
and age – the social determinants of health.
The time for action is now: not just because better health makes economic sense,
but because it is right and just. The outcry against inequity has been intensifying
for many years from country to country around the world. These cries are forming
a global movement. The Commission on Social Determinants of Health places
action to ensure fair health at the head and the heart of that movement (WHO
CSDH, 2007, p. 3).

18

The CSDH’s final report was expected in May 2008. As of this writing, it has not
been released.
At the grassroots level there are also multiple efforts to develop place- and
specific neighborhood-based solutions for achieving health equity. In Southern
California, for instance, the California Center for Public Health Advocacy worked with
the Los Angeles County Health Department, the local school district, as well as other
organizations and city officials to pass and enforce a moratorium on the building of fast
food restaurant drive-thrus in the poor and predominantly Latino and African American
community of Baldwin Hills. The moratorium, says Rosa Soto of the California Center
for Public Health Advocacy, has helped block the establishment of even more fast food
restaurants in the area, in favor of restaurants where residents can find healthier food
options (Soto, 2008).
2

Another non-profit organization, The City Project, also based in Los Angeles, has
been fighting against the closure of a number of California State Parks, considered
necessary by Governor Schwarzenegger because the state is running a deficit of over $16
billion (Halper, 2008).
3
“A lot of these closures,” says City Project CEO Robert Carcia,
“would be taking place in counties that are already at a disadvantage in terms of how
much green space people have access to” (Garcia, 2008). Research has shown that the
absence of parks and recreation centers is linked to the prevalence of obesity and related
health diseases in a community (Booth, Pinkston, & Poston, 2005; Burdette & Whitaker,

2
Prior research has shown that poorer neighborhoods in South Los Angeles, which are predominantly
African American and Latino, have significantly fewer high quality restaurants, compared to more affluent
Los Angeles communities with mostly white residents, and a disproportionate number of fast food
restaurants. Regular consumption of fast food has been connected to obesity and related illnesses, including
Type 2 diabetes (Block, Scribner, & Desalvo, 2004; Lewis et al., 2005; Thompson et al., 2004).
3
According to a report in the Los Angeles Times on February 21, 2008, the state’s chief budget analysts
estimated California’s deficit at $16.5 billion (Halper, 2008).

19

2005; Gordon-Larsen, Nelson, Page, & Popken, 2006). Therefore, closing down parks in
already disadvantaged communities is likely to put the health of residents at even higher
risk.
Some of these grassroots efforts to raise awareness around how the social, natural,
and built environment affects our health have captured the attention of and also mobilized
media organizations to action. Recently, a four-hour long documentary series, with the
title Unnecessary Causes… Is Inequality Making Us Sick? was produced and aired across
the U.S. via the Public Broadcasting System (PBS). In this series, the producers explore
the racial and socioeconomic inequalities in health. Larry Edelman, the Executive
Producer of the documentary, explains what motivated his production team:
We produced Unnatural Causes to draw attention to the root causes of health and
illness and to help reframe the debate about health in America. Economic and
racial inequality are not abstract concepts but hospitalize and kill even more
people than cigarettes. The wages and benefits we're paid, the neighborhoods we
live in, the schools we attend, our access to resources and even our tax policies
are health issues every bit as critical as diet, smoking and exercise. The unequal
distribution of these social conditions - and their health consequences - are not
natural or inevitable. They are the result of choices that we as a community, as
states, and as a nation have made, and can make differently. Other nations already
have, and they live longer, healthier lives as a result (Edelman, 2008).

The producers of Unnatural Causes have summarized the findings of the research
they have brought into the spotlight in their series into a list of “10 things to know about
health”:
Research shows that social conditions – the jobs we do, the money we’re paid,
the schools we attend, the neighborhoods we live in – are as important to our
health as our genes, our behaviors and even our medical care.
Health is tied to the distribution of resources.
Racism imposes an added health burden.

20

The choices we make (e.g., if we smoke or eat healthy food) are shaped by the
choices we have (e.g., if the neighborhood we live in has grocery stores with
fresh produce).
High demand plus low control leads to chronic stress: it’s not the CEOs who
are dying of heart attacks, it’s their subordinates.
Chronic stress can be deadly.
Inequality – economic and political – is bad for our health.
Social policy is health policy.
Health inequalities are not natural; that is to say health disparities that arise
from our social and class inequities result from decisions we as a society have
made – and can make differently.
We all pay the price for poor health. Americans spend approximately $2
trillion a year on health care. That is double the amount citizens of other
industrialized countries spend on average. Yet the U.S. ranks 30
th
in the world
with regard to life expectancy and 31
st
with respect to infant mortality. In
addition, the cost businesses incur due to lost productivity because of illness is
estimated at $1 trillion a year (California Newsreel, 2008).
The existing literature, the efforts made to bridge interdisciplinary divides, and
the gaps identified in the neighborhood effects literature and public health created the
impetus for this study. But it is out of the reports of policymakers, like the WHO, and the
work of activist groups and non-profit organizations – including those that participated in
this study – that the moral argument for pursuing this line of inquiry arises and becomes
crystal clear. A better understanding of the mechanisms through which neighborhood

21

effects are produced and the ability to identify ‘junctures of intervention,’ may contribute
to the development and implementation of policy solutions that can lead to greater health
equity. This is certainly the hope and overarching goal of this project.
Organization of the dissertation
This introduction to the key research problems that drive this project is followed
by two chapters, where I unfold the theoretical framework that guides the study. Chapter
2 tackles four challenges related to the question of how communication fits into the
broader literature of neighborhood effects and presents a communication-centered model
of neighborhood effects. In Chapter 2, I also introduce the research site and attempt to
sketch Greater Crenshaw’s health profile. Chapter 3 applies the communication-based
theoretical framework of neighborhood effects developed in Chapter 2 to health, and
particularly to the study of health literacy and access to health care; both these issues are
particularly salient to the residents of Greater Crenshaw neighborhoods. Chapter 3
concludes with the research questions and hypotheses examined in this study. In Chapter
4, I present the multi-methodological, multi-level analytical framework employed to
conduct this research. I present the key findings of the study in Chapter 5. The study
concludes with Chapter 6, where I discuss the significance of the findings, their
theoretical implications, as well as their implications for future research, policymaking,
and health-oriented intervention campaigns.

22

CHAPTER 2
THEORETICAL FRAMEWORK I:
COMMUNICATION AS A MECHANISM OF NEIGHBORHOOD EFFECTS
Around the mid-1800s, technological innovation in transportation broke, says
Melvin (1987), the “casement of the walking city” (p. 258). As people moved further
away from the densely populated and geographically limited urban centers, they created
and settled into neighborhoods. Researchers began to see cities, counties, states, and the
entire country, more and more, as a union of many identifiable units, a quilt made up of
distinct neighborhoods (Philips, 1940; Woods, 1923). In the early days of the Chicago
School of urban community ecology, between 1915 and 1925, Park, Burgess, McKenzie,
and others pointed to transportation, but also to communication as two of the key
mechanisms of social organization shaping the American urban communities of their
time. Park in particular, perhaps because he had been trained as a journalist, argued that
“transportation and communication [italics added] are primary factors in the ecological
organization of the city” (Park, 1925/1967, p. 2).
1

The Chicago School of the early 20
th
century carved a path of research focused on
understanding how where people live impacts their life and that of their neighbors. This
path of ‘neighborhood effects’ research (Aneshensel & Sucoff, 1996; Brooks-Gunn,
Duncan, Klebakov, & Sealand, 1993; Brooks-Gunn, Duncan, & Aber, 1997; Coulton,
Korbin, & Su, 1999; Elliott, Wilson, Huizinga, Sampson, Elliott, & Rankin, 1996;
Gephart, 1997; Jencks & Mayer, 1990; Leventhal & Brooks-Gunn, 2000; Morenoff,
Sampson, & Raudenbush, 1999; Sampson, Morenoff, & Gannon-Rowley, 2002),

1
For a historical and critical account of the relationship between communication and urban community
ecology studies, see also Burd (2004).

23

however, is marked by discontinuity – at least through the first half of the 1990s. The
inconsistency in the development of the literature can be attributed to three factors: (a)
debates over and shifts in the perceived salience of the concept of neighborhood, (b)
methodological constraints, and (c) the general historical context.
In the 1930s and 1940s, says Smith (1982, in Melvin, 1987), “the general crisis
over the apparent loss of community… was accompanied by a shift away from viewing
the neighborhood as a central element of American cities toward an increasing concern
about the metropolitan or regional unit” (p. 261). A decade later, in the 1950s, the
decreasing importance placed on the neighborhood as a territorial unit was captured in
the idea that neighborhoods represent communities of limited liability (Janowitz, 1967).
Janowitz suggested that residents maintain ties to a variety of different communities, such
as religious or occupational groups, and that they interact with their immediate
surroundings and their neighbors only to the extent that it serves their individual lifestyle.
Ten years later, however, the pendulum again swung. In the turbulent 1960s, a
decade marked by the struggle of civil rights activists and political assassinations,
policymakers and researchers looked at neighborhoods as the appropriate staging grounds
for interventions that would allow them to tackle poverty, inter-ethnic group and racial
conflict, and a host of related social problems. This is the time when President Lyndon B.
Johnson established the Office of Economic Opportunity to administer initiatives such as
the Community Action Programs (CAPS). CAPS were meant to serve his War on Poverty
agenda (Brauer, 1982; Jeffres, 2002). In this context and throughout the 1960s and the
1970s, there were spurts of publishing activity in the area of neighborhood effects.

24

However, a decline in interest with respect to neighborhood studies followed,
once again. This time it was attributed largely to methodological constraints, especially
with regard to modeling social processes through which neighborhood effects were
thought to be produced (Sampson et al., 2002). However, progress on this and other
fronts, such as the development of operational neighborhood definitions that are closer to
residents’ perceptions of their residential community’s boundaries, led to a resurgence of
interest in this line of inquiry in the early 1990s (see Figure 2.1).
2

Figure 2.1. Articles with ‘Neighborhood’ or ‘Social Capital’ in title
(Social Citation Index)

Source: Sampson, Morenoff, & Gannon-Rowley (2002)

2
Sampson et al. (2002) estimate that by the year 2000 there were about 100 neighborhood effects studies
published per year. However, this number could be much larger, as their review was focused primarily on
literature developed in sociology and education, and it neglects studies published in other fields (e.g.,
economics).

25

Communication in the neighborhood effects literature
Despite the growth in the neighborhood effects literature, communication has
been, at least since the mid-1990s, largely absent from the list of mechanisms or social
processes through which neighborhood effects are thought to manifest. Often,
communication is masked behind terms like ‘social ties’ or ‘social interaction.’
Communication, however, as conceptualized by the Chicago School sociologists, is more
than the interaction between neighbors that share a backyard fence or that meet at the
local grocery store.
Park and others linked social interaction (or interpersonal communication) to the
emergence of public opinion, conceptualized as a mechanism of social control in the
urban community. He argued that in urban neighborhoods founded on secondary
relationships (as opposed to primary relationships that connect family members), where
residents’ interactions with one another are “immediate and unreflecting” (Park,
1925/1967, p. 24), it is “public opinion – and not mores – that becomes the dominant
force in social control” (p. 38). Park believed that public opinion may be controlled,
enlightened, and exploited through the media. The first and the most important of
“agencies and devices” to influence public opinion, says Park, is the press. “The
newspaper,” he argues, “is the great medium of communication within the city, and it is
on the basis of the information which it supplies that public opinion rests (Park,
1925/1967, p. 38-39).”
Park’s reflections on the role of the media illustrate that, for the early community
ecology theorists, communication is a process that involves multiple neighborhood
actors, including individual residents and institutional actors (e.g., media organizations).

26

This train of thought supports the notion that communication is (a) more than just social
interaction between residents and (b) that communication should be conceived as a social
process of neighborhood effects that takes place across multiple levels of analysis (i.e.,
the micro or individual level, the meso or organizational/institutional level, and the macro
or community level).
In the 75 plus years since the heyday of the Chicago School of social ecology,
neighborhood effects literature and communication as a discipline have come a long way.
On one hand, today the study of neighborhood effects is not within the purview of
sociology alone. Economists, epidemiologists, anthropologists, education scholars, and
social geographers, among others, have published a multitude of studies that have
brought a diverse set of theories and methodological approaches to bear on the question
of how a community’s physical and social environment affects the lives of residents.
These studies reflect progress but also highlight persistent and new challenges, some of
which will be addressed later in this chapter and this study, more generally speaking. For
now, suffice it to say that the fact that so many disciplines, outside sociology, have
engaged the central question of neighborhood effects work, “how do our social and
physical environs impact us?” offers a partial explanation for why communication is not
found in the neighborhood effects research. Communication is not a term that one is
likely to find in the conceptual toolbox of economists and epidemiologists, among others.
But it is only a partial explanation, because communication is missing even in the
neighborhood effects work produced in sociology.

27

The role of the residential community in communication research
In the past five decades, a number of communication scholars have demonstrated
an interest in trying to understand how (a) people’s interpersonal and mass
communication networks are constrained by the community and environment in which
they live, and (b) how communication processes impact the development of a residential
community.
(a) Communication behavior and media studies in social context
Huckfeldt and Sprague (1987), for instance, found that people tend to include
politically like-minded neighbors in their personal information networks, but that the
configuration of these networks also depends significantly on the diversity of political
views available in their social environment. The authors concluded that information-
transmitting processes, like talking with your neighbors, interact with the social context
in a way that “favors partisan majorities while undermining political minorities” (p.
1197). In one of a series of studies, Wilkin, Ball-Rokeach, Matsaganis, and Cheong
(2007) found that how heavily people rely on particular media to get information that
would help them stay on top of what is happening in their community depends on (a) the
variety of media that are available to them and to which they have access (e.g., having or
not having access to the Internet matters), (b) their ethnic background, and (c) the
geographic location of their neighborhood. Residents of Hispanic origin living in
Southeast Los Angeles, for example, rely more heavily on ethnically-targeted television.
To accomplish the same goal, Hispanics living in the city of Glendale (also in the broader
Los Angeles area) connect mostly to mainstream, English-language television. The
population of Glendale is very diverse, ethnicity-wise, and the community is more

28

established and stable compared to the fairly transient neighborhoods of Southeast Los
Angeles, where the population is predominantly Hispanic of Mexican-origin (see also:
Ball-Rokeach, Cheong, Wilkin, & Matsaganis, 2004).
The impact of social context on media content has been the subject of a number of
mass communication studies. In the work of Tichenor, Donohue, and Olien (1980),
Viswanath and Demers (1999), and others, the mass media are thought to reflect the
communities they serve. They are seen, as Park argued several decades earlier, as agents
of social control. In smaller and more homogeneous communities, media (especially
newspapers) tend to avoid conflict and promote consensus. On the contrary, in more
heterogeneous communities, the media are expected to focus on the ongoing conflict, its
causes and the key stakeholders, thereby fostering the dialogue necessary to eventually
resolve the community’s problems (Downwoody & Griffin, 1999; Hindman, 1996;
Hindman, Littlefield, Preston, & Neumann, 1999; Tichenor et al., 1970; Viall, 1992).
(b) The role of communication in shaping the social neighborhood environment
Because communication is generally conceptualized as a recursive social process
(i.e., it includes feedback loops) that involves one or more senders and receivers, it is
perhaps not surprising that communication researchers have been sensitive to and have
investigated the bidirectional relationship between communication and social context.
Greer (1962) and Keller (1977) both have argued that communication, interpersonal and
mass communication, is the set of processes through which interdependent social groups
in urban communities coordinate their behavior. Communication, said Keller, “is not only
necessary for the formation of human communities, it is also indispensable for sustaining
them” (p. 282). People’s sense of belonging to a community develops as a result of their

29

everyday communication with neighbors at public celebrations, school board meetings, or
other occasions where different interest groups represented in the neighborhood’s
population are forced to interact with one another (Doheny-Farina, 1996).
In a different study, applying Fishbein and Ajzen’s (1975) theory of reasoned
action, Jeffres, Dobos, & Sweeney (1987) examined what determines residents’ decision
to continue living in a neighborhood. Their findings complement the work of Keller and
others. Jeffres et al. found that it is through communication with neighbors that residents
form beliefs and attitudes about their community, which in turn determine their decision
to leave or stay in a community.
This brief review suggests that communication scholars have developed an
interest in better understanding what influence social context particularities have on
individual and community-level communication behaviors. However, with few
exceptions which will be discussed later in this chapter, communication researchers have
generally not accounted for the impact of a community’s physical makeup on residents’
attitudes and communication behaviors. Does having safe parks in a neighborhood, for
instance, have a significant effect on how often residents talk to one another? If so, why
is that? In addition, while social context and communication processes are frequently
linked, it is not often that specific effects of the interaction between context and process
are examined. If having safe parks in a neighborhood encourages residents to talk to each
other more, are these same people also better informed about where they can go in the
community to find medical help, the best child care services, or the freshest produce?

30

Re-introducing communication as a social mechanism of neighborhood effects
The discussion above foreshadows the challenges addressed in the rest of Chapter
2. The first challenge is to re-introduce communication as a social mechanism through
which neighborhood effects operate. The early 20
th
century work in community ecology
and the foregoing brief review of communication research that touches on the dynamic
relationship of neighborhood context and communication suggests that such an
undertaking is warranted. To address the challenge successfully though requires (a)
defining communication in the setting of the residential community, (b) specifying how it
impacts the lives of neighborhood residents, and (c) determining how it relates to the
social mechanisms and structural neighborhood dimensions that are commonly found in
neighborhood effects models. The second, related challenge is to develop and outline a
theoretical model of neighborhood effects that incorporates communication as a social
process and explicates its relationship to other processes that have been examined to date.
In addition, if, as Park (1925/1967) suggested, communication is to be treated as a
process of effects that involves multiple community actors – i.e., individual and
institutional – it is necessary to elaborate the relationship between the micro level (i.e.,
residents) and the meso level (i.e., organizations/institutions) actors and the ways in
which their interaction may impact neighborhood life. The density of institutional
resources (or actors), for instance, has been found in several studies to be linked to
specific effects. Peterson, Krivo, and Harris (2000) found that residents in neighborhoods
with more recreation centers and fewer bars tend to experience a lower incidence of
violent crimes. However, as Sampson et al. (2002) note, rarely do these sorts of studies
account for the impact of the physical presence of institutional resources in a community

31

as well as their influence as actors or agents that partake in the everyday activities of the
residential community. I argue that a communication theory-based approach addresses
this gap in the existing literature. This represents the third challenge that is addressed in
this chapter.
To take up all three challenges though necessitates laying out first the key
concepts that are found in the burgeoning neighborhood effects literature and the
relationships between these core ideas, variables, and processes.
The general model of neighborhood effects
As Figure 2.1 illustrates, from the mid 1990s to the year 2000 the number of
neighborhood studies doubled, “to the level of about 100 papers per year” (Sampson et
al., 2002, p. 444). The development and application of new methods, such as GIS-
assisted socio-spatial mapping (see Goodchild, 1992; Mark et al., 1996; Williamson,
2004) played an important role. Socio-spatial mapping has allowed researchers to
examine, among other things, what difference it makes to operationalize ‘neighborhood’
in different ways (e.g., as a census tract or as an elementary school district – see Hipp,
2007). From a theoretical point of view, what led to the development of more studies
focused on the neighborhood were new ideas “for understanding what makes places more
or less healthy, particularly for young people” (Sampson et al., 2002, p. 444). Up until the
early to mid-1990s though, what was missing was a clear articulation of hypotheses for
and suitable data on the social processes and mechanisms, such as peer-group influence,
collective socialization, or institutional capacity that presumably could explain why it
matters, for example, if a child grows up in a poor neighborhood (Jencks & Mayer,
1990).

32

Structural neighborhood characteristics
(a) The effects of concentrated disadvantage
Occasionally one study or scholar can rejuvenate a field of research. Wilson’s
(1987) book, The Truly Disadvantaged, is one of those seminal works, which generated
new interest and new questions that propelled neighborhood effects research forward in
the 1990s. Wilson and others have argued that the structural transformation of the
American economy resulting from de-industrialization, globalization, and technological
change (Frisbie & Kasarda, 1988; Gephart, 1997), the migration of low-wage jobs from
urban centers to the suburbs (Freeman, 1991), and the flight of middle-class families
from the inner cities (see Gramlich, Laren, & Sealand, 1992) have resulted in severe
social dislocations in many urban neighborhoods. Moreover, Wilson suggested that,
people residing in neighborhoods of concentrated poverty had become isolated
from job networks, mainstream institutions, and role models, and that a variety of
social dislocations resulted from this isolation, including school dropout and the
proliferation of single-parent families (Gephart, 1997, p. 1).
Gephart added that the problems related to concentrated disadvantage are, since
the 1940s, compounded by increases in urban residential segregation attributed to racial
prejudice and racial discrimination in the housing markets, as well as a number of federal
and local government policies. This pattern “led to conditions of ‘hypersegregation’ for
blacks in twenty major metropolitan areas in the 1980s and 1990s (Massey & Denton,
1989)” (Gephart, 1997, p. 4).
While influential, Wilson’s work on the impact of concentrated poverty was only
one of the forces that led to the spike in the number of studies seeking to find and

33

understand effects of neighborhood structural characteristics on individuals. Renewed
interest in the study of immigration in sociology and other social sciences in the 1990s,
and especially since September 11, 2001 (Bean & Stevens, 2003), seems to have
contributed to the growth exhibited in the literature as well. Numerous studies have been
published in recent years on how the labor structure, wages, population stability, and
ethno-racial isolation are changing in cities around the United States (and elsewhere) as a
result of the influx of immigrants and changes in immigration policy (e.g., Gorman,
2002; Maggs & Baumann, 2001; Meissner, 2001; Waldinger, 1996; Bailey & Waldinger,
1991). At the same time, a number of projects have focused on the effects of the changing
environment on the opportunities and chances for upward social mobility of immigrant
populations and the native-born lower class (and underclass), consisting primarily of
African-Americans (see: Bean & Stevens, 2003; Clark, 2003; Waldinger, 2003).
The changes that Wilson describes and the changes in the flows, dynamics of and
policy on immigration are related. The majority of “the residents of neighborhoods in
which 40 percent or more of the residents are poor,” says Gephart (1997), “are members
of minority groups” (p. 3; also: Jargowksi & Bane, 1991; Massey & Eggers, 1990).
“Minorities,” Gephart concludes, “disproportionately experience the effects of
concentrated poverty and its clustering with other forms of disadvantage” (p. 3).
The research spawned by Wilson’s work on concentrated disadvantage has
focused on a variety of outcomes. Concentrated poverty has been shown to be related to a
number of child and adolescent outcomes, such as child abuse (Coulton, Korbin, Su, &
Chow, 1995), infant mortality and low birth-weight (Coulton & Pandey, 1992), dropping-
out of school (Brooks-Gunn et al., 1993; Clark, 1992), poor educational attainment

34

(Garner & Raudenbush, 1991), and frequency and seriousness of delinquent behavior
(Peeples & Loeber, 1994)
3
. According to Sampson (2001), there is evidence to suggest
that a number of health-related indicators, such as homicide, infant mortality, low birth-
weight, and suicide, cluster spatially, therefore lending support to the hypothesis that
there are “geographic ‘hot spots’ for crime and problem-related behaviors, and that such
hot spots are characterized by the concentration of multiple forms of disadvantage”
(Sampson et al., 2002, p. 446).
(b) Social-ecological aspects of neighborhood structure
Neighborhoods have a variety of structural characteristics, other than concentrated
disadvantage. Residential tenure and population stability, home ownership, population
density and ethnic diversity are a few of the most commonly explored. According to
Brooks-Gunn et al. (1997) residential instability (e.g., percentage of residents that moved
in the past 5 years, percent of households in their homes less than 10 years) and low rates
of home ownership are associated with a variety of problem behaviors, such as lifetime
alcohol use (Ennett, Flewelling, Lindrooth, & Norton, 1997), and juvenile delinquency
and crime (Sampson & Groves, 1989). The literature, overall, suggests that concentrated
poverty and other structural neighborhood characteristics will retain their centrality in
neighborhood effects research because: (a) they remain direct predictors of many
outcomes (Morenoff, Sampson, & Raudenbush, 2001), (b) they persist over time
(Sampson & Morenoff, 2004b), and (c) they tend to cluster together (Sampson et al.,
2002).

3
For an overview, refer to Brooks-Gunn, Duncan, & Aber, 1997.

35

Beyond structural characteristics: processes and mechanisms of effects
Despite the fact that many studies have produced evidence suggesting that
structural neighborhood characteristics matter, they do not explain how the neighborhood
affects the lives of its residents. That is to say, that to study structural characteristics of a
neighborhood helps understand how two or more residential communities differ, but it
remains unclear how these differences impact various aspects of residents’ lives (e.g.,
their health, their children’s school performance). A complementary line of research
suggests that there are four types of neighborhood mechanisms, which are related
(Sampson et al., 1999) “yet appear to have independent validity” in explaining how
effects are produced (Sampson et al., 2002, p. 457; Leventhal & Brooks-Gunn, 2003):
(i) Social ties and interaction: Much of the research on the effects of social ties
and interaction in a community has relied on the concept of social capital. Social capital
is usually viewed as a resource that is realized through social relationships (Coleman,
1988; Coleman, 1990; Portes, 1998; Putnam, 1995; Putnam, 2000; Leventhal & Brooks-
Gunn, 2000). In communities blessed with rich stores of social capital, says Putnam
(1993), it is easier for residents to get things done. Social capital is operationalized in
several ways, including the following: (a) level or density of social ties between
neighbors (see also Morenoff et al., 2001), (b) frequency of social interaction among
neighbors, and (c) patterns of neighboring (Warner & Rountree, 1997). With regard to
patterns of neighboring, in particular, Warner and Rountree (1997) studied Seattle
neighborhoods and sought to understand how racial composition mattered with respect to
crime incidence. They found that “social ties significantly and negatively affect assault
rates [i.e., more social ties are associated with a decline in assault rates] in predominantly

36

white neighborhoods, while they have no significant effect in predominantly minority or
racially mixed neighborhoods” (p. 520).
(ii) Norms and collective efficacy: Social ties are important, but the “willingness
of residents to intervene on behalf of children may depend,” say Sampson et al. (2002),
“on conditions of mutual trust and shared expectations among residents” (p. 457).
Sampson, Raudenbush, and Earls (1997) describe a neighborhood, in which mutual trust
and shared willingness to intervene is high, as a community with a high level of collective
efficacy. The key hypothesis for studies that apply this mechanism of neighborhood
effects is that,
neighborhood influences are accounted for by the extent of formal and informal
social institutions in the community and the degree to which they monitor and
control behavior in accordance with socially accepted practices and the goal of
maintaining public order (Leventhal & Brooks-Gunn, 2003b, p. 30).
4

(iii) Institutional resources: Theoretically speaking, models that focus on
institutional resources suggest that neighborhood influences operate by means of the
quality, quantity, and diversity of educational (e.g., schools, libraries), recreational (e.g.,
sports clubs), social, health (e.g., clinics and other medical facilities), family support
centers, child care centers, religious and employment-related institutions available to the
community (Leventhal & Brooks-Gunn, 2000; Leventhal & Brooks-Gunn, 2003b).
Several studies, for example, have suggested that the availability of high quality child
care in a neighborhood is tied to children’s cognitive and socio-emotional well-being

4
The norms and collective efficacy model of effects is rooted in a social control theoretical approach;
researchers applying such a perspective also consider peer group influences (see, for example, Veysey &
Messner, 1999) and the effects of physical threats in the neighborhood (e.g., violence; see Sampson et al.,
1997).

37

(Benasich, Brooks-Gunn, & Clewell, 1992; Lee, Brooks-Gunn, Schnur, & Liaw, 1990).
However, in most cases, empirical measures capture only the presence and quantity of
neighborhood institutional resources based on survey reports (Coulton et al., 1999; Elliott
et al., 1996) and archival records (Peterson et al., 2000).
(iv) Routine activities: Finally, the availability of public transportation, the
location of schools, the mix of land uses (e.g., residential and commercial), and the flow
of nighttime visitors (i.e., non-residents), are additional variables that are considered
relevant to organizing, for instance, “how and when children come into contact with their
peers, adults, and nonresident activity” (Sampson, 2002, p. 458). The nature and
distribution of daily routine activities in a neighborhood can conceivably have many
effects, although the literature has focused primarily on the impact they have on crime
incidence (e.g., Felson & Cohen, 1980; Felson, 1998). In one study, LaGrange (1999)
found that the occurrence of vandalism in Edmonton, Canada was more prevalent in
neighborhoods with high schools or malls, and high unemployment. High schools and
malls are thought to increase the flow of non-residents through a neighborhood,
especially of individuals that have been found to engage more often in vandalism: teens
(LaGrange, 1999). However, as the foregoing example illustrates, direct measures of how
routine activities, like the flow of non-residents are distributed throughout a community,
have not been developed. Moreover, studies that have applied a routine activities
perspective have not, to date, considered the role residents’ daily communication patterns
play in neighborhood life.
Figure 2.2 presents a general model of neighborhood effects, summarizing the
relationships among concepts found in the literature to date.

38

Figure 2.2. How the neighborhood impacts the lives of residents

Communication as an antecedent process of neighborhood effects
The mechanisms through which neighborhood effects are thought to manifest are
complex; they are not set in motion in a vacuum and they depend on more elementary
social processes. Knowing the socio-economic status, ethnic composition of a
community, or what the public transportation grid of a neighborhood looks like (i.e.,
structural characteristics) helps us begin to understand the social and physical context in
which neighborhood-specific patterns of social ties develop, for example. In addition,
structural elements of a community are related to the types of institutional resources that
play a critical role in the lives of neighborhood residents. The centrality of the church in
African American communities, for instance, has been well documented. Historically,
faith-based organizations have acted as mobilizing forces in African American
communities. But they also have been found to act as resources for job-seekers, as well as
informal credit bureaus (Putnam, 1993). Similarly, there is substantial research focusing

39

on the significance of hometown associations for neighborhoods with large Mexican-
origin immigrant populations. These organizations act as loci of social exchange, but also
as engines of development for the villages and towns immigrants come from (Orozco,
2003).
What is it though that makes churches and hometown associations mechanisms of
neighborhood change? Or, to ask a similar question, how is collective efficacy created
and why does it increase or decline? The neighborhood effects literature, to date, has not
answered these questions (Sampson et al., 2002). To answer them would require breaking
the four general social processes articulated earlier into more elementary mechanisms.
Echoing Park (1925/1967), I argue that communication is such an elementary
process through which life in the urban community is organized and transformed. And it
is a process that has been largely overlooked. Through communication we can explain (a)
how social ties are forged and sustained; (b) how public opinion is formed and a
consensus is reached about whether or not the community ‘can do’ something to better its
collective life (i.e., how collective efficacy is built); (c) how norms are constructed,
negotiated, and reconstituted; and (d) how institutional actors reach out to residents and
impact their lives (e.g., via mass media-based outreach efforts or grassroots
communication campaigns). In addition, communication can explain why certain patterns
of everyday neighborhood activities emerge. Communication between parents in a
neighborhood may alter, for instance, the time of day children meet their friends and play
in the local park.
To develop and to test this argument, however, necessitates deploying an
appropriate, communication-based theoretical framework. An adequate theory should be

40

able to account for: (1) the role of the social and (2) the physical environment on
communication behavior and patterns, and vice versa. It should also be able to (3) explain
the nature of the relationship between communication and the aforementioned four social
processes of neighborhood effects. In addition, it should allow for predictions and
account for specific effects we observe in people’s everyday lives. Finally, it should be
able to account for the relationship between micro, meso, and macro levels of analysis,
all of which are relevant in neighborhood effects research.
Communication infrastructure theory and neighborhood effects
Communication infrastructure theory (CIT; e.g., Ball-Rokeach et al., 2001; Kim
& Ball-Rokeach, 2006a) is deployed here to complement sociology-based theories and
concepts that have been used to explain neighborhood effects. CIT is an ecological
approach rooted in communication research. The communication infrastructure consists
of a neighborhood storytelling network (STN) situated in its residential communication
action context (CAC – See Figure 2.3 for a model of CIT).
(i) The storytelling network of neighborhood actors
The storytelling network is created through a storytelling process in which
residents, community organizations, and local/geo-ethnic media work with each other to
construct a vision and a reality of their neighborhoods as places where they belong and
engage shared concerns. They are the key agents of civic engagement. Neighborhood
storytelling is broadly defined as any kind of communicative action that addresses the
residents, their communities, and that relates to their lives in those communities (Ball-
Rokeach et al., 2001). Within the CIT framework, communities are seen as “being
composed of individuals, family units, and neighborhoods that cohere to varying degrees

41

around issues of place, ethnicity, interests, culture, and organization” (Matei et al., 2001).
This definition incorporates elements of a variety of other approaches (e.g., natural
communities, communities of limited liability, organizational dependent community; for
an extended discussion of these terms, see Jeffres, 2002).
(ii) The neighborhood environment as a communication action context
The vitality and durability of this network of storytellers can be strengthened or
weakened by the socio-cultural geography of the residential community (Ball-Rokeach,
2003). The boundaries of the communication action context are the boundaries of a
residential area as defined by shared conventions (e.g., major cross streets, incorporated
area, real estate sections, geographic labels). It is in the CAC that we find the discourse
preconditions for storytelling in a specific neighborhood. They include the elements that
constrain or encourage communication (Ball-Rokeach et al., 2001; Matei & Ball-
Rokeach, 2005). In the parlance of the neighborhood effects literature, these contextual
factors may include neighborhood socio-economic status (e.g., Sampson, Morenoff, &
Earls, 1999), ethnic heterogeneity (e.g., Iyer, Kitson, & Toh, 2005; Kim & Ball-Rokeach,
2006b), population or residential (in)stability (Shah, McLeod, & Yoon, 2001), density of
community organizations (e.g., Sampson et al., 2005), land uses (e.g., Kurtz, Koons, &
Taylor, 1998), cultural and language barriers (e.g., Orellana, Dorner, & Pulido, 2003),
collective memories (Small, 2003).

42

Figure 2.3. Components of the neighborhood communication infrastructure

Figure 2.3. The relationship of the storytelling network and the communication action context is
dynamic. The elements of the communication action context can enable or constrain the
storytelling network, which is composed of neighborhood residents in their familial or other
interpersonal networks, as well as meso-level neighborhood actors, such as local and ethnic
media and community-based organizations. Neighborhood storytellers have the capacity to
transform their environment and the conditions of their everyday life.

CIT and neighborhood effects mechanisms
The overview of CIT suggests that two of the four hypothesized mechanisms of
neighborhood effects – social ties/interaction and norms/collective efficacy – are more
closely linked to the neighborhood storytelling network, while the other two, namely
institutional resources and routine activities, can be most readily considered as elements
of the communication action context. I examine how CIT relates communication to each
of the four neighborhood effects mechanisms, in turn.

43

(a) Social ties and interaction among family members, friends and peers, are
considered to be one way though which neighborhood level variables influence residents.
Some neighborhood effects studies consider the density of social ties between neighbors
(Rountree & Warner, 1999
5
; Veysey & Messner, 1999), others consider the frequency of
social interaction among neighbors (Bellair, cited in Sampson et al., 2002), and still
others focus on the patterns of neighboring (Warner & Rountree, 1997). However, what
is missing is an explanation of why and how these ties are forged, as well as how they are
maintained. Were ties created as a result of a spike, for instance, in the number of violent
crimes in the neighborhood? A violent crime could act as a turning event (Ball-Rokeach
et al, 2001); the exigency that makes residents, local/geo-ethnic media, and community
organizations talk more to each other about what is happening in the community. Ball-
Rokeach et al. (2001) contend that “communication infrastructures are usually invisible
until something happens to impair their functioning” (p.394). In such a case, viewing the
neighborhood through CIT’s lenses, a more or less integrated storytelling network may
emerge (Ball-Rokeach et al., 2001; Kim & Ball-Rokeach, 2006a).
Focusing on the storytelling network would help answer the question of how the
ties are forged and how they are maintained; but only in part. It is in the communication
action context that one will find what is enabling or constraining the neighbors to
interact/communicate with one another. For example, if the construction of a new
highway cuts through a neighborhood, it may become impossible for residents living on
opposite sides of the new road to stay in touch. Moreover, the communication action

5
In this study, based on data from the Seattle Victim Survey, Rountree & Warner (1999) found that the
density of ties between women were negatively related to violence (in census tracts with low levels of
female households), while density of ties between men had no significant relationship to violent crime in
the neighborhood.

44

context may be responsible for the level of the storytelling network’s integration. Lack of
resources in the area may prevent the founding of geo-ethnic media and community
organizations. Through a communication infrastructure approach, it becomes clearer how
social ties between residents, but also between residents and neighborhood meso level
actors are created, maintained, constrained, and dissolved. In a fashion similar to how
social capital is understood as a community resource that is realized through social ties
(e.g., Bourdieu, 1986; Coleman, 1988; Putnam, 2000), when residents, geo-ethnic media,
and community organizations are connected to one another and work together to better
the life of the entire community they build communication capital (Ball-Rokeach, 2003).
The idea and significance of a community’s communication capital for its residents will
discussed further in Chapters 3 and 4.
(b) Collective efficacy according to Sampson et al. (1997) is conceptualized as the
linkage between mutual trust and the shared willingness to intervene for the public good.
Once again, however, neighborhood effects studies do not tell us how collective efficacy
is constructed. In the CIT framework, collective efficacy, just as self-efficacy, “is
constructed in symbolic environments through communication (Bandura, 1977)” (Kim &
Ball-Rokeach, 2006a). In research carried out as part of the multiyear Metamorphosis
Project housed at the Annenberg School for Communication at the University of
Southern California, residential communities with a more integrated storytelling network
have also exhibited higher levels of neighborhood belonging, conceived as subjective
attachment toward neighborhood and objective exchange with neighbors (feelings and
behaviors; see Kim & Ball-Rokeach, 2006a), higher levels of collective efficacy (the
sense that neighbors can come together to solve problems; see Kim, 2003; Kim & Ball-

45

Rokeach, 2006b), and higher levels of civic participation (residents involved in local
policy and opinion making processes) (Ball-Rokeach et al., 2001; Matei, Ball-Rokeach,
& Qiu, 2001; Wilson, 2001). That is to say that when residents, geo-ethnic media, and
community organizations are connected and encourage one another to talk about
neighborhood problems, residents are more likely to develop a sense of belonging and a
shared expectation that they “can do something” to solve those problems. In turn, this
leads them to work with neighbors to effect change (Ball-Rokeach et al., 2001; Kim,
2003; Cheong, Wilkin, & Ball-Rokeach, 2004).
Still, context makes a difference. For instance, Kim (2003) found that residents
living in old immigrant neighborhoods as opposed to areas with high residential turnover
have higher degrees of collective efficacy. A communication perspective, therefore, helps
answer the question of how collective efficacy is formed and how it may be maintained,
while taking into account the constraints of a particular community context.
(c) Institutional resources: Although institutional resources are theoretically
deemed important as mechanisms of neighborhood effects, most studies are limited in
that they only account for the density of resources, such as libraries, medical facilities,
recreation centers, and child care facilities. As Sampson et al. (2002) note, most studies,
“do not distinguish well between structural dimensions of institutions (e.g., density) and
mediating institutional processes” (p. 458, footnote 8). A few studies in the neighborhood
effects literature have attempted to understand the role of participation in neighborhood
organizations (e.g., Veysey & Messner, 1999; Morenoff et al., 2001). However,
residents’ participation in neighborhood-based organizations – indicated, for instance, by
the number of volunteers working on one or more projects – does not capture directly the

46

extent to which an organization’s mission is to improve the lives of a community’s
residents, nor does it provide much insight with respect to the ways through which a
particular organization is actively pursuing an agenda that reflects such a mission.
In communication infrastructure theory, institutions can (a) be part of the
storytelling network and also (b) be implicated in the communication action context as
resources. This, of course, causes analytical problems, but this is a reflection of the partial
disconnect between analytical models and lived experience.
Institutions, like community organizations, non-profits, or health care providers,
are founded by individuals who either are already part of a residential community or who
invest time, money, and effort to build relationships with members of the community. In
the course of everyday life, institutional actors become connected to residents and other
organizations in the community. Health and social service professionals talk to families
about problems they are dealing with, the local Women, Infants, and Children (WIC)
office offers counseling to young mothers, journalists for local media go out into the
community to talk to residents and report on the stories affecting their lives. To the extent
that institutional actors focus their energy on serving the residential community, they
contribute to the stores of social and communication capital available to community
members. They become resources residents depend on to accomplish everyday goals and
the presence of the institution is felt far beyond the walls of its physical location. The
effects of an organization’s actions ripple throughout the community, traveling across the
communicative ties that bind neighborhood actors.
This would suggest that institutions have to be considered as part of the
storytelling network. Yet, it is also true that clinics, child care centers, schools, and

47

recreation facilities may only have the capacity, the human and material resources, to
serve residents that live in areas smaller than any given residential community. Hence,
variables like the density of institutional resources in the neighborhood environment (i.e.,
the communication action context) are also very important.
Communication research, in this case, elucidates the dual role of institutions in
neighborhoods. Insofar as institutions are part of the storytelling network, they do not act
alone and their influence is conditioned by their relationship to residents and each other.
As part of the communication action context, they become problem-solving or problem-
causing hubs. Peterson et al. (2000), for instance, found that the density of recreation
centers in a community was associated with lower crime incidence.
(d) Routine activities: Neighborhood effects studies suggest that the location of
schools, the mix of land uses, the public transportation grid, the inflow and outflow of
vehicular traffic and neighborhood outsiders/visitors (e.g., LaGrange, 1999; Sampson &
Raudenbush, 1999; Peterson, 2000), are likely to influence residents daily activities (e.g.,
when and where children go out to play with their friends). These elements are included
in the formulation of the communication action context.
Limitations of communication infrastructure theory
While amenable to the study of neighborhood effects, communication
infrastructure theory and studies that have applied it to date have not addressed the felt
analytical difficulty, which is related to the dual role of neighborhood institutional
resources. Two of the questions that this reality raises, and which this study sets out to
address, are: (i) how can institutional resources be part of both the storytelling network
and the communication action context? And (ii) how and under what conditions do the

48

two roles that organizations play in the residential community, as actors and as elements
of neighborhood structure, complement each other?
In addition, collective efficacy and civic participation have been cited as social
mechanisms of neighborhood effects. Several studies have also found that the
communication infrastructure – i.e., storytelling network integration and communication
action context conditions – impacts collective efficacy and civic participation (Kim &
Ball-Rokeach, 2006a, Kim & Ball-Rokeach, 2006b; Kim & Ball-Rokeach, 2006c).
However, if communication is indeed a more elementary process of neighborhood
effects, then the communication infrastructure should also be related, directly and
indirectly through collective efficacy and civic participation, to other outcomes. These
theoretical propositions will also be addressed in this study.
The research site
The study unfolds in the area of Greater Crenshaw, a tremendously interesting
area of Los Angeles, both for its history and increasing population diversity. It has
historically been a majority African American community. However, increased middle-
class African American migration to newer neighborhoods, such as the Antelope Valley
and Moreno Valley, and with the increase in Latino immigration, the population
composition of the area has changed significantly in recent years.
Greater Crenshaw is one of the 11 residential communities investigated by the
USC Metamorphosis Project, of which I have been a member since 2003. The study area,
located in South Los Angeles roughly between the Civic Center and University Park,
covers 27.4 square miles, nine zip codes (i.e., 90007, 90008, 90016. 90018, 90019,
90037, 90043, 90056, and 90062), and 87 Census tracts. This rather large geographic area

49

includes the residential areas of Arlington, the Crenshaw District, Jefferson Park, Ladera
Heights, Leimert Park, Windsor Hills, and View Park. It is home to over 376,400 people.
Forty-four percent of the population identifies as Hispanic, while just below 44% is
African American. Of the rest of the population, just over 5% identifies as White and 5%
is Asian.
Table 2.1. Population profile of Greater Crenshaw and Los Angeles County
Population Profile
Greater Crenshaw L.A. County
Population size 376,436 9,519,338
Ethnic background
a. African American (%) 43.7 9.6
b. Asian (%) 5.0 11.9
c. Hispanic/Latino (%) 43.9 44.6
d. White (Non‐Hispanic) (%) 5.4 30.9
Foreign‐born (%) 35.6 36.2
Speak language other than English at home (%) 43.0 54.0
Median age 32 32
Under 17 years of age (%) 29.2 28.0
Education: High School and above (25+ y/o) (%) 58.9 67.8
Median income 31,770 42,189
Households under 200% poverty level (%) 38.2 23.0
Residents that own their home (%) 33.8 47.9
Data source: U.S. Census Bureau (2000)

According to the 2000 U.S. Census, about a third of the population in the area has
been born outside the United States (36%) and 43% of the residents over five years of age
speak a language other than English at home. The median age in Greater Crenshaw is 32,
while 29% of the residents are under 17. Fifty-nine percent of the population over the age
of 25 has a high school diploma or a more advanced academic degree, and over 34% of
the residents own their homes. The median income in the area is near $32,000, while, in
2000, 38% of the households in the area reported incomes that were under the 200%

50

federal poverty level (i.e., under $20,000). Table 2.1 presents how Greater Crenshaw
compares to the larger Los Angeles County area with respect to the aforementioned
statistics.
In 2006, the Metamorphosis Project conducted a random digit dial telephone
survey with African American (N = 304) and Hispanic (N = 303) residents in the Greater
Crenshaw area. Among other questions, residents in the area were asked what they
thought were “the most pressing issues facing the area” they lived in. Table 2.2 presents
the ten issues that residents mentioned most frequently.
Table 2.2. Top-10 most pressing issues in Greater Crenshaw, 2006 (N=607)
Most Pressing Issues (Top‐10) % of respondents
Crime and gangs 55.70%
Drugs 16.10%
Quality of Public Schools 10.00%
Neighborhood Appearance/Development 9.70%
Poor Street Condition and Maintenance 8.20%
Traffic Congestions/Parking Problems 6.60%
Housing Prices 5.90%
Juvenile Delinquency 5.30%
Work Availability/Work Problems 4.80%
Homelessness 4.80%
Data source: Metamorphosis Project (2006)

Greater Crenshaw residents also face a variety of health related problems. Most of
the neighborhoods of Greater Crenshaw are part of Service Planning Area 6 (SPA 6) of
Los Angeles County. SPA 6 is one of eight SPAs in Los Angeles County. They are
administrative units designed for health service planning purposes. According to 2007
County Department of Public Health data, SPA 6 fared worse (i.e., last or second to last)
than the average for all SPAs with respect to several indicators of (a) residents’ general

51

health status, (b) residents’ access to health care, and (c) prominence of unhealthy
behaviors. Table 2.3 presents the relevant data for SPA 6 juxtaposed to L.A. County-
wide and national data.
Table 2.3. Health status, access to health care, and unhealthy behaviors in SPA 6 (2007)
SPA 6
L.A.
County U.S.
Health Status
Percent of children 0‐17 years old who are perceived by their
parents to be in fair or poor health
17.6 12.7 N/A
Percent of adults reporting their health to be fair or poor
33.4 20.6 12.3
Average days in past month adults reported regular activities
were limited due to poor physical/mental health
3.3 2.4 N/A
Average number of unhealth days (due to poor mental or
physical health) in past month reported by adults
7.9 6.4 N/A
Health Care Access

Regular source of care

Percent of adults with no regular source of health care
26.9 19.8 N/A
Percent of children 0‐17 with no regular source of health care
12.0 8.2 5.3
Access to care

Percent of adults who reported difficulty accessing medical
care 43.9 30.1 N/A
Percent of children 0‐17 years old who have difficulty
accessing medical care
20.8 14.5 N/A
Access to dental care

Percent of adults who did not obtain dental care (including
check‐ups) in the past year because they couldn't afford it
35.1 25.6 N/A
Health Insurance

Percent of adults ages 18‐64 who are uninsured
31.7 21.8 18.9
Percent of children ages 0‐17 who are uninsured
11.3 8.3 8.9
Health Behaviors

Percent of adults who consume five or more servings of fruit
and vegetables a day
10.7 14.6 23.2
Percent of adults who obtain recommended amount of
exercise per week (>= 20 minutes of vigorous activity, >=3 days/week; >=
30 minutes of moderate activity, >=5 days/week)
45.6 51.8 49.1
Percent of adults who are minimally active or inactive
44.5 37.5 N/A
Percent of adults who smoke cigarettes
17.3 14.6 20.6
Data source: Los Angeles County Department of Public Health (2007)

52

SPA 6 residents also fared worse than the rest of Los Angeles County with regard
to several health outcomes. Table 2.4 presents the relevant data for SPA 6 and compares
them to L.A. County-wide and national data.
Table 2.4. Health outcomes in SPA 6 (2007)
SPA 6
L.A.
County U.S.
Prevention and Health Outcomes
Overweight and Obesity

Percent of adults who are obese (BMI >= 30.0)
30.0 20.9 24.9
Diabetes

Percent of adults diagnosed with diabetes
9.7 8.1 7.3
Diabetes death rate (age‐adjusted per 100,000 of population)
39.2 20.9 24.9
Cardiovascular Disease

Percent of adults diagnosed with hypertension
29.0 23.4 22.4
Coronary heart disease death rate (age‐adjusted per 100,000)
229.7 176.1 172.3
Stroke death rate (age‐adjusted per 100,000)
64.8 47.6 53.5
Injury

Death attributed to motor vehicle crashes
10.7 9.6 15.3
Cancer

Lung cancer death rate (age‐adjuster per 100,000)
46.0 35.3 54.1
Breast cancer death rate among females (age‐adjusted/100,000)
27.8 23.1 25.2
Colorectal cancer death rate (age‐adjusted per 100,000)
23.2 16.2 19.1
Respiratory Disease

Percent of adults 50+ y/o vaccinated for influenza in the past year
32.5 40.7 N/A
Influenza/Pneumonia mortality rate (age‐adjusted per 100,000)
29.6 26.4 22.0
Data source: Los Angeles County Department of Public Health (2007)

Because of the bleak picture of SPA 6 and Greater Crenshaw these data paint, but
also due to the fact that the public health literature has become increasingly interested in
understanding how a neighborhood’s social, built, and natural environment impacts
health (see, e.g., Oakes, 2004), the decision was made to focus on investigating the role
that the communication infrastructure is playing in the neighborhoods of Greater

53

Crenshaw with regard to health. The key question that drives this research is: how can
knowledge about the communication infrastructure of the Greater Crenshaw communities
be leveraged to improve residents’ health? To do so though requires that the
communication infrastructure theory-based model of neighborhood effects – outlined
briefly in this chapter – is articulated to address health concerns. I do so in Chapter 3.
54

CHAPTER 3
THEORETICAL FRAMEWORK II & HYPOTHESES: THE NEIGHBORHOOD
COMMUNICATION INFRASTRUCTURE AND HEALTH
As part of the Healthy People 2010 initiative, the U.S. Department of Health and
Human Services has set an array of benchmarks against which the health of local
communities across the country is evaluated. Federal, state, and local government
agencies, as well as community based-organizations and health service providers use
these benchmarks to measure the impact of their efforts towards minimizing health
disparities among the U.S. population. Health disparities occur when certain segments of
the population (e.g., people of different socio-economic status, gender, ethnic
background) are disproportionately affected by particular health problems.
As indicated in Tables 2.3 and 2.4 (see Chapter 2), residents in the Greater
Crenshaw communities that are part of Service Planning Area 6 (SPA 6) fare worse than
Los Angeles County residents do on average with respect to a variety of health-related
outcomes. Adults and children in SPA 6 neighborhoods are more likely, for instance, to
lack a regular source of medical and dental care. They also tend to exercise less and
consume fewer fresh fruits and vegetables than L.A. County residents do on average.
Both physical activity and consumption of fruits and vegetables are considered important
in preventing or managing diseases like diabetes and hypertension (Gordon-Larsen,
Nelson, Page, & Popken, 2006; Haapanen et al., 1997; Horowitz, 2004). Not surprisingly,
a larger percentage of residents in SPA 6 neighborhoods have been diagnosed with
diabetes compared to L.A. County as a whole, and the diabetes death rate in SPA 6 is
higher than it is in L.A. County overall. In addition, the percentage of adults in SPA 6
55

diagnosed with hypertension is 29%, whereas the L.A. County average is 23% (the
national average is 22%), and the coronary disease and stroke death rates for SPA 6 are
also higher than they are for both L.A. County and the U.S. as a whole (L.A. County
Department of Public Health, 2007). Similar disparities between SPA 6 and L.A. County
are also evident with regard to lung, breast, and colorectal cancer, as well as respiratory
diseases (see Table 2.4 in Chapter 2).
These health disparities have generated considerable interest among public health
professionals. Many studies have emerged in recent years that seek to explain what leads
to these disparities. Socio-economic status and ethno-racial background are two of the
most commonly explored factors. Not without reason. Low income individuals and
families, as well as certain ethnic groups, especially African Americans and Latinos, are
consistently found to fare worse than other segments of the U.S. population with respect
to a wide variety of health outcomes (Institute of Medicine of the National Academies,
2003). McMahon et al. (1999), for instance, found significant differences between whites
and African Americans in terms of receiving proper diagnostic tests for cancer.
1
Similar
studies detected racial disparities with respect to the quality of health care provided for
diabetes (Chin, Zhang, & Merrell, 1998), HIV (Moore et al., 1994), and cardiovascular
diseases (Schneider et al., 2001), among others.

1
McMahon (1999) found that African Americans received 28% fewer sigmoidoscopic examinations, even
though among African Americans the incidence of colon cancer is 20% higher. Sigmoidoscopic
examinations are generally considered to be “more technically advanced diagnostic procedures than barium
enema” (Institute of Medicine of the National Academies, 2003). More recent research, however, suggests
that a colonoscopy is even more effective as a preventative measure against colorectal cancer (e.g., Brenner
et al., 2007).
56

Disparities in access to health care and health literacy
Disparities in access to health care
Improved access to health care is thought to be linked to better health outcomes.
That is why a number of studies have focused on understanding why health care access
disparities exist. Albeit establishing causal links between access and health outcomes is
challenging (Andrulis, 2000), research has documented, for instance, positive effects of
better access to prenatal services on birth outcomes (Baldwin et al., 1998).
Health care access is frequently associated with having health insurance. A
number of studies suggest that people with coverage tend to seek medical help more
frequently and feel that it is easier for them to get the services they need (e.g., Braveman
et al., 1989; Lillie-Blanton, 1999; Sayer & Peterfreund, 1993). Having both health
insurance and access to medical care has been associated with better health outcomes.
Improved access to innovative treatments, in addition to enrollment in Medicaid
2
, for
example, has been linked to a significant decrease in opportunistic HIV infections
(Andrulis, 1998).
Health care access disparities, however, are not only the result of lacking health
insurance. A person’s level of education, proximity to health care providers, availability
of public transportation, language and culture, and health-related beliefs (e.g., regarding
the effectiveness of a remedy, cause of a disease), are but some of the factors that have
been examined as determinants of health care access disparities (Cunningham et al.,
1998; Flores, Abreu, Olivar, & Kastner, 1998; Haas et al., 2004; Katz, 2007; Rask,

2
Medicaid is a state and federal government partnership that provides health insurance for specific
categories of people with low incomes.
57

Williams, Parker, & McNagny, 1994; Schultz, Williams, Israel, & Lempert, 2002; Song,
2007).
Because the level of health care access in the Greater Crenshaw (SPA 6)
neighborhoods is lower than it is in other parts of Los Angeles, and because the residents
of Greater Crenshaw area are predominantly African American and Latino, I have chosen
health access as one of the outcomes I will focus on for the purposes of this study.
Access to health care and health literacy disparities
Research suggests that health care access also depends on health literacy
(Andrulis, 2000; Morley, 1997). Health literacy is a composite term that describes the
ability to obtain, process, and comprehend basic health information, as well as the
capacity to function appropriately in the health care environment (Council on Scientific
Affairs for the American Medical Association, 1999; Ratzan & Parker, 1999; U.S.
Department of Health and Human Services, 2000). Apart from access to care, health
literacy has been found to be a good predictor of health behaviors and outcomes.
Diabetes patients, for example, with higher levels of health literacy know more about
their disease. And, among Type 2 diabetes patients, lower levels of health literacy are
linked to worse glycemic control and higher levels of retinopathy (Schillinger et al.,
2002; Williams, Baker, Parker, & Nurss, 1998).
Significant disparities exist among U.S. populations with regard to health literacy.
Those that are poor and have less than a high school education, those who belong to
ethnic and cultural minorities, people living in the southern and western regions of the
U.S., and seniors over the age of 65, tend to be at a health literacy disadvantage when
58

they need to deal with personal or family health issues (Arnold et al., 2001; Beers et al.,
2003; Benson & Forman, 2002; Institute of Medicine, 2004; Weiss et al., 1994).
Health literacy is the second, health-related outcome that I will focus on in this
study. That is because of its relationship to health care access and the increased attention
it has received in the public health literature in recent years (see Ratzan, 2001). In
addition, the definition of health literacy implies that it depends heavily on the sources of
health information that individuals have in their environment and that they access. Prior
research has shown, for instance, that Latinos in Los Angeles may have multiple media in
their environment, but they rely most heavily on Spanish-language media (i.e., television,
radio, and print media) for health information (Wilkin, 2005). That is, at least in part,
because many Latinos in Los Angeles are not fluent in English. Therefore health literacy
among Latino communities of Los Angeles is likely to depend greatly on how many
health stories Spanish-language media broadcast and publish, as well as the depth,
breadth, and accuracy of those stories. Wilkin and Gonzalez (2006) evaluated Los
Angeles Spanish-language television news and talk shows for health content. They
reported that Spanish-language networks do not connect their viewers to other available
health information resources and service providers, or otherwise disseminate information
that could build health literacy among their viewers and also assist them in overcoming
health care access barriers.
Arguably, trusted media that are aware of residents’ health concerns and connect
or direct them to health resources, also help individuals and families become more
integrated into the neighborhood storytelling network. In this case, as the focus of the
storytelling is on health, residents are integrated into a neighborhood health-oriented
59

storytelling network. This network is comprised of residents, media, and other
organizational actors in the community, including but not limited to those with an
explicitly health-oriented mission (e.g., health service providers). In the process of
reporting on health-related issues salient to the neighborhood, media themselves connect
to community organizations of various types. Organizations that prove to be good sources
are featured in articles published or stories broadcasted. In addition, often media stories
highlight links between various community-based organizations, as is the case when two
organizations co-sponsor an event (e.g., a health fair), for instance, collaborate on a
particular health study, or launch some other kind of health-related initiative together.
These are all ways through which media, as one type of meso-level storytelling network
actors, can contribute to the integration of the neighborhood storytelling network.
In this study, I investigate if and how a neighborhood’s storytelling network can
influence residents’ health literacy. In addition, I question under what conditions such
effects may manifest. As noted in Chapter 2, the impact of the storytelling network
depends on the configuration of the communication action context. It is there where we
find the factors that enable and constrain the network.
Neighborhood determinants of health disparities
Many studies on health disparities are a-theoretical. Moreover, much health
disparities research focuses on individual risk factors and does not account for social
processes and contextual factors that may impact individuals’ health. This has been the
trend for much of the past century (Pickett & Pearl, 2001; Susser & Susser, 1996).
However, recent advances on the methodological front have encouraged investigators to
pursue research intent on understanding the influence of the social and physical
60

environment on health. Neighborhood variations with respect to health, for example,
could not be studied adequately until more sophisticated statistical techniques were
developed, due to what Pickett and Pearl (2001) call the “intractability of the ecological
fallacy” (p. 111). As they explain, that is “when group level data are used to infer
individual disease risk” (p. 111).
Multilevel analytical techniques, such as hierarchical linear modeling (HLM)
developed by Raudenbush and Bryk (2002), enable researchers to effectively separate out
effects attributed to neighborhood or group level variables and individual level
characteristics. Research on the social determinants of health has also been propelled
forward thanks to the extensive work done in sociology and related fields of inquiry,
since the mid-1990s, towards understanding: (a) the social mechanisms of neighborhood
effects, as well as (b) the interaction of these social processes with the physical and built
environment of residential communities (see Chapter 2; also: Diez-Roux, 2001). “The
notion that ‘place’ maybe be important to health,” says Diez-Roux (2001), emerged in the
literature in the 1980s and 1990s, and has “increased sharply in recent years” (p. 1783).
Pathways of neighborhood influences on health
I. Socioeconomic status
Socioeconomic context is the most frequently studied variable in neighborhood
effects studies focused on health. And despite some inconsistent results, Ellen,
Mijanovich, and Dillman (2001) conclude that the evidence supports the notion that “the
socioeconomic status of a community has an independent effect on [health] behaviors”
(p. 396). There are numerous studies that explore the links between socio-economic
status (SES) and health. In one project aiming to uncover predictors of coronary heart
61

disease factors Hart, Ecob, and Smith (1997) found that area-level socioeconomic status
did not predict smoking patterns in communities of Scotland, but it did account for
significant variation across neighborhoods in terms of alcohol consumption and
cholesterol levels; the variation remained clear, even after controlling for individuals’
education level and their occupation. In another project done in the U.S., Robert (1999a)
found that SES differences between neighborhoods accounted for the level of physical
activity residents engaged in and their body-mass index (BMI).
Additionally, several other studies report that residents of poor neighborhoods
tend to suffer from higher rates of heart disease, respiratory problems, and cancer (Adler,
Boyce, Chesney, Folkman, & Syme, 1993; Crombie, Kenicer, Smith, & Tunstall-Pedoe,
1989; Devesa & Diamond, 1983; Jenkins, 1983). Overall, the general mortality rates (i.e.,
all-cause mortality) in disadvantaged residential communities are higher than they are in
more affluent neighborhoods (Ellen et al., 2001). Moreover, other studies indicate that in
poorer neighborhoods there are more deliveries of babies with suboptimal birth weight,
infants are less likely to survive their first year of life, and children are more likely to be
hospitalized (Coulton & Pandey, 1992; Guest, Almgen, & Hussey, 1998).
In a review of the literature, Robert (1999b) describes eloquently the two general
pathways through which a neighborhood’s socioeconomic status affects health. First, she
says, socioeconomic context shapes the socioeconomic position of individuals. This
suggests that “the opportunities and constraints present in communities with different
socioeconomic contexts can shape the education attainment, job prospects, and income
level of individuals (Foster & McLanahan, 1996; Garner & Raudenbush, 1991; Jencks &
Mayer, 1990; Wilson, 1987)” (p. 493). Secondly, Robert adds, the socioeconomic context
62

directly affects “the social, service, and physical environments of communities shared by
residents, which then affect the individual characteristics, conditions, and experiences of
individuals that more directly impact health” (Robert, 1999b, p. 493). Robert and others
have begun to explore how SES impacts the social, service, and physical environment of
a community, which in turn influence individuals’ life conditions and health. But the
research is still limited. The existing literature has proposed mechanisms of influence
similar to those neighborhood effects studies that are not focused on health, including: (a)
social capital and neighborhood-based networks, (b) neighborhood institutions and
resources, and (c) norms and collective efficacy (Diez-Roux, 2001; Ellen, Mijanovich, &
Dillman, 2001; Wanderman & Nation, 1998).
3

II. Mechanisms of neighborhood influence on health
(a) Social capital and neighborhood networks: Research linking social capital and
health is growing quickly, but is still sparse. This is because social capital was introduced
into health research recently (Kawachi, 1999), and because social capital is an elusive
term (Altschuler et al., 2004; Carpiano, 2006; Portes, 1997). Many different definitions of
social capital have been proposed. As a concept social capital can encompass one or more
of the following dimensions: (a) density of social ties, (b) degree of individual
participation in civic organizations, (c) level of interpersonal trust (e.g., Coleman, 1988;
Putnam, 1995; Putnam 2000). In a neighborhood rich in social capital, neighbors are
more likely to be able to rely on others for support. They may be able to borrow money,
for example, if they need help paying for a prescription or ask a neighbor for a ride to the

3
Non-health-oriented neighborhood effects studies reviewed in Chapter 2 focus on the following four
social processes: (a) social ties and interaction, (b) collective efficacy and norms, (c) institutional resources,
and (d) routine activities (Sampson et al., 2002). Routine activities do not figure into health-oriented
neighborhood effects models.
63

doctor’s office. In a study focused on how socioeconomic inequality and social capital
together affect health, Kawachi, Kennedy, Lochner, and Prothrow-Stith (1997) found that
“income inequality leads to increased mortality via disinvestment in social capital” (p.
1491). That means that, in impoverished communities, a person may have a harder time
trusting a neighbor to repay a loan, especially when lending even a small amount can
have a severe, negative effect on the lender’s own budget and life conditions.
(b) Norms & collective efficacy: The research linking collective efficacy and
health is also limited, although several investigators have found that increased collective
efficacy is associated with lower crime rates in a community (e.g., Klein & Maxson,
2006; Sampson, Raudenbush, & Earls, 1997), and less crime means that fewer people
suffer bodily harm. Moreover, collective efficacy can indirectly influence one’s mental
health, as crime can induce fear, cause increased levels of stress, curb residents’ physical
activity, and prevent them from utilizing health services in their area (Minkler, 1992;
Robert, 1999b)
(c) Neighborhood institutions and resources: Poorer neighborhoods tend to be at
a disadvantage with regard to the number and quality of medical professionals available
to their residents (e.g., McKnight, 1995). Therefore residents in poorer communities often
have to travel longer distances to find the medical help they need. In addition, the same
residents are more likely to have to go outside their neighborhood to find better quality
food (Horowitz, 2004; Lewis et al. 2005; Troutt, 1993).
III. Socioeconomic status and other neighborhood structural characteristics
While socioeconomic status is the most frequently investigated structural
neighborhood characteristic in the literature on neighborhood health effects, it is not the
64

only one. In one project, for instance, Jackson, Anderson, Johnson, & Sorlie (2000) found
that residential segregation was associated with higher rates of mortality from all causes.
Moreover, several studies have found that a neighborhood’s ethno-racial
composition impacts individuals’ access to health care services. In a study that included
all counties across the U.S., Haas et al. (2004) found that African Americans and Latinos
may perceive fewer barriers to care when they live in a county with a high percentage of
people of the same ethno-racial background. In addition, Haas and her colleagues
reported that whites living in an area with a large percentage of Latinos, feel that it is
more difficult for them to get health care.
Similarly, the ethnic make-up of a community has been linked to access to “social
and material resources that promote health and avoid disease” (Schultz et al., 2002, p.
677), including grocery stores and pharmacies. In Detroit, during the 1960s, ethnic
tensions led to the exodus of major grocery stores from poor, predominantly African
American neighborhoods. They have not returned since then, leaving the residents of
these communities with only convenience and liquor stores (Schultz, Parker, Israel, &
Fisher, 2001). Other studies have demonstrated that in neighborhoods with a prevalence
of African Americans and Latinos, residents find it hard to get all the medications they
need (Morrison, 2000) and pay more to fill their prescriptions than residents of mostly
white neighborhoods (Schultz & Lempert, 2000).
IV. Neighborhood social environment stressors
Much of the neighborhood effects work on social disorganization has focused on
crime and violence as outcomes (e.g., Sampson & Groves, 1989; Veysey & Messner,
1999). In the health-oriented literature, crime, violence, and social disorganization, more
65

generally speaking, are treated as social environment variables that negatively impact
residents’ health. Crime and violence can have a direct, short-term impact on residents’
physical and mental health, and long-term effects, through a process Geronimus (1992)
calls “weathering.” She suggests that accumulated stress caused by violence and crime
may, over time, weaken individuals’ resistance to diseases. Moreover, Ganz (2000)
suggests that as crime and violence reduce the expected life span of residents in particular
neighborhoods, they also alter individuals’ perceptions about how detrimental to their
health certain behaviors known to have long-term effects, like smoking, can be.
V. Neighborhood physical environment stressors
Finally, there is a voluminous literature that addresses the impact of the physical
environment on residents’ health. In fact and perhaps unsurprisingly, physical
environment stressors are the most commonly studied pathway of neighborhood effects
in the public health literature (Ellen et al., 2001). Proximity to toxic waste sites and
polluting factories, as well as the quality of the air and water in a residential area, are
often considered factors that impact individuals’ health. These environmental stressors
are usually found in poorer neighborhoods (Anderton et al., 1993; Bullard, 1990; Schultz
et al., 2002).
Neighborhood effects and health: limitations of the existing research
Population health researchers have only recently begun to investigate the ways in
which where a person lives matters for their health. In part due to the history of
epidemiological research, which has emphasized individual level variables, but also due
to the particularities associated with the study of health, neighborhood effects health-
oriented studies differ from the neighborhood effects literature on civic engagement,
66

social disorganization, and educational attainment in at least two important ways. First,
health-focused studies are more likely to account for the impact of physical environment
stressors that are not elements of a community’s built environment, such as air and water
quality. And second, health-oriented neighborhood effects studies focus much more
heavily on the impact of structural neighborhood characteristics, and especially
socioeconomic status, than they do on the influence of specific social mechanisms on
health. Of the social mechanisms that have been studied to date, social capital has
received the most attention, followed by collective efficacy. The impact of mechanisms,
such as routine activities (LaGrange, 1999; Sampson et al., 2002; see also Chapter 2) is
noticeably absent.
Health-oriented neighborhood effects studies also suffer some of the same
limitations found in the neighborhood effects literature developed in sociology,
economics, and education. One of the most evident paths of development extends in the
direction of further investigating the role of institutional resources in the neighborhood.
To date their impact has been examined only insofar as their presence in or absence from
a community is thought to influence residents’ health (Robert, 1999b). The dual role they
can perform in the community, as both elements of the neighborhood structure and as
neighborhood actors has not been considered. This is one of the key goals of this project.
In addition, as in non-health-oriented neighborhood studies, the role of
communication as a social process has been largely overlooked. In the remainder of the
chapter, I set out to articulate a communication infrastructure model of neighborhood
effects that can be applied in health-oriented research and in efforts intended to limit
disparities in health literacy and health care access.
67

Applying communication infrastructure theory to the study of neighborhood health
The integrated storytelling network (STN)
(i.) Actors and links: Communication infrastructure theory (CIT; e.g., Ball-
Rokeach et al., 2001; Kim & Ball-Rokeach, 2006b) distinguishes between three levels of
storytelling agents, from micro, to meso, to macro, based on their primary storytelling
referent and their imagined audience. At the macro level, the mass media tell stories
about the city, the nation, and the world at large. Their imagined audience is the
population of the city, the state, the region, or the country. At the micro-level storytellers
include residents in their networks of family, friends, and neighbors. With regard to
meso-level storytellers, studies that have applied CIT have considered the impact of two
key actors, (a) geo-ethnic media (e.g., Lin & Song, 2006) and (b) community
organizations (e.g., Wilson, 2002a; Wilson, 2002b), on neighborhood belonging,
collective efficacy, and civic participation. I briefly discuss the roles each one of these
neighborhood actors plays as integral parts of a neighborhood’s communication
infrastructure next. In addition, I elaborate on why and how health care service providers
can be incorporated as meso-level neighborhood storytellers in a CIT model applied to
neighborhood health-oriented research.
a. Geo-ethnic media: Geo-ethnic media refer to those media organizations that
either target specific geographical areas or specific populations (such as new immigrant
minorities). These can be newsletters, free publications, or more sophisticated
commercial, print and electronic, media. Prior work has shown that geo-ethnic media, as
community storytellers, do a better job than large-scale, mainstream media that target the
larger metropolitan area, the region, or the nation as a whole – i.e., macro-level
68

storytellers – in promoting individual residents’ civic participation and sense of
neighborhood belonging (e.g., Finnegan & Viswanath, 1988; McLeod et al., 1996; Moy,
McCluskey, McCoy, & Spratt, 2004; Rothenbuhler, Mullen, DeLaurell, & Ryu, 1996).
Finnegan and Viswanath (1998) reported that use of community media such as
community weekly, metro daily, and community cable system is related to various
aspects of community ties. In particular, they found that a connection to community
newspapers is more likely to increase the level of integration into local communities than
a connection to metro papers. In a more recent study, Moy et al. (2004) also found that it
was a connection to local newspapers and not network television stations (which, as a
rule, address non-local news) that
promotes political participation in the
Seattle areas they studied.
The research done to date
suggests that geo-ethnic media have
the capacity to strengthen the ties that
bind neighbors to each other and to the
place they live. They accomplish this
by telling the stories that matter most
to the neighborhoods they serve.
Residents, but also community-based
organizations are part of these stories;
not only as audiences, but also as
sources of information; in a sense, they
Figure 3.1. Geo-ethnic media and their
relationship to other neighborhood
storytellers in the storytelling network
Figure 3.1. Geo-ethnic media include all media
that focus on local issues and/or are targeted to a
particular ethnic group living in the residential
community. Relationships among geo-ethnic
media, as well as between these media and other
neighborhood actors can be unidirectional or bi-
directional.

69

are involved as co-producers. In a neighborhood where certain health issues are
prominent, geo-ethnic media can raise awareness within and beyond the community; they
can generate discussion among residents that may eventually lead to mobilization efforts
(e.g., to urge city council members to address the impact of freeway traffic on children at
a local school); and they can generate conversation among residents and community-
based organizations that have the capacity to help address problems directly (e.g.,
through hosting a health fair, or workshops, provide free immunizations) or advocate on
behalf of the neighborhood. Community-based organizations can push city and state
officials to improve, for example, the quality of food residents can find at local
restaurants or the availability of recreation areas (e.g., Soto, 2008; Garcia, 2008).
b. Community-based Organizations: Community-based organizations are the
second key meso-level storyteller in
the neighborhood storytelling network.
A wide variety of organizations can be
classified as such, ranging from
informal grassroots formations to large
and formal non-profit organizations.
Consistent with previous research,
community organizations have been
found to play storytelling linkage roles
in the overall communication
infrastructure model, particularly with
regard to building civic engagement.
Figure 3.2. Community-based organizations
and their relationship to other neighborhood
actors in the storytelling network
Figure 3.2. Community-based organizations
include religious CBOs, shelters, educational
CBOs, recreation centers, libraries, schools, and
others.
70

Wilson (2001), for instance, indicates that the most important role of community
organizations in the community-building process is to be “the precipitators and sustainers
of neighborhood conversations” (p. 12) that imagine viable community.
Figure 3.3. The health storytelling network
of the neighborhood communication infrastructure

Figure 3.3. The health-oriented neighborhood storytelling network set in its communication
action context. A more detailed depiction of the relationships among particular neighborhood
storytellers can be found in Figures 3.1, 3.2, 3.4, and 3.5.

Several scholars have focused on the functions of storytelling in organizations
(see, e.g., Boje 1991; Boyce, 1990; Gephart, 1991). For example, Boje introduced and
defined the concept of a “storytelling organization” as a “collective storytelling system in
which the performance of stories is a key part of members’ sense making and a means to
71

allow them to supplement individual memories with institutional memory” (p. 106). Kim
and Ball-Rokeach (2006a) note that,
[t] he storytelling function of an organization may apply not only to the
transactions among internal stakeholders (i.e., employees at different levels) but
also to the relationships with outside stakeholders such as clients, suppliers,
competitors, consumers, or even the larger community (p. 180).

Through their relations with individual members, clients, or participants,
community organizations can tell salient stories about institutional and community
history, pressing issues (including health issues), opportunities, threats, and so forth, and
these stories may be passed along in neighbor-to-neighbor conversations. However,
organizations are not just conduits of information. As neighborhood stakeholders, they
also have “their own story to tell, and they can do this not only in communication with
residents but also in communication with local media” (Kim & Ball-Rokeach, 2006a, p.
180) and health service providers. The overlap between the organization’s stories and the
stories of concern to residents is likely to vary. The extent of overlap depends on the
mission of the organization and the degree to which its efforts are truly geared towards
the betterment of the local community.
To accomplish their goals community organizations do not only reach out to
media though; they also recruit the support of and try to establish collaborative
relationships with a variety of other organizations. If their goals, for example, include
addressing neighborhood health concerns, forging ties with health service providers can
be very important.
c. Health service providers as meso-level storytellers: Health care service
providers are not formally considered part of the storytelling network, but rather as
72

elements of the communication action context. Their presence or absence may affect the
strength of the ties between neighborhood storytellers. Neighborhood health clinics, for
example, are places where parents are likely to engage one another in conversation about
their health or the health of their children. Health care professionals (e.g., doctors) and
health service providers (e.g., clinics, hospitals) are important resources for people
seeking information about personal and family health issues. While their presence as
communication action context features is important, I argue that their agentic role as
actors in the community is also critical. Similarly to media and community-based
organizations, health service providers play a dual role in the neighborhoods they serve.
Because the focus of this study is on health-related neighborhood outcomes, they are
included in the storytelling network.
They can be a source of information for
residents. Media and community
organizations may seek information
from them, but they also may push
information towards all three other
storytellers (i.e., media, organizations,
and residents), as part of a community
outreach campaign. Conceptually, the
links between health service providers
and the other three neighborhood
actors are bidirectional.
Figure 3.4. Health service providers and their
relationship to other neighborhood actors in
the storytelling network
Figure 3.4. HSPs Include specialized and
emergency care providers, clinics, etc.
73

d. Residents, as micro-level storytellers: In CIT, residents as constituted in family,
friend, or neighborhood networks are essential community storytellers (micro-level).
Ball-Rokeach, Kim, & Matei (2001) found that individuals talking about the
neighborhood with their neighbors is
the most potent storytelling force in
constructing neighborhood belonging.
Interpersonal communication among
neighbors has also been found to be
positively related to civic participation
(Ball-Rokeach et al., 2001; Kim, Ball-
Rokeach, Cohen, & Jung, 2002;
McLeod, Scheufele, & Moy, 1999;
Wyatt et al., 2000).
In addition, a number of studies have demonstrated that interpersonal
communication is extremely important for the dissemination of health information
(Griffin & Dunwoody, 2000; Valente & Saba, 2001; Wilkin, 2005). In this context,
interpersonal communication might involve residents talking to neighbors, friends,
colleagues, family members, but also residents connecting to health service providers and
talking with health care professionals.
Figures 3.1, 3.2, 3.4, and 3.5 illustrate the relationships that may emerge (a)
among meso-level actors of the same type (e.g., community-based organizations), (b)
across different types of meso-level actors (e.g., community-based organizations and
health service providers), and (c) among meso and micro-level actors (e.g., health service
Figure 3.5. Residents and their relationship to
other neighborhood actors in the storytelling
network

74

providers and residents). Figure 3.3 presents the entire health storytelling network
situated in the neighborhood communication infrastructure.
(ii.) Integration across levels and the added value of the network: The stronger
the connection between neighborhood actors and the more likely it is that a link to one
prompts building links to others, the more integrated the storytelling network is (Ball-
Rokeach et al., 2001; Kim & Ball-Rokeach, 2006b). A communication campaign about
child obesity, for example, will bring health services providers who launched the
campaign in touch with the media. The message will reach residents, but also community
organizations. Community organizations reach out to their members and may attempt to
get in touch directly with the health service providers to receive more information or to
organize seminars and workshops. In this case, the initial action taken by one actor (in
this case a health service provider) creates the impetus for many more links to be
established across the residential community.
As the web of connections among neighborhood actors becomes more tightly
woven, a common understanding of what the neighborhood looks like, but also what its
problems are, and a vision of what it could be emerge. The strengthening of the
storytelling network reflects the degree of co-orientation among neighborhood
stakeholders around the definition of their shared space, problems, and resources. Co-
orientation is the pre-requisite for mobilization, taking action, and effecting change in a
community. It is in this sense that (a) the impact of a strong storytelling network is more
significant than the effect(s) every additional link created among community actors
produces, and (b) that an integrated storytelling network can become an ‘engine’ of
community change.
75

Synergy, however, among neighborhood storytellers may not always be possible
and links between neighborhood storytellers may not exist due, for instance, to ethno-
racial tensions. Or links may exist, but be weak. In such cases the degree of storytelling
integration at the neighborhood level is lower. In communities with a weaker storytelling
network, residents and organizations find it harder to overcome the hurdles that stand in
the way of accomplishing their goals (e.g., help their children do better in school or make
sure they are getting proper nutrition).
In previous work, Kim (2003) and Kim and Ball-Rokeach (2006a, 2006b)
developed a measure of storytelling integration, which has been found to be a critical
factor in neighborhood belonging, collective efficacy, and civic participation – three
components of civic engagement. While the methodological details pertaining to the
creation of the measure will be discussed at greater length in the following chapter on
methodology, it is important here to (a) lay out the logic that links the measure to the
concept of storytelling network integration, and (b) to explain how this project adds to
prior work done on communication infrastructure theory.
The measure of integrated connectedness to a storytelling network (ICSN)
proposed by Kim and Ball-Rokeach captures the extent to which other neighbors, local
and ethnically-targeted media, as well as community-based organizations are integrated
into individual residents’ everyday life. ICSN is calculated as a summation of the three
interaction terms between the scope of residents’ connections to local and ethnically-
targeted media, the scope of residents’ connectedness to community organizations, and
the intensity of interpersonal neighborhood storytelling [See Formula (1)]. In Formula
(1), LC refers to geo-ethnic media connectedness, INS is intensity of interpersonal
76

neighborhood storytelling, and OC refers to the scope of connection to community
organizations.

The ICSN measure represents an important step towards capturing the degree of
integration of the storytelling network. However, it only captures the degree to which
meso level actors (i.e., institutional actors like media and community organizations) and
neighbors are integrated into the lives of residents. It does not tell the organizations’ side
of the story. How important are media, for instance, in the lives of community
organizations? Moreover, ICSN is limited in terms of what it can tell us about how
integrated the neighborhood storytelling network is as a whole, and which actors are
playing a more active role in brokering relationships among neighborhood actors.
(iii.) Community communication capital (CCC): To address these two
shortcomings of ICSN, this study proposes a new measure that directly considers the
impact of meso-level actors. It is intended to capture the extent to which individual and
meso-level neighborhood actors both are integrated into each other’s everyday life and
the degree to which the neighborhood storytelling network is integrated overall. As the
focus is on health-related outcomes, namely health care access and health literacy, along
with local and ethnically-targeted media and community-based organizations I also
consider health service providers as neighborhood storytelling actors. The new measure is
introduced as an indicator of community communication capital created through the
communicative ties that bind micro and meso-level neighborhood storytellers in the
process of everyday life. Communication capital is akin to the idea of social capital
ICSN = (LC x INS) (INS x OC) (OC x LC) + + (1)
77

(Bourdieu, 1986; Coleman, 1988; Putnam, 2000). It is a resource that residents can tap
into in search of critical information that will help them protect their own health, and that
of their families and communities. Social capital though is usually conceptualized and
operationalized at the individual level (i.e., interpersonal networks), and does not account
for the integration of the multi-level network of neighborhood actors (or storytellers). In
addition, the storytelling aspect incorporates the communicative dynamics, which are not
considered in social capital, or in most neighborhood effects studies. Thus, the CCC
concept differs in being both multi-level and focused upon a communicative process.
(iv.) The ideal scenario: Neighborhoods where the storytelling network is truly
integrated should have healthier residents. In such neighborhoods, residents talk to
trusted neighbors about health issues, they exchange information about resources, and
refer each other to health providers. The media report on health issues, including new
medical breakthroughs and the health problems of at-risk groups of residents. They also
report on events hosted by community organizations on child health, for example, and
they publicize the latest findings of research conducted by health providers on the
community. Community organizations may get involved in health campaigns, by hosting
seminars (e.g., on child obesity) for residents to attend. And, finally, health providers
undertake campaigns to alleviate specific community health-related problems (e.g., drug
abuse). In doing so, they enlist the help of community organizations and they approach
the media to get their message across to as many residents as possible. In such a
neighborhood, health problems are more likely to be prevented and residents are more
likely to feel able to cope with the ones that arise.
78

Neighborhood effects
I. Direct effects
Consistent with the literature linking social capital to health, I investigate the
direct link between community communication capital (CCC) and two health outcomes:
health literacy and access to health care.
(1) Health literacy
According to the U.S. Department of Health and Human Services (2000), health
literacy reflects the capacity of individuals to “obtain, interpret and understand basic
health information and services and the competence to use such information and services
to enhance health” (in Kickbusch, 2001, p. 293). There is a dearth of studies that
empirically research the link between social capital and health literacy, although there is
no lack of articles suggesting that exploring the link between the two would be extremely
useful in improving people’s health (e.g., Kickbusch, 2001; Ratzan, 2001). While not
identical, social capital and community communication capital are akin in that they are
resources realized through the ties that emerge and connect residents in the process of
their everyday life.
In this study, I examine the link between community communication capital and
health literacy. In a neighborhood with a highly integrated storytelling network, health
information should travel faster and be made available to more residents. In such a
community, residents should be in a position to obtain the knowledge and develop the
ability to prevent, detect, and solve health-related problems. Therefore, I hypothesize
that:
79

H
1
: The degree of community communication capital will predict the degree of
health literacy in a neighborhood, after controlling for individual level covariates
(e.g., socio-economic status, residential tenure).
(2) Access to health care resources
In a neighborhood rich in communication capital, residents should also feel more
capable about being able to find medical care getting access to healthcare resources for
themselves and their families:
RQ
1
: Is the degree of community communication capital positively related to how
easy residents find that it is for them to access information and services they need
to safeguard their health?
(3) Civic engagement
Individual residents’ integrated connectedness to a storytelling network (ICSN) is
consistently found to be an important factor in civic engagement, conceptualized as
having three dimensions: (a) neighborhood belonging, (b) collective efficacy, and (c)
political participation. Higher ICSN scores are positively related to residents’ attachment
to the neighborhood, their sense that they can come together to solve common problems,
and their active participation in local policy-making and neighborhood opinion-making
processes (Kim, 2003; Kim & Ball-Rokeach, 2006b). A similar relationship is predicted
to exist between community communication capital and the three dimensions of civic
engagement, as media, community organizations, and heath service providers contribute
to the integration of the storytelling network. Therefore:
H
2-1, 2-2, 2-3
: Community communication capital will be positively related to
neighborhood belonging (H
2-1
), collective efficacy (H
2-2
), and civic participation
80

(H
2-3
), after controlling for individual level (i.e, socio-economic status and
residential tenure) and neighborhood level covariates (i.e., ethnic
diversity/heterogeneity, residential stability, and density of institutional
resources).
II. Indirect effects
In a neighborhood with a highly integrated storytelling network, institutional
actors are part of the everyday life of the community. They know the problems residents
have to deal with, they talk about them, and they strive to find solutions working with one
another. In such a community, residents are more likely to feel that they are able to find
the support and information they need to prevent, detect, and cope with health problems.
They are also more likely to feel that they can access neighborhood resources (e.g.,
medical facilities) necessary to address health care issues.
H
3a-1, 3a-2, 3a-3
: There is an indirect path of influence from community
communication capital to health literacy via neighborhood belonging (H
3a-1
),
collective efficacy (H
3a-2
), and civic participation (H
3a-3
).
H
3b-1, 3b-2, 3b-3
: There is an indirect path of influence from community
communication capital to residents’ perceived ability to access information and
services they need to safeguard their health via neighborhood belonging (H
3b-1
),
collective efficacy (H
3b-2
), and civic participation (H
3b-3
).
III. The impact of the communication action context: direct and indirect effects
According to communication infrastructure theory, the configuration of the
communication action context (CAC) enables and constrains the integration of the
storytelling network. In this study I consider the impact of four CAC characteristics –
81

socioeconomic status, density, ethnic heterogeneity, and residential stability – on
community communication capital. Figure 3.6 illustrates the relationship between the
CAC elements included in this study and the neighborhood storytelling network.
Figure 3.6. The communication action context and storytelling network relationship

(1) Density of institutional resources
The density of institutional resources has been found to be a good predictor of
civic engagement (e.g., Sampson et al., 2006) and of health-related outcomes (see, e.g.,
Peterson, 2000). The following direct and indirect links are investigated:
H
4-1, 4-2, 4-3
: Density of institutional resources will predict neighborhood belonging
(H
4-1
), collective efficacy (H
4-2
), and civic participation (H
4-3
), after controlling
for individual level covariates (e.g., socio-economic status and residential tenure)
82

and neighborhood level covariates (i.e., ethnic diversity/heterogeneity, residential
stability, and community communication capital).
H
5a
: Density of institutional resources will predict health literacy, after controlling
for individual level and neighborhood level covariates.
H
5b
: Density of institutional resources will predict residents’ perceived degree of
access to health-related services, after controlling for individual level and
neighborhood level covariates.
In addition and based on the theoretical framework elaborated earlier according to
which the communication action context (or neighborhood environment) is expected to
affect the impact of the social mechanisms of neighborhood effects, I pose two more
research questions:
RQ
2a-1, 2a-2, 2a-3
: Does the density of institutional resources have a larger, positive
effect on health literacy as the levels of neighborhood belonging (RQ
2a-1
),
collective efficacy (RQ
2a-2
), and civic participation (RQ
2a-3
) in a neighborhood
increase?
RQ
2b-1, 2b-2, 2b-3
: Does the density of institutional resources have a larger, positive
effect on residents’ perceived ability to access health care resources as the levels
of neighborhood belonging (RQ
2a-1
), collective efficacy (RQ
2a-2
), and civic
participation (RQ
2a-3
) in a neighborhood increase?
Prior neighborhood effects research indicates that characteristics of a
community’s social environment, such as ethnic diversity and residential stability, are
common and reliable independent predictors of neighborhood outcomes (e.g., Sampson et
al., 2002). Moreover, the public health literature indicates that the ethnic composition of a
83

neighborhood is linked to health access and health literacy disparities. Therefore, apart
from the density of institutional resources available in a community, I also investigate the
impact of the two abovementioned CAC variables on health literacy and access to health
care resources.
(2) The impact of the ethnic heterogeneity
Earlier studies have found a negative impact of ethnic heterogeneity on civic
engagement (e.g., Rice & Steele, 2001; Okten & Osili, 2004). From a CIT point of view,
increased ethnic heterogeneity in a neighborhood might make residents less likely to
engage each other in conversation in the course of every day life. In addition, residents
with different ethno-cultural backgrounds may choose to connect mostly to media that are
targeted to them (e.g., ethnic newspapers), and they may attend meetings of community
organizations that cater to people with the same ethno-cultural background. If ethnic
heterogeneity makes it more difficult for neighbors to communicate with one another,
then the relative absence of resident-organization connections that cut across ethnic lines
are likely to acerbate the problem of disconnectedness among residents.
Organizations have the capacity of bridging the divide between ethnic populations
in shared community space. Institutional actors though may also choose to target
particular segments of the population. This is a common practice for media organizations
that bank on niche marketing strategies. The narrower the scope of connections that
institutional actors forge and maintain in the residential community, however, the more
limited their reach is likely to be.
Overall, the strong presence of connections across the neighborhood storytelling
network (from residents to organizations and vice versa) that privilege ethnic
84

commonality and the absence of ties that reflect efforts to bridge ethno-racial divides
signals the presence of a fragmented neighborhood storytelling network. The more
ethnically heterogeneous a neighborhood is, the more likely it is that the storytelling
network of a neighborhood will be fragmented and that the community’s communication
capital will be lower. Therefore:
RQ
3
: Is ethnic heterogeneity negatively related to community communication
capital?
A bifurcated or even more fragmented storytelling network could mean that a
neighborhood’s stores of community communication capital are low. The distinct,
ethnically-oriented storytelling networks though may be vibrant. Conceptually, at the
neighborhood level, there is no ‘averaging effect’ necessarily. Having two strong
ethnically-oriented storytelling networks in a community does not mean that the overall
neighborhood network will be strong, nor does it necessarily imply that one strong
ethnically-oriented network and another weaker one will make for an overall
neighborhood network that falls somewhere around the middle of a theoretical weak to
strong continuum. The strength of a neighborhood’s STN will depend on:
a) the degree to which there are institutional resources that serve the entire
community (and not one ethnic group); and
b) the extent to which all residents, regardless of ethnic background, connect
to a number of the same institutional actors present in their neighborhood.
Table 3.1 illustrates possible scenarios one might observe with respect to the
storytelling network in a diverse community with two ethnically-distinct populations.
85

RQ
4a
: Is there a significant difference among organizations in Greater Crenshaw
in terms of ethnic reach?
RQ
4b
: Is the number of organizations that all Greater Crenshaw residents connect
to significantly different from the number of organizations to which only residents
of a particular ethnic background connect?
Table 3.1. Scenarios for a bifurcated neighborhood storytelling
network along ethnic lines

Based on the extant research on health disparities, it is reasonable to inquire
whether ethnicity is a factor in predicting health literacy and health care access. In
addition, if residents’ ability to get access to important health information resources
depends on their integration into the neighborhood storytelling network and the degree of
STN integration is influenced by ethnic heterogeneity, then it is possible that ethnic
heterogeneity may be responsible for observed differences in health literacy scores and
access to health care resources among Greater Crenshaw residents.
RQ
5a-5b
: Is ethnicity a predictor of health literacy (RQ
5a
) and access to health care
resources (RQ
5b
) controlling for individual level covariates?
86

RQ
6a-6b
: Is ethnic heterogeneity negatively related to health literacy (RQ
6a
) and
access to health care resources (RQ
6b
) controlling for community communication
capital?
(3) Residential stability
Many neighborhood effects studies and research employing CIT have found
support for the link between residential stability and civic engagement (e.g., Putnam,
2000; Shah et al., 2001; Kim & Ball-Rokeach, 2006b). Residential instability and low
rates of home ownership have also been associated with a variety of problem behaviors
(Brooks-Gunn et al., 1997). This study extends these findings and investigates the
structural relationship between communication capital, civic engagement, and two health
outcomes: health literacy and residents access to health services.
In more unstable communities, residents come and go and the relationships
between neighbors are tenuous. Individuals are likely to feel less attached to the
community and feel that they can rely only on themselves to accomplish their goals. In
such neighborhoods, residents are likely to feel that they expend too much effort to get
access to critical health-resources. Their knowledge about health risks, disease prevention
and detection is likely to be lower than health literacy of residents living in communities
where information that helps residents make important decisions in the course of
everyday life is readily available.
Active institutional actors can counter the effects of residential instability. Geo-
ethnic media, community based organizations, and health-service providers may be in a
position to compensate for the weakness in the relationships between residents, by
reaching out and integrating new residents into the community and keeping older
87

residents from moving away. If meso-level actors engage in such efforts and are
successful, then residential instability will have a less negative effect in neighborhoods
with higher community communication capital.
H
6a
: Residential stability will have larger, positive effects on health literacy, as
the level of a community’s communication capital increases.
H
6b
: Residential stability will have larger, positive effects on residents’ perceived
ability to access information and services they need to safeguard their health, as
the level of a community’s communication capital increases.
Figure 3.7 summarizes the conceptual model of the study.
Figure 3.7. The theoretical model and hypotheses guiding the study

(*) Note: All the hypotheses and research questions presented here reflect the theoretical
framework developed earlier in the dissertation. However, for reasons that will be
discussed in greater detail in the next chapter on methodology (Chapter 4), some of the
88

hypotheses had to be revised to reflect the analytical strategy employed as an alternative
to the one originally designed. The revised hypotheses are articulated in Chapter 5.

89

CHAPTER 4
METHODOLOGY
Data from four different sources were used in this study: (a) a telephone survey of
Greater Crenshaw residents, African Americans and Latinos, conducted through the
multi-year Metamorphosis Project at the Annenberg School for Communication at the
University of Southern California (see http://www.metamorph.org) during the period of
November 2005-January 2006
1
; (b) in-person, structured interviews with staff members
and leaders of organizations located in Greater Crenshaw, which I conducted with the
help of two research assistants under my supervision from November 2007 through May
2008; (c) the U.S. Census Bureau; and (d) two secondary data directories,
HealthCity.org
2
and the Rainbow Resource Directory
3
mined for information on
institutional resources available in the communities studied. I will begin by discussing the
procedure of collecting data from the residents of Greater Crenshaw and the measures
developed based on those data. Subsequently, I will introduce the sampling scheme for
the collection of data from institutional resources in the study area, as well as measures
that were developed to capture attributes of the organizations considered essential for the
purposes of this study. Then, I will describe the relevant measures that capture
dimensions of the study area’s communication action context (i.e., dimensions of the
neighborhood environment). These measures were constructed based on data from the
2000 U.S. Census as well as data from secondary data sources. In the final sections of

1
The author has been a member of the Metamorphosis research team since 2003 and was engaged in the
development of the survey instrument that was used to collect data from residents in the Greater Crenshaw
area.
2
HealthyCity.org is an online community service and policy research tool for all of Los Angeles County.
Healthy City provides access to a database of community resources and localized demographic and health
data on a GIS mapping platform. HealthCity.org is accessible via www.healthycity.org.
3
The Rainbow Resource Directory is a social services referral guide published for Southern California.

90

this chapter, I will lay out the analytical scheme deployed to investigate the hypotheses
that were articulated in Chapter 3.
The multi-level neighborhood storytelling network: data collection and measures
I. Residents
a. Sampling & Data Collection
The individual-level data were gathered from six hundred and seven households
(N = 607) selected by random-digit dialing for participation in a telephone survey. Three
hundred and three (N = 303) of those households self-identified as Latino and the
remaining three hundred and four (N = 304) as African American. The survey was
administered in the respondents’ choice of language (i.e., English or Spanish).
4

Based on the formulae provided by the American Association for Public Opinion
Research (AAPOR) our survey response rate was 13.5% when calculated by dividing the
number of completed interviews by the number of theoretically eligible phone numbers.
Eligible phone numbers were calculated by examining the total number of study area
phone numbers excluding phone numbers for which eligibility could not be determined,
inappropriate/duplicate phone numbers, non-qualified household phone numbers (e.g.,
outside the study area), and the estimated number of initial refusals not likely to qualify
for our study (for more details on calculating the response rate, see: Cohen, Ball-
Rokeach, Jung, & Kim, 2002; Kim, Ball-Rokeach, Cohen, & Jung, 2002; Kim, Jung,
Cohen, & Ball-Rokeach, 2003).

4
A survey research firm was employed using trained bilingual interviewers (the survey was translated and
back translated in each language) programmed for Computer Assisted Telephone Interview (CATI)
administration.

91

b. Measures
Social mechanisms of neighborhood effects: independent variables
(i) Integrated Connection to the Storytelling Network (ICSN)
The measure has three components, each of which is discussed in turn.
Intensity of Interpersonal Neighborhood Storytelling: The intensity of
participation in interpersonal storytelling was measured by asking respondents to
indicate, on a scale from 1 to 10, where 1 represents “never” and 10 “all the time,” how
often they “have discussions with other people about things happening in their
neighborhood.”
Scope of Connections to Community Organizations:
5
Assessing the scope of
connections to community organizations involved a two-step process. Respondents were
asked if they belonged to any of five different types of organizations: (1) sport or
recreational, (2) cultural, ethnic, or religious, (3) neighborhood or homeowner’s, (4)
political or educational, and 5) other. Having a membership in each type was coded as
“1,” and responses to five types were summated to form a synthetic variable ranging from
0 to 5. However, initial examination of the data revealed that although many people
indicated that they did not have a membership in a religious organization, they reported
regularly attending church, temple, or other religious services. For these respondents, we
credited “1” to their religious organization scores if they attended a religious service
more often than once every few weeks. By summating these scores, we created a final

5
One study (Wollebaek, & Selle, 2002) found that scope of organizational participation was a more
effective factor of social capital than intensity of organizational participation.

92

synthetic variable assessing the scope of the respondents’ connections to community
organizations, with possible scores ranging from 0 to 5.
Scope of Connections to the Local Media: To measure the respondents’ scope of
connections to the local media, we examined whether respondents spent any time
connecting to local newspapers, radio, or television in the prior week. Local media were
defined in our survey as either community media targeted to a particular ethnic group
(i.e., Armenians, Latinos or whites) in the study area or public/noncommercial media
oriented to the study area. We summated the number of affirmative connections to create
a scope variable that reflected the breadth of connectedness to the different types of local
media (range = 0-3).
Integrated Connection to a Storytelling Network (ICSN): This concept was
discussed in earlier chapters. ICSN was a measure developed first by Kim and Ball-
Rokeach (e.g., Kim, 2003; Kim & Ball-Rokeach, 2006b). It is calculated as a weighted
summation of three interaction terms between scope of connections to the local media,
scope of connections to community organizations, and intensity of interpersonal
neighborhood storytelling. Each of the three storytelling connection variables was first
standardized into z-scores. Then, the standardized z scores were re-coded into 6 ranked
categories so that all the values became positive numbers. This is the formula for the
ICSN calculation, as presented by Kim and Ball-Rokeach:

In Formula 1 LC is the standardized z score of scope of connection to local media,
INS is the standardized z score of intensity of interpersonal neighborhood storytelling,
W
1
( LC x INS ) W
2
( INS x OC ) W
3
( OC x LC ) + +
(1)

93

and OC is the standardized z score of scope of connection to community organizations.
W
1
,

W
2
, and W
3
are weights for each of the three interaction terms. Since there is no
theory determining a weight for each interaction term in the formula above, a weight of 1
was assigned to each of them.
In the present study, one of the primary goals is to assess the integration of
organizational neighborhood actors into the storytelling network and its impact on
residents. For this purpose organization-level variables of integration were developed
with new data collected directly from organizations. To avoid having the potential impact
of organizations become confounded by the indirect measure of interaction between geo-
ethnic media and community organizations that is part of ICSN, the interaction term was
removed from the measure. For the current study, the ICSN was calculated using a
modified version of Formula 1. To differentiate between the measure constructed by Kim
and Ball-Rokeach and the modified version used for this study, I will refer to it as ICSN-
2:

(ii) Neighborhood belonging
The respondent’s neighborhood belonging was measured with the ‘belonging
index’ (Ball-Rokeach et al., 2001; Cohen et al., 2002; Kim & Jung, 2003). The 8-item
belonging index captures “residents’ feeling of attachment to a residential area that
motivates everyday acts of neighborliness” (Ball-Rokeach et al., 2001). It includes an
equal numbers of items measuring subjective and objective dimensions. The specific
items are listed below.
W
1
(LC x INS) W
2
(INS x OC) +
(2)

94

Do you strongly agree, agree, neither agree nor disagree, disagree, or strongly
disagree with the statement (response on a 5-point Likert scale):
1. You are interested in knowing what your neighbors are like.
2. You enjoy meeting and talking with your neighbors.
3. It’s easy to become friends with your neighbors.
4. Your neighbors always borrow things from you and your family.
How many of your neighbors do you know well enough to ask them to
(respondent specifies a number):
5. Keep watch on your house or apartment?
6. Ask for a ride?
7. Talk with them about a personal problem?
8. Ask for their assistance in making a repair?
The Cronbach’s alpha test for index scalability was .78.
(iii) Perceived collective efficacy
Perceived collective efficacy (or simply collective efficacy, hereafter) was measured as a
composite variable containing six items about individual residents’ confidence in their
neighbors’ willingness to participate in neighborhood problem solving processes. The six
items are listed below.
How many of your neighbors do you feel could be counted on to “do something”
if:
1. A stop sign or speed bump was needed to prevent people from driving too fast
through your neighborhood.
2. There were dangerous potholes on the streets where you live.

95

3. The sports field that neighborhood kids want to play on has become unsafe
due to poor maintenance or gangs.
4. You asked them to help you organize a holiday block party.
5. A child in your neighborhood is showing clear evidence of being in trouble, or
getting into big trouble.
6. The trees along the streets in your neighborhood were uprooting the
sidewalks, making them unsafe.
For each of the six items, the respondents were given five response options:
“none,” “few,” “some,” “most,” and “all.” The average score of the six items were used
as a composite indicator of perceived collective efficacy (range 0-5).
The Cronbach alpha test for index scalability was .89.
(iv) Scope of civic participation
The scope of civic participation is captured in a synthetic variable where the
number of different activities individuals have participated in among five different civic
actions were counted and added up (McLeod et al., 1996): “attending a city council
meeting, public hearing, or legislative meeting,” “writing a letter to the editor of a
newspaper, television station, or magazine,” “contacting an elected official about a
problem,” “circulating a petition,” and “taking part in any political demonstration or
protest.” All of these civic participation items were binary measures where respondents
were supposed to answer either “yes (value=1)” or “no (value=0).” By summating these
five dummy scores, we created a final synthetic variable assessing the scope of civic
participation (range 0-5).

96

Neighborhood effects: dependent variables
(i) Health literacy
The Metamorphosis Greater Crenshaw survey was conducted in a residential area
that is largely African American and Latino. The sample (N = 607) consists of almost
equal numbers of people from both ethnic backgrounds (i.e., 304 African American and
303 Latinos). The literature suggests that African Americans and Latinos are at higher
risk than other populations for developing and suffering from high blood pressure
(hypertension) and diabetes, respectively (NCHS, 2005). The survey included items
meant to capture individuals’ knowledge of how to best prevent diabetes and
hypertension, but also items that reflect a person’s knowledge of the symptoms
associated with these diseases. Latinos were asked those items pertaining to diabetes and
African Americans answered those that were about preventing and recognizing symptoms
of hypertension. Over 55 different answers were given by Latinos when asked about how
they may prevent diabetes and over 35 different strategies or best practices were offered
as answers by African Americans for preventing hypertension. In addition, Latino
residents provided 20 different answers when they were asked about what they thought
are the warning signs or symptoms of diabetes. African Americans gave 25 different
responses to a similar question about hypertension or high blood pressure (see Appendix
A for the complete lists of answers that survey respondents provided).
The next step involved identifying what answers ought to be classified as correct.
For the diabetes items, a diabetes expert (Dr. Francine Kaufman, director of the
Comprehensive Childhood Diabetes Center, and head of the Center for Endocrinology,
Diabetes and Metabolism at Children’s Hospital Los Angeles) was consulted by

97

Metamorphosis as part of a NIH grant-writing effort to determine which answers to
consider correct and incorrect. Correct answers were tabulated and summed for every
respondent. The largest number of correct answers given was used to calculate a ratio of
correct answers provided over the maximum of total correct answers for every individual
(i.e., every Latino respondent). A similar process was employed to develop the
hypertension prevention and detection indices. African Americans’ responses were
compared against the information on how to prevent and detect hypertension provided by
WebMD and the Cleveland Clinic Heart Center. Doing so resulted in ratio measures of
prevention and symptoms detection health literacy for both African American and
Latinos (see Table 5.1 in Chapter 4 for mean and standard deviation scores).
The two indices created are considered measures of health literacy for the disease
a resident runs a higher risk of developing based on his or her ethnic background and
compared to the entire U.S. population. Therefore residents’ scores are treated as
comparable for the purposes of this study.
6

(ii) Access to health care
Two indicators were used to measure health care access, based on the survey data.
Perceived difficulty in getting medical care: Respondents to the survey were
asked: “Overall, how easy or difficult is it for you to get medical care when you need it?
Would you say it is very difficult, somewhat difficult, somewhat easy, or very easy?”
(Range: 1-4).

6
A third measure of health literacy was constructed by combining the prevention and detection dimensions
for each ethnic group. The measure was calculated by summing the number of correct answers every
individual resident provided for both preventing and detecting a disease and dividing the sum by the
maximum number of correct answers. The measure, however, does not appear in the analyses conducted
because it was thought more interesting, conceptually, to examine if there were different predictors of
prevention-oriented and symptoms detection-focused health literacy.

98

Distance to get health care: Residents were asked: “Thinking of the place you go
most often for medical care, how long does it usually take you to get there?” Responses
were reverse-coded so that a lengthier trip and larger distance could be used as an
indication of lack of necessary health services in the respondent’s local area.
Individual-level covariates
Residential tenure is a continuous measure of years of neighborhood residence.
Home ownership was a dichotomous measure, marking owner status by “1” and renter
status by “0.” Respondents were asked their age on their last birthday, and their
household income from the previous year (in ranges staggered from less than $20,000 to
more than $100,000). The highest grade or year of school that the respondents completed
was used to indicate their educational level, ranging from eighth grade or less to a
graduate degree.

II. Organizational neighborhood actors
a. Sampling & Data Collection
The study area of Greater Crenshaw was defined in 2005 as the area of nine
contiguous zip codes: 90007, 90008, 90016, 90018, 90019, 90037, 90043, 90056, and
90062. It extends from as far West as the I-110 highway to as far East as Kenneth Hahn
State Park. Its northern-most boundary rests just North of Venice Blvd., while its
southern-most boundary is Slauson Ave (for a more detailed description of the study area,
please see Chapter 2). Mining the data of HealthyCity.org and the Rainbow Resource

99

Directory, I compiled an inventory of institutional resources available in the area.
7
The
list included organizations from the following nine categories: (a) education-oriented, (b)
religious, and (c) other community-based organizations
8
, (d) schools, (e) libraries, (f)
sports and recreation centers, (g) neighborhood councils, (h) city council district offices,
and (i) health service providers.
A tenth category included media organizations. For the list of media available to
residents of Greater Crenshaw neighborhoods I did not rely on secondary data sources,
but rather on the responses of residents that participated in the Metamorphosis survey to
questions about what media they depend on the most for information that helps them: (a)
stay on top of what is happening in their community, (b) address health concerns, and (c)
make purchasing decisions. The media mentioned by residents most frequently can be
found in Table 4.1. The list of media was created this way because (a) media do not tend
to target populations at the zip code area level, but usually across much larger areas, (b)
the capacity of media to serve a population is not limited in the same way that, for
instance, the capacity of a health center or a school can be.
The total number of organizations that appear to be active in the Greater
Crenshaw area (without counting the media) is 333. Table 4.2 presents summaries of the
available institutional resources in the area by type and by zip code and Figure 4.1
illustrates the distribution of these resources across Greater Crenshaw.

7
There were no respondents to the Metamorphosis Greater Crenshaw survey from zip code 90056. As one
of the goals of this study was to better understand the impact of organizations, as neighborhood storytelling
actors in the everyday lives of residents, organizations based in 90056 were excluded.
8
The other categories of community-based organizations included were (based on the HealthyCity.org
classification system): (a) organizations providing shelter, food, or for other basic needs; (b) organizations
providing public assistance to individuals searching for jobs or dealing with financial problems, (c)
organizations that provide a variety of resources to families (e.g., WIC centers), and (d) others.

100

Table 4.1. Top media choices for African American and Latino residents
of Greater Crenshaw
Ethnic Group Television Radio Newspapers
African Americans KABC (Channel 7) KJLH 102.3 L.A. Times
KTTV (Fox, Channel 11) KKBT 100.3 (THE BEAT) L.A. Sentinel
KCAL (Channel 9) KFWB‐AM 980 The Wave

KCBS (Channel 2) KNX‐AM 1070
L.A. Watts
Times
KNBC (Channel 4)
KTLA (Channel 5)

Latinos
KMEX (Univision, Channel
34) KLAX 97.9 (LA RAZA) L.A. Times
Telemundo (Channel 52) KLVE 107.5 La Opinion
KWHY (Channel 22, Spanish) KSCA‐FM 101 (LA NUEVA) The Wave
KTTV (Fox, Channel 11) KTNQ‐AM 1020
KABC (Channel 7)
KCAL (Channel 9)
Legend
Top 3 media mentioned
Overlap between ethnic groups

Table 4.2. Summary of institutional resources in Greater Crenshaw (media not included)
RESOURCES BY ZIP CODE RESOURCES BY TYPE
Zip No. Resources Org Type No. Resources
90007 55 Schools 147
90008 34 Educational CBOs 10
90016 36 Libraries 8
90018 54 Recreation 25
90019 44 Religious 23
90037 41 CBOs (other) 56
90043 47 N‐Councils 8
90062 22 Political/District 4
Health Providers 52
Total: 333 Total: 333

101

Figure 4.1. Distribution of population of institutional resources
in Greater Crenshaw (N=333)

102

Interviewing all these organizations would be unwieldy and require a large
research team, in order for the data collection process to be completed in a relatively
short amount of time. As an alternative, a structured random sampling procedure was
employed to select organizations that we would attempt to recruit and interview. The
inventory created reflecting the populations of the different types of organizations
available in the area were sorted by zip code and organizations were randomly selected in
every zip code. The rules applied for selecting organizations in every zip code are
summarized in Table 4.3. They were based on the size of the overall population of a
particular type of organization in the study area. Four schools were selected per zip code,
for example, because the population of schools in the study area was the largest (N=147).
Table 4.3. Rules applied for sampling organizations in Greater Crenshaw
Decision Rules
for Sampling
Type of
Organization
a. 4 Schools per zip code (1 elementary, 1 middle school, 1 high school,
1 w/ religious affiliation)
b. 1 Education‐oriented CBO per zip code
c. 1 Library per zip code
d. 1 Recreation‐oriented organization per zip code
e. 1 Religious organization per zip code
f. 2 CBOs per zip code (out of the 4 classes that are clustered together:
i) Basic needs/Shelter/Food, ii) Income Security/Public Assistance,
iii) Individual and Family Life, iv) Other CBOs)
g. 1 Political/District office per zip code
h. 1 Neighborhood Council
i. 2 Health providers (focus on ones that have a more general mission,
larger, likely to serve more people)
Total 14 per zip code (provided an adequately large number of
organizations of every type exists every zip code area)

103

As some zip codes did not have any organizations of one type or another within
their boundaries, the final sample (excluding the media) consisted of 101 organizations
(instead of 112).
From the compiled list of media, we focused on the top three electronic media
(i.e., television and radio in this case) and newspapers mentioned by the survey
respondents. The overlap in terms of the organizations mentioned by African Americans
and Latinos, as evident in Table 4.1, was limited. Therefore the list of media that we tried
to recruit was 16 (the Los Angeles Times and the The Wave were mentioned by both
ethnic groups).
Four different structured interview schedules were created for the different types
of organizations: one for community-based organizations, one for schools, one for health
service providers, and one for media organizations. Once the final sample was
constructed, potential study participants were contacted via phone. After five
unsuccessful attempts to make contact via the phone (and in some cases through e-mail),
9

we planned a site visit hoping that we would find someone to talk to on location and
attempt to recruit them. After two unsuccessful site visits and 10 unsuccessful attempts to
make contact via phone, an organization was dropped from the sample and we attempted
to replace it, again through random selection, with another organization of the same type,
in the same zip code. However, replacement was not possible in every case; usually
because only a small number of organizations of a certain kind, were available in a
particular zip code to begin with. In addition, six organizations from the original sample

9
‘No contact’ means that every telephone call went unanswered or forwarded to voicemail. It does not
account for the many cases in which we had to call back a number of times and speak with several
organization staff members and gatekeepers before talking to someone who we could interview.

104

had either seized to exist or had moved out of the area since the HealthyCity.org and
Rainbow Resource Directory were last updated.
Between November 2007 and May 2008, we successfully interviewed 80
organizations. Of those, 44 were community-based organizations, 25 were schools, 8
were health service providers, and 3 were media (i.e., an African American radio station,
a Spanish-language television network, and a local newspaper whose audience tends to
be African American). The duration of the interviews was between 25 minutes and 90
minutes. The interviews were recorded so that interviewers could double-check the
accuracy of the answers marked on the coding sheets after the interview and in order to
minimize the occurrence of missing data.
10

Figure 4.2. Organizations interviewed in Greater Crenshaw by type (N=80)

Figure 4.1. 44 of the organizations interviewed (55%) were community-based organizations of
various types, 25 were schools, 8 were health service providers, and 3 were media.

10
All interviews except 3 were recorded. Two were not recorded because of technical difficulties and the
third because the respondent requested that their responses were not audio-taped.

105

b. Measures (Table 4.4 summarizes measures created using the organizational data)
Capturing the integration of organizations into the neighborhood storytelling network
New measures were created as part of this study to capture: (a) the degree of STN
integration at the meso-level; (b) the extent to which individual organizations are
integrated into the neighborhood storytelling network (STN); (c) the interaction of
individual residents’ integration into the STN with the integration of organizations in
those residents’ neighborhood; and, finally, (d) community communication capital
(CCC). The work done in organizational science and social network analysis has
informed the thinking and development of these measures (e.g.: Marsden & Campbell,
1984; Marsden, 1990; Monge & Contractor, 2003; Reagans & McEvily, 2003;
Wasserman & Faust, 1994). However, and even though the data collected might lend
themselves to social network analysis in the future, the analytical focus of the work
presented here is not on understanding network structures per se, but rather providing a
better understanding of what difference it makes in a community with particular
structural characteristics if both individual residents and meso-level organizational actors
engage each other in the communicative processes of everyday life.
I. Intensity of Inter-Organizational Connectedness (IIOC)
IIOC is meant to capture the average intensity of the relationships between an
organization participating in the study and the organizations it communicates with on a
regular basis. Intensity is conceptualized as having four different dimensions: (a)
frequency of contact between two organizations, (b) dependency of one organization on

106

another for completing everyday operations, (c) the number of projects/collaborations
between two organizations, (d) and the duration of their relationship.
a. Frequency:
Every organization was asked to give the interviewer the names of all other
organizations with which it was in contact regularly, starting with the one it was in
contact with the most. Subsequently, the interviewee was asked for every one of the
organizations mentioned: “How often are you in contact with this organization?”
Answers to this question were scaled from 1-10. These scores were averaged across the
organizations the interviewee mentioned. The average reflects the average frequency of
contact between the organization that participated in the study and organizations it is in
communication with the most.
b. Dependency:
Participants were also asked: “How important is this organization to you in doing
your work on a day-to-day basis?” This is a 5-point Likert type scale ranging from 1,
meaning ‘not important at all,’ to 5, meaning ‘extremely important.’ Once again
responses for all the organizations the participant mentioned were averaged.
c. Number of Collaborations:
Organizations were also asked about the number of projects they had worked on (or
were currently engaged in) with those organizations they are in contact with most
frequently. Responses were averaged to calculate a single score for the participating
organization.

107

d. Duration of relationship:
Finally, participants were asked about the duration of the relationship between their
organization and the organizations they are in contact with most frequently. We asked
them: “how long ago did you first contact this organization?”
The final measure of the average intensity of inter-organizational connectedness
(IIOC) for every organization interviewed was conceptualized as the interaction term
among these four different dimensions of inter-organizational relationships.
Organizations that contact others more frequently, depend on others more for achieving
their goals, work with other organizations on more projects and initiatives, and have been
in touch with others for a longer period of time are thought of as being more enmeshed,
integrated, or connected into a community network of organizations (meso-level actors).
Scores for the four dimensions were first standardized. Then, the standardized z scores
were re-coded into eight ranked categories so that all the values became positive
numbers.
This is the formula for the calculation of IIOC:
IIOC = Frequency x Dependency x Collaboration x Duration (3)

IIOC for the organizations interviewed ranged from 8 to 29, with a mean of
M=18.10 and a standard deviation of SD=3.78 (skewness of the distribution= -.01,
kurtosis=.55).
II. Scope of Inter-Organizational Connectedness (SIOC)
Participants were also asked to describe the mission of the organizations they
mentioned as ones they are in communication with most frequently. Based on that
information we classified every organization they made reference to into a category (see

108

Appendix B for a list of all the possible categories) and then counted the number of
different categories of organizations the participants reported as being in communication
with.
III. Meso-level Storytelling Network Connectedness (MSNC)
Next, I wanted to get a sense of how connected meso-level neighborhood
storytellers are to other organizations. I hypothesized that a neighborhood in which meso-
level actors score high on the intensity of connectedness measure (i.e., IIOC) and that
have a wide-range of relationships, as captured by the scope of connectedness measure
(i.e., SIOC) can create a stronger, more supportive web of resources on which residents
can rely on to achieve goals, including dealing with health concerns. Therefore I created a
measure of meso-level storytelling network connectedness (MSNC), which was
conceptualized as the interaction term between intensity and scope of inter-organizational
connectedness for every participant organization. Formula (4) was used for the
calculation. For the Greater Crenshaw organizations interviewed, MSNC ranged from 14
to 118, it had a mean of M=52.30 and a standard deviation of SD=22.75 (skewness=.56,
kurtosis=-.22)
MSNI = IIOC x SIOC (4)

IV. Reach or Cross-Storytelling Network Integration (CSNI)
In creating the measure of individual residents’ integration into the storytelling
network (i.e., ICSN and ICSN-2), measures of individuals’ connections to neighbors
were combined with measures of their connectedness to meso-level actors (i.e., media
and community organizations). Therefore it was considered appropriate to create a
measure that would capture the participant organization’s connectedness across levels as

109

well; that is, to create a measure of the degree to which organizations connect to residents
in the neighborhoods under study. We call this measure reach or cross-level storytelling
network integration (CSNI). Two indicators of reach were created based on two different
sets of items.
a. Geo-Demographic Reach (CSNI
GD
)
In the third part of the survey we asked the organizations a series of questions
meant to help us understand what their audience or constituency looked like. We asked
them to specify how many different age brackets their members/clients fall under,
whether they target both men and women, how many different ethnic groups are
represented in their audience, and about the variability among their members/clientele in
terms of income.
In addition we asked them where their membership or audience is located
geographically. This is the question and the response options they were given to select
from: “Would you say that the people you serve or your members are located: (a) Across
the nation? (b) In California? (c) In Southern California? (d) In Los Angeles County? (e)
In the City of Los Angeles? (f) Particular areas or neighborhoods of Los Angeles? (g)
Other? Please specify: ___.” Each organization received a score ranging from 1 to 6
depending on their response. A score of 6 indicated that the organization focused its
efforts on more local communities and specific neighborhoods across Los Angeles, while
a score of 1 indicated that the organization’s audience/member base was spread out
across the U.S.
Within the Greater Crenshaw study area there are 11 different real estate
designated areas. The names of some of these has a historical basis (e.g., West Adams),

110

some though do not (e.g., Vermont Square). Apart from where the audience of
organizations is located generally speaking, we also asked participants how many of
these real estate-designated areas that are part of Greater Crenshaw they recognized as
areas in which they were active. They received a score equal to the number of these
smaller areas within Greater Crenshaw that they checked. The areas included were:
Arlington Heights, Baldwin Hills/Baldwin Vista, Baldwin Village, Chesterfield Square,
Crenshaw, Jefferson Park, Ladera Heights, Leimert Park, Vermont Square, View
Park/Windsor Hills, and (Historic) West Adams (Range of scores: 0-11).
From a communication infrastructure point of view, neighborhoods in which
organizations and residents focus their effort on bettering their local community are
better off. Therefore, conceptually, organizations that reach the broadest possible
demographic base in a local community should receive higher reach scores. Geo-
demographic reach or CSNI
GD
was calculated as the interaction term among the
following items capturing reach: age, gender, income, ethnicity, geographic reach, and
reach in the local area (see Formula 5). The product of these terms was log-transformed
to correct for skewness (see Wuench, 2005 regarding acceptable transformations of
continuous variables to correct for skewness and kurtosis). CSNI
GD
for organizations in
Crenshaw ranges from 1.30 to 4.15. The mean score was M=2.52 with a standard
deviation of SD=.55 (skewness=.54, kurtosis=.58).
CSNI
GD
= Age x Gender x Income x Ethnicity x Geog. Reach x Local Reach (5)

b. Reach as the degree of engaging the local community (CSNI
CE
)
In the fourth part of the interview, we asked organizations a series of questions to
gauge the extent to which they are active in their local community and the extent to

111

which they create opportunities to communicate and work with residents. One set of
items asked whether the organization receives correspondence from residents through
regular mail, whether it has a telephone line for residents to call in, a publicly available e-
mail address, a Web page, and a listserv. Another set of items tried to capture if the
participating organizations sponsor events in the local community, open forums, and
whether they have local community members amongst their leadership. A synthetic
variable was created by adding up organizations’ ‘yes’ and ‘no’ answers to these
questions (‘yes’=1, ‘no’=0). The range for this reach indicator (referenced also as CSNI
CE

where
CE
means ‘community engagement’ was 1-7.
V. Organizational Integration into the Storytelling Network (OISN)
Once again, referring back to the conceptualization and creation of the measure of
an individual residents’ integration into the neighborhood storytelling network [see
Formulae (1) and (2)], we have tried to account for the interaction of the intensity of
connectedness among residents and the extent to which they connect to meso-level actors
in their community. Following a similar logic, we created a measure intended to capture
the degree of an organization’s integration into the neighborhood storytelling network as
the interaction term between meso-level storytelling network connectedness (MSNC) and
reach (CSNI). As there are two indicators of reach (distinct conceptually and from an
operationalization point of view), there are also two versions of organizational integration
into the storytelling network (OISN). The two formulae via which these indicators are
calculated can be found below. For OISN
1
we used geo-demographic reach and for
OISN
2
we used the reach as community engagement measure. For OISN
1
the mean for
the participant organizations was M=12.30 with a standard deviation of SD=8.90

112

(range=1-36, skewness=.74, and kurtosis=-.49), while for OISN
2
the mean for the
participant organizations was M=12.54 with a standard deviation of SD=9.13 (range=1-
30, skewness=.84, and kurtosis=-.88).
OISN
1
= MSNC x CSNI
GD
(6)

OISN
2
= MSNC x CSNI
CE
(7)

VI. Intensity of Health Storytelling (IHS)
I also attempted to create a measure that would capture the extent to which
organizations recruited focused their efforts on addressing health concerns of their
members, clients, and of the communities in which they are based. Unfortunately,
however, this effort was not very successful, in part because the number of media and
health service providers we did succeed in recruiting was fairly limited. Notwithstanding
the limitations of the data collected from media and health service providers with respect
to IHS, the items intended to capture the intensity of health storytelling by community-
based organizations (CBOs) and schools specifically show some promise. We asked
CBOs and schools to tell us how frequently they had sent out health-related information
to their constituents (members or clients) over the course of the past three months, how
frequently they participated in health-oriented events in their local neighborhood (e.g.,
health fairs), how often they became involved in projects or initiatives focused on the
health and the well-being of residents in their local area, and if brochures with health
information were available at their offices for visitors and members to pick up and read.
As more work is required, however, to determine if the IHS measure can be refined and
used as a means to qualify the significance of organizations’ connectedness into the

113

neighborhood storytelling network for health purposes, the measure was dropped from
the analyses conducted for this study.
Two more measures were created based on the individual-level data and the
organizational level storytelling network-related data.
VII. Individual-level Communication Capital (ICC)
In an integrated neighborhood storytelling network, meso-level actors will impact
the capacity of residents as neighborhood storytellers; they will either enable residents or
hamper their efforts. Theoretically, in a neighborhood where organizations are highly
integrated at the meso-level and that score high on reach (on both measures, geo-
demographic reach and reach as community engagement), the residents’ storytelling
capacity should be amplified. That is to say, that a highly integrated meso-level can
contribute to the efforts of residents and strengthen the storytelling network, or may
compensate for the disconnectedness observed at the micro-level. Therefore I tried to
account for the impact of organizations in a community on every individual’s level of
integration into the neighborhood storytelling network by creating the ICC measure,
which is calculated as the square root (to correct for skewness; see, e.g., Wuensch, 2005)
product of the mean OISN for the organizations in a particular resident’s neighborhood
and that resident’s ICSN-2 score. Formula 8 reflects this calculation.
11
The ICC measure
reflects a resident’s communication capital as accounting the influence of meso-level
actors in his or her community. For the residents in Greater Crenshaw overall, the mean

11
The Census tract is too small to realistically be considered as the area of reach for most organizations.
Many community-based organizations, for instance, may define their area of activity based on a
combination of zip codes but never as one census tract. Therefore, for the purpose of ‘assigning’
organizations to Census tracts and thereby to residents, we assumed that organizations in a particular zip
code of Greater Crenshaw would reach residents in all the census tracts that fall within that zip code.

114

ICC was M=25.17 with a standard deviation of SD=7.62 (range= 8.04-41.70,
skewness=.14, and kurtosis=-.69).

ICC

= OISN x ICSN (8)

VIII. Community Communication Capital (CCC)
Following the same rational guiding the creation of ICC, a measure was created to
reflect the degree of storytelling network integration across a neighborhood. CCC is
calculated as the square root of the interaction term between the average OISN in a
particular neighborhood and the average ICSN-2 in the same area. This is considered an
index of the entire community’s communication capital. Formula 9 indicates how CCC is
calculated.

CCC

= OISN x ICSN (9)

Table 4.4 presents a summary of all the storytelling network-related measures
discussed.
The neighborhood communication action context: data and related measures
I. Density of institutional resources
The study area of Greater Crenshaw consists of 8 zip codes. These zip codes
intersect with a total of 87 Census tracts. For the purposes of this study, the Census tract
was defined as the neighborhood unit of analysis. Density of institutional resources in a
neighborhood was defined as the number of organizations available in a census tract per
capita. For the Census tracts of Greater Crenshaw, mean density was M=1.31 with a
standard deviation of SD=2.30 (range=0-16.88).

115

Table 4.4. Storytelling network-related variables created and used in the present study

Storytelling Network Variables Created
from Individual and Meso‐Level Data

Variables Index reference
Individual‐level Variables (Residents)
1 Integrated Connection into the Storytelling
Network
ICSN‐2
Meso‐level Variables (Organizations)
2 Intensity of Inter‐Organizational
Connectedness
IIOC
3 Scope of Inter‐Organizational
Connectedness
SIOC
4 Meso‐level Storytelling Network
Connectedness
MSNC
5 Geo‐Demographic Reach CSNI
GD

6 Reach as Community Engagement CSNI
CE

7 Organizational Integration into the
Storytelling Network
OISN
Cross‐level Variables (Interaction)
8 Individual‐level Communication Capital ICC
9 Community Communication Capital CCC

To determine the number of organizational resources in the Census tracts, we
started with the full inventory of organizations available in the area compiled as
described earlier in the study (N=333). The inventory contained detailed information
regarding the organizations’ location. Their addresses were input into ArcGIS 9.2 (a
spatial data analysis software package by developer ESRI) and geo-coded. Geo-coding is
a process by which physical addresses are placed on a map in real-world coordinates and
represented as points. Using spatial data from the U.S. Census on the boundaries of Los
Angeles County Census tracts in ArcGIS 9.2 we were able to determine in what Census
tracts the organizations are located in. After that the measurement of density was

116

straightforward (see Formula 10). See Figure 4.1, presented earlier, for an illustration of
where all 333 organizations are located in Greater Crenshaw.
Density per Capita = No. Orgs in Census Tract / C.T. Population x 1000

(10)

II. Population composition-related variables
a. Ethnic Heterogeneity
Ethnic heterogeneity is calculated in each Census tract of the study area by using
Formula 11 (Alesina & La Ferrara, 2000), where S
i
is the proportion of ethnic group i in
an area. If we use this formula, the community of a single ethnic group will score 0. As
more and more ethnic groups share a local area evenly in terms of the percentage of
population, the ethnic heterogeneity score of the area will get closer to 1. In the current
study, we considered Whites, Latinos, African Americans, Asians, and ‘others’ to
calculate ethnic heterogeneity scores for each Census tract in our study area (range 0.00–
1.00, M=.49, SD =.15).

Ethnic Heterogeneity = 1 - S
i
2
(11)

b. Residential Stability
We calculated the percentage of individuals for whom ‘yes’ was an appropriate
response to the question in the 2000 U.S. Census: “Did this person live in the same house
five years ago?” Each Census tract in our study area is assigned a percentage score as its
residential stability value (range 1%–100%, M=52.50%, SD= 0.72).
Σ
i

117

Analysis
I. Preliminary analyses
a. Overlaying spatial data
We asked our telephone respondents to identify the closest major intersections to
their residences. Five hundred and twenty-two out of 607 (86%) provided valid
intersection names.
12
ArcGIS 9.2 was used once again to geo-code these addresses and
place the residents on the map. Doing so allowed us to join the residents’ data with the
data collected from the organizations, as well as Census data (necessary to calculate
ethnic heterogeneity and residential stability). As a result, we had neighborhood-level
data for a total of 63 census tracts. As the Greater Crenshaw area consists of 87 Census
tracts, we had complete information for 72.4% of the area.
Before proceeding with the main analyses two more procedures took place.
b. Hierarchical Linear Modeling (HLM)
When researchers have multi-level data with a nested structure such as students
(level-1 data) within schools (level-2 data) or residents (level-1 data) within
neighborhoods (level-2 data), it may be inappropriate to look at only one level–either the
resident-level or neighborhood-level–to examine factors in resident-level performance
variables. It will be difficult for researchers to demonstrate the portion of variance that is
accounted for only by neighborhood-level variables if they consider only individual-level
characteristics (e.g., SES or media connection patterns, etc.) in their analysis. On the
other hand, if researchers use only neighborhood-level variables (e.g., ethnic diversity or

12
In some instances residents provided not an intersection but rather two parallel streets between which
they lived, which made it impossible to pinpoint their location on the map; in other cases some people
provided an incorrect zip code. While we were able to find the correct zip code for many of these
addresses, some ended up being zip codes that were not part of the Greater Crenshaw study area, and
therefore were excluded. Few other survey respondents declined to provide any address information.

118

population density) to explain residents’ attitudes, perceptions or behaviors, an ecological
fallacy or a phenomenon known as within-area homogeneity arises (Holt, Steel, Tranmer,
& Wrigley, 1996). This ecological fallacy problem is related to the problems of “the non-
random composition of the groups” and “the correlations between individuals within the
same group” (Steel, Tranmer, & Holt, 1997) that make it difficult to use the ordinary least
squares (OLS) estimation method in a regression analysis.
The hierarchical linear modeling (HLM) technique was introduced by
Raudenbush & Bryk (1992, 2002) to social research to solve this ecological fallacy
problem related to analyses that involve nested data structures. HLM addresses this
problem by considering both level-1 (e.g., individual-level) variance and level-2 (e.g.,
neighborhoods-level) variance in a single model.
However, HLM requires fairly large samples at both the individual level (level-1)
and the neighborhood level (level-2) (e.g., Hox, 1998). Having geo-coded the addresses
provided by the residents, we discovered that only 48 out of the 63 Census tracts, in
which there were residents with valid addresses lived had more than 3 respondents in
them. Only 32 Census tracts had more than 5 residents in them. These relatively small
numbers of individuals within Census tracts made it impossible to benefit from HLM’s
strengths. Hence the HLM method had to be abandoned as an analytical tool for the
investigation of several of the hypotheses and research questions posed in Chapter 3 and
alternatives were explored (see below).
In doing so, however, it was important to determine first, if in fact there were
significant patterns of spatial dependence among residents’ responses. That was because
we wanted to ensure that we controlled for the possibility that individuals’ scores were

119

spatially determined, and not determined by the predictors of interest. These
considerations led to one more preliminary set of analyses, which were conducted
through ArcGIS 9.2.
c. Spatial Dependence (Autocorrelation)
Applying a spatial statistics tool that is part of ArcGIS 9.2, we conducted a series
of tests to show whether or not there was a pattern of clustering in the civic engagement
and health-related outcome scores of residents’ living across Greater Crenshaw. The
software produces a statistic known as Moran’s I index of spatial autocorrelation, which
ranges from -1 to 1. It is read in a fashion similar to how we interpret correlations. A
score of -1 signifies that scores are dispersed, while a positive score of 1 indicates that
there is a significant pattern of clustering in the data; a score of 0 indicates that there is no
spatial dependence. Table 4.5 shows the results of the analyses. There was no significant
pattern of clustering detected in the data for both the civic engagement variables and the
health-related outcomes.
Table 4.5. Moran’s I test of spatial autocorrelation for variables used as outcomes
Test for Spatial Autocorrelation Patterns in Greater Crenshaw
Variables Moran's I score Z‐Score Sig. Level Pattern
Civic Engagement
Neighborhood Belonging 0.000 0.65 p>.10 Random
Collective Efficacy 0.002 1.45 p>.10 Random
Civic Participation 0.001 1.92 p<.05 Random
Health Literacy
Prevention 0.000 ‐0.03 p>.10 Random
Detection 0.000 1.59 p>.10 Random
Health Care Access
Difficulty Getting Health Care 0.000 1.74 p>.05 Random
Distance to Health Care 0.000 0.30 p>.10 Random

Table 4.5. The z-score indicates whether we can reject the null hypothesis that there is no spatial
clustering detected in the data.

120

II. Analysis plan by hypothesis
As it became evident that it would be impossible to use Hierarchical Linear
Modeling for several of the analyses necessary, an alternative analytical plan was
developed. For some hypotheses similar analytical procedures were employed, therefore I
present the analytical strategy grouping together hypotheses (a) thematically and,
whenever possible, (b) by the type of analyses that were conducted.
a. Storytelling network-focused hypotheses and research questions
(i) Direct Effects
Hypotheses 1, 2-1, 2-2, 2-3, as well as Research Question 1 focus on
understanding the direct effects of the storytelling network on health literacy (H1), health
care access (RQ1), and civic engagement (H2-1: neighborhood belonging, H2-2:
collective efficacy, and H2-3: civic participation). As it was not possible to look at the
interaction between level-2 variables and level-1 variables, instead of using the measure
of community communication capital, we used both the individual level variable of an
integrated connection to a storytelling network (ICSN-2) and the new measure of
individual-level communication capital (ICC), which captures the interaction of
individual-level integration into the neighborhood storytelling network and the
integration of meso-level neighborhood actors into the STN.
13
The analyses run to
investigate the foregoing hypotheses and research question unfolded in four steps:

13
We attempted to run statistical analyses at level-2 only as well, where we could use the measure of
community communication capital as described earlier in this chapter as the independent variable of
interest, but the relatively small number of Census tracts that could be used for such analyses (N=48)
compromised the statistical power of the tests employed.

121

Step 1
Two series of hierarchical multiple regressions were run for every one of the
outcomes of interest. Progressively, we included in the models (and controlled for)
socioeconomic status, ethnic background, residential tenure, home ownership status, and
neighborhood storytelling variables. The first series of regressions was run using ICSN-2
in the final model; ICSN-2 captures the degree of a resident’s integration into the
storytelling network. In the second series of regressions, the exact same models were run,
but this time the final model included not ICSN-2 but ICC instead.
We compared the performance of the final models produced by the hierarchical
multiple regression procedures, using the adjusted R
2
and R
2
change scores as gauges. The
first question we asked was whether storytelling improved the regression model’s fit to
the data. The second question was whether residents’ integration (ICSN-2), or the
interaction of residents’ integration into the storytelling network with the degree of
integration of meso-level actors (ICC) improved the fit of the regression model the most.
In those cases where in the second series of regressions ICC was found to produce a
significant improvement, that improvement (R
2
change) was compared to the R
2
change
produced by ICSN-2 in the final model of the first sequence of regressions to evaluate if
ICC performed better. The adjusted R
2
scores were used to gauge if the final model with
ICC explained more of the variance in the dependent variable than the final model with
ICSN-2.
Step 2
Subsequently, we examined whether the findings from Step 1 held up when the
models where run separately for African American residents and Latinos. Doing so

122

offered additional evidence that could help answer the questions of whether the
storytelling network in Greater Crenshaw is fragmented along the lines of ethnicity or
not, and what such fragmentation could mean with respect to civic engagement, health
literacy, and health access in the community.
Step 3
The measures pertaining to the organizations could not be used in multi-level
models as neighborhood-level predictors. Instead neighborhoods (i.e., Census tracts)
were assigned a score of ‘0=Low’ or ‘1=High’ depending on how the organizations that
were located within them scored, on average, on the measures developed based on the
organizational data. Essentially, the following variables created for the organizations
were dummy-coded:
Intensity of Inter-Organizational Connectedness IIOC
Scope of Inter-Organizational Connectedness SIOC
Meso-level Storytelling Network Connectedness MSNC
Reach or Cross-level Storytelling Network Integration CSNI
Organizational Integration into the Storytelling Network OISN
The idea was to find a way to test whether it made a difference if a resident lived
in a neighborhood where organizations varied across these dimensions or not, with
respect to health literacy, health care access, and civic engagement. To examine if this
was indeed the case, we ran one-way ANOVAs.
Step 4
Finally, for those cases in which the ANOVAs conducted in Step 3 indicated that
an organizational attribute made a difference, a hierarchical multiple regression was

123

conducted as a follow up procedure. The purpose was to determine if the impact of the
organizations, to which the ANOVAs pointed, persisted after controlling for individual
level variables, including residents’ integration into the storytelling network (ICSN-2).
The organizational attribute of interest was added to the final model of the hierarchical
regression as a dummy variable.
(ii) Testing for Direct and Indirect Effects via Structural Equation Modeling
We used structural equation modeling (SEM) (Bollen, 1989; Jöreskog & Sörbom,
1989) to test the theoretical model presented in Chapter 3. The model suggests that the
degree of storytelling network integration will have direct effects with respect to civic
engagement, health literacy, and health care access, but also indirect impact on the
health-related variables through civic engagement (H3a-1 through H3a-3 and H3b-1
through H3b-3). SEM represents a fairly innovative procedure for evaluating the
consistency of theoretical models with empirical data. Complex theoretical models can be
statistically represented and evaluated based on how well these models reproduce the
observed pattern of empirical relationships in the data. The analysis via SEM unfolded in
two stages.
Stage 1: Evaluating the theoretical model
The hypotheses represented in the theoretical path model were tested using
maximum likelihood procedures in LISREL 8.80 (Bollen, 1979; Jöreskog & Sörbom,
1979). The fit of the hypothesized relationships compared to the observed correlations
were assessed through examining the following: first, fit was assessed by examining the
significance of the chi-square statistic; secondly, goodness-of-fit (GFI and AGFI) indices
were examined; and third, the root squared mean error of approximation (RMSEA).

124

The significance of individual paths was assessed using t ratios.
(ii) Revised model
After evaluating the theoretical model, modification indices were examined to see
whether improvements could be made to the model. Following the strategy described by
Schmitz and Fulk (1991; see also Hayduk, 1987), all non-significant paths were deleted
first.
b. Communication action context-centered hypotheses and research questions
A second set of hypotheses and research questions posed earlier in this study (i.e.,
H4 through H6 and RQ2 through RQ6) focuses on the examination of communication
action context (or neighborhood environment) effects on (a) the integration of the
neighborhood storytelling network, (b) residents’ civic engagement, as well as (c) health
literacy and health care access. Again, a multi-pronged approach had to be deployed to
investigate these questions, as an alternative strategy to multi-level modeling. The
following steps were carried out:
Step 1
The variables of density of institutional resources in a particular Census tract,
ethnic heterogeneity, and residential stability were dummy-coded (i.e., 0=Low and
1=High) and assigned to every Census tract in which respondents to the Metamorphosis
survey lived. Subsequently, hierarchical multiple regressions were run, in which the
dummy-coded variables of interest (depending on the hypothesis) were entered in the
final model, to evaluate whether they accounted for a significant portion of the variability
in residents’ scores, over and above socio-economic status, neighborhood experience,
ethnicity, and residents’ integrated connection to a storytelling network (ICSN-2). The

125

hierarchical multiple regression allowed us to answer that portion of the original
hypotheses that addressed the effects of the communication action context controlling for
individual-level covariates.
Step 2
To control for neighborhood-level covariates (i.e., to answer the second part of
the original hypotheses), we conducted three-way ANOVAs. To examine, for instance,
the effect of density of organizational resources on health literacy, controlling for ethnic
heterogeneity and residential stability, all three dummy-coded variables were entered as
factors into the ANOVA and health literacy was defined as the dependent variable. The
three-way ANOVA does not allow us to examine the effect of density of resources and
‘control’ for the effects of ethnic heterogeneity and residential stability in the exact same
sense that a regression analysis does, but it does allow us to examine whether the
interaction of density of resources with one or with both of the other communication
action context variables has any effect on health literacy. Doing so we accomplish the
same goal; we are able to gauge the impact of density of resources net of and in
combination with other neighborhood-level variables.
14

Next, in Chapter 5, I present the results of the analyses conducted.

14
For hypotheses that focus on the effects of the interaction between the communication action context and
civic engagement or the storytelling network on health literacy and health access, a simpler version of this
strategy was employed, as only two-way ANOVAs were necessary. The details of these analyses will be
discussed at greater length and as needed in Chapter 5.
126

CHAPTER 5
RESULTS
Three larger questions underlie the hypotheses driving this project. The first has
to do with the relationship between the degree of integration of a neighborhood’s
storytelling network (STN) across levels of analysis (micro and meso), and the level of
residents’ civic engagement. The second bigger issue is whether or not the integrated
multi-level STN acts as a significant mechanism of neighborhood health effects; namely
health literacy and access to health care. In addition, if the STN is in fact an important
mechanism of neighborhood health effects, how does it interact with more commonly
studied social processes of neighborhood effects, including neighborhood belonging,
collective efficacy, and civic participation? Finally, the third larger question this research
aims to address is whether the communication action context (or the neighborhood
environment) residents live in makes a difference with respect to their level of civic
engagement, their health literacy and their access to health care resources. To the extent
that it does, are the neighborhood environment’s effects amplified or diminished by the
strength of a residential community’s multi-level storytelling network? A common thread
that ties these three questions (and the more specific hypotheses investigated) together is
a preoccupation with understanding the particular role organizations play as
neighborhood meso-level storytellers (or STN actors) in building civic engagement and
health literacy, and enabling access to health care resources.
After providing basic descriptive statistics to gauge how Greater Crenshaw
residents fare and differ along the lines of the independent variables and outcome
127

variables, I present findings from the analyses conducted for every hypothesis and
research question. Hypotheses are grouped thematically.
Descriptive Statistics
Independent sample t-tests were conducted to test for differences among African
Americans and Latinos in Greater Crenshaw with respect to their integration into the
indigenous storytelling network, their civic engagement, health literacy, and health care
access. African Americans are, on average, more integrated into their neighborhood
storytelling network (this is reflected in the ICSN-2 measure
1
) than Latino residents,
t(605)=3.49, p<.001. This is also true when accounting for the interaction of meso-level
actors’ degree of connectedness to the storytelling network (OISN) and individual
residents’ ICSN-2 score (the interaction is captured by the ICC variable). For ICC, the
difference between the two populations is significant, as t(602)=7.59, p<.001.
African Americans and Latinos also differ significantly along all three dimensions
of civic engagement. Once again, African Americans score higher. For belonging the t
value was t(605)=7.59, p<.001, for collective efficacy t(605)=7.59, p<.05, and for civic
participation t(600)=7.59, p<.001.
With respect to health literacy the residents of Greater Crenshaw do not have
significant differences. All three t-tests for health literacy - i.e., for the prevention
dimension and the symptoms detection dimension measures – were non-significant. With
regard to health care access, African Americans reported that it is significantly more
difficult for them to find health care compared to Latinos, t(605)=-2.86, p<.01. However
there seems to be no difference among Greater Crenshaw residents with respect to how

1
Details on the construction of the ICSN-2 measure, please see Chapter 4, p. 4-5.
128

far they have to travel to get health care, t(605)=-29, p>.10. Table 5.1 summarizes the
results of the descriptive statistics and the tests conducted to assess differences between
Latinos and African Americans.

The role of the storytelling network in civic engagement and health
Due to the methodological constraints discussed in Chapter 4, some of the
hypotheses and research questions had to be reformulated. The revised hypotheses are, in
essence, asking the same questions that the original hypotheses were (see Chapter 3), but
are more congruent with the alternative analytical plan we deployed (see details in
Chapter 4).
Table 5.1. Independent and outcome variable descriptive statistics for Greater Crenshaw
129

Hypotheses 1, 2-1 through 2-3, and Research Question 1 pertain to the
relationship between the degree of storytelling network (STN) integration and civic
engagement (H2-1, 2-2 and 2-3) and the relationship between STN integration and health
literacy (H1) and health care access (RQ1). Because the main independent variable in
these hypotheses, STN integration, is the same, and because the procedures used to
investigate them, as described in Chapter 4, are similar, I examine them together, starting
with Hypothesis 1. It stated:

Hypothesis 1
H1a: An individual resident’s integrated connection into the storytelling network (i.e.,
ICSN-2) will be positively related to health literacy after controlling for individual level
covariates, including socio-economic status, residential tenure, homeownership status,
and ethnicity.
H1b: Individual-level communication capital (ICC) will be positively related to health
literacy and will explain more of the variance among residents’ scores than ICSN-2.

The goal set by the hypothesis was to determine if an integrated neighborhood
storytelling network (STN) is in fact a social mechanism that can account for the level of
a resident’s health literacy, above and beyond socio-economic status and neighborhood
experience (operationalized as residential tenure and home ownership status).
Theoretically, the STN is stronger when both residents and meso-level organizational
actors are highly integrated into the network. To be able to account for the degree of
130

integration of actors at both levels, we relied on the ICSN-2 measure, which reflects
individual residents’ integration into the STN, as well as:
ICC, which captures the interaction of residents’ integration into the STN with that of
the meso-level actors that are active in residents’ neighborhoods.
Measures created from the data produced through the interviews with organizations in
the Greater Crenshaw study area: (a) intensity of organizational inter-connectedness
(IIOC), (b) scope of organizational inter-connectedness (SIOC), (c) meso-level
storytelling network connectedness (MSNC), (d) geo-demographic reach (CSNI
GD
)
and reach as community engagement (CSNI
CE
), as well as organizational integration
into the STN (OISN).
This is how the analysis unfolded step-by-step:
STEP 1

First, a hierarchical multiple regression was conducted to investigate whether
ICSN-2 predicted health literacy, controlling for socio-economic factors, residents’
neighborhood experience, and ethnicity. Four models were produced by this analysis.
Then, the same sequence of models were run again as part of a second hierarchical
multiple regression, this time replacing ICSN-2 in the final step (Model 4) with ICC. This
alternative final model (see alternative Model 4 in Table 5.2 that follows) was compared
to the model which included ICSN-2 to determine whether there was an indication of
improvement, based on the R
2
, R
2
change, and F scores produced. Table 5.2 shows the
results of the regression analyses run, reports the coefficients, and notes those coefficients
that were found to be significant in every model.
131

Both ICSN-2 and ICC were positively related to prevention-oriented health
literacy. Adding ICSN-2 to Model 4 and ICC to the alternative Model 4 explained more
of the variance in health literacy than socio-economic status variables and neighborhood
experience did. In addition, introducing ICC to the regression model explained more of
the variability in health literacy than ICSN-2. The R
2
change produced by ICC was .014,
F(1, 558)=8.53, p<.01. That was double the R
2
change created by the addition of ICSN-2
to the model. The R
2
change for ICSN-2 was .007, F(1, 558)=4.37, p<.05. Overall, the
adjusted R
2
for the model including ICSN-2 was .05, while for the model with ICC the
adjusted R
2
was .06.
Table 5.2. Hierarchical multiple regression: STN integration
& health literacy (prevention)

132

The final model (see Alternative Model 4 with ICC in Table 5.2) suggests that
gender, education, and ethnicity are significant predictors of prevention-oriented health
literacy. Females have higher health literacy scores than males do, on average. Also,
residents with higher education have higher health literacy scores, as do Latinos living in
Greater Crenshaw when compared to African Americans.
STEP 2

In Step 2 we ran the same series of models for African Americans and Latinos
separately to see if the STN plays the same role in both populations. Tables 5.3a and 5.3b
summarize the results of these hierarchical multiple regressions.
a. African Americans
When looking at African Americans separately, individual residents’ integration
into the STN (i.e., ICSN-2) does not predict prevention health literacy. Model 1 (see
Table 5.3a), containing only socio-demographic characteristics is the best fitting model
until we add the interaction of meso and micro-level neighborhood actors (i.e., ICC).
Contrary to ICSN-2, ICC does improve the regression model, suggesting that
organizations in Greater Crenshaw are helping strengthen the neighborhood storytelling
network as a whole and playing a positive role in building health literacy. For the model
including ICC the adjusted R
2
is equal to .08, the R
2
change=.02, and F(1, 269)=6.24,
p<.01. Among African Americans, women have higher health literacy (for preventing
hypertension), while a higher level of education is also associated with a higher degree of
health literacy.
133

Table 5.3a. STN & prevention health literacy for Greater
Crenshaw African Americans

a. Latinos
Among Latinos in Greater Crenshaw (see Table 5.3b) the neighborhood
storytelling network does not predict health literacy (for preventing diabetes). ICSN-2
and ICC both do not predict health literacy. It appears that meso-level actors are unable to
compensate for the weakness of the network at the individual level. The only significant
factor in predicting health literacy among Latinos for preventing diabetes is age.
134

Table 5.3b. STN & prevention health literacy for Greater Crenshaw Latinos

STEP 3

Analyses from Steps 1 and 2 indicate that meso-level actors’ integration into the
STN may be important in predicting prevention-oriented health literacy. However it is
unclear what it is about organizations’ STN integration that makes a difference. To
investigate this question further, we turn to the measures created from the data collected
from organizations. The IIOC, SIOC, MSNC, Reach (or CSNI), and OISN measures
described earlier were dummy-coded (0=Low, 1=High). Every individual resident was
135

assigned a score of 0 or 1 depending on how well organizations in their neighborhood
scored on the aforementioned measures. Subsequently 1-way ANOVAs were conducted
to examine whether it made a difference with respect to health literacy if a resident lived
in a neighborhood where meso-level actors had, for example, a broad (=1, high) scope of
inter-organizational connectedness (SIOC) or a narrow one (=0, low). One-way
ANOVAs were run for every one of the dummy-coded organization-related measures. As
similar tests were conducted for Hypothesis 1, Research Question 1, and Hypotheses 2-1,
2-2, and 2-3, the results of the ANOVAs are presented together, grouped by the
organization-related measure of interest. I will refer to them as necessary is subsequent
sections reporting results.
(i) Intensity of Inter-Organizational Connectedness (IIOC)
Residents’ health literacy for preventing diabetes (for Latinos) and hypertension
(for African Americans) does not appear to vary significantly depending on whether one
lives in a neighborhood where organizations are highly connected to other organizations
or not, F(1, 517)=2.00, p>.10 (See Table 5.4).
Table 5.4. The role of IIOC in civic engagement & health outcomes (1-way ANOVAs)

136

(ii) Scope of Inter-Organizational Connectedness (SIOC)
A one-way ANOVA was also conducted to examine if scope of inter-
organizational connectedness mattered with respect to health literacy (for prevention). A
non-significant F(1, 517)=.27, p>.10 indicated that residents’ health literacy does not
differ significantly depending on whether they live in a neighborhood where
organizations have a broad scope of inter-organizational connections as opposed to a
narrow one (see Table 5.5).
Table 5.5. The role of SIOC in civic engagement & health outcomes
(results of 1-way ANOVAs)

(iii) Meso-level Storytelling Network Connectedness (MSNC)
Through a one-way ANOVA we also examined the degree to which the
interaction of the intensity and scope of connections at the meso-level explain differences
in residents’ health literacy scores (for preventing diabetes and hypertension). Do CBOs,
for instance, that are connected strongly to a wide variety of other organizations have a
larger impact on residents’ health literacy, compared to CBOs that have fewer, weaker,
and less diverse ties to other organizations? A non-significant F(1, 517)=.00, p>.10
indicated that this is not the case (see Table 5.6).
137

Table 5.6. The role of MSNC in civic engagement & health outcomes
(results of 1-way ANOVAs)

(iv) Geo-Demographic Reach (CSNI
GD
)
Next, I examined the role of geo-demographic reach. Based on communication
infrastructure theory, I hypothesized that organizations that target a broader socio-
demographic base in a local community would have a larger impact on residents’ health
literacy, as opposed to organizations that target particular demographics only across a
much broader geographic area. A one-way ANOVA indicated that it does matter if a
resident lives in a neighborhood where organizations have high or low geo-demographic
reach, F(1, 517)=.06, p<.05 (see Table 5.7). The effect size of geo-demographic reach is
small however, as η
2
is approximately equal to .01. The comparison of the mean health
literacy scores of residents in areas where organizations have high CSNI
GD
and residents
in areas where organizations have low CSNI
GD
, suggests, however, that, contrary to our
expectations, higher geo-demographic reach is associated with lower levels of health
literacy. The mean score for residents in neighborhoods with organizations scoring high
on CSNI
GD
was M=.16 (SD=.12), while the mean for residents in neighborhoods with
organizations scoring low on reach was M=.19 (SD=.12).
138

Table 5.7. The role of CSNI
GD
in civic engagement & health outcomes:
1-way ANOVA results

(v) Reach as Community Engagement (CSNI
CE
)
A one-way ANOVA also indicated that reach as community engagement is not a
factor that accounts for differences in Greater Crenshaw residents’ health literacy scores,
F(1, 518)=.30, p=.30 (See Table 5.8). The hypothesis was that the more an organization
engages residents and becomes an integral part of a neighborhood’s storytelling network,
the more likely it would be to have an impact on those residents’ lives.
Table 5.8. The role of CSNI
CE
in civic engagement & health outcomes:
1-way ANOVA results

139

(vi) Organizations’ Integration into a Storytelling Network (OISN
1
and OISN
2
)
The hierarchical multiple regressions that were reported earlier indicated that
individual-level communication capital (ICC) is a good predictor of residents’ health
literacy scores. ICC reflects the interaction of ICSN-2 and the organizations’ integration
into a storytelling network (OISN). Hence, a one-way ANOVA was run to explore if the
integration of organizations into the storytelling network alone is related to differences in
health literacy scores among residents. The results of the ANOVAs for both versions of
OISN (i.e., OISN
1
and OISN
2
) appear in Tables 5.9a and 5.9b; they indicate that OISN
alone does not account for differences in health literacy.
2

Table 5.9a-5.9b. The role of OISN in civic engagement & health literacy:
one-way ANOVAs

(a) As described in Chapter 4 OISN
1
is the interaction term of the individual's integration into the
storytelling network and geo‐demographic reach (CSNI
GD
)

2
As discussed in Chapter 4, OISN
1
reflects the interaction of meso-level storytelling network
connectedness (MSNC) and geo-demographic reach, whereas OISN
2
captures the interaction of MSNC and
reach as community engagement.
140

Table 5.9a-5.9b. The role of OISN in civic engagement & health literacy:
one-way ANOVAs (Continued)

(a) As described in Chapter 4 OISN
2
is the interaction term of the individual's integration into the
storytelling network and reach as community engagement (CSNI
GD
)

STEP 4

Finally and based on the results of Step 3, which showed that geo-demographic
reach was related to differences in residents’ health literacy scores, a hierarchical
multiple regression was conducted, in which geo-demographic reach was incorporated as
a dummy-coded predictor variable. The goal was to see if CSNI
GD
accounted for a
significant amount of the variance in prevention-oriented health literacy scores among
residents, above and beyond socio-economic status, neighborhood experience, and ICSN-
2. The analysis produced 5 models. The first included only socio-economic status
variables: age, education, gender, and income. The second included all the variables that
were part of Model 1 plus the variables capturing a resident’s neighborhood experience:
residential tenure and home ownership. Model 3 added ethnicity and Model 4 added the
residents’ degree of integration into their neighborhood storytelling network (ICSN-2).
The final model, Model 5, included the dummy-coded geo-demographic reach variable.
141

The regression analysis indicated that geo-demographic reach by itself did in fact
improve the regression model’s fit. Table 5.10a summarizes the hierarchical regression
analysis results and Table 5.10b reports the coefficients for the predictor variables
entered into the final model (i.e., including geo-demographic reach). The adjusted R
2
for
the final model (Model 5) was .05. The R
2
change produced by adding geo-demographic
reach was .01, F(1, 510)=4.02, p<.05. As the ANOVA indicated earlier, however, the
effect of CSNI
GD
on health literacy is not in the direction expected. The regression
coefficient is Β=-.09, t(513)=-2.15, p<.05).
Table 5.10a. Geo-demographic reach & prevention health literacy:
hierarchical regression

142

Table 5.10b. Impact of predictors including CSNIGD on health literacy
(regression coefficients)

Following the exact same series of procedures we examined the role of the STN in
health literacy for detection of diabetes (among Latinos) and hypertension-related
(among African Americans) symptoms. In none of the models run did the degree of
storytelling network integration (ICSN-2 and ICC) account for more of the variance
in symptoms detection-specific health literacy than already accounted for by socio-
economic status variables. Therefore, the models and the detailed results are not
presented here.
143

SUMMARY FOR HYPOTHESIS 1

The hierarchical multiple regressions conducted at Step 1 indicated that the
degree of storytelling network integration plays an important role in health literacy, at
least with respect to health literacy for preventing the diseases the two populations of
Greater Crenshaw are at high risk for developing: diabetes in the case of Latinos and
hypertension in the case of African Americans. Moreover, the interaction of residents’
integration in the STN and the degree of meso-level actors’ integration in the
neighborhood storytelling network appears to explain more of the variance with respect
to health literacy than just the measure of residents’ integration, even after controlling for
socioeconomic status variables, residential tenure and homeownership.
Running these analyses for the two different populations in Greater Crenshaw –
African Americans and Latinos – indicated that the optimal model for the entire
population applies to the African American population. Interestingly, in the case of
African Americans, while ICSN-2 did not improve the regression model’s fit
significantly, ICC did. In the case of Latinos, only socio-demographic variables explained
variability in residents’ health literacy scores.
As the interaction of residents and organizations, as reflected in the ICC measure,
appeared to be a relatively strong predictor of health literacy, stronger at least than ICSN-
2, we followed up with one-way ANOVAs to investigate if any of the organization-
related variables on their own explained health literacy differences among Greater
Crenshaw residents. The only organization-level variable associated with health literacy
differences was geo-demographic reach. Contrary to what was expected, the ANOVA
144

indicated that residents living in neighborhoods where organizations have higher CSNI
GD

have lower health literacy scores. A hierarchical multiple regression, in which CSNI
GD

was added in the final model in order to determine if its effect persisted, controlling for
socio-economic status, neighborhood experience, ethnicity, and ICSN-2, confirmed this
finding.

Research Question 1
RQ1a: Is an individual resident’s integrated connection into the storytelling network
(i.e., ICSN-2) positively related to how easy they find that it is for them to get medical
care?
RQ1b: Is individual-level communication capital (ICC) positively related to how easy
residents find that it is for them to get medical care and does it explain more of the
variance among residents’ scores than ICSN-2?

The goal set by the RQ1 is similar to that set in the case of H1: to determine if an
integrated STN makes a difference with respect to how easy residents find it to get health
care, above and beyond socio-economic variables, residential tenure, and home
ownership. To investigate RQ1 I employed the same procedures as in the case of H1.
Two measures of health care access were introduced in Chapter 4: (a) perceived difficulty
in getting health care and (b) distance to health care. The same analyses were conducted
for both outcome measures.
145

STEP 1

First, a hierarchical linear regression was conducted to investigate whether ICSN-
2 predicted residents’ perceived difficulty in getting health care access, controlling for
socio-economic factors, residents’ neighborhood experience (i.e., residential tenure and
home ownership), as well as their ethnic background. Four models were produced by this
analysis. Then, the same sequence of models were run again as part of a second
hierarchical multiple regression, this time replacing ICSN-2 in the final step (Model 4)
with ICC. This alternative Model 4 was compared to the model which included ICSN-2
to determine whether there was an indication of improvement, based on the R
2
scores
produced (i.e.: adjusted R
2
and R
2
change). Table 5.11 shows the results of the regression
analyses conducted.
None of the models performed particularly well. Model 4 is the first model where
a storytelling network variable is included; that is residents’ integrated connection to the
storytelling network (ICSN-2). The model represents no significant improvement
compared to Model 3, in which ethnicity is found to be a significant predictor of
perceived difficulty getting health care. According to Model 3 Latinos find it easier to get
health care than African Americans do in Greater Crenshaw. Model 3 produces a
significant R
2
change compared to Model 1, which included only socio-economic status
variables, residential tenure, and home ownership status, R
2
change=.01, F(1, 559)=3.89,
p<.05.
146

Table 5.11. Hierarchical multiple regression: STN integration & difficulty
getting health care

The alternative Model 4, however, which incorporates ICC in the place of ICSN-
2, appears to represent a significant improvement compared to all previous models. The
adjusted R
2
is .02 and R
2
change is equal to .005. It is significant, F(1, 555)=5.38, p<.05.
Interestingly the ICC coefficient is not in the direction we would expect. It suggests that
contrary to expectations, a higher STN integration score that accounts for the interaction
of residents and meso-level actors in the STN is related to poorer perceived health care
access scores.
147

STEP 2

In Step 2 we ran the same series of models for African Americans and Latinos
separately to see if the STN plays the same role in both populations. The results indicate
that when the two groups are looked at separately the role of STN integration (i.e., both
ICSN and ICC) becomes non-significant. Therefore further results from the analyses run
at this stage are not presented.
STEP 3

Results from Step 1 indicate that meso-level actors’ integration into the STN may
be important in predicting difficulty getting health care. That is why a series of one-way
ANOVAs were conducted (as in the case of Hypothesis 1 and prevention health literacy)
to determine if living in a neighborhood where organizations relate to other organizations
and residents differently matters with respect to how hard residents perceive getting
health care to be. The analyses indicated that scope of inter-organizational connectedness
(SIOC), meso-level storytelling network connectedness (MSNC), and reach as
community engagement (CSNI
CE
) make a difference when it comes to health care access.
On the other hand, intensity of inter-organizational connectedness (IIOC) and
organizational integration into the storytelling network (OISN) do not. Below are only
the findings of the analyses that produced significant results.
(i) Scope of Inter-Organizational Connectedness (SIOC)
A one-way ANOVA was conducted to determine if there is a significant
difference with respect to perceived difficulty getting health care between residents who
148

live in neighborhoods where organizations have a broader scope of inter-organizational
connections and residents in areas where organizations have a more narrow scope of
connections. The underlying hypothesis was that the more broad the scope of
collaborations among organizations in a community, the more likely it would be for
residents to receive, through their daily communication/interaction with organizations in
their community, useful information to accomplish goals – including, that is, information
related to getting health care. The analysis indicated that the scope of organization’s
connections does indeed make a significant difference, F(1, 517)=6.87, p<.01. The effect
size was η
2
=.013. Interestingly, however, comparing the means for residents in
neighborhoods with organizations scoring low on SIOC and neighborhoods with meso-
level actors scoring high indicated that broader scope was associated with residents
thinking it was harder for them to get the health care they needed. In neighborhoods with
organizations high on SIOC, the residents’ mean score was M=2.71, SD=1.09, whereas in
neighborhoods with organizations low on SIOC residents scored, on average, M=2.96,
SD=1.06.
(ii) Meso-level Storytelling Network Connectedness (MSNC)
A one-way ANOVA for MSNC produced similar results. The analysis indicated a
significant difference between residents living in a neighborhood where organizations
shared intense and broad communication connections with other organizations, and
people living in neighborhoods where the degree of meso-level storytelling network
connectedness was low, F(1, 517)=8.53, p<.01. The effect size was η
2
=.014. As was the
case with SIOC, low MSNC was associated with a higher mean score among residents,
149

M=2.93, SD=1.08, and high MSNC was associated with a lower mean score, M=2.66,
SD=1.06.
(iii) Reach as Community Engagement (CSNI
CE
)
Higher reach was also found to be associated with higher perceived difficulty in
accessing health care services in Greater Crenshaw, whereas lower reach, conceptualized
as community engagement, was associated with lower perceived difficulty in getting
health care. The one-way ANOVA conducted for CNSI
CE
produced F(1, 517)=5.67,
p<.05, and an effect size of η
2
=.011. Residents in neighborhoods with organizations
scoring higher on reach had a mean difficulty getting health care score of M=2.69,
SD=1.08. Residents in neighborhoods with organizations scoring lower on reach had a
mean score of M=2.92, SD=1.07.
STEP 4

Finally and based on the results of Step 3, we ran hierarchical multiple
regressions, which included, after successively adding socio-economic status variables,
residential tenure, home ownership, and residents’ ICSN-2, the three dummy variables
that the one-way ANOVAs indicated made a difference with respect to access to health
care; namely, scope of inter-organizational connectedness (SIOC), meso-level
storytelling network connectedness (MSNC), and reach as community engagement
(CSNI
CE
).
(i) Regression analysis #1: Scope of inter-organizational connectedness (SIOC)
The regression analysis indicated that the final model including SIOC accounted
for significantly more of the variation in perceived difficulty getting health care, even
after controlling for socio-economic status, residential tenure, homeownership, ethnicity,
150

and residents’ integrated connection to the STN (ICSN-2). The R
2
change between
Models 4 (without reach) and Model 5 (with reach) was .01, F(1, 477) = 4.51, p<.05.
(ii) Regression analysis #2: Meso-level storytelling network connectedness (MSNC)
As in the case of SIOC, meso-level storytelling network integration was also
found to account for a significant amount of the variability in perceived difficulty to
access health care, after controlling for socio-economic status, residential tenure,
homeownership, ethnicity, and ICSN-2. The R
2
change between Models 4 and 5 (the
model containing MSNC) was .01, F(1, 477) = 5.11, p<.05.
(iii) Regression analysis #3: Reach as community engagement (CSNI
CE
)
Similarly to SIOC and MSNC, the reach of organizations in a neighborhood
accounted for significant variation among residents’ scores on health care access above
and beyond socio-economic status, neighborhood experience variables, ethnicity, and
ICSN-2. The R
2
change between Models 4 and 5 (the model containing reach) was .01,
F(1, 477) = 4.90, p<.05.
Following the same sequence of procedures we examined the relationship of STN and
health care access, operationalized as the distance a resident has to travel to get health
care. In none of the models run did the degree of storytelling network integration
account for more of the variance in the distance to health care variable than already
accounted for by socio-economic status. Therefore, the models and the detailed
results are not discussed here further.
SUMMARY FOR RESEARCH QUESTION 1

The initial hierarchical linear regression run to investigate RQ1 suggested that the
interaction of residents’ storytelling network connection and the strength of the STN
151

connection of meso-level actors accounted for a significant amount of the variation
among residents’ scores on perceived difficulty getting health care, above and beyond
socio-economic status, residential tenure and home ownership, ethnicity, as well as
individual residents’ degree of connection to the storytelling network considered alone.
Probing further into why this is the case, one-way ANOVAs were run to investigate the
role of organization-level variables in perceived access to health care resources. The
ANOVAs indicated that scope of inter-organizational connectedness, meso-level
storytelling network connectedness, and reach as community engagement, were all
associated with significant differences in residents’ health care access scores. However,
contrary to what was expected, in all three analyses, the mean score for residents living in
neighborhoods where organizations scored higher on all three aforementioned measures
was lower than for those residing in communities where meso-level actors scored lower
on SIOC, MSNC, and CSNI
CE
. Hierarchical multiple regressions that incorporated
dummy-coded versions of these variables confirmed these findings. SIOC, MSNC, and
CSNI
CE
accounted for a significant amount of variability in the health care access scores
of residents, after controlling for SES, neighborhood experience, ethnicity, and residents’
individual-level measure of integration into the STN (ICSN-2). They all, however,
carried negative beta weights.

Hypotheses 2-1, 2-2, 2-3
H2-1a, 2-2a, 2-3a: An individual resident’s integrated connection into the storytelling
network (i.e., ICSN-2) will be positively related to neighborhood belonging (H2-1),
collective efficacy (H2-2), and civic participation (H2-3), after controlling for individual
152

level covariates, including socio-economic status, residential tenure, homeownership
status, and ethnicity.
H2-1b, 2-2b, 2-3b: Individual-level communication capital (ICC) will be positively
related to neighborhood belonging (H2-1), collective efficacy (H2-2), and civic
participation (H2-3), after controlling for individual level covariates and will explain
more of the variance among residents’ scores than ICSN-2.

Hypotheses 2-1, 2-2, and 2-3 call for the investigation into the role of storytelling
network integration in civic engagement. Prior research has found that individuals’
integrated connection into the storytelling network is a positive predictor of belonging,
collective efficacy, and civic participation (Kim, 2003; Kim & Ball-Rokeach 2006a; Kim
& Ball-Rokeach, 2006b). In this study I seek to understand if and how meso-level actors
contribute to the storytelling network-civic engagement relationship. The analysis
proceeded in four steps, similar to how it unfolded in the cases of Hypothesis 1 and
Research Question 1.
STEP 1

First, hierarchical multiple regressions were conducted to examine if individuals’
connection to the storytelling network by itself (ICSN-2) and the interaction of residents’
connection and meso-level neighborhood actors’ connection to the storytelling network
predicted neighborhood belonging (H2-1a), collective efficacy (H2-2a), and civic
participation (H2-3a), after controlling for socioeconomic status, residential tenure and
home ownership, and ethnicity. For each one of the outcomes four models were produced
by this analysis. Then, the same sequence of models were run again as part of a second
153

hierarchical multiple regression, this time replacing ICSN-2 in the final step (Model 4)
with ICC (so as to test H2-1b, H2-2b, H2-3b). This alternative final model was compared
to the model which included ICSN-2 to determine whether there was an indication of
improvement, based on the R
2
scores produced (i.e., adjusted R
2
and R
2
change). Tables
5.13a, 5.13b, and 5.13c show the results of the regression analyses conducted and signal
the coefficients that were found to be significant in every model.
(a) Neighborhood Belonging
In the case of the neighborhood belonging, ICC was found to be (i) a significant
predictor, while it also (ii) produced a larger improvement in the overall model fit than
ICSN-2. The R
2
change for the model with ICSN-2 was .06, F(1, 558)=33.79, p<.001,
whereas for the alternative model, where ICSN-2 was replaced by ICC, the R
2
change
was .10, F(1, 558)=64.33, p<.001.
In addition, the final model (see Alternative Model 4 with ICC in Table 5.12a)
suggests that ethnicity and gender play a significant role with respect to belonging. Men
are more attached to their residential community than women are, on average. Belonging
is also higher among African Americans than it is among Latinos in Greater Crenshaw.
154

Table 5.12a. Hierarchical multiple regression: STN integration
& neighborhood belonging

Collective Efficacy
ICC was also found to be (i) a significant predictor of collective efficacy and (ii)
it produced a larger change in the overall model fit than ICSN-2. The R
2
change for the
model with ICSN-2 was .007, F(1, 557)=4.28, p<.05, while for the alternative model,
which included ICC, R
2
change was .01, F(1, 557)=5.62, p<.05. Once again the measure
of storytelling network integration that captures the interaction of micro and meso-level
storytellers accounted for more of the variability in collective efficacy scores than the
155

measure accounting for individuals’ integration into the STN only. Table 5.12b
summarizes the results of the analysis.
Table 5.12b. Hierarchical multiple regression: STN integration & collective efficacy

Civic Participation
The results of the hierarchical multiple regression for civic participation appear in
Table 5.13c. The R
2
change for the model with ICSN-2 (Model 4) was .04, F(1,
558)=13.37, p<.001, while for the alternative Model 4, which included ICC, the R
2
change was .06, F(1, 558)=21.18, p<.001. In the final model, residential tenure,
education, and income are also noted as significant predictors of civic participation.
Residents that are more highly educated are more involved in civic activities and, as
156

might be expected, residents that have lived in a neighborhood longer exhibit higher civic
participation. In addition, residents with higher incomes also have higher participation
scores.
Table 5.12c. Hierarchical multiple regression: STN integration & civic participation

STEP 2

In Step 2, we ran the same series of models for African Americans and Latinos
separately to see if the storytelling network plays the same role in both populations
(Tables are used to summarize the results of the analyses).
157

Table 5.13a. Hierarchical regression: STN & belonging for African Americans
in Crenshaw

Neighborhood Belonging
a. African Americans
The results for the African American population of Greater Crenshaw are similar
to those produced by the regression analysis for the entire sample (see Table 5.13a); ICC
explained more of the variance in belonging than ICSN-2 did. The R
2
change for ICC was
.10, F(1, 271)=31.55, p<.001,as opposed to R
2
change=.05, F(1, 271)=15.30, p<.001 for
ICSN-2. Moreover, gender and residential tenure were found to be important predictors
158

of belonging. African American men exhibit higher belonging than women and African
Americans that have lived in the area longer feel more strongly attached to their
neighborhood.
Table 5.13b. Hierarchical regression: STN & belonging for Latinos in Crenshaw

b. Latinos
The patterns of relationships among the variables in the case of Latinos in Greater
Crenshaw mirror those observed in the overall sample and those found in the case of
African Americans. For ICSN-2, R
2
change=.06, F(1, 280)=19.88, p<.001, whereas for
ICC, R
2
change=.05, F(1, 280)=32.86, p<.001. Table 5.13b provides the detailed results.
159

Interestingly, the adjusted R
2
for the alternative Model 3 that includes ICC in both
the African American case and that of Latinos is exactly the same and equal to .15. The
adjusted R
2
for Model 3, which includes ICSN-2 instead of the ICC, for the respective
populations, is not the same. In fact, it is higher for Latinos at .11, compared to .09 for
African Americans. This may be an indication that the organizations are playing an
equalizing effect. However, based on these data and analyses alone we do not want to
conclude too much.
Collective Efficacy
a. African Americans
Both ICSN-2 and ICC do not predict collective efficacy among African
Americans in Greater Crenshaw. The models that included the storytelling variables
performed worse than those models only accounting for socio-economic status. Among
African Americans it appears that only age and income predict collective efficacy. Older
and financially better off African Americans exhibit higher levels of collective efficacy
(detailed results not reported).
b. Latinos
In the case of Latinos, however, while ICSN-2 did not better the fit of the
regression model to the data, the final model including ICC did account for significantly
more of the variance in collective efficacy among the Hispanic residents than the model
that included only socioeconomic status, residential tenure and homeownership status
did. The R
2
change produced by adding ICC was .013, F(1, 279)=3.93, p<.05. The
adjusted R
2
for the final model was .04. The beta weight for ICC was Β=.12, t(279)=1.98,
160

p<.05. Homeownership was also found to be a predictor of collective efficacy among
Latinos, with Β=.17, t(279)=2.45, p<.05.
Table 5.14. Hierarchical regression: STN & collective efficacy for Latinos in Crenshaw

Civic Participation
As in the case of neighborhood belonging, the models that predicted civic
participation best for the entire sample held up when applied to each ethnic group
separately.
a. African Americans
ICC explained more of the variance in belonging than ICSN-2 did. The R
2
change
for ICC was .06, F(1, 267)=21.18, p<.001,as opposed to R
2
change=.04, F(1, 269)=13.37,
161

p<.001 for ICSN-2. Moreover, education and residential tenure were found to be
important predictors of civic participation. More educated African Americans and
African Americans that have lived in Greater Crenshaw longer are more engaged in civic
activities (see Table 5.15).
Table 5.15a. Hierarchical regression: STN & civic participation for African Americans

c. Latinos
The patterns of relationships among the variables in the case of Latinos in Greater
Crenshaw mirror those observed in the overall sample and those found in the case of
African Americans. For ICSN-2, R
2
change=.05, F(1, 278)=21.08, p<.001, whereas for
162

ICC, R
2
change=.06, F(1, 277)=21.75, p<.001. Table 5.15b (see next page) provides the
detailed results.
STEP 3

Analyses from Steps 1 and 2 indicate that meso-level actors’ integration into the
STN may be important in predicting neighborhood belonging, collective efficacy, and
civic participation. To investigate further if there are specific organizational attributes
that play an important role in particular neighborhoods, as in the case of Hypothesis 1 and
Research Question 1, we conducted one-way ANOVAs to examine whether IIOC, SIOC,
MSNC, reach (CSNI
GD
and CSNI
CE
), and OISN make a difference with respect to civic
engagement.
a. Neighborhood belonging
None of the variables pertaining to the connections among meso-level actors and
between meso and micro-level actors were found to be associated with a significant
difference in belonging among Greater Crenshaw residents, except OISN (i.e., OISN
1
and
OISN
2
; see Tables 5.4-5.9b earlier in this chapter). The concept captures the extent to
which organizations in a particular neighborhood are, on average, integrated into a
neighborhood storytelling network, and the measure reflects the interaction between the
degree of connectedness among meso-level actors and their reach in the neighborhoods
they serve. There are two versions of OISN, as there are two versions of reach: geo-
demographic reach (CSNI
GD
) and reach as community engagement (CSNI
CE
).
163

Table 5.15b. Hierarchical regression: STN & civic participation for Latinos

(i) Organizational Integration into a Storytelling Network 1 (OISN
1
)
A one-way ANOVA was run to examine if the extent to which organizations in a
particular neighborhood are connected to the indigenous storytelling network makes a
difference with respect to residents feelings of attachment to their community. Residents
living in neighborhood where OISN
1
was high felt more attached to their neighborhood
than residents living in neighborhood where organizations were less integrated into the
indigenous storytelling network, as F(1, 518)=4.55, p<.05. The effect size was small at
η
2
=.01. The mean belonging score for residents in neighborhoods where OISN
1
is high
164

was M=25.69, SD=10.71, while the mean belonging score for residents where OISN
1
is
low was M=23.69, SD=10.12
(ii) Organizational Integration into a Storytelling Network 2 (OISN
2
)
The one-way ANOVA conducted to test the same hypothesis as in the previous
case, this time with OSNI
2
produced similar results, as F(1, 518)=4.97, p<.05. The effect
size was small at η
2
=.01. The mean belonging score for residents in neighborhoods where
OISN
2
is high was M=26.06, SD=10.94, while the mean belonging score for residents
where OISN
2
is low was M=23.93, SD=10.10.
b. Collective Efficacy
None of the variables pertaining to the connections among meso-level actors and
between meso and micro-level actors were found to be associated with a significant
difference in collective efficacy among Greater Crenshaw residents. The results of the
ANOVA-associated F tests can be found in Tables 5.4 through 5.9b.
c. Civic Participation
One-way ANOVAs were also conducted to examine the relationship between the
organizational-level variables and the degree of residents’ civic participation. As in the
case of belonging, residents living in neighborhoods where organizations were more
integrated into the indigenous storytelling network had higher civic participation scores
than residents in areas where organizations were less connected to the neighborhood
STN. Both versions of OISN were associated with significant differences among
residents with respect to civic participation.

165

(i) Organizational Integration into a Storytelling Network 1 (OISN
1
)
The one-way ANOVA conducted for OISN
1
and civic participation yielded F(1,
514)=11.09, p<.001. The effect size was η
2
=.02. Mean civic participation for residents
living in neighborhoods where organizations were highly integrated into the STN was
M=1.27, SD=1.46, while for residents where organization were not well integrated
M=.87, SD=1.27.
(ii) Organizational Integration into a Storytelling Network 1 (OISN
2
)
The one-way ANOVA conducted for OSNI
2
and civic participation yielded F(1,
514)=10.04, p<.01. The effect size was η
2
=.02. Mean civic participation for residents
living in neighborhood where organizations were highly integrated into the STN was
M=1.33, SD=1.46, while for residents where organizations scored low on STN
integration was M=.83, SD=1.30.
STEP 4

Based on the results of Step 3, a series of hierarchical multiple regressions were
conducted for civic engagement dimensions of belonging and civic participation. The
analyses produced 5 models for each dependent variable. The first included only socio-
economic status variables: age, education, gender, and income. The second included all
the variables that were part of Model 1 plus the variables capturing a resident’s
neighborhood experience: residential tenure and home ownership. Model 3 added
ethnicity and Model 4 added the residents’ degree of integration into their neighborhood
storytelling network (ICSN-2). The final model, Model 5, included the dummy-coded
organizational integration into the storytelling network variable (both versions, OISN
1
,
OISN
2
).
166

We ran these analyses to investigate if the significant differences in residents’
belonging and civic participation, which the ANOVAs in Step 3 attributed to the
organizations’ integration into the storytelling network (OISN) could account for
variance in these two dimensions of civic engagement, controlling for socio-economic
status, residential tenure, home ownership, ethnicity, but also individuals’ degree of
integration into the STN.
Neighborhood Belonging
(i) Organizational Integration into a Storytelling Network 1 (OISN
1
)
The results of the analysis indicate that the integration of organizations into the
neighborhood STN can account for a significant amount of the variance beyond what has
already been explained by SES, neighborhood experience, ethnicity, and ICSN-2. The R
2

change produced by adding OISN
1
to the model was .01, F(1, 514)=4.85, p<.05. OISN
1

carries a positive beta weight of Β=.09, t(514)=2.20, p<.05.
(iii) Organizational Integration into a Storytelling Network 2 (OISN
2
)
As in the case OISN
1
, OISN
2
can also account for a significant amount of the
variance beyond what has already been explained by SES, neighborhood experience,
ethnicity, and ICSN-2. The R
2
change produced by adding OISN
2
to the model was .01,
F(1, 514)=4.00, p<.05. OISN
2
carries a positive beta weight of Β=.08, t(514)=2.00,
p<.05.
Civic Participation
(i) Organizational Integration into a Storytelling Network 1 (OISN
1
)
Contrary to what the analyses showed in the case of belonging, the regression for
civic participation suggested that OISN
1
cannot account for a significant amount of the
167

variance beyond what has already been explained by SES, neighborhood experience,
ethnicity, and individuals’ integration into their neighborhood STN. The R
2
change from
Model 4 to Model 5, in which OISN
1
is included in non-significant, R
2
change=.00, F(1,
480)=.07, p>.50).
(ii) Organizational Integration into a Storytelling Network 2 (OISN
2
)
Similar results are produced with OISN
2
in Model 5.
SUMMARY FOR HYPOTHESES 2-1, 2-2, 2-3

Hierarchical multiple regressions indicated that the interaction of meso-level
actors’ integration into the neighborhood storytelling network and individuals’
integration into their community STN contributes more to explaining the degree of civic
engagement (particularly neighborhood belonging and civic participation) in a residential
community than degree of residents’ integration alone.
In addition, follow up ANOVAs suggest that the integration of organizations into
the STN accounts for significant differences among residents in terms of belonging and
civic participation (not collective efficacy). Increased integration of organizations into the
network was associated with higher belonging and civic participation. These results seem
to suggest that organizations may be important generators of civic engagement and
mechanisms through which neighborhood belonging is built and sustained.
However, regression analyses, in which the degree of organizations’ integration
into the STN was included as a predictor of civic engagement, suggested that the positive
impact of organizations’ integration into the storytelling network persists only in the case
of neighborhood belonging, after controlling for socio-economic status, neighborhood
experience and an individual resident’s integration into the storytelling network (i.e.,
168

ICSN-2). In the case of civic participation, the positive influence of the organizations’
integration disappears, once we have accounted for individual level socio-economic
variables, residential tenure, homeownership, and individual-level integration into the
STN. Considering the results of analyses from all four stages described earlier, we
surmise that organizations in the Greater Crenshaw area can act as amplifiers of
residents’ efforts (as indicated by the significance of the micro-meso level interaction
captured in the individual-level communication capital variable), but that their unique
effect as distinct storytelling network actors in building civic engagement is tenuous.

The storytelling network as a mechanism of neighborhood health effects

Hypotheses 3a(1-3) and 3b(1-3)
3

H3a-1, 3a-2, 3a-3: There is an indirect path of influence from individual-level
communication capital (ICC) to health literacy via neighborhood belonging (H3a-1),
collective efficacy (H3a-2), and civic participation (H3a-3).
H3b-1, 3b-2, 3b-3: There is an indirect path of influence from individual-level
communication capital to residents’ perceived ability to access information and services
they need to safeguard their health via neighborhood belonging (H3b-1), collective
efficacy (H3b-2), and civic participation (H3b-3).

3
These hypotheses are aimed primarily towards examining the indirect effects of the storytelling network
on health literacy and health care access, because the direct effects were studied in much more detail
through the analyses conducted for Hypothesis 1 and 2, and Research Question 1. However, structural
equation modeling, which is employed to investigate the indirect effects, also produces results that allow us
to confirm earlier findings regarding direct effects. Hence, results pertaining to the direct links between the
STN and the health-related outcome variables are reported here as well.
169

While Hypotheses 1 and 2, as well as Research Question 1 called for the
investigation of the direct influence of storytelling network integration on health literacy
and civic engagement, the objective set for Hypotheses 3a and 3b is to understand what
the relationships are among STN integration, civic engagement, health literacy, and
health care access. Earlier in this study we posited that communication can be a more
elementary social process through which neighborhood effects manifest. Therefore the
goal now is to investigate if communication has not only direct influence on health
literacy and health care access, but also indirect effects through social processes that have
been studied more extensively as mechanisms of neighborhood effects – i.e., collective
efficacy, neighborhood belonging, and civic participation. The correlations among the
variables involved in the structural equation modeling analyses appear in Table 5.16.
Testing the overall theoretical model
Figure 5.1 presents the results of the LISREL analysis for the combined
hypotheses. The chi-square was significant, χ
2
(9, N = 599) = 176.65, p < .001. The Root
Mean Square Error of Approximation is over .10 (RMSEA=.18). The goodness-of-fit
index is relatively high at GFI=.93, although the adjusted GFI score is low at .73. These
results suggest a poor overall fit.

Local level hypotheses
a. Direct Effect
(i) Individual-level communication capital (ICC) and civic engagement:
Consistent with the findings of the analyses conducted for Hypotheses 2-1, 2-2, and 2-3
as well as previous research, the interaction of individuals’ integrated connection into the
neighborhood storytelling network and the level of organizations’ integration into the
170

storytelling network predicted neighborhood belonging, collective efficacy, and civic
participation.
Table 5.16. Correlations among exogenous, endogenous variables in the SEM analyses
Correlations among exogenous and endogenous variables
(ξ1) (η1) (η2) (η3) (η4) (η5) (η6) (η7)
ICC (ξ1) ‐ ‐

Neighborhood
Belonging (η1)
0.34 ‐ ‐

p‐value
**
Collective Efficacy
(η2)
0.13 0.36 ‐ ‐

p‐value
** **
Civic Participation
(η3)
0.37 0.30 0.14 ‐ ‐

p‐value
** ** **
Health Literacy:
Prevention (η4)
0.15 0.03 0.03 0.15 ‐ ‐

p‐value
** **
Health Literacy:
Detection (η5)
0.07 0.01 0.08 0.09 0.35 ‐ ‐

p‐value
* **
Difficulty Getting
Medical Care (η6)
‐0.12 0.00 0.00 ‐0.05 0.06 ‐0.02 ‐ ‐
p‐value
**
Distance to
medical care (η7)
0.03 0.08 0.05 0.04 0.08 0.09 0.01
‐ ‐
p‐value
* * *
Note
** p<.01
* p<.05

(ii) ICC and prevention health literacy: The SEM analysis confirmed the findings
from analyses conducted to test Hypothesis 1, indicating that the degree of storytelling
network integration predicts prevention health literacy, with a positive sign. However, the
path from STN integration to detection health literacy was found to be non-significant.
171

(iii) ICC and perceived difficulty accessing medical care: Structural equation
modeling analysis corroborates what was found in earlier analyses; that is, that the degree
of integration of the neighborhood storytelling network predicts perceived difficulty in
accessing medical care services, with a negative sign.
(iv) ICC and distance traveled to health care: The analysis suggests that there is
no direct link between individual-level communication capital and the distance residents
have to travel to find the health care they need.
b. Indirect Effects
(i) Indirect paths of influence to health literacy: Hypotheses 3a-1 through 3a-3
proposed indirect paths of positive influence from communication capital to health
literacy via belonging, collective efficacy and civic participation. The t-values produced
by the initial analysis provided support for the hypothesis that civic participation predicts
health literacy. That is only true, however, for prevention health literacy. The paths from
belonging and collective efficacy, however, to prevention health literacy were non-
significant. In addition, none of indirect effects links from ICC to symptoms detection-
related health literacy via civic engagement were significant.
(ii) Indirect paths of influence to medical care access: As in the case of health
literacy, none of the indirect effects were significant. Figure 5.1 shows the theoretical
model and Table 5.17 shows the direct and indirect effects for the hypothesized paths.
172

Figure 5.1. Theoretical model of neighborhood health effects from a CIT perspective

Model revision
Table 5.18 outlines the process of revision and the final model is presented in
Figure 5.2. The first revision of the model, which entailed deleting all non-significant
paths resulted in a model with only slightly better fit scores. The chi-square was still
significant and higher than that of the first model, χ
2
(22, N = 602) = 193.83, p < .001. As
the chi-square statistic is sensitive to sample size, I followed the example of Joreskog and
Sorbom (1989) and Schmitz and Fulk (1991) and also calculated the ratio of chi-square
over degrees of freedom. A result under 5 is noted as an appropriate standard (Wheaton,
Muthen, Alwin, and Summers, 1977). Based on this indicator of fit, the first revision did
in fact yield a slightly better model. Whereas χ
2
/d.f. in the original model was equal to 20,
it dropped to 9 after the first revision. It still, however, was over 5. The RMSEA was
slightly better at .11, still indicating a bad fit. The goodness of fit and adjusted goodness
of fit indices indicated the same minor improvement, at .93 and .88 respectively.
173

Table 5.17. Theoretical model: significant direct and indirect effects

EFFECTS on
*
:
Health Literacy
Prevention Detection

Direct Indirect Total Direct Indirect Total
ICC (ξ1) .002 ‐‐‐ .002 .001 ‐‐‐ .001
Neighborhood
Belonging (η1)
‐.001

‐.001 .000

.000
Collective
Efficacy (η2)
.002

.002 .008

.008
Civic
Participation (η3)
.005 .005 .005 .005

Access to Health Care
Difficulty Getting
Health Care
Distance to
Health Care

Direct Indirect Total Direct Indirect Total
ICC (ξ1)
‐.015
‐.002
(.001)
‐.017 .001 ‐‐‐ .001
Neighborhood
Belonging (η1)
.005

.005 .002

.002
Collective
Efficacy (η2)
.016

.016 .008

.008
Civic
Participation (η3)
‐.012 ‐.012 .002 .002
*Standard errors for indirect effects in parentheses

Examination of the modification indices suggested that it was necessary to add
one antecedent to collective efficacy and civic participation, other than individual-level
communication capital (ICC): neighborhood belonging. Modification indices also
suggested adding a path from the prevention dimension of health literacy to the
symptoms detection dimension. The analysis indicated that the new model resulting from
revision 2 (see Figure 5.2) was a significant improvement. The chi-square was non-
174

significant with a χ
2
(20, N = 602) = 23.42, p > .10. The χ
2
to d.f. ratio was equal to 1.20.
Goodness of fit was estimated at .99, adjusted goodness of fit at .98. The RMSEA
dropped to .02, indicating a good fit for the model. It is noteworthy that in this final
model, the path from storytelling network integration to collective efficacy drops out as
non-significant. The direct and indirect effects in the new model are shown in Table 5.19.
Table 5.18. Summary of LISREL model revisions

Path changes χ
2
Goodness Adjusted Root Mean
(Modification (d.f.) of Fit (GFI) Goodness Squared
Indices) of Fit Error
(AGFI) of
Approximatio
n
(RMSEA)

Hypothesized model 176.65 .93 .73 .18
(9)
p < .001
χ
2
/d.f. = 20

Revision 1 193.83 .93 .88 .11
Non-significant (20)
paths deleted p < .001
χ
2
/d.f. = 9

Revision 2 23.42 .99 .98 .02
Belonging (20)
Æ Collective Efficacy p = .27
(t = 9.73) χ
2
/d.f. = 1.2

Belonging
Æ Civic Participation
(t = 4.71)

Prevention Health Literacy
Æ Detection
(t = 8.92)

175

Figure 5.2. Revised model of neighborhood health effects from a CIT perspective

176

Table 5.19. Revised model: significant direct and indirect effects

EFFECTS on
*
:
Health Literacy
Prevention Detection

Direct Indirect Total Direct Indirect Total
ICC (ξ1)
.001
.001
(.001)
.002 ‐‐‐ ‐‐‐ ‐‐‐
Neighborhood
Belonging (η1)
‐‐‐

‐‐‐ ‐‐‐

‐‐‐
Collective
Efficacy (η2)
‐‐‐

‐‐‐ ‐‐‐

‐‐‐
Civic
Participation (η3)
.009 .009 ‐‐‐ ‐‐‐

Access to Health Care
Difficulty Getting
Health Care
Distance to
Health Care

Direct Indirect Total Direct Indirect Total
ICC (ξ1) ‐.017 ‐‐‐ ‐.017 ‐‐‐ ‐‐‐ ‐‐‐
Neighborhood
Belonging (η1)
‐‐‐

‐‐‐ ‐‐‐

‐‐‐
Collective
Efficacy (η2)
‐‐‐

‐‐‐ ‐‐‐

‐‐‐
Civic
Participation (η3)
‐‐‐ ‐‐‐ ‐‐‐ .‐‐‐
*Standard errors for indirect effects in parentheses

African Americans and Latinos in Greater Crenshaw
Subsequently we examined whether the revised model applied both to the African
American and Latino populations of Greater Crenshaw by repeating the process
described above. The final models for both populations are presented in Figures 5.3a and
5.3b.

177

Figure 5.3a. Final model of neighborhood health effects for African Americans

Figure 5.3b. Final model of neighborhood health effects for Latinos

From the analyses in which the two ethnic groups in Greater Crenshaw were
considered separately, three results are worth noting: first, in the case of African
Americans, there’s no direct path of influence from civic participation and indirect path
178

from individual-level communication capital (ICC). This is a path that was significant for
the population in the area as a whole. In addition, in the case of African Americans, the
path from collective efficacy to detection health literacy is significant. That was not the
case for the entire sample. Second, in the case of Latinos the direct link from civic
participation to prevention health literacy and from storytelling network integration to
health literacy through civic participation is significant, as shown in the analysis for the
entire sample. Finally, the negative link between the storytelling network and perceived
difficulty in accessing health care seems to dissipate when the two ethnic groups are
treated separately.

The role of the communication action context in neighborhood health effects
The third larger question underlying the research questions driving this project is
related to how the communication action context influences the storytelling network and
what the role of their interaction is in civic engagement and the manifestation of
particular effects, such as health literacy and access to health care. Three elements of the
communication action context are considered in this study: the density of institutional
resources in residents’ neighborhoods, ethnic heterogeneity, and residential stability.

Density of institutional resources, health literacy, and health care access
The density of institutional resources is the key independent variable in
Hypotheses H4-1, H4-2, H4-3, H5a and H5b, as well as Research Questions RQ2a1,
RQ2a2, and RQ2a3.
179

Hypotheses 4-1, 4-2, and 4-3
H4-1a, 4-2a, and 4-3a: Density of institutional resources will predict neighborhood
belonging (H
4-1
), collective efficacy (H
4-2
), and civic participation (H
4-3
), after controlling
for individual level covariates.
H4-1b, 4-2b, and 4-3b: Density of institutional resources will predict neighborhood
belonging (H
4-1
), collective efficacy (H
4-2
), and civic participation (H
4-3
), after controlling
for neighborhood level ethnic heterogeneity and residential stability.

The challenge set by Hypothesis 4 is to determine whether density of institutional
resources in a resident’s community impacts their level of neighborhood belonging,
collective efficacy, and civic participation. Two series of analyses were conducted for
each one of the hypotheses, as described in the analysis plan, in Chapter 4. First,
hierarchical multiple regression analyses were run to see if density (as a dummy-coded
variable) impacted residents’ level of civic engagement, controlling for socio-economic
status, residential tenure, and homeownership status, but also their degree of integration
into their neighborhood storytelling network (ICSN-2).
The second series of analyses were aimed at figuring out if residents that lived in
neighborhoods with high density of organizational resources exhibited higher levels of
civic engagement compared to those individuals that lived in areas with low density of
organizational resources. As controlling for ethnic heterogeneity and residential stability
was impossible in the fashion in which it would have been done through a regression
model (multi-level or other), we chose to proceed with a 3-way ANOVA. We opted for
180

this procedure because it would tell us if the main (unique) effect of density of
organization resources on civic engagement was significant, as well as if any interaction
among density of resources, ethnic heterogeneity, and residential stability had a
significant effect on civic engagement. By being able to account for unique and
interaction effects we would, in essence be accomplishing the goal of detecting the effect
density of resources ‘controlling’ for ethnic heterogeneity and residential stability.
STEP 1

The hierarchical multiple regressions conducted suggest the density of
organizational resources does not explain more of the variance in belonging and civic
participation scores among Greater Crenshaw residents than is already accounted for by
socio-economic status variables, neighborhood experience, and their integration into the
storytelling network (ICSN-2). The R
2
change produced in both analyses by adding
organizational density to the final model was non-significant. However, in the case of
collective efficacy, organizational density did contribute to the explanation of the
variability of residents’ scores. The addition of density of institutional resources to the
final model (see Model 5 in Table 5.20) produced a significant R
2
change=.02, with F(1,
398)=7.20, p<.01. The adjusted R
2
for the final model was .05. In the final model, the
coefficient for organizational density was Β=.13, t(398)=2.64, p<.01.
Table 5.20 presents the results of the analysis for collective efficacy.
181

Table 5.20. Density of institutional resources and collective efficacy (hierarchical
regression)

STEP 2

In step 2, we investigated if the interaction of density of institutional resources in
a neighborhood and other neighborhood environment characteristics – i.e., ethnic
heterogeneity and residential stability – make a difference with respect to residents’ civic
engagement. A three-way ANOVA was conducted, using dummy-coded versions of the
three communication action context variables: density of institutional resources, ethnic
heterogeneity, and residential stability. The results suggest that the main effects of all
variables, except organizational density, on collective efficacy are non-significant. In
addition, the interaction of density of organizations in a neighborhood, ethnic
heterogeneity, and residential stability also makes no difference with respect to residents’
level of civic engagement.

182

Hypotheses 5a, 5b
H5a-1: Density of institutional resources will predict health literacy, after controlling for
individual level covariates.
H5a-2: Density of institutional resources will predict health literacy, after controlling for
neighborhood level ethnic heterogeneity and residential stability.
H5b-1: Density of institutional resources will predict residents’ perceived degree of
access to health-related services, after controlling for individual level covariates.
H5b-2: Density of institutional resources will residents’ perceived degree of access to
health-related services after controlling for neighborhood level ethnic heterogeneity and
residential stability.

The goal set by Hypotheses 5a and 5b is to determine whether density of
institutional resources in a resident’s community impacts their level of health literacy
(prevention and symptoms detection health literacy), as well as health care access. Two
series of analyses, similar to those conducted for Hypothesis 4, were run. First, multiple
regression analyses were run to see if density (as a dummy-coded variable) impacted
residents’ level of health literacy and health care access, controlling for socio-economic
status, residential tenure, and homeownership status, but also their degree of integration
into their neighborhood storytelling network. Second, through a 3-way ANOVA we
investigated the possible effects of the interactions among institutional resource density,
ethnic heterogeneity, and residential stability on residents’ health literacy and health care
access.

183

STEP 1

Both series of hierarchical multiple regressions conducted testing the link between
organizational resource density and health literacy yielded results indicating that density
of resources does not account for variance unexplained by socioeconomic status,
neighborhood experience, ethnicity, and a resident’s degree of integration into the
storytelling network. Likewise, organizational resource density does not seem to make a
difference with respect to how difficult Greater Crenshaw residents believe it to be to get
health care in their neighborhood, or how far they have to travel to get medical care.
STEP 2

In addition, the 3-way ANOVA conducted suggested that organizational density
does not interact with residential stability and ethnic heterogeneity in a way that explains
differences among residents with regard to health literacy and health care access.

Density of institutional resources, civic engagement, health literacy, and health care
access
Research Questions 2a-1 through 2b-3 investigate if the density of institutional
resources influences health literacy and health care access in interaction with civic
engagement.

Research Questions 2a and 2b
RQ2a: Are there significant differences in health literacy scores among residents living
in neighborhoods with a high density of organizational resources and residents living in
184

neighborhoods where organizational resources are sparse, which may be explained by
variation in neighborhood levels of civic engagement?
RQ2b: Are there significant differences in perceived access to health care among
residents living in neighborhoods with a high density of organizational resources and
residents living in neighborhood where organizational resources are sparse, which may
be explained by variation in neighborhood levels of civic engagement?

For the purposes of answering these questions two-way ANOVAs were conducted
to examine if the interaction of density of institutional resources and civic engagement
makes a difference with respect to (a) health literacy (prevention and detection), (b)
difficulty getting health care, and (c) distance travelled to get health care. In each case,
dummy versions of the density and civic engagement variable of interest were utilized.
a. Health literacy
Prevention: In the case of prevention health care literacy, density of resources and
belonging, density and collective efficacy, as well as the interaction of density of
resources and civic participation do not account for differences among residents’ scores.
Detection: While 2-way ANOVAs yielded similar results for the interaction of
density of resources and belonging, and density and civic participation, the interaction of
collective efficacy and density of resources seems to matter when it comes to the
detection dimension of health literacy, F(1, 443)=4.50, p<.05, partial η
2
=.01.
An additional analysis was conducted to address the hypothesis that health
literacy would be higher in communities that are richer in organizational resources and
that enjoy a higher level of collective efficacy, compared to communities where
185

organizational resources are sparse and collective efficacy is low. We subtracted the
mean health literacy score of residents living in a neighborhood low on collective
efficacy and low on organizational density from the mean score of residents living in a
community low on collective efficacy and high on organizational density. From that
score we subtracted the mean health literacy score of residents living in a neighborhood
high on collective efficacy and low on organizational density from the mean score of
residents living in a community high on collective efficacy and high on organizational
density. The result was .04, F(1, 439)=4.50, p<.05. The results of this comparison
supported the hypothesis driving the post-hoc analysis. The means and standard
deviations for symptoms detection-related health literacy are presented in Table 5.21.
Table 5.21. Interaction of organizational density & collective efficacy (post-hoc analysis)

Access to health care
Two-way ANOVAs indicated that the interaction of density of organizational
resources and civic engagement does not explain differences among Greater Crenshaw
residents with regard to perceived difficulty in getting health care and distance travelled
to receive health care.
186

Ethnic Heterogeneity
Research Question 3
RQ3: Are there significant differences in individual-level communication capital among
residents that are accounted for by ethnic heterogeneity?

A t-test was conducted to examine if residents’ individual-level communication
capital scores differed significantly depending on whether they lived in a highly
ethnically heterogeneous neighborhood or a community low on ethnic heterogeneity. The
t-test indicated that there was no significant difference in communication capital between
residents living in ethnically heterogeneous and ethnically homogeneous neighborhoods.

Research Questions 4a and 4b
RQ4a: Is there a significant difference among organizations in Greater Crenshaw in
terms of ethnic reach?
RQ4b: Is the number of organizations in Greater Crenshaw that all residents connect to
significantly different from the number of organizations to which only residents of a
particular ethnic background connect?

The purpose of these hypotheses was to determine the extent to which there is a
storytelling network in Greater Crenshaw that is bifurcated along ethnic lines. The
significant difference among Latinos and African Americans in terms of their integration
into a neighborhood storytelling network (ICSN-2), which was reported earlier, as well as
their significant difference in terms of individual-level communication capital (see results
187

of t-tests for ICC in Table 5.1) suggest that this may be the case. The mean ICSN-2 score
for Latinos was significantly lower than it was for African Americans. In addition, the
fact that the mean ICC score for Latinos was also significantly lower than the mean ICC
for African Americans indicates that organizations are not compensating for the
differences between the two ethnic populations. It also indicates that (a) organizations’
ethnic reach may be narrow (i.e., their efforts are targeted to one or the other group more)
and that (b) Latinos and African Americans may be connecting to very different meso-
level actors.
Figure 5.4. Ethnic reach of organizations interviewed in Greater Crenshaw (N=80)

As indicated in Figure 5.4, among the organizations interviewed for this project,
52% of them stress that they target or appeal to both African Americans and Latinos in
Greater Crenshaw, while a little over half of that number, 27%, stated that their
188

constituency was either primarily Latino or primarily African American. A one-sample
chi-square test was run to evaluate whether the difference in ethnic reach that is
illustrated by the aforementioned percentages is significant. The chi-square score
produced was significant, χ
2
(2, N=71)=11.60, p<.001 the effect size, calculated by
dividing the χ
2
by the product of total sample size across categories and the number of
categories minus 1, was .08. The effect size indicates that the observed frequencies
deviate moderately from the expected frequencies. These results suggest that
organizations do differ significantly in terms of ethnic reach, but most of them appeal to
both African Americans and Latinos about the same, and not more to one or the other
ethnic group. Therefore, we would expect that organizations could be playing a bridging
role between the two ethnic populations in Greater Crenshaw.
However the significant differences in ICSN-2 and ICC along ethnic lines as well
as the amount of overlap in the organizations that residents consider to be most important
indicates that meso-level actors are not terribly successful as bridging agents in the
community. The Metamorphosis survey of residents included four questions, which asked
residents to mention the names of (a) sports and recreation-focused organizations, (b)
religious, ethnic, and cultural organizations, (c) neighborhood or homeowners’
associations, as well as (d) political and educational organizations they found to be most
important. The full list of organizations reported by African American and Latino
residents can be found in Appendix C. Out of a total of 149 organizations mentioned only
6 were clearly mentioned by both ethnic groups; that is only 4% of the total. A one-
sample chi-square indicated that the obviously large difference between the number of
organizations mentioned by both African Americans and Latinos and the number of those
189

mentioned by only one of the two ethnic groups was indeed significant, χ
2
(2,
N=71)=125.97, p<.001. The effect size was strong at .85.

Research Questions 5a-b and 6a-b
RQ5a-5b: Is ethnicity a predictor of health literacy (RQ5a) and access to health care
resources (RQ5a) controlling for individual level covariates?
RQ6a-6b: Are there significant differences in health literacy (RQ6a) and perceived
access to health care (RQ6b) among residents living in neighborhoods with high ethnic
heterogeneity and residents living in neighborhood where ethnic heterogeneity is low,
which may be explained by variation in neighborhood levels of community
communication capital?

The objective of Research Questions 5a and 5b was accomplished earlier through
the hierarchical multiple regression analyses conducted to evaluate the effect of
storytelling network integration on health literacy and health care access, controlling for
individual level variables, including ethnicity. Ethnicity was shown to be a significant
predictor of prevention health literacy even in the final models. The coefficient for
ethnicity (dummy-coded, where being Latino=1) was Β=.13, t(555)=2.92, p<.01,
indicating that health literacy among Latinos was higher (see Table 5.2 for more details).
However, ethnicity was not found to be a predictor of symptoms detection health literacy,
or of health care access.
To investigate Research Questions 6a and 6b two-way ANOVAs were conducted
to explore if living in a highly ethnically heterogeneous neighborhood or in a more
190

homogeneous community mattered with regard to health literacy and health care access,
and if such differences are explained by the interaction of ethnic heterogeneity and
community communication capital.
a. Health Literacy
The 2-way ANOVA confirmed the main effect of ethnic heterogeneity on
symptoms detection-oriented health literacy, F(1, 443)=6.83, p<.01. The main effect of
community communication capital was non-significant, as was, more importantly for the
purpose of these research questions, the effect of the interaction of the two factors. The
main effect of ethnic heterogeneity and the effect of the interaction between ethnic
heterogeneity with community communication capital on prevention-specific health
literacy were found to be non-significant.
b. Health Care Access
A 2-way ANOVA also confirmed the main effect of ethnic heterogeneity with
regard to distance travelled for medical care, F(1, 443)=12.16, p<.001. The main effect of
community communication capital, however, as well as the effect of the interaction
between community communication capital and ethnic heterogeneity was non-significant.
A two-way ANOVA indicated that ethnic heterogeneity alone and in interaction with
CCC has no significant effect on perceived difficulty getting health care.

Residential Stability
Hypotheses 6a and 6b
H6a: Differences in health literacy scores among residents living in stable and unstable
communities will vary as a function of community communication capital.
191

H6b: Differences in health care access scores among residents living in stable and
unstable communities will vary as a function of community communication capital.

Hypothesis H6a calls for an investigation into whether or not community
communication capital interacts with residential stability in a way that residential
instability will be associated with lower health literacy scores in neighborhoods where the
STN is less integrated. Hypothesis H6b asks a similar question, this time with respect to
access to health care resources. To address them, we conducted 2-way ANOVAs in
which residential stability and community communication capital were dummy-coded, so
that, for instance, every resident who lived in a neighborhood with low a residential
stability score was assigned a ‘0’ and everyone living in more stable community received
a ‘1.’
a. Health Literacy
Prevention: In the case of prevention health literacy, the 2-way ANOVA
indicated that the degree of storytelling network integration at the neighborhood level had
a significant main effect, as F(1,443)=6.11, p<.01. The effect size associated with STN
integration was η
2
=.01. However, the main effect of residential stability was non-
significant, as was the effect of the two variables’ interaction. That is to say that living in
a community in which meso and micro level actors are more integrated into the
indigenous storytelling network matters with respect the health literacy (for prevention of
diabetes and hypertension), albeit living in a more or less stable community does not
make a difference.
192

Detection: Neither of the main effects nor the effect of the interaction of
community communication capital and residential stability were significant in the 2-way
ANOVA conducted using the symptoms detection-oriented health literacy measure as the
dependent variable.
b. Health Care Access
Difficulty getting health care: The two-way ANOVA conducted suggested that
residential stability matters with respect to how easy or how hard residents perceive it to
be to get health care. The main effect of residential stability was significant, F(1,
443)=5.68, p<.05. The effect size for residential stability was η
2
=.01. The analysis
indicates that residents that live in more stable communities find it easier to get health
care. The mean for residents in stable communities was M=2.95, SD=.08, while for
residents in unstable communities, M=2.72, SD=.08. The main effect of community
communication capital was also significant. Residents living in communities with a
highly integrated multi-level storytelling network had a mean score of M=2.69, SD=.09,
whereas residents in communities with a much less integrated storytelling network had a
mean score of M=2.98, SD=.07. This finding confirms what was shown earlier in the
analyses conducted for Research Question 1 and the structural equation modeling-based
analyses, in which case individual-level communication capital was negatively related to
perceived health care access.
Distance to health care: Neither of the main effects nor the effect of the
interaction of community communication capital and residential stability were significant
in the 2-way ANOVA conducted using the distance traveled to get medical care measure
as the dependent variable.
193

Table 5.22 summarizes the results of the analyses by hypothesis, while in Chapter
6 I discuss these findings further and elaborate on their implications.

Figure 5.22. Summary of hypotheses, research questions, and results
SUMMARY OF RESULTS

STEP 1 STEP 2 STEP 3 STEP 4

Overall Test By Ethnic Group Role of Organizations Role of Organizations
Hypotheses and Research Questions Hierarchical Multiple
Regression
Hierarchical
Multiple Regression
ANOVAs Hierarchical Multiple
Regression
Follow up on significant
ANOVAs
H1a: An individual resident’s integrated
connection into the storytelling network
(i.e., ICSN‐2) will be positively related to
health literacy after controlling for
individual level covariates, including socio‐
economic status, residential tenure,
homeownership status, and ethnicity.
Confirmedfor
prevention‐oriented
health literacy only
Final model does
not fit African
Americans or
Latinos when
examined
separately
Geo‐demographic
reach (CSNI
GD
) of
organizations explains
differences in health
literacy mean scores
Geo‐demographic reach
(CSNIGD) is a significant
predictor, controlling for
individual‐level covariates,
and ICSN‐2
H1b: Individual‐level communication capital
(ICC) will be positively related to health
literacy and will explain more of the
variance among residents’ scores than ICSN‐
2.
Confirmedfor
prevention‐oriented
health literacy only
Final model
confirmed for
African Americans
only

RQ1a: Is an individual resident’s integrated
connection into the storytelling network
(i.e., ICSN‐2) positively related to how easy
they find that it is for them to get medical
care?
Rejected Final model does
not fit African
Americans or
Latinos when
examined
separately
Scope of Inter‐
Organizaitonal
Connectedness (SIOC),
Meso‐level Storytelling
Network
Connectedness
(MSNC), and Reach as
community
engagement (CSNI
CE
)
emerge as significant
factors
Role of SIOC, MSNC, and
CSNI
CE
as predictors
confirmed as significant,
controlling for individual‐
level covariates, and ICSN‐
2

Figure 5.22. Summary of hypotheses, research questions, and results (Continued)

RQ1b: Is individual‐level communication
capital (ICC) positively related to how easy
residents find that it is for them to get
medical care and does it explain more of the
variance among residents’ scores than ICSN‐
2?
Confirmed Final model does
not fit African
Americans or
Latinos when
examined
separately

H2‐1a, 2‐2a, 2‐3a: An individual resident’s
integrated connection into the storytelling
network (i.e., ICSN‐2) will be positively
related to neighborhood belonging (H2‐1),
collective efficacy (H2‐2), and civic
participation (H2‐3), after controlling for
individual level covariates, including socio‐
economic status, residential tenure,
homeownership status, and ethnicity.
Confirmedfor all
three dimensions of
civic engagement:
neighborhood
belonging, collective
efficacy, and civic
participation
Final model fits both
African Americans
or Latinos when
examined
separately in the
case of belonging,
civic participation;
not in the case of
collective efficacy
Organizational
integration into the
STN (OISN
1
and OISN
2
)
explain differences in
belonging and civic
participation; not
collective efficacy
ANOVA results confirmed
only for belonging (and
after controlling for
individual level covariates
and ICSN‐2); not for civic
participation
H2‐1b, 2‐2b, 2‐3b: Individual‐level
communication capital (ICC) will positively
related to neighborhood belonging (H2‐1),
collective efficacy (H2‐2), and civic
participation (H2‐3), after controlling for
individual level covariates and will explain
more of the variance among residents’
scores than ICSN‐2.
Confirmed for all
three dimensions of
civic engagement:
neighborhood
belonging, collective
efficacy, and civic
participation
Final model fits both
African Americans
or Latinos when
examined
separately in the
case of belonging,
civic participation;
fits Latinos only in
the case of
collective efficacy

Figure 5.22. Summary of hypotheses, research questions, and results (Continued)

Hypotheses and Research Questions STEP 1 NOTES for STEP 1 STEP 2
H3a‐1, 3a‐2, 3a‐3: There is an indirect path
of influence from individual‐level
communication capital (ICC) to health
literacy via neighborhood belonging (H3a‐
1), collective efficacy (H3a‐2), and civic
participation (H3a‐3).
H3a‐1: Rejected;
H3a‐2: Rejected;
H3a‐3: Confirmed
*Analyses
confirmed direct
effect of ICC on
prevention health
literacy
Final model produced
in STEP 1 applies to
African Americans and
Latinos. In the case of
African Americans
only, collective efficacy
predicts detection‐
oriented health
literacy
H3b‐1, 3b‐2, 3b‐3: There is an indirect path
of influence from individual‐level
communication capital to residents’
perceived ability to access information and
services they need to safeguard their health
via neighborhood belonging (H3b‐1),
collective efficacy (H3b‐2), and civic
participation (H3b‐3).
H3b‐1: Rejected;
H3b‐2: Rejected;
H3b‐3: Rejected
*Analyses
confirmed direct
effect of ICC on
health care access
When the two ethnic
groups are examined
separately, the ICC to
health care access path
dissipates.
Hypotheses and Research Questions STEP 1 STEP 2

Hierarchical Multiple
Regression
3‐Way ANOVA

H4‐1a, 4‐2a, and 4‐3a: Density of
institutional resources will predict
neighborhood belonging (H
4‐1
), collective
efficacy (H
4‐2
), and civic participation (H
4‐3
),
after controlling for individual level covaria
tes.
H4‐1a: Rejected;H4‐
2a: Confirmed; H4‐
3a: Rejected (in all
cases controlling for
individual level
covariates and ICSN‐
2)

Figure 5.22. Summary of hypotheses, research questions, and results (Continued)

H4‐1b, 4‐2b, and 4‐3b: Density of
institutional resources will predict
neighborhood belonging (H
4‐1
), collective
efficacy (H
4‐2
), and civic participation (H
4‐3
),
after controlling for neighborhood level
ethnic heterogeneity and residential
stability.
Interaction of
organizational
density, ethnic
heterogeneity, and
residential stability
not related to civic
engagement
differences among
residents.
Hypotheses:
Rejected

H5a‐1: Density of institutional resources will
predict health literacy, after controlling for
individual level covariates.
Rejected
H5a‐2: Density of institutional resources will
predict health literacy, after controlling for
neighborhood level ethnic heterogeneity
and residential stability.
Rejected
H5b‐1: Density of institutional resources will
predict residents’ perceived degree of access
to health‐related services, after controlling
for individual level covariates.
Rejected
H5b‐2: Density of institutional resources will
residents’ perceived degree of access to
health‐related services after controlling for
neighborhood level ethnic heterogeneity
and residential stability.
Rejected

Figure 5.22. Summary of hypotheses, research questions, and results (Continued)
Hypotheses and Research Questions STEP 1 NOTES

2‐Way ANOVA

RQ2a: Are there significant differences in
health literacy scores among residents living
in neighborhoods with a high density of
organizational resources and residents living
in neighborhoods where organizational
resources are sparse, which may be
explained by variation in neighborhood
levels of civic engagement?
No differences in
prevention health
literacy explained by
neighborhood
variation in terms of
density of
insitutional
resources.
Interaction of
collective efficacy
(only) and density of
resources accounts
for differences in
detection health
literacy.
*A post‐hoc
analysis indicated a
significant
difference in
detection health
literacy among
residents living in
less efficacious
neighborhoods with
fewer resources and
residents living in
more efficacious
neighborhoods with
higher density of
resources.

RQ2b: Are there significant differences in
perceived access to health care among
residents living in neighborhoods with a
high density of organizational resources and
residents living in neighborhood where
organizational resources are sparse, which
may be explained by variation in
neighborhood levels of civic engagement?
No differences in
health care access
explained by
neighborhood
variation in terms of
density of
insitutional
resources, or the
interaction of density
of resources and
neighborhood levels
of civic engagement.

Figure 5.22. Summary of hypotheses, research questions, and results (Continued)
Hypotheses and Research Questions STEP 1

t‐test

RQ3: Are there significant differences in
individual‐level communication capital
among residents that are accounted for by
ethnic heterogeneity?
No significant
differences

Hypotheses and Research Questions STEP 1 NOTES

χ
2
‐tests

RQ4a: Is there a significant difference
among organizations in Greater Crenshaw
in terms of ethnic reach?
Yes, significant
difference detected.
*The number of
organizations
reporting that they
cater to both
African Americans
and Latinos
significantly larger

RQ4b: Is the number of organizations in
Greater Crenshaw that all residents connect
to significantly different from the number of
organizations to which only residents of a
particular ethnic background connect?
Yes, significant
difference detected.
*Results indicate
that connections to
organizations are
determined by
ethnicity

Figure 5.22. Summary of hypotheses, research questions, and results (Continued)
Hypotheses and Research Questions STEP 1

Hierarchical Multiple
Regression

RQ5a‐5b: Is ethnicity a predictor of health
literacy (RQ5a) and access to health care
resources (RQ5a) controlling for individual
level covariates?
Earlier analysis (see
H1, RQ1) indicated
ethnicity is as
significant predictor
only for prevention
health literacy.
Latinos score higher.
Hypotheses and Research Questions STEP 1 NOTES

2‐Way ANOVA

RQ6a‐6b: Are there significant differences in
health literacy (RQ6a) and perceived access
to health care (RQ6b) among residents
living in neighborhoods with high ethnic
heterogeneity and residents living in
neighborhood where ethnic heterogeneity is
low, which may be explained by variation in
neighborhood levels of community
communication capital?
Interaction of ethnic
heterogeneity and
CCC was found to be
non‐significant.
*Main effect forethnic
heterogeneity was significant,
though. Residents in more
homogeneous communities
travel further to get medical care.
H6a: Differences in health literacy scores
among residents living in stable and
unstable communities will vary as a function
of community communication capital.
Interaction of
residential stability
and CCC was found
to be non‐
significant.

Figure 5.22. Summary of hypotheses, research questions, and results (Continued)
H6b: Differences in health care access
scores among residents living in stable and
unstable communities will vary as a function
of community communication capital.
Interaction of
residential stability
and CCC was found
to be non‐
significant.
*Main effect forstability was
significant, though. Residents in
more stable communities
believed it was easier for them to
get health care.

202

CHAPTER 6
DISCUSSION AND IMPLICATIONS
There were three overarching questions driving this project: Does where we live
matter for our health? What are the mechanisms through which place impacts our health?
And, finally, what can communication research contribute to the development of place-
based solutions for achieving health equity across increasingly ethnically diverse
communities? These questions led to the review of two growing - exponentially it seems
– bodies of literature that are (once again) clearly converging.
1

The first, rooted primarily in sociology, is preoccupied with understanding
neighborhood effects. This research investigates the causal links between a
neighborhood’s structural characteristics and outcomes, such as educational attainment,
crime incidence, and child development. In addition, this work tries to uncover those
social mechanisms or processes through which the neighborhood influences our attitudes,
our behaviors, the beliefs and norms to which we subscribe, our children’s success at
school, our ability to find well-paid, fulfilling jobs, and so forth.
The second line of research that was clearly relevant to this quest is based in
public health research. The influence of epidemiology, as might be expected, is very
strong in this line of inquiry. In recent years, however, public health researchers and
major American agencies funding health-related research have been tuning into the

1
The review of the public health-oriented literature (see, for example, MacIntyre & Ellaway, 2003)
indicates that Hippocrates, in the 5
th
century B.C., was in all likelihood the first to write in his work on Airs,
Waters, and Places, about the connection between environment and the prevalence of certain diseases
among the populations of towns developed in one part of the then known world or another. An account of
more recent history tells us that as early as the 1600s there were researchers seeking to understand if there
was something about living in a city (versus the country) that made a difference to people’s health.
MacIntyre and Ellaway refer, for instance, to the work published in 1662 by Petty and John Graunt,
Natural and Political Observations upon the Bills of Mortality. This was a quantitative analysis of the
weekly bills of mortality compiled in London by parish clerks, which gave numbers and causes of death.

203

findings of sociology, education, economics, and social geography on the links between
place and various aspects of human behavior and life. As a result, they have been
investing an increasingly larger amount of resources into investigating if there are certain
environmental factors – physical, but most importantly from the point of view of this
project, social factors – that impact health. If the role of the environment is important, the
next, obvious question is, how do these influences manifest? And, can we (a) measure
them, (b) control for the undesirable ones, and (c) amplify those influences that we
believe can be beneficial in the short and mid-term, as well as over our entire life cycle?
Theoretical and methodological contributions
The review of this literature suggested, rather surprisingly, that communication is
largely absent as a social mechanism of neighborhood effects. Surprisingly, because the
academic roots of current neighborhood effects literature is found in the research of early
20
th
century Chicago School urban sociologists, such as Park, Burgess, and McKenzie. In
their writings communication (along with transportation) is featured as a primary
mechanism of social organization in the metropolitan areas emerging around the U.S. in
the early1900s. But communication is also surprisingly absent from the work done from a
public health perspective on neighborhood effects, despite the evident interest of public
health researchers and agencies in developing effective health communication campaigns.
The five challenges
These realizations set two initial challenges for this project, which are discussed
in greater detail in Chapters 2 and 3: (a) to re-introduce communication as a social
mechanism through which neighborhood effects operate, and (b) to develop, outline and
test a theoretical model of neighborhood effects that: (i) incorporates communication as a

204

primary social process of influence, and (ii) defines the relationship of communication to
other processes that have been studied to date.
A third goal became part of the endeavor as a result of reviewing the
neighborhood effects literature on how institutional resources available in the
communities we live in influence our lives. There is more extensive research done on this
front that focuses on the impact of institutional resources’ physical presence (i.e., density,
diversity) in neighborhoods, but little work has been done on understanding the role of
these resources as agents or actors that partake in the everyday activities of residential
communities. From a communication point of view, this role seems equally important, if
not more so, than physical presence. For what good is a community-based organization to
the residents of a neighborhood, for example, if its staff does little to engage the
community, understand its problems, and help it devise strategies to tackle them? As
Robert Sampson et al. (2002) put it, the problem with the work done on institutional
resources is that it falls short of “distinguish[ing] well between structural dimensions of
institutions and mediating institutional processes” (p. 458, footnote 8). That provided the
impetus for this study to advance a theoretical framework that incorporated both
dimensions of institutional resources and examined their impact as elements of the
neighborhood environment and as neighborhood actors.
The fourth challenge was both theoretical and methodological in nature. Various
scholars have noted that all too often neighborhood effects studies rely on individual-
level data to create measures that purportedly capture neighborhood characteristics. This
usually involves averaging scores of residents living in a particular area. Doing so
presents fewer problems when measures reflect the average age in a community, for

205

instance, or the median income. But they seem less defensible when they are meant to
capture the robustness of social mechanisms; that is, their capacity to produce effects or
change in a community. Measures that are meant to reflect the significance of a social
process are particularly problematic when the impact of that process depends not solely
on the population of residents, but also on the interaction between residents (i.e.,
individual-level actors) and institutional (meso-level) neighborhood actors.
2
Who should
be considered the holder as well as the building force of a community’s social capital, for
instance? A measure created from survey, interview, or otherwise collected individual
residents’ responses does not account for the role played by neighborhood community-
based organizations that can help create the connective tissue that binds residents together
and creates a sense of belonging to a community.
3

To address this problem, Sampson and Raudenbush (1999) have suggested that it
is time to develop what they have termed ecometric measures (as opposed to
psychometric) that reflect neighborhood properties. Heeding their call, the fourth
challenge for this project involved the development of an ecometric measure of
communication capital as a mechanism of neighborhood effects.
Last but not least, this study set out to determine how a communication-based
model of neighborhood effects could conceptually and practically benefit public health

2
Conceptually it would be correct to assume that macro-level actors can play an important role in how
neighborhood effects manifest in a community. The role of the media as macro-level agents that can affect
the way residents perceive their surroundings, interpret the world around them, and relate to each other has
been researched extensively. The work of Matei & Ball-Rokeach (2005) documenting how the mass media
affect people’s views of Watts as the “epicenter of fear” in Los Angeles, even though over 40 years have
passed since the Watts riots of 1965, illustrates this reality.
3
Some neighborhood or place-based organizations may in fact counter any individual-level efforts to build
a strong sense of community among residents. That is arguably the case with gangs, whose actions can
instill fear in residents and lead to disinvestment in the community by individuals and institutional actors
(e.g., Klein & Maxson, 2006).

206

research, particularly research aimed towards improving health literacy and health care
access among diverse ethnic populations.
Towards a communication-based theoretical approach to neighborhood effects
To address the foregoing five challenges I employed communication
infrastructure theory (or CIT; see Ball-Rokeach et al., 2001; Kim & Ball-Rokeach,
2006a) to complement sociology-based theories of neighborhood effects. CIT represents
an ecological approach to the study of urban communities that is rooted in
communication research. It has been developed through a series of studies conducted by
the research team of the ongoing Metamorphosis Project. Over the course of a decade,
Metamorphosis has focused its efforts on better understanding the transformations
occurring in urban communities under the forces of globalization, population diversity,
and new communication technologies. CIT allows us to study how communication can be
a mechanism of neighborhood effects and specifies the conditions under which its impact
is magnified or diminished. It also defines the communication process as one that takes
place at multiple levels of analysis and across levels of analysis; it therefore enables us to
account for the interaction of micro- and meso-level neighborhood actors.
The communication infrastructure consists of a neighborhood storytelling network
(STN) set in a particular residential communication action context (CAC). The STN
comprises of individuals and institutional level actors that interact in the course of
everyday life. Formally, to date, the actors or storytellers included in the conceptual
model of CIT have included local and ethnically-targeted media, community-based
organizations, and individual residents. As the STN becomes stronger, it also becomes a
more valuable resource to neighborhood actors; one that residents can tap into to get

207

critical information, like where to find a doctor after hours, but also one that community-
based organizations can leverage to mobilize residents and recruit volunteers.
The capacity of the STN to become a valuable community resource depends on
the configuration of the CAC. The CAC consists of all the tangible and intangible
elements of the natural and built neighborhood environment; that is, the physical makeup
of the community (e.g., the layout of the street grid and the availability of grocery stores),
social characteristics such as ethnic diversity, and residents’ feelings and perceptions
about the place they live in (e.g., is it a safe neighborhood for kids to play in). All these
elements can either enable the strengthening of the STN or hamper its development.
Understanding the dual role of institutional neighborhood actors
In this project, I argued that the composition of the STN (as well as the CAC) is
conceptually different depending on what problem, or what rhetorical scholar Lloyd
Bitzer (1968) would call the exigency the storytelling network is turned to or mobilized to
address. For the purposes of this study, which is focused on understanding the role of
communication as a determinant of health literacy and health care access, it was
necessary to consider community-based organizations of a wide variety and health
service providers as neighborhood storytellers. In addition, I suggested that organizations
can play a dual role in a community. They can effectively be considered meso-level
neighborhood storytellers and elements of the communication action context. In many
cases, for example, the density of organizations in an area can be critical. The closing of a
clinic or hospital in a community could have dramatic effects in terms of how easy it is
for residents to get health care. The results of this study indicate that this very likely the
case in Greater Crenshaw. Therefore it is not necessary to consider institutional actors

208

strictly as part of the STN or just as elements of the CAC. They may be considered
simultaneously as neighborhood storytellers and elements of neighborhood structure.
Acknowledging the dual role of organizations in community life helps address
conceptually the shortcomings of the neighborhood effects theoretical frameworks that
try to account for the role of institutional resources in a community, but fail to theorize
(and operationalize) their role as neighborhood actors.
The pragmatic benefits of understanding the dual role
of institutional community resources
From a pragmatic point of view, for researchers and policy-makers interested in
evaluating the performance of grantee organizations or evaluating the resource needs of a
community, the idea of a dual role highlights the need for evaluations to account for:
a) The impact of institutional actors as they become more integrated into the
storytelling network of communities they serve; and for
b) The constraints they face given the environment they are in (e.g., a very
unstable community population-wise).
Incorporating these elements into evaluation would show if, for example,
organizations are in danger of becoming burned out, as they are facing a general lack of
other organizations they can work with. Alternatively, it would also show if there is too
much competition for few important resources, such as funding and volunteers.
Extending CIT: diagnosing a fragmented storytelling network at the micro and meso level
Prior work done based on CIT has noted the possibility for a neighborhood
storytelling network to be fragmented along the lines of ethnicity (see, e.g., Kim, 2003;
Wilkin, 2005). Having data coming directly from the organizations that are active in

209

Greater Crenshaw (in addition to data from a survey of residents) made it possible to
investigate through this project:
a) If the storytelling network in the area is bifurcated; that is, in the case of
Greater Crenshaw, if there is one network constructed by and oriented towards
Latino residents and one built by and focused on African Americans; and
b) If organizations are bridging the divide, strengthening the overall
neighborhood storytelling network or preserving the fragmentation.
A bifurcated neighborhood storytelling network may be weaker, equally vibrant,
or more vibrant than the two ethnically-oriented STNs at work in the same community.
There is no ‘averaging effect’ necessarily, in which case one strong ethnically-oriented
STN and another, weaker one, create a neighborhood-wide STN that falls somewhere
along a weak to strong continuum. The extent to which the neighborhood-wide STN will
be weak depends on:
a) The degree to which there are institutional actors that serve the entire
community (not one or another group more).
b) The extent to which all residents, regardless of ethnic background, connect to
a number of the same institutional resources present in the neighborhood; and
on
c) The strength and quality of the ties that exist among institutional resources in
the community.
4

From a practical standpoint, being able to diagnose a fragmented STN is
important, for at least three reasons:

4
This condition was not directly assessed in this study but should be pursued in future research.

210

(i) It signifies that a community as a whole is likely to be less able to address
and solve problems that affect the entire community and not just one
ethnic group or another.
(ii) It indicates ongoing intergroup relationship tensions or problems that may
require attention in any campaign or effort that is targeted to the entire
area, or particular ethnic groups in the community.
(iii) It suggests to policy-makers and researchers that to effect change in the
community, they must find ways to leverage the two ethnically-oriented
STNs and to bridge the gap between them.
Methodological contributions:
assessing organizations’ integration into the storytelling network
From a methodological point of view, this study generated data that made
possible the creation of measures capturing attributes of community-based organizations,
health service providers, and media as meso-level actors in the neighborhood STN.
Three of the measures created reflect characteristics of the relationships among
actors at the meso-level:
IIOC is a measure of the intensity of inter-organizational connectedness;
that is the average strength of the communication connections between
one organization and all others it is in communication/contact with on a
regular basis.
SIOC reflects the scope of connections that organizations have to other
organizations; or, in other words, the diversity of communication ties with
other organizations that one meso-level actor sustains.

211

MSNC, as the interaction term of IIOC and SIOC, is meant to indicate the
degree of connectedness of one organization to the meso-level of the
storytelling network.
Two measures indicate what audiences/populations an organization connects to
and the degree of connectedness to them. Those indicators are:
CSNI
GD
, which reflects the extent of an organization’s reach in terms of
demographics and geographic areas served or targeted; and
CSNI
CE,
which captures the degree of an organization’s involvement or
engagement in a community.
In addition, one measure, OISN, was created to reflect an organization’s
integration into the neighborhood storytelling network.
OISN is the equivalent of ICSN, proposed and applied first by Kim (2003)
and Kim and Ball-Rokeach (2006b) to capture an individual’s integrated
connection to the neighborhood STN (for a more detailed discussion of
ICSN and the other measures, please see Chapter 3 and 4).
Finally, two more measures were developed to capture the interaction of meso
and micro-level actors:
ICC is a measure of individual-level community capital. It is constructed
as the interaction term of an individual’s connection to the storytelling
network with the average integration of all the organizations that are
active in the individual’s neighborhood (i.e., the average OISN of
organizations in a neighborhood); and

212

CCC, which is a measure of community level communication capital. It is
the interaction term of the average integration score of organizations in a
neighborhood and the average score of integrated connectedness to the
STN of residents (i.e., the interaction of the average OISN and the average
ICSN).
Community communication capital: an ecometric measure
The CCC measure developed was proposed as an ecometric measure of a
community’s communication capital. Communication capital as described in earlier
chapters is a concept akin to that of social capital. Both are realized through ties that
emerge and connect residents in the process of their everyday life. However, social
capital is usually limited to the individual level (i.e., interpersonal networks), and does
not account for the integration of a multi-level network of neighborhood actors (or
storytellers). In addition, the storytelling aspect incorporates the communicative
dynamics not considered in social capital, or in most neighborhood effects studies. Thus,
the CCC differs in being both multi-level and focused upon a communicative process.
While CCC is a community-level characteristic, ICC reflects how much (or how
little) an individual’s communication capital ‘appreciates’ as a result of being integrated
into a STN with a particular composition of meso-level actors.
Major findings and implications
The research questions and hypotheses articulated in this study were aimed
towards accomplishing three goals:
(a) To investigate the role of individual and community-level communication
capital in building health literacy and access to health care in diverse urban communities;

213

(b) To assess if the integrated multi-level storytelling network is a significant
mechanism of neighborhood effects, how it performs, and what its relationship is to other
more commonly studied social processes of neighborhood effects (i.e., in this case,
neighborhood belonging, collective efficacy, and civic participation); and
(c) To understand if and how three elements of the communication action context
– density of institutional resources, ethnic heterogeneity, and residential stability – relate
to health literacy and health care access, but also if and how they enable and constrain the
storytelling network and civic engagement considered as mechanisms of neighborhood
effects. The impact of organizational density was of particular interest, because we hoped
to gain further insight into how the dual role of institutional actors as elements of the
neighborhood environment and as meso-level storytellers manifests in a community.
Residents’ integration into the STN, a significant predictor of prevention health literacy
The degree to which residents connect to their neighbors, local and ethnically-
targeted media, as well as community-based organizations – i.e., the degree to which they
are integrated into their neighborhood’s storytelling network – emerged as an important
factor in predicting prevention-oriented health literacy (even after controlling for
individual-level covariates). This suggests that the community communication resource,
which the STN represents, can and should be leveraged by policy-makers, health-oriented
institutional actors, and activists groups in developing place-based solutions for the
improvement of health literacy.
The fact that individuals’ integration into the STN was positively related to
prevention-oriented health literacy and not symptoms-detection health literacy is curious
however. It could potentially be attributed to the fact that knowledge about particular

214

symptoms of diseases is harder to develop and sustain, because they can be very specific
to a disease. In comparison, some of the preventive measures recommended by
physicians (such as increasing the intake of fruits and vegetables, for instance) can be
applied to a wider spectrum of diseases. This may make them easier to digest, remember,
and share with other people. Alternatively, the lack of a link between the storytelling
network and health literacy oriented towards detecting symptoms of a disease could mean
that a particular storytelling network is turned or focused more on prevention rather than
detection. Community-based organizations may be, for instance, dedicating much more
of their time and efforts into educating parents on how to prevent diseases their children
or themselves may be at risk for, but less effort on how to detect (or manage) a disease.
Individual-level communication capital and health literacy:
the power is in the interaction
In creating the measure of individual-level communication capital, it was assumed
that the organizations have the capacity to support individuals’ ability to effectively
address problems, in this case pertaining to their health. The analyses reported in Chapter
5 indicate that organizations can in fact contribute to a resident’s communication capital.
They can bolster individuals’ ability to connect to and stay integrated into those
community information resources that will help them get the pertinent information they
need to make important decisions about their everyday life.
While the integrated connection of residents to the STN (reflected in the ICSN-2
measure) was generally found to be positively related to prevention health literacy, the
interaction of organizations’ connectedness to the STN with that of residents’ was found
to be a stronger predictor of health literacy than ICSN alone. These findings suggested to

215

us that the degree of organizations’ integration into the STN might itself predict health
literacy; meaning that organizations may have an independent impact on the health
literacy levels of the neighborhoods they serve. If that was the case, then we could
conclude that organizations are in a position to compensate for the low levels of
connectedness to the STN of some residents, which might be expected in relatively
unstable, transient communities. That led to a series of follow up analyses to examine if
there were particular STN-related attributes of organizations that made them especially
effective as meso-level components of the STN mechanism of health effects.
Organizations matter, but the onus is on the residents: the ideal scenario versus reality
ANOVAs indicated to us that only organizations’ geo-demographic reach
mattered; it explained, said the test, differences in health literacy among people living in
neighborhoods where organizations reached out to the broadest demographic base in a
geographically limited, local area and people living in neighborhoods where
organizations served a narrow demographic base that lived in a geographic area that
extended past the boundaries of a neighborhood (e.g., Los Angeles County, Southern
California). We were predicting a difference, and the results suggested that there was in
fact a difference; yet, it was not the one we anticipated. Residents in neighborhoods
where organizations scored lower on geo-demographic reach by the standards we set (i.e.,
they had a more circumscribed demographic base and broader geographic base) fared
better with respect to health literacy. Before making too much out of the result, geo-
demographic reach was entered into a multiple hierarchical regression to figure out if the
effect the ANOVA was pointing to was significant after accounting for socio-economic

216

factors, residents’ neighborhood experience, but also their personal integration into the
storytelling network. The results of the regression analysis confirmed the finding.
a. What is driving the geo-demographic reach effect?
If geo-demographic reach is a significant factor in predicting health literacy,
independent of individuals’ integration into the storytelling network and in the direction
indicated by the results, there could me interesting implications for policy-makers and
health communication professional. The obvious question is: why do organizations with a
narrower demographic base and wider geographic base have a larger impact on residents’
health literacy scores? One possible explanation is that the effect of geo-demographic
reach is driven mainly by larger organizations that, per their mission, have a larger
audience or target population; that would include organizations such as the Los Angeles
County Department of Public health, for instance. Their impact might be more significant
because they have the specialized knowledge to deal with particular at-risk populations
and diseases, and they also have substantial enough financial resources to fund disease
prevention-oriented health campaigns. In an ideal scenario, these organizations would
play their role, but they would also link up to smaller, grassroots organizations that can
work more closely with residents and local neighborhood stakeholders to help translate
the knowledge conveyed by larger organizations into everyday practices.
Because of the methodological constraints imposed on the analysis, as described
in Chapters 4 and 5, we had to rely on a dummy-coded version of the geo-demographic
reach variable. Therefore confirming these findings through an analytical scheme that
would allow for the use of a less crude version of CSNI
GD
is necessary.

217

b. Organizations are unable to compensate, but they still matter
Organizations integration into the storytelling network, overall, does not seem to
have an independent effect on health literacy. The OISN measure accounted for no
differences among residents with respect to health literacy. That means that meso-level
actors are, in all likelihood, unable to compensate for whatever difficulties residents’ are
having connecting to information resources and improving their knowledge about health
risks they may face. The fact the organizations on their own do not seem to have an
impact on health literacy, yet the interaction of their integration into the storytelling
network and that of residents does (i.e., individual-level communication capital matters),
suggests that organizations become critical resources for activist residents who seek out
and connect to organizations for support. This does explain the fact that ICC amplifies
the predictive power of ICSN-2 in terms of prevention-health literacy, whereas OISN and
other organizational-level specific variables (with the potential exception of CSNI
GD
)
make no difference.
Communities are not monolithic: dealing with a fragmented STN
Given the results of the analyses pertaining to prevention health literacy and the
question of whether ethnicity matters when it comes to storytelling network integration,
we were curious as to whether the general pattern observed across Greater Crenshaw
neighborhoods was the same when looking at African Americans and Latinos separately.
In the case of African Americans, the predictors of health literacy were the same as the
ones that emerged from the analyses focused on the entire sample of Greater Crenshaw
residents. That was not the case, however, in the case of Latino residents. In their case,
the analysis suggested that neither an individual’s integration into the storytelling

218

network nor the interaction of organizations integration with that of residents could
predict prevention-oriented health literacy. In fact, out of the variables included in the
hierarchical regression models, only age emerged as a significant predictor, with older
Latinos exhibiting higher levels of health literacy.
This is a clear example of where it would be important for organizations to be
able to play a compensatory role. The Latino population in Greater Crenshaw is fairly
new, especially compared to the African American population which has been in the area
for many decades. Being ‘new’ means that Latinos may not have established as of yet
those communicative ties that are essential in building a strong storytelling network. In
this context, organizations could function as support and stabilizing forces, by helping
residents forge those ties and bridges to the institutions that are in the area, as well as the
communication resources that African Americans have built and relied on over the course
of years. Moreover, in creating bridges between the old and new residents, institutional
actors could offset the negative effects of a bifurcated neighborhood storytelling network.
However, they seem to be failing to do so.
Based on these results, we hypothesized that either (a) Latinos are not connecting
to institutional actors in the area (at least ones with a greater capacity to effect change in
the community) and/or that (b) institutional actors are not oriented towards them.
The effects of an ethnic group divide and a bifurcated STN on health literacy
Prima facie the hypothesis that organizations in Greater Crenshaw are not
oriented towards Latinos seems to be rejected by an analysis done to evaluate what is the
ethnic reach of organizations in the area. Analyzing the organizational data suggests that
most organizations (over half, i.e., 52%) reported that it is oriented towards both ethnic

219

groups in the area. Twenty seven percent said they target primarily African Americans or
Latinos, while another 21% reported targeting everyone in the study area, regardless of
ethnic background. However, an analysis of survey data on what sports and recreation,
cultural, ethnic and religious organizations, as well as neighborhood associations, and
political and educational organizations Latinos and African Americans in Greater
Crenshaw perceive as being the most important for them, showed us that only 6 out of
149 organizations mentioned by all residents in the sample were mentioned by both
Latinos and African Americans. That is 4% of the total number reported. These findings:
Suggest that organizations are not (at least not yet) succeeding in bridging
the ethnic divide, even though that might be their desire.
Confirm the bifurcation of the storytelling network in Greater Crenshaw.
At least with respect to health literacy, African Americans can rely on the
information resources of the STN to gain knowledge about how to prevent
a disease, whereas Latinos cannot.
Positive signs of change and the challenge of inclusivity
On a positive note, it seems likely that the influx of Latinos in Greater Crenshaw
in recent years may have led to the creation of a number of new organizations, which
may need time to mature and establish themselves before their presence can be felt. The
fact that in the case of both Latinos and African Americans, the role of (a) organizational
integration into the storytelling network on its own, as well as the interaction of
organizations’ and residents’ STN integration (reflected in the measure of individual-
level communication capital) was a significant predictor of neighborhood belonging may

220

be suggesting that either (i) the efforts of new organizations are beginning to have some
effect, or that (b) older ones are slowly becoming more inclusive.
The case of SCOPE is illustrative of how challenging it can be for an organization
to be inclusive in the context of a very diverse community. Founded after the 1992 civil
unrest in Los Angeles, SCOPE's mission is to build power of communities most impacted
by poverty, racism, and discrimination in order to achieve social and economic justice
(www.scopela.org). SCOPE achieves its mission by combining community organizing,
strategic alliance-building, community-based research and analysis, training and
capacity-building to build effective movements for social change at the local, state, and
national levels. The organizations seeks to address the underlying economic and political
disadvantages in Los Angeles' underserved communities – that is, the lack of economic
opportunity, civic participation, and policy solutions based on community needs.
While initially developed as an African American organization, SCOPE has
changed along with the demographics in South Los Angeles over the years. A very
conscious and strategic decision was made to incorporate Latino residents, but the shift
has not been an easy one to make. The decision to incorporate Latinos into SCOPE's
membership was very carefully planned; leaders had to think through potential obstacles
as they worked to build a multiethnic organization. Working with two very different
populations who spoke different languages meant that a special effort would have to be
implemented to ensure that all new members felt welcomed and accepted. Today,
membership and staff meetings are supplemented with active translations from English to
Spanish and vice versa in order to overcome language barriers. SCOPE believes that
Latino and African American residents can improve their communities and their

221

standards of living if they work together for a common cause. While cultural and social
divides may exist between these two populations, SCOPE makes an effort to bridge these
divides through active interaction and dialogue. Rather than focusing on issues such as
competition over limited resources, the organization asks its members to question why
their entire community is underserved and how to change the economic and political
disadvantages of all South Los Angeles residents.
The example of SCOPE demonstrates the amount of resources that need to be
dedicated to truly accomplish inclusivity and greater ethnic reach in a diverse
community. SCOPE has a fairly long history in South Los Angeles, dedicated staff, and
considerable resources to draw on to pursue these goals. Many organizations though do
not. As they struggle, sometimes on a monthly basis, more often on a year-by-year basis
to stay afloat financially, finding the time, financial, and human resources to achieve
greater ethnic reach may be out of the question.
This reality should be instructive to researchers for one additional reason.
Organizations may report that they do target, serve and reach out to multiple ethnic
communities, but self-reported reach may only reflect a mission statement, a vision, and
good intentions. Looking deeper into how a particular organization is achieving
inclusivity is a must for understanding the capacity of organizations to bridge ethnic
divides, but it is also necessary to inform intervention campaigns and policy-making.
Health care access and the STN
Examining the links of the neighborhood storytelling network to health care
access, once again, both individuals’ integration into the storytelling network and

222

individual-level communication capital predict residents’ perceived ability to access
health care resources. There is a twist, however and a rather unexpected one at that.
Both the multiple regression analyses conducted to test Research Question 1, as
well as the structural equation models run for investigating Hypothesis 3 suggest that the
integration of individuals into the STN and the interaction of individual integration with
that of organizations have a significant and negative effect on health care access. The
more connected residents are to the storytelling network, the more difficult they feel it is
for them to get the health care they need. The integration of organizations into the
storytelling network only amplifies this negative effect. The question is why? We would
expect that by becoming part of the STN, residents would have more access to better
information about where to go to find health care they need for themselves and their
families.
Do your homework: the importance of knowing your research site well
In the months prior to us conducting the survey of residents in Greater Crenshaw
the potential and then eventual closure of the Martin Luther King, Jr./Drew Medical
Center located in South Los Angeles was among the top stories reported by most Los
Angeles news media. The African American community of South Los Angeles seemed
the most vocal in expressing concern that the closure of the medical center would leave
the already under-resourced community, with even fewer health resources.
From a CIT point of view, residents that are most connected to the STN are the
ones most likely to have heard the news and be aware of the changing conditions in the
community. Knowing that health care resources in the community were about to become
even more scarce, after the closure of a major medical center, could have influenced how

223

difficult residents integrated into the STN perceived accessing health care to be. A bad
story could have had a much larger negative impact than expected.
To examine this post-hoc hypothesis, we ran two analyses using an item on the
Metamorphosis Greater Crenshaw survey that asked residents: “How much will the
closure of this trauma center affect you and your family on a 10 point scale, where ‘10’
means a great deal and ‘1’ means it would not affect you at all?” The first analysis was a
correlation, in which we predicted that the more strongly residents would feel that the
King/Drew facility closure would affect them, the more difficult they would feel it would
be to access health care resources. The correlation was significant and negative as
predicted, with r=-.29, p<.001.
5

The second analysis was to determine if the King/Drew closure effect was
significant only for African Americans, who seemed to have been the most vocal in their
opposition to the closure of the facility, or if it was equally significant for Latinos. An
independent-samples t-test was conducted to evaluate if there was a significant difference
between Latinos and African Americans with respect to how much they believed the
closure of King/Drew would affect them. The test indicated no significant difference, as
t(564)=2.76, p>.50. That led us to conclude that, most likely, the negative relationship
between STN integration and perceived difficulty accessing health care is a
neighborhood-wide issue and not specific to one ethnic group.
Negative STN effects and using the STN to turn things around
The results from the analysis of the relationship between the STN and health care
access in Greater Crenshaw are instructive for policy-makers who may need to counter

5
Negative, because the perceived difficulty accessing health care measure ranges from 1 to 4, where
1=very difficult and 4=very easy.

224

negative perceptions acerbated by events like the closing of the King/Drew trauma
facility. Engaging the storytelling network, telling different stories, empowering
organizations to help residents address shared concerns are all necessary parts of any
prescription to better health care access.
The role of organizations in the STN with respect to health access
Analyses of how specific organizational-level variables were related to health care
access confirmed that institutional actors can indeed amplify what individual integration
into the storytelling network can do, but they also provided evidence of organizations
having independent effects on health care access. Scope of inter-organizational
connectedness (SIOC), degree of integration among actors at the meso-level (MSNC), as
well as reach as community engagement (CSNI
CE
), were all associated with expected
differences in perceived access to health care. The broader the scope of inter-
organizational connectedness, the stronger the connections among storytellers at the meso
level, and the greater the reach (as community engagement) of organizations into a
community, the more difficult residents believed it was to get health care services. The
disturbing network news on King/Drew could have travelled more rapidly through a
strong inter-organizational network and impacted organizations’ members, clients, staff,
and residents who just connect to these organizations for various services. The challenge
that the King/Drew trauma center closure presents, of course, is how to harness the power
of this powerful storytelling network to turn things around.
The integrated multilevel STN & other neighborhood effects mechanisms
One of the main objectives of this study was to further our understanding of
communication and particularly how the communication infrastructure (Ball-Rokeach et

225

al., 2001; Kim & Ball-Rokeach, 2006a) is related as a social process to other more
commonly studied processes of neighborhood effects. Three dimensions of civic
engagement were selected for the purposes of this study: neighborhood belonging,
collective efficacy and civic participation. Prior research has indicated that the degree to
which an individual is integrated into their neighborhood storytelling network is a strong
predictor of civic engagement (e.g., Kim & Ball-Rokeach, 2006b). The question raised at
the onset of this study was two-pronged:
(a) Can organizational actors contribute to the relationship between STN
integration and civic engagement? And,
(b) Can communication be considered a more elementary process of
neighborhood effects, an antecedent to processes like neighborhood
belonging, collective efficacy, and civic participation?
The role of organizations’ integration into the storytelling network in civic engagement
The analyses confirm prior research on the link between the STN integration and
civic engagement. Results also indicate that the interaction of individuals’ integration into
the storytelling network and that of organizations in the community strengthens residents’
attachment to their neighborhood, makes residents feel that they can come together to
solve common problems, and increases residents’ participation in civic activities.
Generally the same patterns observed for the entire sample from Greater Crenshaw also
applied to the two separate ethnic groups. Only in the case of African Americas did
ICSN-2 and ICC not predict collective efficacy.
What is more clearly gleaned from the results is that the degree of organizations’
integration into the STN (captured in the OISN measure), independent of ICSN, is

226

associated with higher levels of belonging among residents. Residents that live in
neighborhoods where organizations are less integrated into the STN feel significantly less
attached to their community than residents living in neighborhoods where organizations
are highly connected to the STN. A hierarchical multiple regression suggested that this is
the case even after controlling for individual-level socio-economic status, residential
tenure, home ownership, but also individual residents’ connectedness to the STN. In the
case of belonging it appears that organizations can in fact not only amplify, but also
compensate for difficulties that individual residents’ face in their attempts to build
neighborhood belonging.
The impact of individual-level communication capital on health literacy
via civic engagement
The structural equation modeling analyses conducted suggest that individual-level
communication capital does not have an indirect effect on health literacy and access to
health care via neighborhood belonging and collective efficacy, but that it does have an
indirect effect via civic participation. The findings, albeit surprising at first, seem
plausible considering the nature of the outcomes we focused on. Belonging and collective
efficacy are more important mechanisms for solving common neighborhood problems;
issues that is that require a more significant mobilization of concerned citizens. But
health issues, even though they may in fact be of concern to an entire community, they
are most likely first and foremost a personal or family matter, in which case belonging
and collective efficacy are less likely to be relevant. Civic participation may be, however.
Civic activities, especially those that bring together residents in shared space,
create opportunities for people to mingle with neighbors and friends. There, on a one-to-

227

one basis, individuals are more likely to share information about a health problem they
are facing or ask for advice from a trusted neighbor, friend, or a neighbor they know (or
just discovered) is a doctor. To the extent that civic participation creates opportunities for
interpersonal communication, the indirect link from individual level communication
capital to prevention-oriented health literacy via civic participation makes sense.
Of course this finding also suggests to policy-makers or institutional actors that if
in fact there are health issues that a community should come together to address, then
there is much work to be done in building up that capacity. To accomplish this,
harnessing the power of the integrated multilevel STN will be crucial. The first step
would be to use the STN to develop a definition of a health problem as a shared concern.
Getting media to work with local health service providers to develop stories that would
educate the community, increase the salience of a particular health issue, and urge
residents to work together to deal with the problem would be key in this direction.
The role of the communication action context in neighborhood effects
Three elements of the communication action context were examined in this study
in relation to the STN, but also as determinants of civic engagement, health literacy, and
health care access; those elements were: the density of organizational resources, ethnic
heterogeneity, and residential stability.
Role of organizational density in civic engagement
Earlier, we reported that the interaction of organizational integration into the
storytelling network with individuals’ integration into the STN is good predictor of civic
engagement. Probing further into the question of how the dual role of organizations plays
out in a community, we examined if the density of organizational resources was

228

significantly related to the civic engagement dimensions included in this study:
neighborhood belonging, collective efficacy, and civic participation. Hierarchical
multiple regressions, in which we were able to control for socioeconomic status and other
individual level covariates, as well as individuals’ integration into the storytelling
network, suggest that whether a resident lives in a neighborhood with a dense network of
resources or in a sparse neighborhood environment with respect to institutional resources
matters only in the case of collective efficacy.
While unclear at this point why organizational density did not matter in the case
of neighborhood belonging and civic participation, the finding seems to indicate that the
presence of more institutional resources bolsters residents’ sense that they can come
together to solve common problems. Through a three-way ANOVA we also tested
whether the role of organizational density in collective efficacy was somehow impacted
by ethnic heterogeneity and residential stability. The analysis indicated that the role of
organizational density in collective efficacy is unrelated to whether residents live in less
or more ethnically heterogeneous or stable communities.
Institutional resource density and health care access
The relative presence or absence of institutional resources in a community does
not appear to play a role in predicting health literacy; at least it does not on its own.
However an analysis of the effect of the interaction of organizational density and
collective efficacy on health literacy, suggested that residents living in neighborhoods
more dense in resources and where there is a heightened sense of collective efficacy score
higher on the symptoms-detection related health literacy. It does seem plausible that a
community richer in institutional resources and where residents feel that they can count

229

on their neighbors to deal with shared concerns would also be conducive for the
circulation of important health information that could improve residents’ health literacy.
However, further implications are unclear at this point.
The impact of the CAC on individual-level communication capital in Greater Crenshaw
Organizational resource density, ethnic heterogeneity and residential stability
were shown not to have any significant effect or make a significant difference with
respect to individuals’ integration into the storytelling network or individual-level
communication capital. While the STN may be under other neighborhood constraints not
examined in this study, the three that were do not seem to have a notable impact on it.
The communication action context and health literacy
Two findings warrant mentioning, albeit in both cases more research is required
to better understand the potential implications.
a. Ethnic heterogeneity and distance to medical care
The analysis suggests that residents living in more homogeneous neighborhoods
travel further to receive necessary medical care compared to those living in more
heterogeneous neighborhoods. It is unclear if this is because more ethnically
homogeneous areas in Greater Crenshaw are also more disadvantaged and under-
resourced areas, or if there are very ethnically homogeneous and primarily higher-income
residential areas in which, because of land use regulations, there are no health facilities.
b. Residential stability and difficulty getting health care
In addition, residents living in more stable communities reported less difficulty
getting the health care they need. This could be because residents in more stable
communities have accumulated knowledge about available resources. But it could also be

230

that more stable areas are also neighborhoods with higher income residents, which tend to
have better access to health care services.
Limitations
There are several limitations to this study that future research should try to
account for. The first one is methodological in nature. Having to rely on individual-level
data that had already been collected, it was impossible to impose at the onset of the study
those sampling parameters necessary that would allow us to employ multi-level analytical
procedures (e.g., hierarchical linear modeling, general linear mixed modeling procedures)
without problems. The fact that some residents misinterpreted the survey question asking
them to indicate what the cross-streets nearest their home were, or that some did not
provide an answer at all, limited our ability to identify the census tract within which
every respondent lived. After identifying all the Census tracts that we could, it became
evident that the number of respondents per Census tract was small enough to disallow the
use of hierarchical linear modeling. By abandoning the initial analytical plan, which
included HLM, we had to seek an alternative. The final analytical scheme employed,
while useful, was less direct than the one based primarily on HLM. Multi-level modeling
procedures are also built to handle nested data (as in the case when the researcher is
interested in understanding how residents living in one neighborhood are different from
residents in surrounding neighborhoods) more efficiently than analytical tools used
primarily for analysis at one level only. HLM is more sensitive to the interplay of the
individual and meso-level and can differentiate and account better for changes that occur
in the dependent variables of interest as a result of both sources of variation: between and
within neighborhoods differences.

231

The second limitation stems from the nature of the sample of organizations
recruited. Despite the rather significant efforts to recruit a substantial number of media
and health service providers, very few media agreed to participate. Because of their great
importance as information resources it will be important for future studies attempting to
capture what happens at the nexus of the micro and meso level of storytelling
community, to include as many media as possible. Health service providers were more
accessible, but their staff is often overburdened with work and getting them to participate
can present a considerable challenge. For this study, we had to postpone several
interviews multiple times and, in some cases, cancel them completely.
The third limitation is related to the measures of prevention and symptoms
detection-oriented health literacy used for this study. In the case of Latinos we used
survey items that reflected these residents’ knowledge about how to prevent and detect
symptoms of diabetes. In the case of African Americans we used similar items pertaining
to hypertension. The rationale guiding the construction of the measures (for more details,
please see Chapter 4) was that these were diseases that each group has been found to be
at higher risk for developing, and therefore the scores of respondents’ from both ethnic
groups should be comparable. Future research however may want to consider more
closely if there are differences between diseases like diabetes and hypertension. Not from
a medical perspective, of course, but from the point of view of public health history, in
the sense that some diseases may have been at the center of attention of the public health
community for much longer than others. That could mean that they have also been on the
radar of the general public longer as well. It is likely that hypertension, compared to
diabetes, has been a health concern that the medical and public health community has

232

been trying to educate the public on for a longer period of time. That may be an important
factor in hypertension health literacy, even though the salience of diabetes in recent years
appears to have grown significantly.
Suggestions for future research
Being the first study that tried to directly measure the integration of
organizational/meso-level actors into the storytelling network and to capture the impact
of the interaction between meso and micro-level neighborhood storytelling network
actors, much work lies ahead to further test, refine, and establish the conceptual utility
and the reliability of the organizational-level variables created. Having to use dummy-
coded versions of many of these measures made it impossible to test them more
extensively in the context of this study. The analysis indicated that geo-demographic
reach, reach as community engagement, and the measure of organizational integration
into the storytelling network are likely important in understanding the role of
organizations in the communication infrastructure. Those would be the measures that I
would start working on bettering first. Doing so would add not only to the conceptual and
operational advancement of communication infrastructure theory, but also to the
understanding of those “mediating processes,” Sampson et al. (2002) referred to, through
which institutional actors become mechanisms of neighborhood effects.
Moreover, it would be of interest to investigate what the configuration of the
storytelling network looks like at the meso level and whether there are (a) structural
characteristics of the network and (b) attributes of meso-level actors that matter with
respect to organizations’ ability to help build civic engagement in a community, to foster
the improvement of health literacy, and health care access. One question of interest that

233

we could not answer through this study was whether or not the diversity of organizational
types at the meso level has any effect on how neighborhood influences manifest. Does it
matter, for instance, in developing a health campaign focused on improving health
literacy what kind of community-based organizations exist in a community? Does it
matter more what types of links the organizations that do exist in a community have to
other organizations? To answer these sorts of questions employing social network
analysis techniques could be particularly useful.
For this project, I focused on two health-related outcomes, health literacy and
health care access. That is due, in large part, to the fact that Greater Crenshaw
neighborhoods and residents fare worse than any other health services provision area
(i.e., SPAs) of Los Angeles County with respect to health care access and several
prevention and health outcomes, such as: (a) percent of adults who are obese, (b) diabetes
death rate, (c) rate of adults diagnosed with hypertension, and (d) coronary heart disease
death rate. Obviously the importance of pursuing the line of research into how the
storytelling network can be leveraged to create positive change with regard to health is
paramount. But to be able to further validate the role of communication as a primary
social process of neighborhood effects, variations of the communication-based model
presented here should be tested in other contexts as well. Does communication matter, for
example, with respect to educational attainment, child development, or crime incidence in
a community?
Finally, through this study I more formally developed and operationalized the idea
of communication capital, which was first proposed by Ball-Rokeach (2003). I discussed
community communication capital as an ecometric measure; a measure that is that

234

reflects a community-level attribute. Which raises the question: how should community
be defined? It is a big issue that I do not intend to take on right now (for an informative
review, you may refer to Jeffres, 2002). But one point should be made. For the purposes
of this study the residential neighborhood was operationalized as the Census tract. By a
number of accounts this is a defensible operationalization of neighborhood; as designed,
Census tracts are meant to approximate the size of a neighborhood and in designing these
tracts the Census takes into account physical and other neighborhood characteristics. But
the idea of using these administrative units as proxies for neighborhoods (something done
extremely often) has been challenged and a number of researchers have sought alternative
operationalizations of ‘neighborhood.’ Future research focusing on the dual role of
institutional actors in residential communities should investigate further into what the
definition of a neighborhood by an organization’s standards is and whether shifting units
of analysis, from Census tracts to something else (e.g., school districts or a more
subjective definition of neighborhood) would affect the outcomes of a study.

235

REFERENCES
Adler, N.E., Boyce, W. T., Chesney, M. A., Folkman, S., & Syme, S. L. (1993).
Socioeconomic inequalities in health: no easy solution. Journal of the American
Medical Association, 269, 3140-3145.

Ajzen, I., & Fishbein, M. (1977). Attitude-behavior relations: A theoretical analysis and
review of empirical research. Psychological Bulletin, 84, 888-914.

Ajzen, I., & Fishbein, M. (1980). Understanding the attitudes and predicting social
behavior. Englewood Cliffs, NJ: Prentice-Hall, Inc.

Alesina, A., & La Ferrara, E. (2000). Participation in heterogeneous communities.
Quarterly Journal of Economics, 115, 747-904.

Altschuler, A., Somkin, C. P., & Adler, N.E. (2004). Local services and amenities,
neighborhood social capital, and health, Social Science & Medicine, 59 (6),
1219-1229.

Anderton, D. L., Anderson, A., Rossi, P. H., Oakes, J. M., Fraser, M. R., Weber, E. W.,
& Calabrese, E. J. (1993). Hazardous-waste facilities – environment equity issues
in metropolitan areas. Evaluation Review, 18 (2), 123-140.

Andrulis, D. P. (2000). Community, service, and policy strategies to improve health care
access in the changing urban environment. American Journal of Public Health, 90
(6), 858-862.

Andrulis, D. (1998). Access to care is the centerpiece in the elimination of socioeconomic
disparities in health. Annals of Internal Medicine, 129, 412-416.

Aneshensel, C. S., & Sucoff, C. A. (1996). The neighborhood context of adolescent
mental health. Journal of Health and Social Behavior, 37, 293-310.

Arnold, C. L., Davis, T. C., Berkel, H. J., Jackson, R. H., Nandy, I., & London, S. (2001).
Smoking status, reading level, and knowledge of tobacco effects among low-
income pregnant women. Preventive Medicine, 32 (4), 313-320.

Bailey, T., & Waldinger, R. (1991). Primary, secondary, and enclave labor markets: a
training systems approach. American Sociological Review, 56(4), 432-445.

Baldwin, L. M., Larson, E.H., Connell, F.A., Nordlund, K.C. (1998). The effect of
expanding Medicaid prenatal services on birth outcomes. American Journal of
Public Health, 88(11), 1623-1629.

236

Ball-Rokeach, S., Kim, Y.-C., Wilkin, H., Matsaganis, M., & Moran, M. (2008).
Communication environments and health literacy: a multi-site study of diabetes
specific health literacy among African Americans. Grant proposal (R01)
submitted to the National Institutes of Health (NIH).

Ball-Rokeach, S. J. (2003).Report to the Carnegie Mellon Foundation, University of
Southern California. Los Angeles, CA.

Ball-Rokeach, S.J, Kim, Y.C. & Matei, S. (2001). Storytelling neighborhood: Paths to
belonging in diverse urban environments. Communication Research, 27, 4, 392-
427.

Bean, F. D., & Stevens, G (2003). America's newcomers and the dynamics of diversity.
New York: Russell Sage Foundation.

Beers, B. B., McDonald, V. J., Quistberg, D. A., Ravenell, K. L., Asch, D. A., & Shea, J.
A. (2003). Disparities in health literacy between African American and non-
African American primary care patients. Abstract. Journal of General Internal
Medicine, 18 (Supplement 1), 167-169.

Benasich, A. A., Brooks-Gunn, J., & Clewell, B.C. (1992). Journal of Applied
Developmental Psychology, 13(3), 311-362.

Benson, J. G., & Forman, W. B. (2002). Comprehension of written health care
information in an affluent geriatric retirement community: use of the test of
functional health literacy. Gerontology, 48 (2), 93-97.

Bitzer, L. F. (1968). The rhetorical situation. Philosophy and Rhetoric, 1, 1-14.

Boje, D. M. (1991). The storytelling organization: A study of story performance in an
office-supply firm. Administrative Science Quarterly, 36, 106–126.

Bollen, K. A. (1979). Structural equations with latent variables: New York: John Wiley.

Booth, K. M., Pinkston, M. M., & Poston, W. S. C. (2005). Obesity and the built
environment. J Am Diet Assoc., 5 (1),110-117.

Bourdieu, P. (1986). The forms of capital. In J. Richardson (Ed.), Handbook of theory
and research for the sociology of education (pp. 241–258). New York:
Greenwood Press.

Boyce, M. E. (1990). Collective centering and collective sense-making in the stories and
storytelling of one organization. Organization Studies, 16, 107–137.

Brauer, C. M. (1982). Kennedy, Johnson, and the War on Poverty. The Journal of
American History, 69 (1), 98-119.

237

Braveman, P., Oliva, G., Miller, M. G., Reiter, R., & Egerter, S. (1989). Adverse
outcomes and lack of health insurance among newborns in an eight-county area of
California, 1982 to 1986. The New England Journal of Medicine, 321, 508-513.

Brenner, H., Chang-Claude, J., Seiler, C. M., Stürmer, T., & Hoffmeister, M. (2007).
Potential for colorectal cancer prevention of sigmoidoscopy versus colonoscopy:
population-based case control study. Cancer Epidemiology, Biomarkers, &
Prevention, 16, 494-499.

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

Brooks-Gunn, J., Duncan, G. J., & Aber, J. L. (Eds.) (1997). Neighborhoods poverty
(vol.1-2). New York: Russell Sage Foundation Press.

Brooks-Gunn, J., Duncan, P., Klebanov, P. K., & Sealand, N. (1993). Do neighborhoods
influence child and adolescent development? American Journal of Sociology,
99(2), 353-95.

Bullard, R. D. (1990). Dumping in Dixie: race, class, and environmental quality.
Boulder, CO: Westview.

Burdette, H. L., & Whitaker, R. C. (2005). Neighborhood playgrounds, fast food
restaurants, and crime relationships to overweight in low-income preschool
children. Prev Med.,115 (5), 57-63.

California Newsreel (2008). Un-natural causes: is inequality making us sick? Produced
by California Newsreel with Vital Pictures. PBS documentary series.

Carpiano, R. M. 2006). Toward a neighborhood resource-based theory of social capital
for health: can Bourdieu and sociology help? Social Science & Medicine, 62(1),
165-175.

Cheong, P. H., Wilkin, H. A. & Ball-Rokeach, S. J. (2004). Diagnosing the
communication infrastructure in order to reach target audiences: A case study of
Hispanic communities in Los Angeles. In Whitten, P. & D. Cook (Eds.),
Understanding health communications technologies: A case study approach, pp.
101-110. Jossey-Bass Publications.

Chin, M. H., Zhang, J. X., & Merrell, K. (1998). Diabetes in the African-American
Medicare population: morbidity, quality of care, and resource utilization. Diabetes
Care, 21 (7), 1090-1095.

Clark, R. L. (1992, December). Neighborhood effects on dropping out of school among
teenage boys. Discussion paper. Washington, D.C.: The Urban Institute.

238

Coleman, J.S. (1988). Social capital in the creation of human capital. American Journal
of Sociology, 94 (Supplement), S95-S120.

Coulton, C., Cook, T., & Irwin, M. (2004). Aggregation issues in neighborhood
research: a comparison of several levels of census geography and resident
defined neighborhoods. Center on Urban Poverty and Social Change, Mandel
School of Applied Social Sciences. Case Western Reserve University, Cleveland,
OH. Paper retrieved from
http://povertycenter.cwru.edu/urban_poverty/dev/pdf/appam1104.pdf.

Coulton, C. J., Korbin, J. E., Chan, T., & Su, M. (2001). Mapping residents’ perceptions
of neighborhood boundaries: a methodological note. American Journal of
Community Psychology, 29 (2), 371-373.

Coulton, C. J., Korbin, J. E., Su, M. (1999). Neighborhoods and child maltreatment: a
multi-level study. Child Abuse & Neglect, 23, 1019–1040

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

Coulton, C. J., & Pandey, S. (1992). Geographic concentration of poverty and risk to
children in urban neighborhoods. American Behavioral Scientist, 35, 237-257.

Council on Scientific Affairs for the American Medical Association (1999). Health
literacy: report for the AMA council on scientific affairs. Journal of the American
Medical Association, 281, 552–557.

Crombie, I. K., Kenicer, M. B., Smith, W C. S., & Tunstall-Pedoe, H. D. (1989).
Unemployment, socio-environmental factors, and coronary heart disease in
Scotland. British Heart Journal, 61, 172-177.

Cunningham, W., Hays, R., Ettl, M., Dixon, W. J., Liu, R. C-C., Beck, C. K., & Shapiro,
M. D. (1998). The prospective effect of access to medical care on health-related
quality-of-life outcomes in patients with symptomatic HIV disease. Medical Care,
36 (3), 295-306.

Devesa, S. S., & Diamond, E. K. (1983). Socioeconomic and racial differences in lung
cancer incidence. American Journal of Epidemiology, 46 (10), 1141-1149.

Diez-Roux, A. V. (2001). Investigating neighborhood and area effects on health.
American Journal of Public Health, 91 (11), 1783-1789.

Doheny-Farina, S. (1996). The wired neighorhood. New Haven, CT: Yale University
Press.

239

Ellen, I., Mijanovich, T., & Dillman, K.-N. (2001). Neighborhood effects on health:
exploring the links and assessing the evidence. Journal of Urban Affairs, 23 (3-4),
391-408.

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, 379-426.

Ennett, S., T., Flewelling, R. L., Lindrooth, R. C. & Norton, E. C. (1997). School and
neighborhood characteristics associated with school rates of alcohol, cigarette,
and marijuana use. Journal of Health and Social Behavior, 38(1), 55-71.

Felson, M. (1998). Crime and everyday life. Thousand Oaks, CA: Sage.

Felson, M. & Cohen, L.E. (1980). Human ecology and crime: a routine activity approach.
Human Ecology, 8(4), 389-406.

Finnegan, J. R., & Viswanath, K. (1988). Community ties and use of cable TV and
newspapers in a Midwest suburb. Journalism Quarterly, 65, 456–463, 473.

Fishbein, M. & Ajzen, I. (1975). Belief, Attitude, Intention and Behavior: An Introduction
to Theory and Research. Boston: Addison-Wesley.

Flores, G., Abreu, M., Olivar, M. A., Kastner, B. (1998). Access barriers to health care
for Latino children. Archives of Pediatrics & Adolescent Medicine, 152 (11),
1119-1125.

Freeman, R. (1991). Employment and earnings of disadvantaged young men in a labor
shortage economy. In Jencks, C., & P. Peterson (Eds.), The urban underclass,
Washington D.C.: The Brookings Institution.

Frisbie, W. P., & Kasarda, J. D. (1977) Spatial processes. In Handbook of sociology (N.
J. Smelser, ed.), pp. 629–667. Sage Publications, Newbury Park, CA.

Foster, E. M., & McLanahan, S. (1996). An illustration of the use of instrumental
variables: do neighborhood conditions affect a young person’s chance of finishing
high school? Psychological Methodology, 1 (3), 249-260.

Ganz, M. L. (2000). The relationship between external threats and smoking in central
Harlem. American Journal of Public Health, 90 (3), 367-371.

Garcia, R. (2008, April). The City Project. Presentation at Path to Healthy Communities:
Place-Based Solutions to Achieving Health Equity. California Pan-Ethnic Health
Network (CPEHN).

240

Garner, C. L., & Raudenbush, S. W. (1991). Neighborhood effects on educational
attainment: a multilevel analysis. Sociology of Education, 64, 251-262.

Gephart, M. A. (1997). Neighborhoods and communities as contexts for development. In
Brooks-Gunn, J., Duncan, G. J., & Aber, J. L., Neighborhood poverty: context
and consequences for children (Vol. I), pp. 1-43. New York: Russell Sage
Foundation.

Gephart, R. P. (1991). Succession, sensemaking, and organizational change: A story of a
deviant college president. Journal of Organizational Change Management, 4, 35–
44.

Geronimus, A. T. (1992). The weathering hypothesis and the health of African American
women and infants: evidence and speculations. Ethnicity & Disease, 2 (3), 207-
221.

Glossop, R. (1997). Bronfenbrenner’s ecology of human development: A reappreciation.
In A. Pence (Ed.), Ecological research with children and families: From concepts
to methodology, pp. 1-15. New York: Teachers College Press.

Goodchild, M. F. (1992). Geographical Information Science. International Journal of
Geographical Information Science, 6, 31-45.

Gordon-Larsen, P., Nelson, M. C., Page, P., & Popken, B.M. (2006). Inequality in the
built environment underlies key health disparities in physical activity and obesity.
Pediatrics, 117 (2), 417-424.

Gramlich, E., Laren, D., & Sealand, N. (1992). Moving in and out of poor urban areas.
Journal of Policy and Management, 11(2), 273-287.

Griffin, R. J., & Dunwoody, S. (2000). The relation of communication to risk judgment
and preventative behavior related to lead in tap water. Health Communication,
12(1), 81-107.

Greer, S.A. (1962). The emerging city. New York: Free Press

Greer, S.A. (1962). Governing the metropolis. New York: John Wiley and Sons.

Guest, A. M., Almgen, G., & Hussey, J. M. (1998). The ecology of race and
socioeconomic distress: infant and working age mortality in Chicago.
Demography, 35 (1), 23-34.

Haapanen, N., Milunpalo, S., Vuori, I., Oja, P. & Pasanen, M. (1997). International
Journal of Epidemiology, 26, 739-747.

241

Haas, J. S., Phillips, K. A., Sonneborn, D., McCulloch, C. E., Baker, L. C., Kaplan, C. P.,
Pérez-Stable, E. J., & Liang, S.-Y. (2004). Variation in access to health care for
different racial/ethnic groups by the racial/ethnic composition of an individual’s
county of residence. Medical Care, 42 (7), 707-714.

Hart, C. Ecob, R., & Smith, G. D. (1997). People, places and coronary heart disease risk
factors: a multilevel analysis of the Scottish heart health study archive. Social
Science Medicine, 45 (6), 893-902.

Hayduk, L.A. (1987). Structural equation modeling with LISREL: essentials and
advances. Baltimore, MD: John Hopkins University Press.

Hindman, D. B., Littlefield, R., Preston, A., & Neumann, D. (1999). Structural pluralism,
ethnic pluralism, and community newspapers. Journalism and Mass
Communication Quarterly, 76(2), 250-263.

Hipp, J. R. (2007). Block, tract, and levels of aggregation: neighborhood structure and
crime and disorder as a case in point. American Sociological Review, 72 (5), 659-
680.

Horowitz, C.R., Colson, K. A., Hebert, P. L., & Lancaster, K. (2004). Barriers to buying
healthy foods for people with diabetes: evidence of environmental disparities.
American Journal of Public Health, 94 (9),1549-1554.

Hox, J. J. (1998). Multilevel modeling: When and how. In Balderjahn, I. Mathar, R. &
Schader, M.(Eds.). Classification, data analysts, and data highways. New York:
Springer-Verlag.

Huckfeldt, R., & Sprague, J. (1987). Networks in context: the social flow of political
information. American Political Science Review, 81 (4), 1197-1216.

Institute of Medicine of the National Academies (2004). Health literacy: a prescription
to end confusion. Washington, D.C.: The National Academies Press.

Institute of Medicine of the National Academies (2003). Unequal treatment: confronting
racial and ethnic disparities in healthcare. Washington, D.C.: The National
Academies Press.

Iyer, S., Kitson, M., & Toh, B. (2005). Social capital, economic growth and regional
development. Regional Studies, 39, 1015-1040.

Jargowski, P.A., & Bane, M. J. (1991). Ghetto poverty in the United States, 1970–1970.
In Jencks C, & P. Peterson, (Eds), The urban underclass, pp. 235-273.
Washington, DC: Brookings Institute.

242

Jeffres, L. W. (2002). Urban communication systems: Neighborhoods and the search for
community. Cresskill, NJ: Hampton Press, Inc.

Jeffres, L.W., Dobos, J., & Sweeney, M. (1987). Communication and commitment to
community. Communication Research, 14(6), 619-643.

Jencks, C., & Mayer, S. E. (1990). The social consequences of growing up in a poor
neighborhood. In Lynn Jr., L. E., & McGeary, MGH (Eds.), Inner-city poverty in
the United States (pp. 111-186). Washington, D. C.: National Academies Press.

Jenkins, C. D. (1983). Social environment and cancer mortality in men. New England
Journal of Medicine, 308 (7), 395-398.

Jöreskog, K. G., & Sörbom, D. (1996). LISREL 7: user’s reference guide. Chicago, IL:
Scientific Software International.

Kawachi, I., & Berkman, L. F. (Eds.). (2003). Neighbrohoods and health. New York:
Oxford University Press.

Kawachi, I. (1999). Social capital and community effects on population and individual
health. Annals of the New York Academy of Sciences, 869, 120-130.

Kawachi, I., Kennedy, B. P., Lochner, K., & Prothrow-Stith, D. (1997). Social capital,
income inequality, and mortality. American Journal of Public Health, 87 (9),
1491-1498.

Keller, S. (1977). The telephone in new and old communities. In de Sola Pool, I. (Ed.),
The social impact of the telephone (pp. 281-298). Cambridge, MA: MIT Press.

Kickbusch, I. S. (2001). Health literacy: addressing the health and education divide.
Health Promotion International, 16 (3), 279-297.

Kim, Y. & Ball-Rokeach, S. (2006a). Community storytelling network, neighborhood
context, and civic engagement. Human Communication Research, 32 (4), 411-
439.

Kim, Y. & Ball-Rokeach, S. (2006b). Civic engagement from a communication
infrastructure perspective. Communication Theory, 16 (2), 173-197.

Kim, Y.-C., (2003). Storytelling community: communication infrastructure and civic
engagement in urban spaces. Unpublished doctoral dissertation, University of
Southern California, Los Angeles.

243

Kim, Y. C., Ball-Rokeach, S. J., Cohen, E. L., & Jung, J.-Y. (2002). Metamorphosis of
civic actions post September 11th: From local storytelling networks to national
action. In Greenberg, B. (Ed.), Communication and terrorism (pp. 289–304).
Cresskill, NJ: Hampton Press.

Klein, M.W., & Maxson, C. L. (2006). Street gang patterns and policies. New York:
Oxford University Press.

Kurtz, E. M., Koons, B. A., & Taylor, R. B. (1997). Land use, physical deterioration,
resident-based control, and calls for service on urban streetblocks. Justice
Quarterly, 15 (1), 121-149.

Katz, V. S. (2007). From conversation to conversion: children’s efforts to translate their
immigrant families’ social networks into community connections. Unpublished
doctoral dissertation, University of Southern California (Los Angeles, California).

Lee, S.-Y. D., Arozullah, A. M., & Cho, Y. I. (2004). Health literacy, social support, and
health: a research agenda. Social Science & Medicine, 57, 1309-1321.

Leventhal, T., & Brooks-Gunn, J. (2003a). Children and youth in neighborhood contexts.
Current Directions in Psychological Science. American Psychological
Association, 27-31.

Leventhal, T., & Brooks-Gunn, J. (2003b). Moving to opportunity: an experimental study
of neighborhood effects on mental health. American Journal of Public Health, 93
(9), 1576-1582.

Lin, W.-Y., & Song, H. (2006).Geo-ethnic storytelling: An examination of ethnic media
content in contemporary immigrant communities. Journalism, 7(3), 362 – 377.

Lillie-Blanton, M. L. (1999). A review of the nation’s progress and challenges in assuring
access to health care for low-income Americans. In Lillie-Blanton, M. L.,
Martinez, R., Lyons, B., & Rowland, D. (Eds.), Access to health care (pp. 19-37).
Washington, D.C.: Kaiser Family Foundation.

Lee, V.E., Brooks-Gunn, J., Schnur, E., & Liaw, F.R. (1990). Are Head Start effects
sustained? A longitudinal follow-up comparison of disadvantage children
attending Head Start, no preschool, and other preschool programs. Child
Development, 61(2), 495-507.

MacIntyre, S., & Ellaway, A. (2003). Neighborhoods and health: and overview. In
Kawachi, I., & Berkman, L. F. (Eds.), Neighborhoods and Health (pp. 20-42).
New York: Oxford University Press.

244

Mark, D. M., Chrisman, N., Frank, A. U., McHaffe, P. H., Pickles, J. (1996). The GIS
history project. Retrieved on April 22, 2005 from
http://www.geog.buffalo.edu/ncgia/gishist/bar_harbor.html.

Marsden, P. V. (1990). Network data and measurement. Annual Review of Sociology, 16,
435-463.

Marsden, P. V., & Campbell, K. E. (1984). Measuring tie strength. Social Forces, 63 (2),
482-501.

Matei, S, A., & Ball-Rokeach, S. J. (2005). Watts, the 1965 Los Angeles riots, and the
communicative construction of the fear epicenter of Los Angeles. Communication
Monographs, 72 (3), 301-323.

Matei, S., Ball-Rokeach, S.J, & Qiu, J.L. (2001). Fear and misperception of Los Angeles
urban space: A spatial-statistical study of communication-shaped mental maps.
Communication Research, 27, 4, 429-463.
http://matei.org/research/publications/matei_fear_CR.pdf

Massey, D. S., & Eggers, M. L. (1990). The ecology of inequality: minorities and the
concentration of poverty, 1970-1970. American Journal of Sociology, 95, 1153-
1177.

Massey, D. S., & Denton, N. A. (1979). Hypersegregation in U.S. metropolitan areas:
black and Hispanic segregation along five dimensions. Demography, 26, 373-391.

McLeod, J. M., Scheufele, D. A., & Moy, P. (1999). Community, communication and
participation: The role of mass media and interpersonal discussion in local
participation. Political Communication, 16, 315–336.

McLeod, J. M., Daily, K., Guo, Z. S., Eveland, W. P., Jr., Bayer, J., Yang, S. C., et al.
(1996). Community integration, local media use, and democratic-processes.
Communication Research, 23, 179–209.

McMahon, L. F., Wolfe, R.A., Huang, S., Tedeschi, P., Manning, W., Edlund, M. J.
(1999). Racial and gender variation in use of diagnostic colonic procedures in the
Michigan Medicare population. Medical Care, 37 (7), 712-717.

Melvin, P. M. (1987). The organic city: urban definition and community organization,
1880-1920. Lexington: University Press of Kentucky.

Minkler, M. (1992). Community organizing among elderly poor in the United States: a
case study. International Journal of Health Services, 22 (2), 303-316.

Monge, P., & Contractor, N. (2003). Theories of Communication Networks. New York:
Oxford University Press.

245

Moore, R. D., Stanton, D., Gopalan, R., & Chaisson, R. E. (1994). Racial differences in
the use of drug therapy for HIV disease in an urban community. The New
England Journal of Medicine, 330 (11), 763-768.

Morenoff, J., Sampson, R. J., & Raudenbush, S. (2001). Neighborhood inequality,
collective efficacy, and the spatial dynamics of urban violence. Criminology, 39,
517-560.

Morrison, R. S., Wallenstein, S., Natale, D. K., Senzel, R. S., & Huang, L.-L. (2000).
“We don’t carry that” – Failure of pharmacies in predominantly nonwhite
neighborhoods to stock opioid analgesics. New England Journal of Medicine, 342
(14), 1023-1026.

Morley, G. (1997). Functional health literacy: a primer. Princeton, NJ: Center for Health
Care Strategies, Inc.

Moy, P., McCluskey, M. R., McCoy, K., & Spratt, M. A. (2004). Political correlates of
local news media use. Journal of Communication, 54, 532–546.

National Center for Health Statistics (2005). Health, United States. Retrieved April 20,
2007 from www.cdc.gov/nchs/hus.htm.

Okten, C., & Osili, U. (2004). Contributions in heterogeneous communities: evidence
from Indonesia. Journal of Population Economics, 17, 603-626.

Orellana, M. F., Dorner, L., & Pulido, L. (2003). Accessing assets: immigrant youth’s
work as family translators or “paraphrasers.” Social Problems, 50, 505-524.

Orozco, M. (2003, September). Hometown associations and their present and future
partnerships: new development opportunities? Inter-American Dialogue.
Washington, D.C.: USAID.

Park, R. E., Burgess, E., & McKenzie, R. D. (1925/1967). The city. Chicago: University
of Chicago Press.

Peeples, F., & Loeber, R. (1994). Do individual factors and neighborhood context explain
ethnic differences in juvenile delinquency? Journal of Quantitative Criminology
10, 141-157.

Peterson, R. D., Krivo, L.J., & Harris, M. A. (2000). Disadvantage and neighborhood
violent crime: do local institutions matter? Journal of Research in Crime and
Delinquency, 37, 31-63.

Phillips, W. C. (1940). Adventuring democracy. New York: Social Unit Press.

246

Pickett, K. E., & Pearl, M. (2001). Multilevel analyses of neighborhood socioeconomic
context and health outcomes: a critical review. Journal of Epidemiology &
Community Health, 55, 111-122.

Portes, A. (1997). Social capital: origins and applications. Annual Review of Sociology,
24, 1-24.

Putnam, R. (2000). Bowling alone: The collapse and revival of American community.
New York: Simon & Schuster.

Putnam, R. (1995). Bowling alone: America’s declining social capital. Journal of
Democracy(6)1, 65-77.

Rask, K., Williams, M., Parker, R., & McNagny, S. E. (1994). Obstacles in predicting
lack of regular provider and dealys in seeking care for patients at an urban public
hospital. Journal of the American Medical Association, 271, 1931-1933.

Raudenbush, S. W., & Bryk, A. S. (2002). Hierarchical linear models: applications and
data analysis methods (2nd edition). Newbury Park, CA: Sage.

Ratzan, S. C. (2001). Health literacy: communication for the public good. Health
Promotion International, 16 (2), 207-214.

Ratzan, S., & Parker, R., (1999). Introduction—current bibliographies in medicine 2000-
2001: Health literacy, January 1990 through October 1999. In Selden, C., Zorn,
M., Ratzan, S., & Parker, R. (Eds.), Health Literacy [bibliography online] (pp. 1-
19). Bethesda, MD: National Library of Medicine.

Raudenbush, S. (2003). The quantitative assessment of neighborhood social
environments. In Kawachi, I., & Berkman, L.F. (Eds.), Neighborhoods and health
(112-131). New York: Oxford University Press.

Raudenbush, S., & Sampson, R. J. (1999). Ecometrics: toward a science of assessing
ecological settings, with application to the systematic social observation of
neighborhoods, Sociological Methodology, 29, 1-41.

Reagans, R., & McEvily, B (2003). Network structure and knowledge transfer: the effects
of cohesion and range. Administrative Science Quarterly, 48 (2), 240-267.

Reichman, N., & Florio, M. (1996). The effects of enriched prenatal care services on
Medicaid birth outcomes in New Jersey. Journal of Health Economics, 15, 455-
476.

Rice, T. W., & Steele, B. (2001). White ethnic diversity and community attachment in
small Iowa towns. Social Science Quarterly, 72, 397-407.

247

Robert, S. A. (1999a). Neighborhood socioeconomic context and adult health: the
mediating role of individual health behaviors and psychological factors. Annals of
New York Academy of Science, 896, 465-468.

Robert, S. A. (1999b). Socioeconomic position and health: the independent contribution
of socioeconomic context. Annual Review of Sociology, 25, 489-516.

Roberts, M., & McCombs, M. (1994). Agenda setting and political advertising: Origins
of the news agenda. Political Communication, 11, 249-262.

Ross, C. E. (2000). Neighborhood disadvantage and adult depression. Journal of Health
and Social Behavior, 41, 177-177.

Rountree, P. W., & Warner, B. D. (1999). Social ties and crime: is the relationship
gendered? Criminology, 37(4), 779-713.

Sampson, R. J. (2001). How do communities undergird or undermine human
development? Relevant contexts and social mechanisms. In Booth, A., & N.
Crouter (Ed.), Does It Take a Village? Community Effects on Children,
Adolescents, and Families, pp. 3–30. Mahwah, NJ: Lawrence Erlbaum.

Sampson, R. J., McAdam, D., McIndoe, H., & Weffer-Elizondo, S. (2005). Civil society
reconsidered: the durable nature and community structure of collective civic
action. American Journal of Sociology, 111 (3), 673-714.

Sampson, R.J. & Morenoff, J.D. (2004a). Spatial (dis)advantage and homicide in Chicago
Neighborhoods. In M. F. Goodchild & D.G. Janelle (Eds.), Spatially integrated
social science, pp. 145-170. Oxford: Oxford University Press.

Sampson, R. J., & Morenoff, J. D. (2004b). Durable inequality: spatial dynamics, social
processes, and the persistence of poverty in Chicago neighborhoods. Manuscript
prepared for Bowles, S., Durlauf, S., & Hoff, K. (Eds.), Poverty Traps. Princeton
University Press.

Sampson, R., Morenoff, J., & Gannon-Rowley, T. (2002). Assessing “neighborhood
effects”: Social processes and new directions in research. Annual Review of
Sociology 27, pp. 443-477.

Sampson, R. J., Raudenbush, S. (1999). Systematic Social Observation of Public Spaces:
A New Look at Disorder in Urban Neighborhoods, American Journal of
Sociology, 105, 603-651.

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

248

Sayer, B., & Peterfreund, N. (1993). Insurance, income, and access to ambulatory care in
King County, Washington. American Journal of Public Health, 83, 1583-1587.

Schillinger D., Wang, F., Grumbach, K., Daher, C., Palacios, J.; Diaz-Sullivan, G., Piette,
J., & Bindman, A. B. (2002). Association of health literacy with diabetes
outcomes. Journal of the American Medical Association, 288, 475-482.

Schmitz, J., & Fulk, J. (1991). Organizational colleagues, media richness, and electronic
mail: a test of the social influence model of technology use. Communication
Research, 18(4), 487-523.

Schneider, E. C., Leape, L. L., Weissman, J. S., Piana, R. N., Gatsonis, C., & Epstein, A.
M. (2001). Racial differences in cardiac revascularization rates: does “overuse”
explain higher rates among White patients? Annals of Internal Medicine, 135 (5),
328-337.

Schultz, A. J., Williams, D. R., Israel, B. A., & Lempert, L. B. (2002). Racial and spatial
relations as fundamental determinants of health in Detroit. The Milbank
Quarterly, 80 (4), 677-707.

Schultz, A. J., Parker, E. A., Israel, B. A., & Fisher, T. (2001). Social context, stressors,
and disparities in women’s health. Journal of the American Medical Women’s
Association, 56, 143-149.

Schultz, A. J., & Lempert, L. B. (2000, August). “Being part of the world”: Detroit
women’s perceptions of health and the social environment. Paper presented at the
annual meeting of the American Sociological Association, Anaheim, California.

Shah, D.V., McLeod, J. M., & Yoon, S. H. (2001). Communication, context, and
community: An exploration of print, broadcast, and Internet influences.
Communication Research, 27, 464-506.

Small, M. L. (2003). Culture, cohorts, and social organization theory: understanding local
participation in a Latino housing project. American Journal of Sociology, 107, 1-
54.

Song, H. (2007). Acculturation and Latino immigrants' health attitudes and behaviors: a
bi-dimensional acculturation model approach. Unpublished doctoral dissertation.
University of Southern California, Los Angeles, CA.

Soto, R. (2008, April). California Center for Public Health Advocacy. Presentation at
Path to Healthy Communities: Place-Based Solutions to Achieving Health Equity.
California Pan-Ethnic Health Network (CPEHN).

249

Subramanian, S. V., Jones, K., & Duncan, C. (2003). Multilevel methods for public
health research. In Kawachi, I., & Berkman, L.F. (Eds.), Neighborhoods and
Health (pp. 65-11). New York: Oxford University Press.

Susser, M., & Susser, E. (1996). Choosing a future for epidemiology: I. Eras and
paradigms. American Journal of Public Health, 86, 668-673.

Tichenor, P. J., Donohue, G. A., & Olien, C. N. (1970). Mass media flow and differential
growth in knowledge. Public Opinion Quarterly, 34, 159-170.

Troutt, D. D. (1993). The thin red line: how the poor still pay more. San Francisco, CA:
Consumers Union of U.S., Inc., West Coast Regional Office.

U.S. Department of Health and Human Services. (2001). Healthy people 2010:
Understanding and improving health. Washington D.C.

U.S. Department of Health and Human Services (2000, November). Healthy People 2010
(2nd ed). With Understanding and Improving Health and Objectives for
Improving Health (2 Volumes). Washington, DC: U.S. Government Printing
Office, November 2000.

Valente, T. W., & Saba, W. P. (2001). Campaign exposure and interpersonal
communication as factors in contraceptive use in Bolivia. Journal of Health
Communication, 6, 303-322.

Veysey, B. M., & Messner, S. F. (1999). Future testing of social disorganization theory:
an elaboration of Sampson and Groves’s community structure and crime, Journal
of Research in Crime and Delinquency, 36, 156-174.

Waldinger, R. D., & Lichter, M. I. (2003). How the other half works: immigration and
the social organization of labor. Berkeley: University of California Press.

Wanderman, A., & Nation, M. (1998). Urban neighborhoods and mental health:
psychological contributions to understanding toxicity, resilience, and
interventions. American Psychologist, 53 (6), 647-656.

Warner, B. D., & Rountree, P. W. (1997). Local ties in a community and crime model:
questioning the systemic nature of informal social control. Social Problems, 44,
520-536.

Wassernan, S., & Faust, K. (1994). Social network analysis: methods and applications.
New York: Cambridge University Press.

250

Weiss, B. D., Blanchard, J. S., McGee, D. L., Hart, G., Warren, B., Burgoon, M., &
Smith, K. J. (1994). Illiteracy among Medicaid recipients and its relationship to
health care costs. Journal of Health Care for the Poor & Underserved, 5 (2), 99-
111.

Wilkin, H. A., & Gonzalez, C. (2006, June). Are Spanish language television shows
connecting Los Angeles Latino residents’ to their health storytelling networks?
Paper presented at the 56th Annual International Communication Association
Conference, Dresden, Germany.

Wilkin, H.A., Ball-Rokeach, S. J., Matsaganis, M. D., & Cheong, P. H. (2007).
Comparing the communication ecologies of geo-ethnic communities: How people
stay on top of their community. Journal of Electronic Communication, 17 (1-2).

Wilkin, H. A. (2005). Diagnosing communication connections: reaching underserved
communities through existing communication ecologies. Unpublished doctoral
dissertation, University of Southern California, Los Angeles.

Williams, M. V., Baker, D. W., Parker, R. M., & Nurss, J. R. (1998). Relationship of
functional health literacy to patients' knowledge of their chronic disease: A study
of patients with hypertension and diabetes. Archives of Internal Medicine,
158,166-172.

Williamson, C. (2004). Geographic information and society: Course notes and guide.
University of Southern California, Los Angeles.

Wilson, M. (2001). Leading the community chorus: The role of organizations in society.
Unpublished doctoral dissertation. University of Southern California, Los
Angeles.

Wilson, M. (2002a). Communication, organizations, and diverse populations. In F.
Sudweeks & C. Ess (eds.), Third International Conference on Cultural Attitudes
towards Technology and Communication. (pp. 67-77). Murdoch University,
Australia: Murdoch University Press.

Wilson, M. (2002b) Community organizations, communication infrastructure, and
community. Paper presented to the National Communication Association
Conference. New Orleans, LA: November 21-24, 2002.

Wilson, W. J. (1977). The truly disadvantaged: the inner city, the underclass, and public
policy. Chicago, IL: University of Chicago Press.

World Health Organization (2007). Interim statement of the Commission on Social
Determinant of Health. Accessed via
http://www.who.int/social_determinants/resources/interim_statement/en/index.ht
ml on March 2008.

251

Wuensch, K. L. (2005). Using SPSS to screen data. Downloaded from
http://core.ecu.edu/psyc/wuenschk/spss/Screen-SPSS.doc on April 3, 2008.

Wyatt, R. O., Katz, E., & Kim, J. (2000). Bridging the spheres: Politics and personal
conversation in public and private spaces. Journal of Communication, 50, 71–92.

Zenk, S. N., Schultz, A. J., Israel, B. A., James, S. A., Bao, S., & Wilson, M. L. (2005).
Neighborhood racial composition, neighborhood poverty, and the spatial
accessibility of supermarkets in metropolitan Detroit. American Journal of Public
Health, 95 (4), 660-667.
252

APPENDICES
Appendix A
Health Outcomes: Health Literacy-Related Items

Prevention measure (e.g., for diabetes and hypertension)

Q. There are often things that doctors and medical professionals recommend that you do
to reduce your chances of having various health problems. These things include
suggestions for what you eat and drink, exercising, quitting smoking, using condoms,
having medical exams, performing self examinations, etc. Can you tell me what, if
anything, can help you prevent from…

¾ Hispanics/Latinos only

Q. Becoming diabetic?
1. MEDICAL/DENTAL CHECK-UPS/TESTS
2. CHECK BLOOD SUGAR
3. CHECK BLOOD PRESSURE
4. CHEW SUGAR-FREE GUM/EAT SUGAR FREE CANDY
5. DIET PILLS
6. DON’T EAT/SKIP MEALS
7. DRINK LESS ALCOHOL, WINE OR BEER
8. DRINK MORE WINE OR BEER
9. DRINK LESS CAFFEINE (E.G., COFFEE, SODA, TEA, ETC.)
10. DRINK MORE CAFFEINE (E.G., COFFEE, SODA, TEA, ETC.)
11. EAT BREAKFAST
12. EAT LESS CHOCOLATE
13. EAT MORE CHOCOLATE
14. EAT LOW CARB/CARBOHYDRATE FOODS/AVOID CARBS (E.G., ATKINS AND SOUTH
BEACH DIETS)
15. EAT/DRINK LESS FRUITS AND VEGETABLES
16. EAT/DRINK MORE FRUITS AND VEGETABLES
17. EAT/DRINK LOW FAT FOODS/AVOID FATTY FOODS (INCLUDE FRIED FOODS)
18. EAT LESS RED MEAT
19. EAT MORE RED MEAT (E.G., STEAK, HAMBURGERS)
20. EAT LESS SALT/SALTY PRODUCTS (E.G., CHIPS)
21. EAT MORE SALT/SALTY PRODUCTS (E.G., CHIPS)
22. EAT LESS SUGAR/NO SUGAR (E.G., CANDY, CHOCOLATE, COOKIES, CAKES, ETC.)
23. EAT MORE SUGAR (E.G., CANDY, CHOCOLATE, COOKIES, CAKES, ETC.)
24. EXERCISE/BE MORE ACTIVE/WALK EACH DAY
25. LOSE WEIGHT /DIET
26. WEIGH SELF REGULARLY/WATCH WEIGHT/MAINTAIN HEALTHY WEIGHT
27. QUIT/AVOID SMOKING/SMOKEY ENVIRONMENTS
28. QUIT/AVOID CHEWING TOBBACCO
29. REDUCE/AVOID STRESS OR STRESSFUL SITUATIONS
30. TAKE VITAMINS/HERBS
31. TAKE BIRTH CONTROL PILLS
32. DON’T TAKE BIRTH CONTROL/ESTROGEN
33. USE CONDOMS/SPERMACIDES
34. DON’T USE CONDOMS/SPERMACIDES
253

35. DON’T USE DEODORANT
36. FLOSS
37. BRUSH TEETH DAILY/USE TOOTHPASTE
38. USE FLORIDE/MOUTHWASH
39. DRINK MORE WATER
40. SELF-EXAM
41. EAT MORE CHICKEN OR FISH
42. EAT LESS PORK
43. INCREASE CALCIUM/MILK INTAKE
44. LOWER CHOLESTEROL
45. TAKE MEDICINE/FOLLOW DOCTOR’S ORDERS
46. CHURCH/RELIGION/PRAY/POSITIVE OUTLOOK
47. MORE FIBER/HIGH FIBER DIET
48. KNOW FAMILY HISTORY
49. EDUCATE SELF ABOUT HEALTH PROBLEMS
50. SLEEP/REST
51. ACTIVE SEX LIFE/HAVE SEX MORE
52. MODERATE SEX LIFE/LIMIT SEX
53. URINATE FREQUENTLY/DO NOT HOLD IN
54. DON’T GET HIT IN CHEST
55. BREAST FEED
56. WEAR PROPER BRAS/DON’T WEAR BRA OFTEN
57. DON’T GET BREAST IMPLANTS
58. NONE/NO WAY TO PREVENT GETTING DIABETES
59. (66) OTHER1 (SPECIFY ______________________)
60. (77) OTHER2 (SPECIFY ______________________)
61. (88) DON’T KNOW
62. (99) REFUSED

¾ African Americans only

Q. Getting hypertension or high blood pressure?
1. MEDICAL/DENTAL CHECK-UPS/TESTS
2. CHECK BLOOD SUGAR
3. CHECK BLOOD PRESSURE
4. CHEW SUGAR-FREE GUM/EAT SUGAR FREE CANDY
5. DIET PILLS
6. DON’T EAT/SKIP MEALS
7. DRINK LESS ALCOHOL, WINE OR BEER
8. DRINK MORE WINE OR BEER
9. DRINK LESS CAFFEINE (E.G., COFFEE, SODA, TEA, ETC.)
10. DRINK MORE CAFFEINE (E.G., COFFEE, SODA, TEA, ETC.)
11. EAT BREAKFAST
12. EAT LESS CHOCOLATE
13. EAT MORE CHOCOLATE
14. EAT LOW CARB/CARBOHYDRATE FOODS/AVOID CARBS (E.G., ATKINS AND SOUTH
BEACH DIETS)
15. EAT LESS FRUITS AND VEGETABLES
16. EAT MORE FRUITS AND VEGETABLES
17. EAT LOW FAT FOODS/AVOID FATTY FOODS (INCLUDE FRIED FOODS)
18. EAT LESS RED MEAT (E.G., HIGH PROTEIN DIETS – ATKINS)
19. EAT MORE RED MEAT (E.G., STEAK, HAMBURGERS)
20. EAT LESS SALT/SALTY PRODUCTS (E.G., CHIPS)
21. EAT MORE SALT/SALTY PRODUCTS (E.G., CHIPS)
22. EAT LESS SUGAR/NO SUGAR (E.G., CANDY, CHOCOLATE, COOKIES, CAKES, ETC.)
254

23. EAT MORE SUGAR (E.G., CANDY, CHOCOLATE, COOKIES, CAKES, ETC.)
24. EXERCISE/BE MORE ACTIVE/WALK EACH DAY
25. LOSE WEIGHT
26. WEIGH SELF REGULARLY/WATCH WEIGHT
27. QUIT/AVOID SMOKING/SMOKEY ENVIRONMENTS
28. QUIT/AVOID CHEWING TOBBACCO
29. REDUCE/AVOID STRESS OR STRESSFUL SITUATIONS
30. TAKE VITAMINS
31. TAKE BIRTH CONTROL PILLS
32. DON’T TAKE BIRTH CONTROL
33. USE CONDOMS/SPERMACIDES
34. DON’T USE CONDOMS/SPERMACIDES
35. DON’T USE DEODORANT
36. FLOSS
37. BRUSH TEETH DAILY
38. USE FLORIDE
39. NONE/NO WAY TO PREVENT GETTING HIGH BLOOD PRESSSURE OR HYPERTENSION
40. (66) OTHER1 (SPECIFY ______________________)
41. (77) OTHER2 (SPECIFY ______________________)
42. (88) DON’T KNOW
43. (99) REFUSED

Detection measure (e.g., for diabetes and hypertension)

Q. People often get warning signs or symptoms when they are sick or have a health
problem. For example, a sore throat, runny nose, fever, cough, and itchy eyes, might
suggest you have a cold, the flu or allergies. Sometimes these warning signs are picked
up by certain doctors’ tests, for example, when they check your blood, urine or heart for
health problems. What do you think are the warning signs or symptoms…

¾ Hispanics/Latinos only

Q. … of diabetes?

01. BLOOD SUGAR LEVEL
02. FREQUENT URINATION
03. EXCESSIVE THIRST/DRY MOUTH/SORE THROAT
04. EXTREME HUNGER
05. FEELING VERY TIRED MUCH OF THE TIME/INSOMNIA
06. HEADACHES
07. HEART DISEASE/HEART FAILURE/HEART ATTACK/STROKE
08. KIDNEY FAILURE/PROBLEMS
09. MORE INFECTIONS THAN USUAL
10. SORES THAT ARE SLOW TO HEAL
11. SUDDEN VISION CHANGES/BLINDESS
12. TINGLING OR NUMBNESS IN HANDS OR FEET
13. UNEXPLAINED WEIGHT LOSS
14. WEIGHT GAIN/OBESITY
15. VERY DRY SKIN
16. NONE, THERE ARE NO SYMPTOMS FOR DIABETES
17. DIZZINESS/FAINTING/SEIZURES/SHAKING
18. HIGH BLOOD PRESSURE
19. FEVER/HIGH TEMPERATURE
20. CRAVING/EAT MORE SWEETS
255

¾ African Americans only

Q. … of hypertension/high blood pressure?

01. BLOOD PRESSURE LEVEL
02. CONFUSION/TROUBLE SPEAKING/CONCENTRATION PROBLEMS
03. DIZZINESS/LOSS OF BALANCE OR COORDINATION/FAINTING
04. FEELING VERY TIRED MUCH OF THE TIME/INSOMNIA/WEAKNESS
05. HEADACHES
06. HEART DISEASE/HEART FAILURE/HEART ATTACK/STROKE
07. HIGH CHOLESTEROL
08. IMPOTENCE/ ERECTILE DYSFUNCTION
09. KIDNEY FAILURE/PROBLEMS
10. MORE INFECTIONS THAN USUAL
11. SORES THAT ARE SLOW TO HEAL
12. SUDDEN VISION CHANGES/BLINDESS/RED EYES
13. TINGLING OR NUMBNESS IN ARMS, HANDS OR FEET
14. UNEXPLAINED WEIGHT LOSS
15. VAGINITIS
16. WEIGHT GAIN
17. NONE, THERE ARE NO SYMPTOMS FOR HYPERTENSION/HIGH BLOOD PRESSURE
18. SHORTNESS OF BREATH/DIFFICULTY BREATHING
19. SWELLING [E.G., FEET OR HANDS]
20. CHEST PAINS
21. FEVER/SWEATING/CHILLS
22. NAUSEA/VOMITING/OTHER FLU/COLD SYMPTOMS
23. HEART PALPATATIONS/INCREASED HEART RATE
24. STRESS/ANXIETY
25. ANGER/IRRATABILITY

256

Appendix B
Classification Scheme for Organizations Mentioned by Participant Organizations
for the Scope of Inter-Organizational Connectedness Measure

100 MEDIA ORGANIZATIONS
101 Newspaper
102 Magazine
103 Radio
104 Television
105 Online
901 Other [Specify:________________]

200 COMMUNITY-BASED ORGANIZATIONS
201 School
202 College/University
203 Education-oriented organizations (but not a school)
204 Library
205 Sports & Recreation-focused organization (e.g., YMCA, Boys & Girls Clubs)
206 Religious (e.g., churches or church-founded organizations)
207 Provides food and shelter to people in need (not affiliated with a church)
208 Public assistance programs/Career services
(e.g., Pathfinders Job & Career Center, organizations that serve people w/
disabilities)
209 Focuses on family/parenting issues, personal development
(e.g., Parenting Institute, Jeffrey Foundation/Child Training Center)
210 Neighborhood Council
211 Neighborhood Watch group
212 Homeowners’ Association
213 City Council District Office
214 Foundations
902 Other community-based organization/Non-profit:
PLEASE SPECIFY: ___________________________

300 HEALTH CARE SERVICE PROVIDERS
301 Emergency Medical Care
302 General Medical Care
303 Health Screening/Diagnostic Services
304 Family Planning/Sex Education
305 Specialty Medicine
306 Substance Abuse
903 Other [Specify:_________]

257

Appendix C
Organizations African Americans and Latinos in Crenshaw Cited
as Most Important
(Source: Metamorphosis survey of residents in Greater Crenshaw, 2005)
A. Organizations African Americans in Greater Crenshaw
consider most important (N=125)

Sports
& Recreation
Organizations
Cultural, Ethnic,
& Religious
Organizations
Neighborhood
and Homeowners
Associations
Political &
Educational
Organizations
Q. What is the name
of the most
important sport or
recreation group?
Q. What is the name of
the most important
cultural, religious or
church group?
Q. What is the name
of the most
important
neighborhood or
homeowners’
organization or
group?
Q. What is the
name of the most
important
political or
educational
organization or
group?
Amazing Grace
Conservatory
2nd AME church 21st block
organization
48th and
Empowerment
Congress
American Youth
Soccer Organization
(AYSO)
A Breath of Life 4th ave block club ACORN
Athen Park Pirates,
coach
African American
Episcopal church
Baldwin Hills
Homeowners'
Association
African American
Cultural Center
Bally's Fitness Club Agape International
Center of Truth
BALDWIN VILLAGE
NEIGHBORHOOD
ASSN
Agenda
CALIFORNIA BLACK
AVIATION ASSN
Apostolic Faith Home
Assembly
BALDWIN VISTA
BLOCK CLUB
Baypac
CHALLENGERS BOYS
& GIRLS CLUB
BETHANY WEST LOS
ANGELES CHURCH
Century 21 Beyond the Bell
CHURCH OF THE
HARVEST/TRUST
CLUB
Church of the
Transfiguration (Catholic)
Cherrywood Block
Association
Big Brothers
Crenshaw Cougars Chrenshaw Christian
Center
Feed the Children Black Student
Alum Association
Occidental
Crenshaw High
School
Church of Christ Gramercy Park
Homeowners
Association
Boys and Girls
Club
258

CRENSHAW HIGH
SCHOOL MARCHING
BAND
CHURCH OF CHRISTIAN
FELLOWSHIP
Jefferson park
community
association
CAL STATE LOS
ANGELES
East 60th St
Community Center
Church of Deliverance L.A. Urban Round
Table
CCD
ELEGANT FLARE
DANCE COMPANY
Church of God and Christ Ladera Heights Civic
Participation
CEJ Coalition
Education Justice
Encino/Balboa Golf
Course
CHURCH OF THE
HARVEST
Leimert Park
Neighborhood
Watch
CENTER FOR
POSITIVE
OPPORTUNITY
EPIC‐Swimming at
the Coliseum
Church of the Redeemer Neighborhood
Watch
CROSSROADS
ALUMNI
ASSOCIATIONS
Equestrians City of Praise Christian
church
PARK MESA
HEIGHTS
NEIBORHOOD
Council
Democratic Party
GRACE HOSPICE City of Refuge church Second Ave Block
Club
Hancock Park
School PTA
Inner City
Management Club
Crenshaw United
Methodist Church
The Harvard Blvd.
Club
Headstart
L. A. Best Disciples of Christ United
Homeowners
Association
HILLCREST
ELEMENTARY
L.A. Roadruners Eastern Star Missionary Urban League I VOTE
L.A. CITY AQAUATICS Ever Increasing Faith Van Ness Senior
Citizen's Club
Iglewood
Teachers
Association
L.A. Fitness
Basketball Group
Faithful Central Bible
Church
VIEW PARK/WINSOR
HILL HOMEOWNERS
ASSOCIATION
L.A City College
Student
Organization
LAKERS BASKETBALL FAMILY MENNONITE
CHURCH
Vineyard senior
citizen
L.A. Urban Round
Table
Leimert Park Activity
Group
FIRST AME CHURCH West Adams
Historical Society
NAACP
Madera Little League First Church of God/City
of Hope
OMEGA PSI PHI
FRATERNITY INC
Midtown Rollers Liberty Tabernacle SPEAKING OUT
FOR FAIRNESS &
ACCURACY
New Horizon Bowling
League
Mount Calvary St. Bernardette's
High School
NORMANDY
RECREATIONAL PARK
Mt. Sinai Missionary
Baptist Church
The Family
Watch
259

OASIS FOR SENIOR
CITIZENS
New Light Baptist Church United Teachers
of Los Angeles
PAN PACIFIC SPORT
AND RECREATION
St. Mark's Missionary
(Baptist church)
Young Black
Scholars
QUEEN ANN REC
CENTER
St. Mary's Ethiopian
Orthodox Church

Renaissance Runners Second Mount Carmel
Baptist

St. Andrew's Park seventh adventist church
The Den Sister Sister Space
People's Coordinated
Center
St. Bernadette's Catholic
Church

THE RED RIBBON St. Peter Clavier
YMCA trinity baptist church
USC LITTLE LEAGUE West Angeles Cathedral
Church

Victory Adult
Baseball League

West Virginia Tech
Football

WINEGUARD
B. Organizations Latinos in Greater Crenshaw consider
most important (N=49)

Sports &
Recreation
Organizations
Cultural, Ethnic,
& Religious
Organizations
Neighborhood/
Homeowners
Associations
Political &
Educational
Organizations
Q. What is the
name of the most
important sport or
recreation group?
Q. What is the name of
the most important
cultural, religious or
church group?
Q. What is the name of
the most important
neighborhood or
homeowners’
organization or group?
Q. What is the
name of the most
important
political or
educational
organization or
group?
AYSO Adventista Septimo Dia Adams Normandie
Neighborhood Watch
Association de
Padres de Familia
Béisbol—parque
jaba hill—
aeropuerto
Brillen Institute Ana neighborhood CA Democratic
Caucus
260

Club de soccer en
parquet Palms
Case de la cultural
Guatemala
Arlington Heights
Homeowners Association
Central American
Student
Association
Club deportivo
Colima
Catholic church Baldwin Hills Association ECN AN DC
Dinamo Church of JC of Alter
Day Saints
C.RA.LA. High school
boosters
Dover Basketball El sebrador Comunidades unidas por
el progreso
Homies Unidos
Exposition Park
Interacionale
Evangelico christiano Departe de policia Jeffrey
Foundation
Kidsport Headhua Gramercy Block Club Mujeres Libres
L.A. Fitness Iglesia Evangelica Elin Neighborhood watch Radical Women
Liga Centro
Americana Soccer
Iglesia Juveniles Prop de viviendos
Salvador Gonzalez
School site
comité
Liga de futbol La Iglesia de San
Vicente
West Coliseum Block
Club
University
(Riverside)
Los Altos Chivas Misioneros de
Guadalupe

Los Swimmers Pentecostés
Mid City Soccer Puerto al cielo
Normadie
Basketball
Renovación Cristiana
Normandie Rec
Center
Salvation Army
YMCA San Patricio
Santa Agatha
Santa Cecilia
Santa Inez iglesia
Santo Tomas Catholic
Church

Santo Tomas
sembrando jóvenes

Servicio communitaro
Jefferson

Testigos de Jehová

Abstract (if available)

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

Self-perceptions of Aging in the Context of Neighborhood and Their Interplay in Late-life Cognitive Health

Asset Metadata

Creator Matsaganis, Matthew D. (author)

Core Title Rediscovering the communication engine of neighborhood effects: how the interaction of residents and community institutions impacts health literacy and how it can be leveraged to improve health c...

School Annenberg School for Communication

Degree Doctor of Philosophy

Degree Program Communication

Publication Date 10/21/2008

Defense Date 07/22/2008

Publisher University of Southern California (original), University of Southern California. Libraries (digital)

Tag communication environment,health care access,health literacy,institutional resources,neighborhood effects,OAI-PMH Harvest,storytelling network

Language English

Contributor Electronically uploaded by the author (provenance)

Advisor Ball-Rokeach, Sandra (committee chair), Cody, Michael J. (committee member), Murphy, Sheila (committee member), Valente, Thomas W. (committee member)

Creator Email matsagan@usc.edu,mdmatsaganis@yahoo.com

Permanent Link (DOI) https://doi.org/10.25549/usctheses-m1687

Unique identifier UC1216239

Identifier etd-Matsaganis-2425 (filename),usctheses-m40 (legacy collection record id),usctheses-c127-123621 (legacy record id),usctheses-m1687 (legacy record id)

Legacy Identifier etd-Matsaganis-2425.pdf

Dmrecord 123621

Document Type Dissertation

Rights Matsaganis, Matthew D.

Type texts

Source University of Southern California (contributing entity), University of Southern California Dissertations and Theses (collection)

Repository Name Libraries, University of Southern California

Repository Location Los Angeles, California

Repository Email cisadmin@lib.usc.edu