University of Southern California Dissertations and Theses

Social network norms and HIV risk behaviors among homeless youth in Los Angeles, California

(USC Thesis Other)

Social network norms and HIV risk behaviors among homeless youth in Los Angeles, California

PDF

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content

Social Network Norms and HIV Risk Behaviors among Homeless Youth in Los Angeles,
California

By

Anamika Barman-Adhikari

A Dissertation Presented to the
FACULTY OF THE USC GRADUATE SCHOOL
UNIVERSITY OF SOUTHERN CALIFORNIA
In Partial Fulfillment of the
Requirements for the Degree
DOCTOR OF PHILOSOPHY
(SOCIAL WORK)
December 2013

ii

DEDICATION
This dissertation is dedicated to all the people in my life without whom I could not have
imagined reaching this goal. I would first like to thank my parents, Badan Barman and Pritidhara
Barman, who have molded me into the person that I am today. They sacrificed many of their
own needs to invest in the future of their children. Thank you so much for instilling in me the
value of hard work and the perseverance to follow my dreams. You both are my pillars of
strength.
This dissertation is especially dedicated to my husband Diganta Adhikari who has shown
me what true love and dedication looks like. Last, but not the least, I have to acknowledge the
love of my life, my daughter Aahana Barman Adhikari, who continues to inspire me every day to
become a better person so that I can become a better role model for her. Especially since I was
pregnant when I wrote this dissertation, she legitimately deserves half the credit for this work. I
love you my little one.

iii

ACKNOWLEDGEMENTS
The process of getting a PhD is analogous to running a marathon. You have to continue
at a steady pace over a long stretch and hope that your endurance and stamina can carry you
through to the end. However, while you get to the finish line alone, numerous people on the
sidelines are cheering for you, supporting you, and rooting for your success. I am lucky to have
had an amazing group of people who have helped me at every step of this long journey. This
dissertation is a culmination of all their collective efforts.
First, I would like to thank my dissertation advisor and mentor Dr. Eric Rice. You have
been instrumental in shaping my growth as a graduate student, and more importantly as an
academic in countless ways. Your level of commitment and dedication towards your mentees is
unparalleled. It is amazing how you always patiently listened to me week after week, took my
most mundane ideas and helped me turn them into innovative and pioneering projects. Thank
you for gently but steadily nudging me to do my best work. I am hoping that I can someday
model the same relationship with my students.
I also have to take this opportunity to thank my other dissertation committee members,
Dr. Suzanne Wenzel and Dr. Nicole Esparza, who continually provided me with valuable and
insightful feedback on my dissertation drafts, responded to all my questions patiently, and were
always available to me when I needed their advice.
Dr. Wenzel, I am most inspired by your humility and quiet and unassuming ways. It is
very rare to come across somebody who is so accomplished, but modest at the same time. Thank

iv

you for being such a wonderful person and a rigorous scholar at the same time. These are
qualities, which are seldom seen together in one person.
Dr. Esparza, I am thankful that you took me on when you really did not have to. You
always encouraged me by helping me see beyond my weaknesses and highlighting my strengths.
I still remember the email that you sent me after my less than ideal qualifying exam proposal.
Your words of encouragement will forever remain etched in my memory. Without them, I think I
could not have regained my confidence.
A special thanks to Dr. Julie Cederbaum. You helped me understand the value of
work/life balance. Early, during my PhD days, you gave me some advice about work/life balance
that I still follow diligently until this day. I still work only during my office hours and try to be as
focused as possible when I am in my office so that I can spend more quality time with my
family. I have always looked up to you because you are proof that a woman in academia can “do
it all”!
As a social network researcher, I think it is especially important that I acknowledge my
network of family, friends, and fellow students who bore the brunt of being associated with me
during some of the most stressful moments of my life.
My husband in particular could win the Nobel Prize for “patience”. He refers to himself
as my unofficial “therapist”. He always had faith in me even when I completely lost it. He never
questioned my unholy working hours, always telling me that my goals were his. He labored with
me through my panic attacks, celebrated my successes, and assured me that life did not end just

v

because a manuscript had been rejected. Honey, thank you for keeping me grounded and always
keeping things in perspective. To you, I owe my sanity.
I do not think I would have ever embarked on an academic career in the first place if it
were not for my parents. From a young age, my father would sit down with me and we would
read the newspaper together and separate the facts from fiction. I owe my critical thinking skills
to that early childhood experience. Mom, you always were very steadfast in your resolve that
your daughter will not have the same experience as you (being married off at a tender age and
despite being the brightest girl in her class, not being able to finish her education). I hope I have
fulfilled your dreams. Thanks so much for always gently coaxing me never to give up and
helping me believe in my dreams.
Last, but not the least, my daughter Aahana. Even on my worst days at work, when I
come back home and see you smile, all my worries dissipate. From you I have learned to enjoy
the simple things in life and am reliving my childhood all over again. Thank you for making
weekends fun and meaningful again. I do not even mind waking up in the wee hours of the
morning if it means that I get to spend more time with you. I hope you read this someday and
know that mommy was busy for the first year of your life because she was trying to impress you
and hopefully be an inspiration for you.
My special thanks to USC School of Social Work and NIH for funding my graduate
studies and helping me stay out of debt!

vi

TABLE OF CONTENTS
Dedication ii
Acknowledgements iii
List of Tables viii
List of figures x
Abstract xi

Chapter One: Introduction 1
Introduction
Purpose of this Study
Significance of this Study
Organization of Dissertation
Chapter Two: Literature Review 12
Introduction
Homeless Youth and HIV/AIDS
Methamphetamine and Injection Drug Use among homeless youth
Methamphetamine and Injection Drug Use as a risk for HIV/AIDS among homeless
youth
Individual correlates of methamphetamine and injection drug use among homeless youth
Social networks and engagement in substance use behaviors
Structural network characteristics
Normative network characteristics
Compositional Network characteristics
Internet and social media use among homeless youth-implications for compositional
characteristics of networks
Relationship between structural and normative network characteristics
Relationship between compositional and normative network characteristics
Summary
Chapter Three: Theoretical Conceptualization 37
Introduction
Types of Norms
Norms and behavior
Social Network Structures and Perception of Norms
Social Referents and Normative Behavior
Social Capital Theory: Online Relationships and Composition of Networks
Summary
Chapter Four: Methodology 50
Introduction
Sampling
Boundary Specification
Procedures
Name Generator
Measures
Sociodemographic characteristics
Network Variables

vii

Network Norms
Online Social Capital
Behaviors
Data Analyses
Chapter Five: Result Study Aim 1 68
Purpose of the Studies
Major Findings
Practice and Research Implications
Chapter Six: Result Study Aim 2 78
Purpose of the Studies
Major Findings
Practice and Research Implications
Chapter Seven: Result Study Aim 3 93
Purpose of the Studies
Major Findings
Practice and Research Implications
Chapter Eight: Discussion 122
Discussion of Key Findings
Norms and Behavior
Internet and social media use among homeless youth- composition of networks and
associations with norms
Sociometric Positions and Norms
Associations with socio-demographic characteristics
Study Limitations
Implications for Policy, Practice, and Future Research
References 142

viii

LIST OF TABLES
Table 1: Skewness of Perceived Norms Variables 59
Table 2: Descriptive Statistics of Homeless Youth, Los Angeles, CA 69
Table 3: Descriptive Norms for Methamphetamine and IDU 70
Table 4: Injunctive Norms for Methamphetamine and IDU 71
Table 5: Substance Use and Sex Risk Behaviors 72
Table 6: Collinearity among Independent Variables for 73
Study Aim 1
Table 7: Odds Ratios for Associations between Perceived Norms and 75
Methamphetamine Use
Table 8: Odds Ratios for Associations between Perceived Norms and 77
IDU
Table 9: Collinearity among Independent Variables for 79
Study Aim 2
Table 10: Perceived Norms for Methamphetamine and IDU 80
Table 11: Online Social Capital among Homeless Youth 81
Table 12: Odds Ratios for Associations between Social Capital and 83
Descriptive Norms of Methamphetamine Use
Table 13: Odds Ratios for Associations between Social Capital and 85
Injunctive Norms of Methamphetamine Use
Table 14: Odds Ratios for Associations between Social Capital and 86
Injunctive Norms of Methamphetamine Use
Table 15: Odds Ratios for Associations between Social Capital and 88
Descriptive Norms of Injection Drug Use
Table 16: Odds Ratios for Associations between Social Capital and 90
Injunctive Norms of Injection Drug Use
Table 17: Odds Ratios for Associations between Social Capital and 92
Injunctive Norms of Injection Drug Use
Table 18: Descriptive Statistics of Homeless Youth in Hollywood and 94
Santa Monica
Table 19: Perceived Norms for Methamphetamine and Injection Drug 96
Use
Table 20: Sociometric Characteristics of Networks 97
Table 21: Odds Ratios for Associations between Sociometric Properties 112
Descriptive Norms of Methamphetamine Use
Table 22: Odds Ratios for Associations between Sociometric Properties 114
Injunctive Norms of Methamphetamine Use
Table 23: Odds Ratios for Associations between Sociometric Properties 116
Descriptive Norms of Injection Drug Use
Table 24: Odds Ratios for Associations between Sociometric Properties 117
Descriptive Norms of Methamphetamine Use
Table 25: Odds Ratios for Associations between Sociometric Properties 118
Injunctive Norms of Methamphetamine Use

ix

Table 26: Odds Ratios for Associations between Sociometric Properties 119
Descriptive Norms of Injection Drug Use

Table 27: Odds Ratios for Associations between Sociometric Properties 121
Injunctive Norms of Injection Drug Use

x

LIST OF FIGURES
Figure 1: Conceptual Model of the Study Aims 4
Figure 2: Egocentric Relationships 23
Figure 3: Sociometric Relationships 23
Figure 4: Sociometric Network of Homeless Youth in Hollywood, CA 63
Figure 5: Sociometric Network of Homeless Youth in Santa Monica, CA 64
Figure 6: Perception of Meth Use in Hollywood Network 99
Figure 7: Perception of Encouragement of Meth Use in Hollywood 100
Network
Figure 8: Perception of Objection to Meth Use in Hollywood 101
Network
Figure 9: Perception of Injection Drug Use (IDU) in Hollywood 102
Network
Figure 10: Perception of Encouragement of IDU in Hollywood 103
Network
Figure 11: Perception of Objection to IDU in Hollywood 104
Network
Figure 12: Perception of Meth Use in Santa Monica Network 105
Figure 13: Perception of Encouragement of Meth Use in Santa Monica 106
Network
Figure 14: Perception of Objection to Meth Use in Santa Monica 107
Network
Figure 15: Perception of Injection Drug Use (IDU) in Santa Monica 108
Network
Figure 16: Perception of Encouragement of IDU in Santa Monica 109
Network
Figure 17: Perception of Objection to IDU in Santa Monica 110
Network

xi

ABSTRACT
Homeless youth engage in substance use risk behaviors that place them at an increased
risk of HIV infection. Studies on this population have shown that their social networks are
consistently linked with their HIV risk behaviors. However, recent reviews suggest that substance
use prevention programs have had an individualistic focus. This has led leading experts in the
field to call for novel network-level prevention interventions to be developed to reduce HIV
incidence among homeless youth. Social networks influence behavior through several
mechanisms, one of which is through the establishment, and maintenance of social norms. While
several theories suggest that norms offer a potent channel of initiating and sustaining behavioral
change, intervention efforts have been hampered because of the paucity of research examining
clustering of norms within specific risk-taking social network structures. For example, although
much of the theoretical research on norm distribution emphasizes structural elements such as
network position or the cohesiveness or size of the network in the shaping of perceived norms,
research in this area is very limited. The purpose of this proposed study was to utilize sociometric
analyses to understand whether social norms of HIV risk behaviors are clustered with social
network structures, and whether the norms of these network members are associated with these
youth’s risk taking behaviors. Researchers have also suggested that future studies must address
how different relationship roles influence HIV behavioral norms. Therefore, this study also
utilized egocentric social network analyses to understand which referent groups other than street
peers that these youth connect to (especially via new forms of communicative technology), and
whether these differential forms of influences contribute to different types of norms regarding
these behaviors.

xii

Data for this research came from a larger NIMH funded “Youthnet” study (MH R01
903336). Rice and colleagues are collecting multiple panels of egocentric and sociometric
network data over time from homeless youth ages 13-25 in two drop-in centers in Los Angeles,
CA. This proposed study used only the baseline data to accomplish its research objectives. Taken
together, the egocentric and sociometric data was able examine multiple research questions that
have not yet been addressed in the HIV risk norms literature. These data suggest that both social
proximity as well as social positioning within networks is associated with risk norms. The results
therefore overall supported the general proposition that both egocentric and sociometric network
attributes affect substance use among homeless youth. Other findings elaborated on these overall
results. The findings can be used to inform new directions in HIV prevention interventions,
specifically what network-level interventions could be adapted in the context of the homeless
youth population, and the feasibility of online technology as a potential mechanism through
which network interventions can be delivered.

1

CHAPTER ONE
INTRODUCTION
It is estimated that 1.5 to 3 million youth experience homelessness each year in the
United States (Toro, Lesperance, & Braciszewski, 2011), and are considered to be one of the
most marginalized groups in the nation. Homeless youth are known to be vulnerable to a host of
social, psychological, and behavioral problems, including an increased risk of engaging in risky
substance use behaviors (such as injection drug use and meth use), making them more
susceptible to the risks of HIV transmission and acquisition (Arnold & Rotheram-Borus, 2009;
Booth & Zhang, 1996; Greene & Ringwalt, 1996; Marshall, Kerr, Shoveller, Patterson, Buxton,
& Wood, 2009). The prevalence of HIV infection among homeless youth range from 5% to 17%
across various studies (Beech, Myers, Beech, & Kernick, 2003; Noell et al. 2001; Pfeifer &
Oliver 1997; Stricof, Kennedy, Nattell, Weisfuse & Novick 1991). More importantly, compared
to their stably-housed peers, homeless youth are three to nine times as likely to be infected with
HIV (Beech, Myers, Beech, & Kernick, 2003; Tevendale, Lightfoot, & Slocum, 2009) and are
seven times as likely to die from AIDS (Ray, 2006).
Although engagement in risky drug-using behaviors linked to HIV has consistently been
linked to the social networks of homeless youth (Rice, Milburn, & Rotheram-Borus, 2007; Rice,
Stein, & Milburn, 2008; Tyler, 2008; Wenzel, Tucker, Golinelli, Green, & Zhou, 2010), recent
reviews suggest that most substance use prevention programs have had an individualistic focus
(DiClemente & Wingood 2000; DiClemente et al., 2004; Malow, Kershaw, Sipsma, Rosenberg
& Devieux, 2007; Parker, 2004; Sumartojo, 2000), which may partially explain why most HIV
prevention interventions have not had the most effective outcomes. Contextual information is
considered essential for designing tailored interventions that respond to the needs and

2

preferences of people within specific communities. This has led leading experts in the field to
call for novel network-level prevention interventions to be developed to reduce HIV incidence
among homeless youth (Arnold & Rotheram-Borus, 2009; Tyler 2013).
A social network is a set of people or groups of people with some pattern of connections
or interactions between them (Wasserman & Faust, 1994). Social network analysis (SNA) is the
measuring, mapping, analyzing, and interpretation of social network structures, the ties between
nodes, and the flows that occur within and across networks (Wasserman & Faust, 1994). Social
networks influence behavior through several mechanisms, one of which is through the
establishment, and maintenance of social norms (Davey-Rothwell & Latkin, 2007; Friedkin,
2001). Social norms form a crucial element of several common theories of health behaviors
(Bandura, 1977; Fishbein, Middlestadt, & Hitchcock, 1994; Fisher & Fisher, 1992) and have
been linked to both behavior and behavioral intentions (Barrington, 2008).
Norms are defined as perceived rules or properties of a group that characterize specific
beliefs around what behaviors are considered acceptable or common within that group (Kincaid,
2004). Perceived norms have been generally categorized as descriptive or injunctive norms.
Descriptive norms indicate the perceived prevalence of a behavior in a group whereas injunctive
norms refer to perceived approval of a behavior (Davey-Rothwell & Latkin, 2007). While
several theories suggest that norms offer a potent channel of initiating and sustaining behavioral
change (Fisher & Fisher, 1992; Latkin, Forman, Knowlton, & Sherman, 2003), intervention
efforts have been hampered because of the paucity of research examining clustering of norms
within specific risk-taking social network structures (Latkin et al., 2009). For example, although
much of the theoretical research on norm distribution emphasizes structural elements such as

3

network position or the cohesiveness or size of the network in the shaping of perceived norms,
research in this area is very limited (Horne, 2001; Latkin et al., 2009).
Additionally, recent research has revealed that homeless youth tend to be much more
heterogeneous in terms of their social network composition than was previously thought (Rice,
Milburn, & Rotheram-Borus, 2007; Rice, Stein, & Milburn, 2008; Wenzel et al., 2012;
Whitbeck, Rose & Johnson, 2009) and this heterogeneity has implications for norm formation
and engagement in risk behaviors among this population. The multidimensionality of norms has
been emphasized by various scholars (Coleman, 1990; Latkin et al., 2009). It is important to
determine the appropriate reference group or groups in understanding social norms because they
exert different forms of influence depending on who they are (Rimal, Lapinski, Cook, & Real,
2005).
While it might seem surprising to most people, in spite of lacking housing resources,
internet use is pervasive among homeless youth (Karabanow & Naylor, 2010; Rice, Monro,
Barman-Adhikari, & Young, 2010). More importantly, these new forms of communicative
technologies are increasingly being used by these youth to maintain ties to pro-social home-
based peers or supportive family, primarily via the internet and cell phones (Rice et al., 2010;
Barman-Adhikari & Rice, 2011). As a result, internet and social media have afforded homeless
youth with the opportunity to engage in social worlds beyond their street environments.
Therefore, the diverse networks among homeless youth, which are partially facilitated by these
new forms of communicative technology, might be the source of multiple and conflicting
behavioral norms.

4

Network position Descriptive Norms
Cohesiveness HIV Behaviors:
Centrality IDU
Network Composition Meth Use
Injunctive Norms Ties maintained via Social
Networking Technology or
cell phone
Social Network
Characteristics

Figure1: Conceptual Model of the Study Aims
Purpose of this Study
This study draws on theoretical perspectives of social networks, social norms, and social
capital to investigate the processes through which social structural contexts influence substance
use behaviors related to HIV transmission (specifically methamphetamine and injection drug
use) among homeless youth. More specifically, this dissertation examined how social norms of
substance use behaviors are structured within social networks, and how these norms of network
members are associated with substance use behaviors among these youth. In addition, since
internet and social media is pervasive among homeless youth, and these youth are increasingly
using these new communicative technologies to maintain ties out of street life, this dissertation
also investigated whether these relationships maintained via these new avenues facilitated
differential norms regarding substance use among these youth. By examining perceived social
network norms as both an independent and dependent variable, three aims and hypotheses guided
this study (see Figure 1):

5

1. To assess the association between perceived social network norms (descriptive and injunctive)
and HIV risk behaviors (methamphetamine and injection drug use) among homeless youth.
Hypothesis 1: The stronger the perceived network norms approving HIV risk behaviors, the
greater the likelihood that respondents will engage in HIV risk-taking behaviors.
2. To assess how technology facilitates connection to pro-social home-based peers and family, to
explore differential norms (descriptive and injunctive) about HIV risk behaviors
(methamphetamine and injection drug use) among homeless youth.
Hypothesis 2: Connecting and communicating online with pro-social home-based peers and
family will be associated with norms that are preventive of engagement in HIV risk behaviors.
3. To identify social network characteristics that are associated with perceived social network
norms (descriptive and injunctive) regarding the use of methamphetamine and injection drugs
among homeless youth.
Hypothesis 3: Network norms for HIV risk behaviors (substance use) will vary based on the
structural characteristics of the networks (such as social position, cohesiveness etc.).
To accomplish the above aims, the proposed study utilized social network analysis
(egocentric and sociometric) using baseline data from Dr. Rice’s NIMH funded “Youthnet”
study (MH R01 903336) data of 386 homeless youth in Los Angeles, CA.
Successfully addressing these aims provided:
(1) A network based understanding of norms of risk taking in the social networks of homeless
youth,

6

(2) New directions in HIV prevention interventions, specifically what network-level
interventions could be adapted in the context of the homeless youth population, and;
(3) Examining the feasibility of online technology as a potential mechanism through which
network interventions can be delivered.
Significance of this Study
Social network analysis provides a fruitful means through which one can examine the
structure of norms in naturally occurring social groups (Latkin et al., 2009). Networks can be
analyzed based on its structure and function (Latkin, Forman, Knowlton, & Sherman, 2003).
Structural network theory is primarily concerned with characterizing network structures (e.g.
small worldness) and node positions (e.g. core/periphery) and associating it with a multitude of
outcomes, including job performance, mental health, organizational success etc. (Borgatti, 2011).
Norms on the other hand fulfill a very important function in social networks by acting as a
source of social influence and regulation (Horne, 2001). Extant research shows that structure and
function of networks are often intertwined in accounting for HIV risk behaviors (Latkin,
Mandell, & Vlahov, 1996; Sherman, Latkin, & Gielen, 2001; Friedman, Curtis, Neaigus, Jose, &
DesJarlais, 1999; Frey et al., 1995). However, very few studies have looked at the intersection
of social network structure and norms in understanding HIV risk behaviors among vulnerable
homeless populations, especially among homeless youth.
Additionally, social network analyses can be approached at two levels: egocentric (the
direct ties of an index person with all of his or her network members), and sociometric networks
(which refers to the complete set of relations between people in a population, both direct and

7

indirect ties) (Neaigus, 1998) and has implications for the understanding of norms. Norm
emergence is considered both an element of the ego-centered approach where norms are
internalized as a process of imitation or observational modeling (Bandura, 1986) as well as a
sociocentric process where people in similar structural positions endorse similar norms (Horne,
2001). Among homeless youth, network influences have been examined showing HIV risk-
taking youth are embedded in networks with other HIV risk-taking youth, in both egocentric
(Tyler, 2008; Rice et al., 2005; Rice et al., 2007) and sociometric network analyses (Rice et al.,
2012). However, while studies have looked at the relationship between homeless youth’s
perception of norms and engagement in risky substance use behaviors (Rice et al., 2005;
Tevendale et al., 2009; Solorio et al., 2008; Tyler et al., 2012), to the best of my knowledge, no
study has yet tried to assess the structural characteristics of these norms among this population.
This study utilizes a sociometric design, which reveals whether specific norms are
attached to network positions in large inter-connected networks. It is important to know which
network characteristics are associated with corresponding norms because they can provide us
important insight about how HIV prevention interventions can be tailored. For example, being in
the core of a network has been found to be associated with both positive and negative outcomes
(Latkin et al., 2003). This disparity could be a result of the differences in norms among the
groups that were studied. While being in the core of a network of a normative group could reflect
healthy norms, this might not be true for a high-risk group, such as homeless youth. For
example, a recent study found that homeless youth who are peripheral to their networks were less
likely to engage in sexual risk behaviors compared to youth who occupied central positions
within this network (Rice et al., 2012). Therefore, if one looks at the network structure alone,
important contextual information can be missed, leading to inconsistencies in findings.

8

Also, by examining two levels of social processes for the unit of analysis (individual and
group relationships) through both egocentric and sociocentric network analysis, and extending
our network analysis to include different types of relationships (e.g., home-based friend, relative,
caseworker, etc.), we were able to examine multiple research questions that have not yet been
addressed in the substance use norms literature. More specifically, the use of both forms of data
not only permitted examining the structural features of norms (Horne, 2001), but also allowed us
to delineate source-specific types of norms (Rimal et al, 2005). As noted before, norms are not
one-dimensional by nature. More specifically, different kinds of referent groups contribute
different kinds of norms surrounding both risk and protective behaviors. For example, while
one’s family members might promote norms around abstinence, peer groups might promote
norms promoting substance use behaviors (Latkin et al., 2009). In designing and tailoring
interventions, it is important to understand which of these influences are most salient in these
youth’s lives.
One important motivation for many individuals to adopt a behavior is the desire to gain
social status (Rogers, 2003). For a rule to be a norm, it must be accepted by group members
(Horne, 2001). Therefore, interventions are increasingly attempting to change social network
norms as a way through which they can diffuse behavioral change in communities using the
natural forms of influence already prevalent within social networks (Barrington, 2008). Very few
interventions have however used a whole network approach to norm change (Latkin et el., 2009),
especially among homeless youth. However, while extant research has tried to assess how these
norms influence behavior, less attention has been paid to the characteristics of social groups that
shape the development of these norms (Boer & Westhoff, 2006). More importantly, very few
studies have tried to understand whether norms adhere to some kind of social organization, and

9

how this structure can sustain such norms (Horne, 2001). This study is innovative in that it seeks
to understand how norms are clustered or distributed within specific sociometric networks. More
specifically, this study helped us explicate the process through which social status (such as
network positions) affects the clustering of norms in a bounded network.
The Popular Opinion Leader (POL) (Kelly, 1994) is one network-based approach to norm
change that has been successfully used in the gay community to reduce HIV risk. These models
have typically used influential leaders (one who often have greater social status) within the
community as change agents to target risky norms regarding HIV behaviors (Elford, Bolding, &
Sherr, 2004; Kelly et al., 1991; Sikkema et al., 2005). However, its translatability within the
homeless youth community has not yet been assessed. In order to understand whether the people
who homeless youth regard as influential (or opinion leaders) their social networks of homeless
youth are ideal targets for interventions, one has to assess whether these youth who are in these
critical positions endorse risky or protective behaviors, and the extent to which they engage in
drug use themselves (Green, Haye, Tucker, & Golinelli, 2013). The results from this study
provide important insight as to whether this model could be translated in the context of the
homeless youth population.
The Internet and other new media have opened new avenues through which homeless
youth communicate with pro-social peers or family outside of their street networks (Barman-
Adhikari & Rice, 2011; Rice et al., 2010) and could signal a new tool for HIV prevention using
these naturally occurring online networks. Recent studies have found that over 80% of homeless
youth get online more than once per week, and approximately one-quarter use the internet for
over one hour daily (Barman-Adhikari & Rice, 2011; Rice, 2010; Young & Rice, 2011). While

10

there has been a great amount of work on homeless youth’s experiences on the streets, especially
the ways in which they enter, sustain, or exit street life (Milburn et al., 2007; Milburn et al.,
2009; Whitbeck & Hoyt, 1999), relatively little is known about their internet use and its
consequences.
Internet, social media, and cell phone use have become ubiquitous among adolescents,
and how adolescent behavior is being impacted by these technologies is of great concern to
researchers, policy makers, and parents (Greenfield & Yan, 2006; Pascoe, 2012; Subrahmanyam,
Reich, Waechter, & Espinoza, 2008). In previous studies, prevention and intervention programs
that are delivered online to other adolescent populations have proven to be both viable and
effective. These technologies have the potential to influence HIV prevention by providing a cost-
effective, confidential, and viable means to reach a much larger audience; one that is particularly
suited to meet the needs of homeless youth who by nature are unstable and hesitant to or do not
have access to formal services (Alemagno & Kenne, 2012; Barman-Adhikari & Rice, 2011;
Rice, 2010; Young & Rice, 2011). The findings from this study provide us with important
information about the feasibility of translating network-based interventions for homeless youth
that involves technology.
Moreover, most of the research on norms has focused on face-to-face interaction and
influences, limiting our understanding of how norms may be promoted through social
networking technology. In previous studies, internet and social media access among homeless
youth have been linked to increased communication with family, friends, caseworkers, service
organizations, and potential employers. Similarly, in a qualitative study with homeless youth in
Canada (Karabanow & Naylor, 2010), a majority of the youth reported that the use of the internet

11

and social media restored their sense of connection with the outside world. A series of studies
have found that homeless youth who connect with home-based, positive-role-modeling peers,
family, and case managers through internet and social networking technologies are more likely to
use condoms (Rice, 2010), less likely to use drugs and alcohol (Rice, Milburn, & Monro, 2011),
and less likely to experience depressive symptoms (Rice, Kurzban, & Ray, 2011).
Organization of Dissertation
The present chapter (Chapter One) provides an introduction to this dissertation as well as
the purpose and significance of the study. The remainder of this dissertation is organized in six
Chapters. In Chapter Two, I review the existing literature on the associations between social
network norms and drug use behaviors among homeless youth. Specifically, this chapter
examines research that has been published on perceived norms in HIV prevention and how social
network structure and norms are related in accounting for these risk behaviors. This chapter will
also review the literature on internet and social media use among homeless youth, and the
implications it has on norms around HIV risk behaviors in this population. This chapter also
identifies gaps in the literature where further research is needed. Chapter Three reviews the
theories that were used to inform the specific aims and hypotheses for this study. A particular
focus in this section was to understand the theoretical elements that link social network structure,
composition, and norms. In Chapter Four, I describe the methodology used in this study,
including the data/sample, measures, and analysis plan. In Chapters Five, Six, and Seven, I
present the results from the egocentric and sociometric analyses associated with each aim and
hypotheses. Finally, in Chapter Eight, I discuss the implications of the study’s findings for future
research, and interventions. In addition, the study’s limitations are discussed.

12

CHAPTER TWO
LITERATURE REVIEW
This chapter reviews the empirical findings relevant to the specific aims of this study.
First, the chapter briefly discusses the prevalence of HIV/AIDS among homeless youth and how
risky needle use and methamphetamine use has the potential to increase HIV incidence among
this population. Second, I present evidence linking social networks and social norms to substance
use among vulnerable populations, including homeless youth. Third, I discuss the ubiquity of
internet and social media use among homeless youth, and the implications it has on the
heterogeneity of homeless youth’s relationships, and consequently the kind of risk and protective
behaviors that they engage in. In addition, wherever applicable, limitations of the published
literature are also identified.
Homeless Youth and HIV/AIDS
An estimated 784,701 individuals were living with a diagnosis of HIV in the United
States, as of 2009 (National Center for HIV/AIDS, 2010). Of particular concern, is the increase
in the annual number of HIV diagnoses among youth as well as a rising number of young people
living with HIV in recent years. In 2006, more HIV infections occurred among adolescents and
young adults aged 13–29 than any other age group (Hall et al., 2008). The number of people
between the ages of 13 and 24 living with AIDS increased by 41 percent between 2002 and 2006
(CDC, 2008), and in 2008, 12.4 percent of AIDS diagnoses occurred among youth (CDC, 2010).
Even though HIV affects youth of every sociodemographic group, certain sub-groups of young
people in the U.S., particularly homeless youth are at a greater risk of contracting AIDS or HIV-
related illnesses.

13

Research suggests that as many as 1.6 million youth (approximately 7.6 percent) are
homeless or become runaways at some point each year (Ringwalt, 1998; Toro et al., 2011). It has
been difficult to obtain a precise estimate of the number of homeless youths because of the
difficulties in characterizing as well as reaching this diverse population, with figures ranging
between 500,000 and 2 million (Lifson & Halcon, 2001). The measurement and definition of
homelessness is plagued with ambiguity; this particular study followed the definition of
homelessness put forth by Tsemberis, McHugo, Williams, Hanrahan, and Stefancic (2004) in
adopting criteria for characterizing homelessness. The Tsemberis et al. (2004) definition includes
categories of literally homeless, temporary homeless, institutional residence, stable residence,
and functional homeless. This definition is considered more comprehensive in its
representativeness of homelessness than most other definitions. Moreover, unlike the federal
definition of homelessness, which has been often criticized for obscuring the magnitude of the
problem, the Tsemberis et al. (2004) definition acknowledges that homelessness encompasses a
broad spectrum of people. This range of homeless includes not only those living on the streets
and in shelters but also those living in motels or with family and friends because of economic
hardship.
It has been widely documented that youth experiencing homelessness are more likely
than other subpopulations to engage in behaviors associated with HIV risk, including risky
sexual practices, injection drug use and needle sharing, and trading sexual acts for money, drugs,
or a place to stay (Kidder, Wolitski, Campsmith, & Nakamura, 2007). It is therefore not
surprising to find that HIV prevalence rates among homeless youth are disproportionately high
compared to youth in the general population. Among youth experiencing homelessness, HIV
prevalence is estimated to be between 5-11% (Beech, Myers, Beech, & Kernick, 2003; Noell et

14

al. 2001; Pfeifer & Oliver 1997; Stricof, Kennedy, Nattell, Weisfuse & Novick 1991). The
environments in which young people live can exacerbate the impact of behavioral and
interpersonal risk factors for HIV infection (AMFAR, 2010). Homeless youth in particular face
a number of adversities and barriers, both before and after they become homeless (Whitbeck,
2009). A large proportion of homeless youth have a history of sexual or physical abuse, foster
care or juvenile justice involvement, and/or come from dysfunctional families, characterized by
parental violence or substance abuse (Booth & Zhang, 1996; Robertson, 1996; Roman & Wolfe,
1995). Additionally, they are exposed to other deleterious circumstances once they become
homeless, starting from the lack of basic needs such as food, housing and monetary resources, to
more serious physical and emotional health related needs (e.g. Milburn et al., 2006 Ray, 2006;
Whitbeck, & Hoyt, 1999). Generally, it is seen that the longer youth spend on the streets the
more likely they are to experience a wide range of negative behavioral outcomes, including but
not limited to acquiring HIV and other related diseases (Milburn et al., 2006). HIV/AIDS and
homelessness are therefore intricately related.
The President’s National HIV/AIDS Strategy identifies the need to focus prevention and
service delivery programs on youth at elevated risk of HIV/AIDS, which includes youth who are
homeless (AMFAR, 2010). Accordingly, a number of agencies serving homeless youth offer
HIV prevention programs as a part of the services they offer (AMFAR, 2010). However, more
needs to be accomplished to identify new prevention methods and resources that are specifically
tailored to meet the unique needs of the homeless youth population. In particular, homeless youth
use and abuse substances, especially ones that are considered to facilitate the spread of HIV
(such as methamphetamine and injection drug use) with far greater frequency than do their
nonhomeless counterparts (Wenzel, Tucker, Golinelli, Green, & Zhou, 2010) and pose a

15

significant risk for HIV transmission and acquisition. This has led to calls for an updating of the
risk hierarchy so that the prevention of methamphetamine and injection drug use is given greater
attention (Vlahov, Fuller, Ompad, Galea, & Des Jarlais, 2004). Therefore, HIV and STD
prevention and treatment for homeless youth could be enhanced to include assessment for
methamphetamine and injection drug use, with referrals to substance use treatment, primary
testing, and sexual health promotion.
Methamphetamine and Injection Drug Use among homeless youth
It is not clear to what extent substance use by adolescents leads to homelessness. Many
homeless youth report using alcohol and other drugs both prior to and since becoming homeless
(Whitbeck, 2009). Regardless of the timing of drug use, studies continue to show that these
youth engage in substantially higher levels of substance use compared to housed youth (Ennett,
Bailey, & Federman, 1999; Nyamathi, Hudson, Greengold, & Leake, 2012; Rice, Milburn, &
Monro, 2010; Tyler, 2008; Wenzel, Tucker, Golinelli, Green, & Zhou, 2010) lifetime substance
use rates as high as 97% have been reported among homeless young people (Kral, Molnar,
Booth, & Watters, 1997; Martinez, Gleghorn, Marx, Clements, Boman, & Katz, 1998); 66% of
homeless youth in Los Angeles met DSM-III criteria for substance abuse (Kipke, Unger,
O’Connor, Palmer, & Lafrance, 1997).
While alcohol, cigarettes, and marijuana are the most commonly used substances on the
street, the use of other illicit and hard-drugs such as methamphetamine and injection drug use is
on the rise (Greene, Ennett, & Ringwalt, 1997; Wenzel, Tucker, Golinelli, Green & Zhou, 2010).
This is a problem because engagement in meth and injection drug use are known to have more
dire physical and mental health effects than many other commonly used drugs. For example, a

16

recent study found that drug users who used methamphetamine had an 80% greater risk of
attempting suicide than drug users who did not (Marshall, Galea, Wood, & Kerr, 2011). Of
greater concern, however is the consistently higher overdose rates for methamphetamine and
injection drug users compared to people who use other drugs (DARA, 2013). Prevalence rates
for methamphetamine and injection drug use among homeless youth vary by study and physical
setting. Generally, methamphetamine and injection drug use are higher in larger urban
metropolitan areas. For instance, one study conducted among homeless youth in Los Angeles
found that 30% reported ever having injected drugs; among these youths, 59% reported ever
having shared needles (Kipke, Connor, Palmer, & MacKenzie, 1995). Similar rates have been
also found in a much more recent survey of homeless youth in Los Angeles (Nyamathi et al.,
2012). In this particular study, they found that the 28% of the youth surveyed reported ever-
injecting drugs. In a study conducted in Minneapolis, 15% of the youth reported injecting drugs
(Lifson & Halcon, 2001). In San Francisco, studies have reported rates as high as 38-40%
(Martinez et al., 1998; Parriott & Auerswald, 2009). Such high rates of injection drug use have
also been reported by other studies (Clatts, Davis, Sotheran, and Atillasoy, 1998; DeRosa et al.,
2001; Lifson & Halcon, 2001; Marshall, Kerr, Qi, Montaner, & Wood, 2010).
Methamphetamine rates have also gone up exponentially in North American cities,
particularly West Coast cities (such as Los Angeles where this study was conducted) in recent
years (Wood et al., 2008). Rates of methamphetamine use are three to four times higher among
homeless youth compared to their housed counterparts (Greene et al., 1997). The proportion of
methamphetamine users have been found to quadruple in homeless persons over the past decade
(Das-Douglas, Colfax, Moss, Bangsberg, & Hahn, 2008). In one recent study conducted with
homeless youth in Canada, an alarming percentage of the youth (70%) had indicated that they

17

had ever used methamphetamine (Marshall et al., 2011). Similarly, in Los Angeles, more than
half (54%) of homeless youth indicated that they had ever used methamphetamine (Nyamathi et
al., 2012). Even though the prevalence of both methamphetamine and injection drug use has
been documented to be high, evidence regarding the epidemiology of both methamphetamine
and injecting drug use among homeless youth continues to be sparse (Kerr et al., 2009; Marshall
et el., 2011; Parriot & Auerswald, 2009). In particular, although there is a growing literature
regarding the patterns of use and associated risks related to the use of methamphetamine and
injection drug use among gay and bisexual men (Halkitis, Palamar, & Mukherjee, 2007; Reback,
Larkins, & Shoptaw, 2004; Shoptaw, 2006), there are few evaluations of other vulnerable
populations, such as homeless and street-involved youth.
Methamphetamine and Injection Drug Use as a risk for HIV/AIDS among homeless youth
Drug use increases the risk for HIV transmission through its association with high-risk
sexual behaviors such as unprotected sex, sex with an injection drug user, sex with an HIV-
positive partner, and prostitution (Clements, Gleghorn, Garcia, Katz, & Marx, 1997; Huba et al.,
2001; Kipke et al., 1997; Martinez et al., 1998; Whitbeck, Hoyt, Yoder, Cauce, & Paradise,
2001). Homeless youth often use substances to cope with the insecurity and stress associated
with life on the street and/or as a means of mitigating the harmful effects of early exposure to
family violence and abuse experienced by many homeless adolescents (McMorris, Tyler,
Whitbeck, & Hoyt, 2002; Thompson, Rew, Barczyk, McCoy, & Mi-Sedhi, 2009; Tyler &
Johnson, 2006; Tyler, 2008). However, instead of alleviating their problems, sustained drug use
might result in significant and debilitating cognitive, health, and behavioral problems (Nyamathi
et al., 2012; Wenzel et al., 2010).

18

Methamphetamine use increases the risk for HIV transmission and acquisition in a
number of ways. Methamphetamine is known to induce feelings of euphoria, lowered
inhibitions, and heightened sexual experiences, which can increase youth’s risks of acquiring
HIV and other sexually transmitted diseases (Bungay, Malchy, Buxton, Johnson, Macpherson, &
Rosenfeld, 2006; Nyamathi et al., 2012). Methamphetamine users are also more likely to engage
in unprotected anal sex and to have sex with injection drug users, HIV-positive partners, and
those of unknown HIV status; they also tend to report a greater number of sex partners and to
have a history of other sexually transmitted diseases (STDs) (Buchacz et al., 2005; Celentano et
al., 2006; Frosch, Shoptaw, Huber, Rawson, & Ling, 1996). In a Multicenter AIDS Cohort
Study, the relative risk for HIV seroconversion was 1.5 among methamphetamine users
compared with nonusers and was even higher (3.1) among men who used both
methamphetamine and poppers (Plankey et al., 2007). Of greater concern are reports that besides
facilitating the spread of HIV transmission, methamphetamine use is also associated with
harmful behavior changes that can affect the prognosis and overall health of people living with
HIV. For example, studies show that current methamphetamine use decreases adherence to HIV
treatment and medical follow-up (Reback et al., 2004).
Similarly, injection drug use has also been cited as a major risk factor for HIV/AIDS.
Injection Drug Users (IDU’s) are a sub population of drug users who are at risk of HIV from
sharing injection equipment such as needles and syringes (Davey-Rothwell, 2006). In addition,
injection is also an indirect risk factor for HIV; sex with an HIV infected IDU is a major source
of HIV transmission among heterosexuals. According to Centers of Disease Control and
Prevention (CDC) (2003), about one-third of cases of people who were infected with HIV could
be attributed to injection drug use. These cases include IDU’s, their sex partners, as well as

19

children. In 2009, 9% of new U.S. HIV infections occurred among injecting drug users (IDUs)
(Prejean et al., 2011). There are also age and racial/ethnic disparities in HIV prevalence among
injection drug users. More specifically, even though IDU is more prevalent among Caucasians,
prevalence of HIV infection was higher among Hispanics (12%) and non-Hispanic blacks (11%)
than non-Hispanic white (6%) IDU’s (CDC, 2009). Furthermore, prevalence of HIV-associated
risk behaviors appears to be greater among younger users, which is a new trend, and could be
problematic for obvious reasons (CDC, 2009). These results suggest that even within people who
inject drugs, certain groups are more vulnerable than others.
Individual correlates of methamphetamine and injection drug use among homeless youth
A vast majority of the literature on injection drug use and methamphetamine use has been
focused on either adult users (Davey-Rothwell, 2006; Knowlton, Hua, & Latkin, 2005; Latkin et
al., 2009) or young men who have sex with men (Halkitis et al., 2007; Reback et al., 2004;
Shoptaw, 2006). Fewer studies have tried to assess methamphetamine and injection use
specifically among homeless youth. Among youth in the general population, a recent systematic
review found that that ethnicity (being White), risky sexual behavior, a history of heroin or other
opiate use, and a family history of drug use were all associated with greater methamphetamine
use (Russell et al., 2008). In samples of homeless youth, methamphetamine use as been
associated with older age (Nyamathi et al., 2012; Rawson, Gonzales, Obert, McCann, & Brethen,
2005), ethnicity (Nyamathi et al., 2012; Wood et al., 2007), longer duration of homelessness
(Rice et al., 2005; Wood et al., 2007), sexual orientation (identifying as lesbian, gay, or bisexual)
(Douglas, Colfax, Moss, Bansberg, & Hahn, 2008; Salomonsen-Sautel et al., 2008), low level of

20

education (Wood et al., 2007), sexual abuse (Marshall et al., 2011; Nyamathi et al., 2012) and
history of foster care (Nyamathi et al., 2012).
Injection drug use among homeless youth in particular has been correlated with length of
time homeless (Kipke et al., 1993; Rosenthal et al., 2008), sexual orientation (being gay, lesbian,
bisexual, or of “unsure” sexual orientation) (Noell & Ochs, 2001), being a “traveler” (Martino,
Tucker, Ryan, Wenzel, Golinelli, & Munjas, 2011), psychiatric illness such as a history of
suicide attempts (Kipke, Montgomery, &MacKenzie 1993), supporting oneself via the street
economy or survival sex (Fuller et al., 2002; Haley, Roy, Leclerc, Boudreau, & Boivin, 2004;
Kipke et al., 1993; Martinez et al., 1998; Weber, Roy, & Haley, 2002), early deviant or
oppositional behaviors (Dinwiddie, Reich, & Cloninger, 1992) and early onset drug use
(Dinwiddie et al., 1992). However, even though the problems associated with methamphetamine
and injection drug use among homeless young people has been studied; significant gaps exist in
the research literature.
Social networks and engagement in substance use behaviors
Social context, including social norms and the formation of social networks, is an
important determinant of youth HIV risk behavior. Social networks are an important part of
normative adolescent development. These groups are comprised of a set of relationships that link
social actors and mostly include those who are in close proximity to one another (Tyler, 2013).
While the role of individual background characteristics on drug use among homeless youth is
relatively well researched, the role of social-network factors on homeless youth’s HIV risk
behaviors is still being explored (McMorris, Tyler, Whitbeck, & Hoyt, 2002; Rice et al., 2005;
Rice et al., 2008; Rice et al., 2009; Rice, 2010, Rice et al., 2011; Tucker, Edelen, Ellickson, &

21

Klein, 2011; Tucker et al., 2012; Tyler, 2008; Wenzel et al., 2010; Wenzel et al., 2012).
Furthermore, whereas substantial research has linked social networks to HIV sexual behavior
among homeless youth, less attention has been given to the relationship between their social
network properties and methamphetamine and injection drug use behaviors.
Social networks, as noted before, refer to the web of social ties that exist among
individuals (Davey-Rothwell, 2006). Social network analysis (SNA) on the other hand is the
systematic analysis of these social networks (Wasserman & Faust, 1994). Social network
analysis views social relationships in terms of network theory, consisting of nodes (representing
individual actors within the network) and ties (which represent relationships between the
individuals, such as friendship, kinship, organizational position, sexual relationships etc.)
(Wasserman & Faust, 1994). These networks are often depicted in a social network diagram,
where nodes are represented as points and ties are represented as lines. Adolescent social
networks have been generally found to be significant determinants of their HIV risk-taking
behaviors (Alexander et al., 2001; Ennett & Baumann, 1993; Kipke et al., 1997; Rice, Milburn &
Rotheram-Borus, 2007, Tyler, 2008; Wenzel et al., 2010).
A useful framework for examining the role of social networks in understanding
substance use is to view networks according to their structure, composition, and function (Latkin
et al., 2003). The structural aspects of social networks measures characteristics of social
networks such as size of network, centralization vs. isolation, density etc. (Valente, Gallaher, &
Moutapa, 2004). Compositional characteristics refer to relationship traits of network members
such as proportion of social ties (for e.g. family vs. peers) (De, Cox, Boivin, Platt, & Jolly,
2007). Functional characteristics on the other hand refer to the roles that network member’s

22

play. The main functions of social networks have been characterized as social influence, social
engagement, social support, and access to resources (Berkman, Glass, Brissette, & Seeman,
2000). Norms are often the direct result of one’s social ties and are embedded in one’s social
networks. Social norms fulfill a very important function in social networks by acting as a source
of social influence and regulation (Horne, 2001).
Structural network characteristics
Structural network theory is primarily concerned with characterizing network structures
(e.g. small worldness) and node positions (e.g. core/periphery) and associating it with a
multitude of outcomes, including job performance, mental health, organizational success etc.
(Borgatti & Halgin, 2011). As I noted in the previous chapter, social network analysis can be
conducted at two levels: egocentric and sociometric analyses. (I) egocentric (or local) networks
where only the ego (or the respondent) is interviewed about their network contacts and how these
contacts are related, (See Fig 2) and, (II) sociometric (or global) data (see Fig 3) where members
of the entire delineated community is queried regarding their relationships to each other.
Although egocentric networks are easier to investigate, they show only the characteristics of a
network from the perspective of the index subject while sociometric analyses consider the sum of
connected egocentric networks. Both techniques have been used to understand drug use among
housed youth. Work on homeless youth, however, has focused almost exclusively on dyadic or
egocentric influences.

23

Figure 2: Egocentric Relationships Figure 3: Sociometric Relationships
(Circles represent people and lines represent (Circles represent people and lines
relationships) represent relationships)

Among egocentric measures of network structure, network size and density have been
used most frequently; however, results have been less than consistent. For example, Rice and
colleagues (2005) found that a greater density of injection and methamphetamine drug using
peers increases the likelihood of drug use. On the other hand, Ennett and colleagues (1999)
found that youth without a network were more likely to be drug users than youth who named a
network. Wenzel and colleagues (2010) found no statistically significant relationship between
network density and substance use in a probability sample of homeless youth in Los Angeles,
California. However, egocentric network structural measures (such as size and density) are
limited measures of network integration because it only measures the relationship between egos
and alters (Ueno, 2005). Since egocentric data is collected from only the focal index person, the
structure of the entire network cannot be ascertained (Doherty, Padian, Nancy, Marlow, & Aral,
2005). Beyond the egocentric level, it is likely that peer influence also occurs at the level of the
broader peer group, making it important to understand social network structure and dynamics
within groups (Lansford, Killeya-Jones, Ley, Miller, & Costanzo, 2009). Since the sociometric
Ego
Alter
Alter
Alter
Alter
A
D
C
F
B
E

24

design provides for theoretical specification of an entire network, it allows for direct examination
of the relationship between network structure and the distribution of drug use norms in the
network.
One of the significant and unique aspects of sociometric network analyses is its ability to
characterize people in terms of their position in the larger network (Valente et al., 2004). The
prominence of a network member’s position in their network is measured by the member’s
centrality (Scott, 2012). Actors at the center of a network have more linkages within that network
and consequently are more active, in comparison to peripheral actors. In studies of non-homeless
youth, social position has been found to be an important predictor of substance use (Alexander.
Piazza, Mekos, & Valente, 2001; Ennett & Baumann, 1994; Ennett et al., 2008; Flom et al., 2001;
Pearson & West, 2003). For example, in their investigation of IDUs in New York City, Friedman
et al. (1997) found that core members of a network were more likely to engage in drug equipment
sharing and had a higher probability of HIV acquisition and transmission than peripheral network
members. These results are supported by a study in France in which people in the densest and
most central part of injecting networks had the highest likelihood of needle sharing (Lovell,
2002). Interest in social position underscores the argument that adolescent substance use is
embedded in a larger peer context, typically an entire network of youth within a bounded
population (Christakis & Fowler, 2007; Scott & Hofmeyer, 2007).
Unfortunately, this attribute of network analyses has received little empirical attention
among homeless youth. Although peer relationships have held a central role in theoretical and
empirical work on homeless youth, sociometric data depicting the larger web of interconnected
ties have only been collected recently (Rice et al., 2011). To the best of my knowledge, only one

25

study until date has investigated the associations between sociometric status and HIV risk
behaviors among homeless youth. This study suggests that youth who are more influential in their
networks are more risk-taking. Specifically, this study found that youth in the periphery were
significantly less likely to report unprotected sex (Rice et al., 2011). This preliminary finding
indicates that the structure of the network and one’s position within it can determine the kind of
HIV risk behaviors that youth engage in. However, what is not clear is whether these larger
sexual and substance use networks are related to group norms surrounding these behaviors. These
data were therefore used to understand whether these network structures and norms are
interrelated.
Normative network characteristics
The structural network perspective emphasizes the constraining influence of structural
network characteristics on behavior. However, it fails to account for normative forces that
facilitate adaptation to these shared social environments (Lakon, 2004). Norms influence
behavior through modeling (Bandura, 1977), comparison of attitudes behaviors with one’s
referent groups (Marsden & Friedkin, 1994), and through social feedback (Fisher & Misovich,
1990). Norms are defined as perceived rules or properties of a group that characterize specific
beliefs around what behaviors are considered acceptable or common within that group (Kincaid,
2004). Perceived norms have been generally categorized as descriptive or injunctive norms.
Descriptive norms indicate the perceived prevalence of a behavior in a group whereas injunctive
norms refer to perceived approval of a behavior (Davey-Rothwell & Latkin, 2007). Extant
research shows that network norms are strongly associated with substance use behavior.
However, most of the research describing the association between network norms and drug use

26

behaviors has tended to focus on adult IDU’s (Davey-Rothwell, 2006; Kottiri, Friedman,
Neaigus, Curtis, & Des Jarlais, 2002; Knowlton, Hua, & Latkin, 2005 Latkin et al., 2009).
One exception is a study conducted by Ennett and colleagues (1999) which focused on
the perception of network members’ approval/disapproval regarding engagement in illicit drug
use behaviors among homeless youth. Ennett et al. (1999) found that pressure to engage in drug
use in a network was not reported too often (only about 15% of the time), however when
reported, it was positively associated with risk taking behaviors. However, this study did not
specifically focus on methamphetamine and injection drug use and instead focused on illicit drug
use, which could include a number of other drugs. Tyler and colleagues (2013) also studied
social norms among homeless youth, but limited their focus to sexual risk behaviors. In their
qualitative examination, they found that out of the 19 youth who they interviewed, only three
youth reported that their social networks endorse safe sex practices; however, more importantly
these youth were more likely to use condoms more consistently (Tyler et al., 2013). Taken
together, these two studies signify the potent influence of norms on homeless youth’s behaviors.
Among adult IDU’s, research shows that social injecting and supportive injecting
environments produce strong social ties that promote mutual injecting and create norms for risky
behaviors. Despite the lack of focus on homeless youth, these studies investigating associations
in other populations could be helpful in identifying correlates of methamphetamine and injection
drug use behaviors among this population. Generally, these studies have reported that
individuals who believed that their network members engaged or endorsed drug equipment
sharing were also more likely to share injection equipment (Davey-Rothwell, 2006; Friedman et
al., 1997; Lakon, 2004; Latkin et al., 2003; Latkin et al., 2009). Nonetheless, it is also important

27

to note that these studies may not generalize to homeless adolescents because of their
developmental context and differences in their social network properties, which makes it much
more essential to investigate these associations among this population.
Additionally, it has also been found that descriptive and injunctive norms have
differential impact of people’s behavior; one type of norm might be more persuasive than the
other depending on the behavior (Davey-Rothwell, 2006). For example, when investigating why
people choose to exercise, Okun and colleagues (2003) found that people who exercised
indicated that they did so because they believed that other people engaged in it (descriptive
norms), and not because of how they perceived others would react if they did not exercise
(injunctive norms). On the other hand, Rimal and Real (2003) found that college students
drinking behaviors were influenced more by the injunctive norms prevalent within their referent
groups. Other studies have also reported similar results (van Empelen, Pepjin, Herman, Kok, &
Jansen, 2001; Bunk, Bakker, Siero, van den Eijnden, & Yzer, 1998; Latkin et al., 2003). These
findings suggest a need for further understanding on the unique effects of descriptive vs.
injunctive norm in understanding behavior.
Compositional Network characteristics
Compositional characteristics of networks refer to the characteristics of people that form
a part of one’s networks. It is considered important to account for the composition of social
networks, given that patterns of social influence are typically neither random nor equal among all
network members. It is generally believed that having supportive people in one’s networks who
are non-drug users may serve as an avenue for social integration, as these individuals can
discourage involvement in a drug using life-style and change fatalistic attitudes toward risk

28

taking (Lovell, 2002). Typically, research on homeless youth has focused on the problematic
influence of peers on risk-taking behaviors, with little attention given to positive impacts of
social support or affiliation with pro-social relationships (Rice et al., 2007). Conventional
wisdom dictated that homeless youth have strained family relationships and their social networks
are only comprised of other similarly situated street peers (Whitbeck, 2009). However, evidence
from various studies suggest that homeless youth’s relationships are not confined to street
associations alone (Ennett et al., 1999; Johnson, Whitbeck, & Hoyt, 2005; Rice et al., 2005;
Rice et al., 2007; Rice et al., 2008, Rice 2010; Wenzel et al., 2012; Whitbeck & Hoyt, 1999). For
example, in their study, Johnson et al., (2005) found that over 80% of youth reported having at
least one non-street relationship. Likewise, Wenzel and colleagues (2012) reported that on an
average, youth designated that 17.91% of their networks were comprised of relatives or family
members. More significantly, even though relatives comprised less than 20% of their networks,
youth stated that they primarily relied on their relatives for tangible and emotional support
(67.36%).
Presence of relationships that are considered prosocial (defined as family, relatives,
home-based friends, service providers etc.) have implications for engagement in risk behaviors
(Hagan & McCarthy, 1997; Johnson et al. 2005; Rice et al., 2005; Rice et al., 2007; Rice et al.,
2008, Rice 2010; Wenzel et al., 2012; Whitbeck & Hoyt, 1999). Maintaining connections with
home-based peers and supportive family members has a positive function for most homeless
youth as evidenced by existing studies (Rice et al., 2007; Rice et al., 2008; Wenzel et al., 2010;
Tyler, 2008). For example, Rice and colleagues (2007) found that associations with prosocial
peers’ reduced hard drug use (cocaine, methamphetamine and heroin) over time (Rice et al.
2007). Likewise, other studies have found that youth who have connections with family

29

members are less likely to report engaging in any substance use (Ennett et al. 1999; Wenzel et
al., 2010; Tyler, 2008). In light of the emergent evidence of the heterogeneity of homeless
youth’s network relationships and its implications for multiple sources of influence, leading
scholars (Wenzel et al., 2010; Whitbeck, 2009) have called for a more comprehensive
investigation of homeless youth’s networks in relation to their substance use.
Internet and social media use among homeless youth-implications for compositional
characteristics of networks
Internet and social media (ISM) have initiated a new age in the way teens communicate
and socialize and has soon grown into an important part of their developmental transitions
(Pascoe, 2012). About 17 million youth ages 12 through 17 use the Internet. Teenagers’ uses of
ISM play a major role in their relationships with their friends, their families, and their schools
(Lenhart, Rainie & Lewis, 2001). ISM has almost become a proxy for another level of social
environment for adolescents. More specifically social networking sites and technology are
changing the way teens communicate with friends, relatives, romantic partners etc. (Gudelunas,
2012; Greenhow & Robelia 2009).
Social networking sites functionality has gone beyond their original intended purpose of
simply locating and maintaining friendships, or perhaps, providing a way to bide time, but more
pertinently, have become a way of gaining social capital, managing emotional lives and social
positions (Gudelunas, 2012; Papacharissi & Rubin, 2000). A study conducted by Ellison,
Steinfeild, and Lampe (2007) found that college students used Facebook as a way to deal with
low self-esteem and low life satisfaction to bridge social capital and stay connected while
reducing feelings of isolation. Similarly, among homeless youth, ISM have provided these

30

otherwise marginalized and isolated youth with a way to disengage from their street identities
(Karabanow & Naylor, 2010; Mitchell & LaGory, 2002), and connect with relationships from
home (Barman-Adhikari & Rice, 2011; Guadagno, Muscanell, & Pollio, 2012; Pollio, Batey,
Bender, Ferguson, & Thompson, 2013; Rice et al., 2010, Rice, 2010; Rice & Barman-Adhikari,
in press; Young & Rice, 2010).
Homeless youth are resource poor, and lack even the most basic needs. Therefore, most
people assume that they have inadequate to no access to technology and are victims of a “digital
divide”. Surprisingly, new data on technology use by homeless young adults does not support
this notion (Barman-Adhikari & Rice, 2011; Pollio et al., 2013; Rice et al., 2010, Rice, 2010;
Rice & Barman-Adhikari, in press). Emerging evidence suggests that homeless youth use the
internet to keep in touch with friends, family, and employers. Prior to the emergence of ISM,
homeless youth had very limited avenues through which they could connect with their bridging
social ties, and thus leverage such critical social and emotional resources. For example, Barman-
Adhikari & Rice (2011) found that 57% of youth in their study were using the internet to connect
to their parents, and 66% were connecting to home-based peers. More notably, results indicated
that continued connection with parents via the ISM was significantly associated with youth
seeking HIV or STI information. Likewise, other studies have found that these online
relationships are crucial in understanding homeless youth’s behaviors (Rice et al., 2010, Rice et
al., 2011; Rice, 2010; Rice & Barman-Adhikari, in press; Young & Rice, 2010). The role of the
ISM in enabling these relationships is hence of crucial importance, and will be explored further
in this study.

31

Relationship between structural and normative network characteristics
Besides understanding how norms influence risk within this population, one of the main
goals of this study is to understand how the configuration of one’s networks affects not just
behavior, but perceptions of norms around those behaviors. Community level interventions
usually target group norms around risk behaviors (Kelly et al., 1992) and typically recruit
community persons who are influential (or central) to their networks as change agents. Many
eminent scholars in the field have suggested that HIV preventive efforts have been hampered
because of the paucity of research examining clustering of norms within specific risk-taking
social network structures (De et al., 2007; Latkin et al., 2009).
Although the empirical literature reviewed above suggests that being in certain positions
in sociometric/whole networks could be associated with riskier drug-related behaviors, there is
little information on how these network positions influence normative perceptions of substance
use among vulnerable groups (Friedman, Curtis, Neaigus, Jose, & DesJarlais, 1999; Frey &
Meier, 2004; Sherman, Latkin, & Gielen, 2001). This study will give us some important
direction for what kinds of peer-led programs would be most effective for homeless youth. For
example, if this study finds that central youth are more likely to endorse risky behavioral norms,
then the Popular Opinion Leader (POL) model (Kelly et al., 1992) which relies on recruiting the
core, seems less viable than a model like SHIELD (Self-Help in Eliminating Life-Threatening
Diseases) (Latkin et al., 2003) that looks to make change across a network via localized action
around a dispersed group of peer leaders located throughout a risk-taking network.
To the best of my knowledge, this will be the first study to look at associations between
social network structural and normative characteristics among homeless youth. In studies among

32

adult IDU’s, social network structure has been found to be significantly associated with
normative characteristics. For example, both network size and cohesion have been found to be
associated with normative perceptions of HIV risk behaviors among adult IDU’s (Barrington,
2008; Davey-Rothwell, 2006; Latkin, Mandell, Vlahov, & Celentano, 1996; Latkin et al., 2003).
Network cohesion reflects the interconnection or density of ties among network members
(Wasserman & Faust, 1994).
Cohesive networks are not considered to be efficient in terms of receiving new
information because of the inherent dense nature of ties (Barrington, 2008; Latkin et el., 2003),
however they are effective channels of social influence and thus the diffusion of norms. Dense
networks may be prone to what is described as a “spiral of silence” (Latkin et al., 2003);
discussing new behaviors might be perceived as being disruptive to the existence of the group
(Latkin et al., 2003). While the dense nature of ties can prevent new information from coming
in, once new information is introduced, they can also reinforce the adoption of new norms
(Latkin et al., 2009). On the other hand, weaker ties can facilitate the spread of new information,
but these new norms might need to be reinforced more strongly in the absence of strong ties
(Barrington, 2008).
Even though these findings give us important insight into how network structures affect
norms in specific networks, there are still areas that remain understudied. First, all the studies
above collected egocentric data, which as noted before, provide us limited information about
network structures. More importantly, these kind of data do not allow for the identification of
influential people within bounded networks, and the kind of norms that they endorse. Such
information can be crucial when designing community level interventions. Furthermore, most of

33

these studies focused exclusively on IDU networks (i.e. people who only inject drugs).
Homeless youth generally tend to use a wide variety of substances. It is quite possible that
people in different parts of the network will be exposed to norms supportive of different kind of
substances (for e.g. methamphetamine vs. heroin). For example, preliminary results from a
recent study (Rice et al., in progress) revealed that while methamphetamine was concentrated in
the core region of the network of homeless youth, heroin was more prevalent among youth in the
periphery. This has implications for how future interventions can be tailored to these specific
network spaces by targeting norms supportive of these specific drugs.
Relationship between compositional and normative network characteristics
A number of existing studies (Borsari & Carey, 2003; Larimer et al., 2009; Neighbors et
al., 2010; Neighbors, Lee, Lewis, Fossos, & Walter, 2009) support theoretical perspectives (i.e.,
Social Comparison, Festinger, 1954; Social Impact Theory, Latané, 1981; Social Identity
Theory, Hogg, Abrams, Otten, & Hinkle, 2003, Terry & Hogg, 1996) suggesting that the
reference groups to which individuals are closely connected by proximity or identification are
more relevant and have greater influence on individual behavior and attitudes than reference
groups to which individuals do not feel connected. Recent evidence suggests that homeless
youth’s networks are much more heterogeneous than what was previously assumed. This has
implications for the kind of norms that these youth might perceive is prevalent within their
networks. For example, family norms might discourage drug use; peer norms on the other hand
might support use.
Among college students, studies have found that the student body as a whole is typically
not the most salient referent group in understanding their substance use behaviors. For instance,

34

LaBrie and colleagues (2011) found that students were less likely to endorse permissive
marijuana use norms when parents were named. On the other hand, they were more likely to
perceive that their “close friends” would approve of their marijuana use. This pattern has been
also found to be true in other studies; the relationship between norms and behavior has been
shown to differ depending on the nature and influence of the referent (Borsari & Carey, 2003;
Carey, Borsari, Carey, & Maisto, 2006; Cho, 2006). Since people’s perceptions of norms vary
according to the referent, it is necessary to understand how someone will react to a specified
referent (Cox, 2010). For example, among homeless youth, it is possible that if they are
connecting to more prosocial relationships (such as family and home-based peers), they would be
more likely to emulate the norms of these networks instead of their more risky peers.
Several studies report that homeless youth who use illicit drugs and engage in risky
sexual practices generally have friends who engage in similar behaviors (Kipke, Unger, Palmer,
Iverson, & O’Connor, 1998; Rice et al., 2005); however, none of these studies has investigated
whether these perceptions differ based on the referent group. To the best of my knowledge, only
one previous study (Wenzel et al., 2012) has investigated how the different referent groups might
be perceived as contributing to different kinds of norms. Notably, they found that people that
youth met on the streets were more likely to be perceived as being engaged in risky behaviors
than other network members.
Furthermore, little is known whether the presence of these different referent groups could
alter youth’s perceptions of what is considered acceptable behavior. For example, if one is
connected to parents or other prosocial relationships, then his/her perceptions of what is socially
sanctioned behavior will be different than a youth who might not have these protective

35

influences in his/her life. It is important to understand how much influence these different
referent groups yield over these youth, so that interventions can be tailored to both mitigate the
negative influences and amplify the protective ones.
Summary
Homeless youth are a unique subset among the larger group of people experiencing
homelessness. Discussed above are a multitude of individual and social network factors which
influence methamphetamine and injection drug use among homeless youth. Prevention of
methamphetamine and IDU and their sequelae requires not only knowledge of the incidence of
these substances, but also the social context of their use, so that effective interventions can be
developed. One of the most successful examples of community-level behavior change has
involved altering social norms (Latkin et al., 2009). However, even though there is a large body
of work documenting the associations between social norms and behaviors, not many studies
have focused on factors that affect the prevalence of these norms.
It is important to understand how these norms are perpetuated in order to identify factors
that are amenable to modification in the context of a network intervention. Furthermore, there is
a need for a better understanding of how the interplay between network structure, composition,
and function affect risk perceptions for the individual and to identify which combination of these
elements can be targeted most effectively by interventions to combat norms supportive of risky
behavior especially among this vulnerable population (De et al., 2007). In particular, network
structure and composition can be used to identify potential routes for norm diffusion and locate
targets for prevention with the ultimate goal of modifying behavior. Taken together, the

36

egocentric and sociometric data was able examine multiple research questions that have not yet
been addressed in the HIV risk norms literature.

37

CHAPTER THREE
THEORETICAL CONCEPTUALIZATION
Over the past several decades, researchers have increasingly emphasized the use of
ecological models and methods in behavioral and social sciences in general, and public health in
particular (Luke & Harris, 2007). Until recently, much of the research literature has focused on
behavior as a function of individual characteristics. This theoretical perspective is increasingly
being called into question, because it neglects social contexts and interactions linked to behaviors
(Carpentier & White, 2002). The social norms approach provides a theory of human behavior
that has important implications for health promotion and prevention. Social norms are not only
significant in maintenance of social behaviors; they also present barriers to changing social
behaviors (Latkin et al., 2003). The main interest in norms from a public health perspective is
motivated by findings that changing individuals’ actions is best achieved by highlighting and
influencing the behavior of others around them; i.e. focusing on social norms to promote sustain-
able behavior. In this chapter, I draw upon a number of social network and social influence
theories to elucidate the role social norms in understanding behaviors. In particular, the
relationship between social network theory and social norms is discussed to understand how
structure and norms are interrelated. Social Capital Theory is also highlighted to describe how
the dissemination of new technology can create avenues through which homeless youth can
access relationships outside of street life, which could provide opportunities for mobility for
these otherwise, disenfranchised youth.
Social Norms-Definition and Origins
Norms are regarded as cultural and social phenomena that prescribe and proscribe
behavior in specific circumstances (Hechter & Opp, 2001). Seminal studies by Sherif (1966) and

38

Asch (1952) have documented how individual behavior is shaped and influenced by others in the
social environment. Social norms usually include societal expectations of our behavior (Blake &
Davis, 1964; Pepitone, 1976), the expectations of significant others (Fishbein & Azjen, 1975),
people’s own expectations of their behavior (Schwartz, 1977), and standards that develop out of
observing other people’s behavior (Bandura, 1977). Norms usually do not have any legal basis,
and in many situations, can be contrary to the law (for e.g. the use of illicit drugs) (Coleman,
1990). Norms instead represent social regularities, which result in the imposition of certain
informal standards and constraints on human behavior, generally in the pursuit of maintaining
social order (Major, 2000). Norms have therefore long been considered responsible for
regulating social behavior by acting as a source of informal social control.
For years, norms have been used by social scientists to explain a variety of health
behaviors. However, less attention has been paid to the ways that they are sustained within
groups. It is important to understand the societal processes through which norms are maintained
and changed because they could provide important insight into factors that are modifiable within
the context of interventions. Generally, two perspectives have been used widely to account for
the emergence of norms. Theorists influenced by the anthropological traditions of Boas and
Mead believe that norms are cultural and its value is specific to that that particular group; its
power is granted by its acceptance within that culture (Cialdini & Trost, 1998). According to this
perspective, norms persist because they are rewarded within that particular system, either directly
or through vicarious reinforcement (Opp, 1982). For most sociologists, social enforcement is an
essential component of norms and cannot exist without it. While some focus on internalized
sanctions as an enforcement mechanism, a majority of scholars emphasize the role of external

39

sanctions, usually imposed by people who form a part of that person’s social network (Horne,
2001).
An alternative perspective argues that norms are functional and develop in order for
people to adapt to a shared social environment (Campbell, 1975; Sherif, 1936). This is typically
accomplished by imitating behavior that is observed within the immediate environment (Cialdini
& Trost, 1998). The functional approach tends to argue that there is consensus within the social
system. When many people engage in the same behavior, that behavior becomes associated with
a sense of oughtness (Horne, 2001). Consequently, these behaviors become normative. In this
regard, norms almost represent a central tendency of behavior (i.e. what is typical) (McAdams,
1997). Studies show that people mimic each other´s behavior, perhaps as a way of smoothing
social interactions (Berkman et al., 2000). For example, among high-risk youth, if they are
immersed in a peer group that engages in a high level of drug use, then observing and modeling
this standard of drug use might send the message that this behavior is acceptable, and gradually
become a norm for them (Bauman & Ennett, 1996; Tyler, 2013). Studies that have tried to
examine how norms emerge and diffuse from one person to another have found that both
perspectives can account for normative behavior (Cialdini & Trost, 1998).
However, one of the limitations of the functionalist perspective is that not all norms are
necessarily established in order to achieve positive and healthy outcomes (Barrington, 2008).
Therefore, it is important to note that not all norms have desirable consequences; social order can
also be achieved with negative behavior. Portes (1998) for example highlights the concept of
“downward leveling norms” which function to keep people isolated or marginalized. Especially
since norms are used as means of social control, the same social solidarity and trust that can
provide sources of socio-economic mobility can also have the opposite effect (Portes, 1998).

40

Norms running counter to the behaviors of the overarching society or culture may be transmitted
and maintained within small subgroups of society (such as homeless youth). Portes (1998) cites
the mafia families, gambling rings, and youth gangs as examples of how embeddedness in social
structures can facilitate downward leveling norms. Similarly, the Marxist perspective also
focuses on the coercive nature of norms in an effort to clarify the process through which norms
can lead to social exclusion, and exploitation (Barrington, 2008). Therefore, in understanding
how norms are sustained, it is important to consider both functional and structural characteristics
that facilitate the establishment of norms. This is considered important because while some
group structures can engender healthy norms, other structures might do the exact opposite.
Types of Norms
In theorizing about norms, scholars have stressed on the importance of distinguishing
between collective and perceived norms (Lapinski & Rimal, 2005). Collective norms refer to
norms that exist at a collective level (such as a group, community, or culture). Perceived norms
on the other hand refer to the individual’s interpretation of these group norms (Lapinski & Real,
2005). It is important to distinguish between these two norms, because studies have found that
individuals might not infer a collective norm accurately; also known as-pluralistic ignorance
(Berkowitz, 2004). Collective norms generally operate at the level of the social system, which
could be a specific social network, or even the entire society (Lapinski & Real, 2005). On the
other hand, perceived norms exist at the individual psychological level and represent the
individual’s representation of the collective norm (Lapinski & Real, 2005). Studies have found
that in understanding behavior, perceived norms are more accurate predictors of behavior than
collective norms (Berkowitz, 2004). This study therefore focused on perceived norms of
substance use behaviors among homeless youth.

41

Two types of perceived norms are generally described in the literature reflecting the two
perspectives described in the previous section- descriptive norms and injunctive norms (also
referred to as prescriptive and proscriptive norms) (Cialdini, Reno, & Kallgren, 1990).
Descriptive norms refer to the perceptions of the prevalence of a behavior, which conforms
closely to the functional perspective (i.e. whether people engage in a particular behavior)
(Cialdini et al., 1990). On the other hand, injunctive norms consider the outcome of a behavior
within a specific social or behavioral context and are aligned better with the culture specific
perspective of norms (i.e. whether it would meet people’s approval or whether it would be
sanctioned) (Davey-Rothwell, 2006). The crucial difference between the two is that descriptive
norms generally do not involve any social sanction for non-conformity to the norm.
Differentiating between these types of norms is motivated by the long-standing philosophical
difference between “the ought” and “the is” (Christensen, Rothgerber, Wood, & Matz, 2004).
This difference also is reflected in theories in which people are influenced by others to gain
personal and social rewards versus to act effectively (e.g., Deutsch & Gerard, 1955; Cialdini &
Trost, 1998; Wood, 2000). Empirical examination of these two constructs has also revealed that
their significance in determining a behavior varies based on the context and the specific behavior
(Reno, Cialdini, & Kallgren, 1993).
More specifically, in certain circumstances, people are more motivated by the desire to
engage in a typical behavior. On the contrary, there might be other situations where people might
be more motivated by social rewards. For example, studies have found that injunctive norms are
more significant in determining consistent condom use (van Empelen, Kok, Jansen, & Hoebe,
2001), whereas descriptive norms were more salient in understanding the intention to consume
alcohol (Reno, Cialdini, & Callgren, 1993). Condom use is private behavior, which usually

42

occurs between two people, whereas alcohol is often consumed in social settings. Therefore, it is
possible that descriptive norms are more salient in influencing a more social and visible behavior
because of the desire to conform. On the other hand, for private behaviors like condom use,
people might be influenced by how people might perceive that behavior (whether they would
encourage and object to it).
Norms and behavior
Norms are a key component of several theories of health behavior (Latkin et al., 2009).
The tests of these theories have revealed that norms are significant in understanding why people
adopt new behavior. Two popular theories that have included a norm construct as part of their
overall model are the Theory of Reasoned Action (Ajzen & Fishbein, 1980) and the Theory of
Planned Behavior (Ajzen, 1991). These theories propose that the opinion of the important people
in our lives (defined as subjective norm) is associated with the intention to perform that behavior
(Davey-Rothwell, 2006). These theories however only account for injunctive norms and do not
capture the concept of descriptive norm.
Perhaps, one of most seminal theories, which broadly covers both categories of perceived
norms (i.e. descriptive and injunctive norms), is the Social Cognitive Theory (SCT) (Bandura,
1977). According to SCT, the social environment that a person is embedded in provides him/her
with models for behavior. Observational learning occurs when a person replicates the behavior of
another person and receives reinforcement (either positive or negative) for that behavior
(Bandura, 1977).
Festinger’s (1950, 1954) Social Comparison Theory is another theory that can inform
how norms affect behavior. Festinger proposed that individuals form their attitudes by
comparing one’s attitudes and behaviors against other persons (Friedkin, 2001). Festinger’s

43

theory is especially important in understanding the heterogeneity of norms. He proposed that
norms are not based on any objective standards, but are decided based on the cues that are
received from social network members.
Social Network Structures and Perception of Norms
Norms as we have noted before, are predominantly social in nature. Social norms are for
that reason inherently linked to one’s social networks as they usually materialize out of one’s
interaction with others; and the sanctions for deviations come from the people who comprise
one’s networks (Cialdini & Trost, 1998). Researchers have argued that most behavior is closely
embedded in a network of personal relations, and that a theory of norms cannot leave the specific
social context out of consideration (Granovetter 1985). The utility of social network theory in
understanding social norms rests on the assumption that social structure of networks is
responsible for determining individual behavior and attitudes by shaping the kind of people that
one connects with which in turn can determine access to opportunities or place constraints on
behavior (Berkman & Glass, 2000).
As noted in the previous chapter, social network analysis includes analyses of both
egocentric networks (where the individual is at the center) and sociometric networks (where an
entire community is queried, and both direct and indirect ties are studied) (Berkman & Glass,
2000). Norm diffusion is an element of both the ego-centered as well as socio-centered approach
(Horne, 2001). In the ego-centered approach, the focus is on the importance of modeling
(Bandura, 1977) and especially who comprises these networks (e.g. family vs. peers) and who
the individual chooses as the most salient referent (Horne, 2001). This is described in detail in
the proceeding section.

44

Moving from egocentric to sociometric theorizing about peer influences requires that we
include not only dyadic influences, but also the macro level connections among all the actors in
the network. On the sociometric or community level, it has been proposed that people who
occupy similar social positions in a larger social structure are more likely to adopt similar norms
(Burt, 1987; Horne, 2001). A feature of social networks that could explain this is a high degree
of clustering—meaning that two people who both have a link with a third are likely to be linked,
themselves. Higher clustering in a network indicates high “cliqueishness,” so attitudes and norms
also tend to cluster in these networks (Latané & Bourgeois, 1996). For example, studies have
found that people who are too enmeshed within their social networks, then they are subjected to
a lot more social influence than people who are in loosely connected networks (Forsyth, 1983;
Nadler & J. Fisher, 1988). Clustering can also be explained by the concept of homophily
(McPherson, Smith-Lovin, & Cook, 2001). Homophily refers to the tendency for people to have
ties with people who are similar to themselves in socially significant ways. Homophily could in
turn be the result of two processes-social influence- a process that makes the attitudes of linked
individuals become similar or through social selection-a process by which people form links to
others with similar attitudes (Mason, Conrey, & Smith, 2007). Disentangling the effects of social
selection vs. influence is beyond the scope of this study. However, studies have shown that these
two processes often work in tandem to influence norms and behavior.
Social network analysis with homeless youth has been conceptualized as a way of
measuring homeless youth’s emersion into street culture (Whitbeck, Rose & Johnson, 2009) and
sociometric data helps us assess precisely how a youth is positioned vis-à-vis others in a network
of other homeless youth. In theory, youth who are positioned in different social locations in the
network (such as the core or the periphery) might vary in their perception of what is normative

45

because of several reasons. Placement in these social positions reflects varying levels of
interaction with their street-peers (Ennett & Bauman, 1993). Therefore, one can hypothesize that
youth who are in the core or center of their street networks might adhere more closely to the
norms of their street-peer group and engage in greater substance use. On the other hand, youth
who are in the periphery might have access to diverse opportunities for obtaining information,
ideas, and resources from many different sources and be less constrained by the influence of their
street-peers (Ennett & Bauman, 1993).
Social Referents and Normative Behavior
One of the properties of networks that determine the degree to which influence occurs,
and might explain how normative perceptions are formed is the composition of people who make
up the network (Barrington, 2008). Hyman (1980) proposed that an individual’s attitudes and
behaviors are influenced by reference groups. A reference group is defined as a cluster of people
who serve as a reference point for behaviors and attitudes (Davey-Rothwell, 2006). Social
referents wield their influence over peers’ perceptions of collective norms through the
mechanism of everyday social interaction, particularly interaction that is frequent and personally
motivated (Paluck & Shepherd, 2012). However, it is important to remember that each person
has several reference groups, and depending on who constitutes this group of people, they might
have varying levels of influence on the individual depending on the context and the behavior
(Davey-Rothwell, 2006). Accordingly, some groups may hold more weight than others when
evaluating which behavior is considered the norm.
Social Identity Theory (SIT) (Tajfel & Turner, 1986) also provides us with a useful
framework to understand how different referent groups influence people’s perception of what
can be regarded as normative. From a social identity perspective, norms are linked to specific

46

social groups, and the norms of more relevant groups should be more influential (Terry, Hogg, &
White, 2000). Aligned with a cognitive approach, SIT emphasizes that beliefs about appropriate
behavior directly emanate from one’s perception of which group he/she belongs to (Christensen,
Rothberger, Wood, & Matz, 2004).
The significance of the composition of social networks is also reflected in Bandura’s
(1977) Social Learning Theory. According to this theory, attentional processes are crucial to
observational learning (Barrington, 2008). Individuals are exposed to many modeling processes,
and the influence of these processes will depend on whom the individual associates with and the
role of the people in their lives (Barrington, 2008). Therefore, it is important to recognize that
there is a selective aspect of the observational learning process. Therefore, homeless youth who
are still connected to non-street ties could have access to more models of behavior and
consequently have more prosocial norms than youth who have more homophilous street ties.
However, it has also been now widely recognized that individuals generally do not
belong to one social group and might have multiple intersecting identities. Social Identity
Complexity (Roccas & Brewer, 2002) is a theoretical construct that refers to an individual's
subjective representation of the interrelationships among his or her multiple group identities.
Social identity complexity is also associated with a higher inclination for openness to change,
and higher tolerance for diversity of ideas (Roccas & Brewer, 2002). Homeless youth’s networks
are much more heterogeneous that what was previously thought. Among college going youth,
Rimal et al. (2005) found that group identity has a moderating effect on the association between
descriptive norms and behavior. However, one does not know whether this is true for homeless
youth. Therefore, this study hypothesized that homeless youth whose networks are diverse and

47

consequently do not have a strong sense of group identity will be expected to perceive that risky
behaviors are less normative than youth who have strong affiliation with their street networks.
Social Capital Theory: Online Relationships and Composition of Networks
The internet is a critical resource for homeless youth, providing a vast repository of
readily available information (Jones & Fox, 2009). More importantly, internet and social media
use among homeless youth is associated with inclusion in social worlds beyond their street
environments (Barman-Adhikari & Rice, 2011; Roberson & Nardi, 2010; Rice et al., 2009; Rice,
2010; Rice & Barman-Adhikari, in press), making youth less bound to the resources of their
current geographic area or neighborhood. Social Capital Theory provides a succinct framework
for thinking about how online social interactions may facilitate differential norms regarding
substance use behaviors among homeless youth. Social capital is derived from people’s
interactions within social relationships (Bourdieu, 1986; Burt, 1992; Coleman, 1990; Lin, 1999;
Putnam, 2000; Woolcock, 1998). Portes (1998) describes social capital as the ability for people
to secure benefits and resources through memberships in social networks. Lin (1990, 1999) sees
social capital as resources that are activated by individuals through their social ties. Coleman
(1988) conceptualizes social capital as a set of relationships among people that facilitate
productive activity. Studies with youth in the general population support the idea that internet
users can enhance their social capital by engaging and investing in online activities (Pénard &
Poussing, 2010; Quan ‐Haase & Wellman, 2004).
Social capital can be defined as the ability of individuals to accumulate benefits by virtue
of their personal relationships and memberships in particular social networks (Warschauer,
2004). These benefits can be shared through two types of capital: bonding social capital and
bridging social capital (Putnam, 2000). Bonding social capital refers to relationships shared

48

among dense, inward-looking social networks (Warschauer, 2004). Bonding social capital often
plays a dual role; while it strengthens solidarity within a particular group, it can occasionally
come at the cost of distance from other groups, limiting one’s access to other significant social
sources of information and support (Warschauer, 2004). In the case of homeless youth, bonding
social capital translates into connecting with other homeless youth (Stablein, 2011), relationships
often characterized by more conflict, less stability, and failing to offer the opportunity to develop
conventional behaviors (Whitbeck, 2009).
Bridging social capital refers to resources derived from a heterogeneous group of people,
and typically enhances access to social resources outside a given population (Irwin, LaGory,
Ritchey, & Fitzpatrick, 2008). For homeless youth, bridging ties are often non-street
relationships, such as caseworkers or agency staff who assist in linkage to much-needed services,
or family members and non-street friends who can provide an emotional outlet and a means to
disengage from their street identities (Karabanow & Naylor, 2010; Mitchell & LaGory, 2002).
Prior to the internet and social media, homeless youth had very limited avenues through which
they could connect with their bridging social ties. Therefore, one can assume that internet and
social media would expose homeless youth to a range of social relationships that they otherwise
would not have any contact with.
Summary
The study of norms can help us understand a broad range of human behaviors, including
that of methamphetamine and injection drug use behaviors among homeless youth. The theories
reviewed above underscore how norms are best understood when framed within a social network
perspective; they do not operate in a vacuum. While norms are often socially prescribed, they are
also not one-dimensional. As some of the theories delineated above demonstrate, the content of

49

these norms can vary based on the structure and composition of networks. Based on the
theoretical conceptualization elucidated above, I would expect that the normative perceptions of
methamphetamine and injection drug use among homeless youth would differ based on the
structural characteristics of both their egocentric as well as their positioning within whole
networks. Furthermore, the composition of their networks would also have an impact on the
nature of their normative perceptions.

50

CHAPTER FOUR
METHODOLOGY
Sampling
An Event Based Approach (EBA) (Freeman & Webster, 1994) sample of 358 homeless
youth (aged 13-25 years) were recruited between October 2011 and February 2012 from two
drop-in centers, one in Hollywood, CA and one in Santa Monica, CA. The entire populations of
youth accessing these agencies were eligible and were invited to participate. In Santa Monica,
93.3% of the population was interviewed; 180 youth were approached, six youth declined the
survey and an additional five failed to complete the survey. In Hollywood, 80.2% of the
population were interviewed; 303 youth were approached, 18 refused to participate and an
additional 42 failed to complete the survey. Four of the surveys completed in Santa Monica
were incomplete and 20 of the youth interviewed in Hollywood provided incomplete data or had
previously been interviewed in Santa Monica, six youth who identified themselves as
transgender were dropped from the final analyses, yielding the final sample of 358.
Boundary Specification
These data come from a larger multi-year panel study (MH R01 903336), whose primary
aim is to assess the large interconnected networks of homeless youth in Hollywood and Santa
Monica. Event Based Approach (EBA) (Freeman & Webster, 1994) was used to delineate the
boundary of this sociometric network of homeless youth, who were all accessing services at two
drop-in centers in Los Angeles, CA and Santa Monica, CA. Social networks are not random and
simple random sampling procedures are inappropriate for social network studies, which strive to
understand how a population of individuals is interconnected. Rather, the aim of these “whole
network studies,” is to capture as much of a target population as possible. Event Based

51

Approach (EBA) (Freeman & Webster, 1994) was used to delineate the boundary of this
sociometric network of homeless youth, who were all accessing services at a drop-in center in
Los Angeles, CA. Network researchers acknowledge that in practice, the social world is
composed of infinite sets of relations; however, it becomes necessary to impose certain
constraints on who will be included in the sample (Wasserman & Faust, 1994). While there are
several ways to set such inclusion criteria, and no clear consensus exists, EBA is especially
viable for sampling traditionally unbounded populations such as homeless youth.
The EBA binds individuals into a social group based on a set of shared activities or
events and is most relevant for homeless youth because of several reasons (Rice, Barman-
Adhikari, Monro & Milburn, 2012). First, homeless youth are a transient and fluid population,
and therefore, it is unrealistic to expect them to be listed in a formal membership list, a technique
that is often used to delineate boundaries for social network studies, and is known as the
positional approach (Marsden, 2005). Second, there are homeless youth who are often isolated
from other homeless youth, therefore the relational approach (Marsden, 2005), which starts with
an agency roster and then expands to include other people who are nominated by the existing
sample, will exclude social isolates and other peripheral youth. The EBA approach is able to
mitigate these limitations, because using this procedure, participants are sampled from “natural
settings”, where people socialize, and entry, and exit is common. In the original study, this
setting was a beach, and people who visited the beach three or more times in the previous month
were designated as “regulars”. In the present study, youth who attended the two drop in centers
at least once in the prior month were included in the sample. Consistent with the EBA’s
(Freeman and Webster, 1994) original criterion, drop-in-centers serve as natural settings where

52

homeless youth routinely interact with each other and these interactions can be easily observed.
Moreover, these spaces are relatively unregulated and entry and exit is common.
Procedures
Data Collection
Different data-collection procedures were used to collect the individual level and network
level data. Computer assisted interviews were used to collect the individual-level demographic
and behavioral data. All the demographic and behavioral data were based on self-reports.
Any client over 13 years of age receiving services at the respective agency was eligible to
participate. Recruitment was conducted for 19 days at each agency. Recruiters were present at
the agency to approach youth for the duration of service provision hours at each site. Each
agency has one main entrance where youth sign-in for services for the day, allowing recruiters to
ensure that all youth were approached. Youth new to the agency first completed the agency’s
intake process before beginning the study, to ensure they met the eligibility requirements for the
agency (and thus the study). A consistent set of two research staff members were responsible for
all recruitment to prevent youth completing the survey multiple times within each data collection
period per site.
Signed voluntary informed consent was obtained from each youth, with the caveats that
child abuse and suicidal and homicidal intentions would be reported. Informed consent was
obtained from youth 18 years of age and older and informed assent was obtained from youth 13-
to 17-years-old. The Institutional Review Board (IRB) waived parental consent, as homeless
youth under 18 years are unaccompanied minors who may not have a parent or adult guardian
from whom to obtain consent. Interviewers received approximately 40 hours of training,
including lectures, role-playing, mock surveys, ethics training, and emergency procedures.

53

Instruments
All surveys were conducted in a private space at the agency. The survey consisted of two
distinct parts: (Part 1) Audio Computer Assisted Self-Interview (ACASI) and (Part 2) Face-to-
Face Social Network Interview (F2F-SNI). The questionnaire and social network interview could
be completed in English or Spanish. Study participation lasted about 60-90 minutes total. All
participants received $20 in cash or gift cards as compensation for their time. Survey items and
procedures were approved by the Institutional Review Board at University of Southern
California.
Part 1: Audio Computer Assisted Self-Interview (ACASI)
ACASI allows participants to enter answers to questions privately into the computer, as
they read questions silently on the computer screen and/or listen to the questions being read to
them through headphones. After responses are entered, the computer selects the next questions to
be answered based on pre-programmed skip patterns. After responses are entered, the computer
selects the next questions to be answered based on pre-programmed skip patterns. ACASI
reduces non-response rates to sensitive questions about potentially socially undesirable activities,
such as sexual behaviors, illicit substance-using behaviors, and criminal activity (Turner, Ku,
Rogers, Lindberg, Pleck & Sonenstein, 1998; Macalino, Celentano, Latkin, Strathdee & Vlavov,
2002; Ghanem, Hutton, Zenilman, Rimba & Erbelding, 2005; Morrison-Beedy, Carey & Tu,
2006).
Part 2: Face-to-Face Social Network Interview (F2F-SNI)
F2F-SNI yields standard “ego-centric” network data as well as the data needed to match
all actors within the sample, creating the sociometric network model. The F2F-SNI assesses ties

54

maintained face-to-face as well as through other technologies such as the internet and cell phone.
F2F-SNI, however, provides a visual stimulus, which has been documented to enhance the
youth’s ability to focus on providing a large quantity of social network data while simultaneously
reducing participant burden (Rice et al., 2012).
Name Generator
The network data was collected by a trained interviewer from the participants using a
name generator. Participants provided information on up to 50 people they had interacted with in
the past 30 days. Participants were asked, “Think about the last month. Now I am going to draw
a map of your network. We are interested in the people you interact with. We’re interested in the
people you talk to, “hang out”/”kick it”/ “chill” with, have sex with or hook up with, party with
or drink or use drugs with?” Subsequently an extensive list of possible alter roles were read,
“friends; family; people you “hang out with”/ “chill with”/ “kick it with”/ have conversations
with; people you party with – use drugs or alcohol; boyfriend/girlfriend; people you are having
sex with; “baby mama”/ “baby daddy”; case worker or agency staff; people from school; people
from work; old friends from home; people you talk to (on the phone, by email); people from
where you are staying (“squatting with”); people you see at this agency; other people you know
on the streets.” Once the youth had finished nominating persons in their networks, attributes of
each nomination were then collected, including, first name and last initial, aliases, age, gender,
race/ethnicity, and whether the nominee was a client of the agency.

55

Measures
Theoretical and empirical findings informed the selection of items used to measure the
constructs of interest. A wide array of data was collected as part of the overall “Youthnet”
questionnaire, ranging from demographics, family background, social-cognitive processes,
substance use, and sexual risk. The following measures were used for this study:
Sociodemographic characteristics
Information on a number of socio-demographic variables was collected as a part of the
larger study. For these particular analyses, youth’s age, site where they received services,
ethnicity, gender, sexual orientation, time homeless, migratory status (traveler or not),
educational status, time since first homeless, and housing situation were utilized. Single-item
demographic variables represented age and time since first homeless in years, gender (male vs.
female) time homeless (more than two years versus less than two years), race/ethnicity (White
vs. other), site where they were recruited (Hollywood vs. Santa Monica), education (high school
diploma/GED and above or less than high school diploma/GED) and sexual orientation
(heterosexual vs. lesbian/gay/bisexual/unsure). Since there were only six youth who identified
themselves as transgender, they were dropped from the final analyses.
Housing situation was assessed by asking youth to indicate in which of the following
places they currently resided: “(1) Family home, (2) Foster family home, (3) Relative's home, (4)
Friend's home, (5) Home of my boyfriend/girlfriend/person I'm having sex with, (6) Group
home, (7) Shelter (emergency, temporary), (8) Hotel, motel, (9) Sober living facility, (10)
Juvenile detention center, jail, (11) Transitional living program, (12) Own apartment, (13) Street,

56

squat, abandoned building, (14) Car, (15) Bus, (16) Other, please specify.” Consistent with
Tsemberis et al.’s (2007) definition of “homelessness,” this study included youth who were
literally homeless (i.e., sleeping on the streets, or in a car or bus), staying in emergency shelter
services, or temporarily housed (i.e., transitional living, multiple reports of relatives or friends),
or stably-housed. For the purposes of analyses, youth who responded as currently staying on the
streets, in a squat or abandoned building (13) or in a car (14) or bus (15) were categorized as
currently living in a “street-based” environment; all other responses were categorized as “non-
street-based.”
Youth chose the total number of months and years they have been homeless during their
lives; responses were converted into years. Being a traveler was assessed by a yes response to
the following: “Have you ever been a ‘traveler?’ (A ‘traveler’ is someone who moves by
themselves or with friends from city to city after a short period of time).”
Age and time since first homeless was measured continuously; all other demographic
variables were dummy coded with the first category listed above coded “1” and the remaining
category as the reference group.
Network Variables
Sociometric Network Structural Measures
Since the analyses associated with the sociometric data measures perceptions of other
street alters’ behaviors, the analyses excluded all isolates. Isolates are people who are not
connected to anybody within a bounded network (Wasserman & Faust, 1994). Therefore, when

57

operationalizing positioning in networks, these youth and their associated positions within
networks were excluded.
Position in network: Analysis of centrality measures such as K-cores, degree centrality,
and eigenvector centrality will be used to delineate the position of youth in the sociometric
network. Centrality measures address the question, “Who is the most important or central person
in this network?” The following centrality measures were used for this study:
A K-core is a maximal sub graph in which each point is adjacent to K other points; all the
points in the K-core have a degree greater than or equal to K (Wasserman & Faust, 1994). For
example, in a network where everybody is connected to each other is the simplest form of
component and has a 1 core. For both the Hollywood and Santa Monica network, K-core 1
through 4 can be assigned. Periphery membership was defined by K-core 1 and 2 indicating that
a youth either only 1 tie or 2 ties to another network member.
Degree centrality refers to the number of ties a node (or a person) has to other nodes (or
persons) (Wasserman & Faust, 1994). Therefore, a person who has one connection will have a
degree centrality score of 1. In the Hollywood sociometric network, degree centrality scores of 1
through 13 can be assigned. In the Santa Monica network, degree centrality scores of 1 through 9
can be assigned. Based on these scores, youth who had less than two ties (median=2) were
regarded as peripheral to the network.
Eigenvector centrality is like the recursive version of degree centrality. A node is central
to the extent that the node is connected to others who are central (Wasserman & Faust, 1994). In
the Hollywood network, the eigenvector scores ranged from .00 (which signifies no influential

58

connections) to .43 (which signifies the greatest level of influential connections). The median
eigenvector score (.04) was used to dichotomize the assignment of actors as central or peripheral
to the network based on their level of influential connections. In the Santa Monica network, the
eigenvector scores ranged from .00 to .39. The median eigenvector score (.03) was used to
dichotomize the assignment of actors as central or peripheral to the network based on their level
of influential connections.
All the network structural measures were dummy coded with youth in the core coded as
“1” and the other category as the reference group.
Network Norms
Descriptive Norms: To assess descriptive norms regarding drug related behaviors, after
youth finish nominating their network members, they were asked: Out of the people you
nominated, how many of them use- a) Meth b) Injected Drugs. These two variables were
measured as the proportion of people who the respondent thought engaged in either meth or
injection drugs.
Injunctive Norms: Similarly, to assess injunctive norms regarding drug related behaviors,
after youth finish nominating their network members, they were asked: Out of the people you
nominated, how many of them will object you using a) Meth b) Injected Drugs. Similarly, they
were also asked: Out of the people you nominated, how many of them will encourage you to use
a) Meth b) Injected Drugs. The “object to” and “encourage to” questions were asked separately
for each of the alters. These two variables were measured as the proportion of people who the
respondent thought would object or encourage then to engage in either meth or injection drugs.

59

To include the injunctive norm variables in the multivariate model as independent
variables (to investigate Aim 2), the “encourage” and “object” questions were combined. More
specifically, because of a desire to understand how protective influences might mitigate harmful
ones, injunctive norms were operationalized as having greater proportion of alters who object to
them engaging in substance use compared to those who encourage them to engage in such use.
Accordingly, this variable was then dichotomized into “0” if the proportion of alters who
encourage meth or injection drug use was greater than the proportion of people who objected,
and recoded into “1” if vice-versa.
To include the perceived norms variables (both descriptive and injunctive) as dependent
variables in the multivariate models (to investigate Aim 1 and Aim 3), the assumption of normal
distribution for both these variables was assessed (see Table 1). To test the assumption of normal
distribution, skewness should be within the range ±1 (Brown, 1997). However, as the table
indicates, the variables were highly skewed. The median is the preferred measure of determining
a threshold for skewed measures (Wang, Fan, & Willson, 1996). Based on the median, these
norms were then dichotomized as none vs. few/some/all (if median was 0) or all vs.
none/few/some (if median was 1).
Table 1: Skewness of Perceived Norms Variables
Min Max Median SD
Proportion of Meth Using Alters 0.00 1.00 0.0 0.19 2.86 0.13
Proportion of Alters Encourage Meth 0.00 1.00 0.0 0.38 -0.88 0.13
Proportion of Alters Object Meth 0.00 1.00 0.1 0.12 5.24 0.13
Proportion of IDU Alters 0.00 0.91 0.0 0.10 4.17 0.13
Proportion of Alters Encourage IDU 0.00 1.00 0.0 0.36 -1.25 0.13
Proportion of Alters Object IDU 0.00 1.00 0.1 0.09 6.62 0.13
Skewness

60

Online Social Capital
Online Social Capital: Social capital accessed via the internet and social media was
assessed by combining responses to two items: the first, “Who do you use your email to
communicate with? Check all that apply” and “When you use social networking websites like
MySpace, Facebook, or Twitter, who do you communicate with? Check all that apply.”
Answers to both items were: (1) Parents (including foster family or step family), (2) Brothers,
sisters, cousins, or other family members, (3) People you know from the streets of Los Angeles,
(4) People you know from home (before you came to the streets of Los Angeles), (5) Case
workers, social workers, or staff or volunteers at youth agencies, (6) No one, I don't use [relevant
technology].” Responses were dichotomized. Responses were combined across technologies
and dichotomized. Answers to items 1 and 2 were also combined.
Behaviors
Substance Use
Youth Risk Behavior Survey (YRBS) (Eaton et al., 2011) items were used to assess
recent (past 30 days) methamphetamine and injection drug use. Recent use of injection drugs and
methamphetamine were assessed with the question, “During the past 30 days, how many times
have you used (1) meth (2) Injection Drugs. The response options for methamphetamine ranged
from 0 times, 1or 2 times, 3 to 9 times, 20 to 39 times, and 40 times or more. The response
options for injection drugs ranged from 0 times, 1 time, to 2 or more times. They were further
dichotomized with “0” being never used and “1” indicating ever-used within the past 30 days.

61

Sex under the influence of drugs
This was assessed with the question, “Did you drink alcohol or use drugs before you had
sex (vaginal or anal sex) the last time?” The responses were “yes” or “no”.
Data Analyses
Overview
Data analyses were conducted using SAS Version 9.2 (SAS Institute Inc., 2008).
Exploratory data analysis such as calculating frequencies and means and standard deviations and
skewness were conducted to examine the distribution of study variables. The socio-demographic
characteristics are presented as a whole, and additionally stratified by field site (i.e. Hollywood
vs. Santa Monica). Field site was controlled for in all multivariate analyses.
The analyses is presented in three distinct chapters-separated by study aims and
egocentric (Chapter Five and Six) and sociometric analyses (Chapter Seven). This was
necessitated because of the differences in how alters were included for the two different
analyses. As noted in the previous chapters, egocentric data consists of the local network of each
participant, whereas sociometric network is designed to connect all the participants in the
network based on a shared activity. Therefore, the egocentric method allows for the inclusion of
a greater number of alters (both street and non-street), whereas the sociometric method limits
eligible alters to street peers who are also service-seekers at the drop-in centers from where this
data was collected. The sociometric analyses were further stratified by field site (i.e. Hollywood
vs. Santa Monica) because they represent two very distinctive bounded networks.

62

Additionally, since the analyses associated with the sociometric data measures
perceptions of other street alters’ behaviors, the sociometric analyses excluded all isolates.
Isolates are people who are not connected to anybody within a bounded network (Wasserman &
Faust, 1994). Since they did nominate anyone within the sociometric networks, there is no data
on the perception measures. Therefore, these isolates were excluded. There were 65 isolates in
the network sample from Hollywood and 23 isolates in the network sample from Santa Monica.
In order to ensure statistical power and preserve degrees of freedom, a statistically
accepted strategy (Hosmer & Lemeshow, 1989) was used to reduce the number of variables
without compromising the comprehensiveness provided by the conceptual model. Therefore, the
analyses predicting associations with the dependent variables proceeded in two stages. First, a
series of bivariate logistic regressions were run to determine significant associations (p < .05)
between the independent variables and employment services utilization. These bivariate
associations were examined in a pair-wise approach, which is logically equivalent to the
examination of a correlation matrix. Any independent variable that was found to be significantly
associated (i.e., p <.10 level) with any dependent variable was retained in the final multivariate
logistic regression model (Hosmer & Lemeshow, 2000). Variance inflation factor (VIF) was
checked to determine the potential multicollinearity among the independent variables.
Sociomatrix Construction
A sociomatrix was created linking participants in the sample. A directed tie from
participant i to participant j was recorded if participant i nominated participant j in his/her
personal network. Matches were based on name, alias, ethnicity, gender, approximate age, and
agency attendance. If two distinct youth matched on all information, presence of a third common

63

Figure 4: Sociometric network of homeless youth in Hollywood, California (n=160).
tie in each personal network was used to assign adjacency. Questionable matches were left un-
coded (hence a conservative matrix of ties) (see Figure 4 & Figure 5). Two research assistants
each created independent adjacency matrices. The matrices were combined and discrepant ties
dropped. For the sociometric analyses for this study, isolates were excluded from the analyses
and therefore are not visible in the figure below.

64

Legend: Circles represent nodes. Lines indicate connections
between youth. COLOR: Blue= 4core; Red = 3 core, Black
=2 core; Gray =1 core. Isolates are excluded.

Figure 5: Sociometric network of homeless youth in Santa Monica, California (n=130).

Network Visualization
A common first step in network analysis is visualization. These diagrams are an excellent
tool for pattern recognition. Netdraw 2.090 (Analytic Technologies, Irvine, CA) graph
visualization software was used to generate network visualizations using the spring embedder
routine. In this particular study, the visualizations generated were used to understand visually
whether there are positional attributes associated with the clustering of similar norms regarding
methamphetamine and injection drug use norms in the sociometric network.

65

Network Analysis
UCINET (Analytic Technologies, 2002) was utilized to generate sociometric social
network measures (related to network position). These social network metrics provide valuable
information on the structural properties of the sociometric networks. Analysis of centrality
measures such as components, degree centrality, and eigenvector centrality will be used to
delineate the position of youth in the sociometric network.
Statistical Analysis by Aim
Network analysis focuses on relationships and not on attributes. Therefore, network
analysis violates the assumption of independence, which underlies statistical analysis. However,
studies have recently started incorporating social network data into statistical modeling. For
these analyses, sociometric network level measures were assigned to individuals as an attribute
and included in the statistical models. The statistical data analysis for this proposal was
conducted using SAS Version 9.2 (SAS Institute Inc., 2008).
The specific aims were addressed via a series of multivariate logistic statistical models.
Following is a description of the strategy used for testing hypotheses specific to the three
research aims specified for this study. The strategies used in modeling each of the three specific
aims and hypotheses are delineated below:
1. To assess the association between perceived social network norms (descriptive and injunctive)
and HIV risk behaviors (methamphetamine and injection drug use) among homeless youth.
Hypothesis 1: The stronger the perceived network norms approving HIV risk behaviors, the
greater the likelihood that respondents will engage in HIV risk-taking behaviors.

66

In order to address Hypothesis 1, I examined whether self-reported social norms of
homeless youth are associated with their meth and injection drug use behaviors. Bivariate
analyses were first conducted to examine unadjusted associations between study variables and
outcome measures (Hosmer & Lemeshow, 2000). Based on these analyses, multivariate
regression models were constructed. I included the main independent variables, descriptive and
injunctive norms regarding methamphetamine and injection drug use in the model after
controlling for other relevant covariates. Both types of perceived norms (descriptive and
injunctive) were included in the model simultaneously to assess their independent effect on meth
and injection drug use behaviors in the presence of each other.
2. To assess how technology facilitates connection to pro-social home-based peers and family, to
explore differential norms (descriptive and injunctive) about HIV risk behaviors
(methamphetamine and injection drug use) among homeless youth.
Hypothesis 2: Connecting and communicating online with pro-social home-based peers and
family will be associated with norms that are preventive of engagement in HIV risk behaviors.
In order to address Hypothesis 2, I used the social network technology survey data to
assess whether the type of people that youth connect with via internet and social network
technology are reflective of norms that are either facilitative or preventive of HIV risk behaviors.
More specifically, the objective was to understand if youth connected to different kinds of
network relationships, whether that would be associated with different perceptions of descriptive
and injunctive norms. Bivariate analyses were first conducted to examine unadjusted associations
between study variables and outcome measures (Hosmer & Lemeshow, 2000). Based on these
analyses, the multivariate models were constructed. In the multivariate model, the independent

67

variables were the social network ties (with family, home-based peers, or street peers)
maintained via internet (e-mail) and social media and the dependent variables were each of the
descriptive, and the injunctive norms (both encourage and object) regarding meth and injection
drug use behaviors.
3. To identify social network characteristics that are associated with perceived social network
norms (descriptive and injunctive) regarding the use of methamphetamine and injection drugs
among homeless youth.
Hypothesis 3: Network norms for HIV risk behaviors (substance use) will vary based on the
structural characteristics of the networks (such as social position, cohesiveness etc.).
In order to address Hypothesis 3, I examined whether norms regarding HIV risk
behaviors are clustered within certain social network sociometric structures. For the purpose of
this analysis, only the youth who were a part of the sociometric network were included.
Therefore unlike the other two multivariate models, where the descriptive and injunctive norms
represented norms emanating from all members of their network (such as family, home-based
friends etc.), in this model, only the norms of street youth were included. Bivariate analyses were
first conducted to examine unadjusted associations between study variables and outcome
measures (Hosmer & Lemeshow, 2000). Based on these analyses, the multivariate models were
constructed. In the multivariate model, the independent variables were the sociometric network
metrics (such as Kcore, degree centrality etc.), the dependent variable were each of the
descriptive, and the injunctive norms regarding meth and injection drug use behaviors.

68

CHAPTER FIVE
RESULTS STUDY AIM 1
EGOCENTRIC ANALYSES
As noted in the previous chapters, egocentric data consists of the local network of each
participant, whereas sociometric network is designed to connect all the participants in the
network based on a shared activity. Therefore, the egocentric method allows for the inclusion of
a greater and more heterogeneous set of relationships (both street and non-street), while
sociometric data restricts the sample of network alters to other street peers. Therefore, the
analyses is separated for the egocentric and sociometric analyses.
This chapter highlights the key findings derived from the egocentric data for Aim 1. First,
I describe the characteristics of the study sample, and the description of their egocentric
networks. Then I present the bivariate and multivariate statistics.
Descriptive Statistics
Socio-Demographic Characteristics
Data on socio-demographic characteristics of participants are presented in Table 2. The
sample was predominantly composed of males (n= 250, 69.8%). Average age of participants was
21.4 years. A majority of youth self identified as heterosexual (n=266, 75.3%). Racially, a
majority of the youth identified as White (n=126, 35.29%). A high proportion of the youth
indicated that they were precariously housed; i.e. either living on the streets (n= 142, 39.89%) or
couch surfing (n=122, 34.27%). Notably, there were important differences among youth in the
two field sites. Youth in Santa Monica were more likely to be male (p< .01), White (p < .001), to
be living on the streets (p < .001) and be travelers (p < .001).

69

Table 2: Descriptive statistics of homeless youth in Los Angeles, CA (n=358)
Total Santa Monica Hollywood
n=358 n=153 n=205
n % n % n % chi sq/t-test
Gender
Male 250 69.83 120 78.95 130 64.36 8.90 **
Female 104 29.05 32 21.05 72 35.64
Transgender 4 1.12
Sexual Orientation
Heterosexual 266 75.35 121 79.61 145 72.14 2.60
Non-heterosexual 87 24.65 31 20.39 56 27.86
Race
Native American 10 2.80 4 2.63 6 2.93 82.00 ***
Asian 1 0.28 0 0.00 1 0.49
Black 107 29.97 20 13.16 87 42.44
Native Hawaiin or Pacific Islander 2 0.56 1 0.66 1 0.49
White 126 35.29 91 59.87 35 17.07
Latino 54 15.13 11 7.24 43 20.98
Mixed Race 57 15.97 25 16.45 32 15.61
Current Living Situation
Couch Surfing 122 34.27 42 27.63 80 39.22 55.61 ***
Stable Situation 43 12.08 13 8.55 30 14.71
Streets 142 39.89 92 60.53 50 24.51
Emergency Shelter 49 13.76 5 3.29 44 21.57
High School Graduate 148 41.57 59 38.56 89 43.84 1.00
"Traveler" 169 47.74 94 61.84 75 37.13 21.23 ***
Mean Time Homeless (std dev) 2.781 2.79 2.942 3.01 2.651 2.60 1.33
Mean Age (std dev) 21.48 2.09 21.8 2.15 21.24 2.01 2.55 *
*=p<.05; **=p<.01; ***=p<.001

70

Perceived Norms for Methamphetamine and Injection Drug Use
The descriptive norms were dichotomized based on the median (see Table 3).
Variables n %
Proportion of Meth alters
None 248 69.08
Any 110 30.92
Proportion of IDU Alters
None 341 94.99
Any 18 5.01
Table 3: Descriptive Norms for Meth and IDU

As one can see on Table 3, around 30% of the youth indicated that they had at least one
person in their network that used methamphetamine. However, only 5% of youth stated that they
knew anybody in their networks who were injection drug users.
As noted in the previous chapter, in order to include the injunctive norm variables in the
multivariate model as independent variables (to investigate Aim 2), the “encourage” and “object”
questions were combined. There was a high amount of multi-collinearity detected between these
two variables, and therefore could not be included in the same multivariate model. Also
conceptually, because of a desire to understand how protective influences might mitigate harmful
ones, injunctive norms were then dichotomized into “0” if the proportion of alters who
encourage meth or injection drug use was greater than the proportion of people who objected,
and recoded into “1” if vice-versa. The injunctive norms presented in the table above (Table X)
therefore reflect this operationalization.

71

Variables n %
Alters who object to meth use vs. encourage
Lesser 153 42.62
Greater 206 57.38
Alters who object to injection drug use vs. encourage
Lesser 117 32.59
Greater 242 67.41
Table 4 : Injunctive Norms for Meth and IDU

As one can see on Table 4, for both methamphetamine and injection drug use, on an
average youth stated that a greater number of people in their network would object to them
engaging in any of these substances. Specifically, 57.38% of youth indicated that they had a
greater proportion of people in their network who would object to their engagement in
methamphetamine use and almost two-thirds of youth (67%) indicated that a greater proportion of
people in their network who would object to their engagement in injection drug use than people
who encourage it.
Substance use and sex risk behaviors
Table 5 presents rates of methamphetamine and injection drug use rates among youth in
this sample. It also documents the percentage of youth who reported whether they engaged in sex
under the influence of drugs. Approximately 29.71% of the youth indicated they had engaged in
methamphetamine use over the past 30 days. Reported injection drug use was much lower. About
12.5% of the youth indicated that they had engaged in injection drug use in the last 30 days. More
worryingly, almost half (44%) of youth indicated that they had engaged in sex under the influence
of drugs.

72

Variables n %
Meth use 246 70.29
No 104 29.71
Yes
Injection Drug use
No 314 87.47
Yes 45 12.53
Sex under the influence
No 201 55.99
Yes 158 44.01
Table 5: Substance use and sex risk behaviors

Collinearity among Independent Variables
Table 6 presents collinearity diagnostics among key independent variables. A tolerance of
less than 0.20 or 0.10 and/or a VIF of 5 and above indicates a multicollinearity problem
(Mansfield & Helms, 1982). Based on these threshold measures, as seen on Table,
multicollinearity can be ruled out.

73

Variables Tolerance VIF
Time Homeless 0.96 1.05
Sexual Orientation 0.91 1.10
Living Situation 0.74 1.35
Traveler 0.83 1.20
Sex Under the Influence 0.85 1.18
White 0.71 1.41
Fieldsite 0.65 1.54
Male 0.83 1.20
Descriptive Meth Norms 0.71 1.40
Injunctive Meth Norms 0.55 1.81
Descriptive IDU Norms 0.82 1.21
Injunctive IDU Norms 0.64 1.57
Note: IDU=Injection Drug Use
Table 6 : Collinearity among independent variables for Study Aim 1

Bivariate and Multivariate Analyses
The first aim of the study was to assess the association between perceived social network
norms (descriptive and injunctive) and HIV risk behaviors (methamphetamine and injection drug
use) among homeless youth. Described below are the findings from the bivariate and the
multivariate analyses. The findings are further separated by outcomes (methamphetamine and
injection drug use).
Methamphetamine Use
Table 7 presents associations between perceived norms of methamphetamine use (both
descriptive and injunctive) and actual methamphetamine use among homeless youth.
Bivariate results revealed that field site, time since first homeless, gender, sexual
orientation, race, current living situation, being a traveler, engaging in sex under the influence of
drugs, descriptive norms, and injunctive norms regarding methamphetamine use were all
significantly associated with self-reported methamphetamine use. Specifically, youth recruited in

74

Santa Monica were 2.3 times more likely to report engaging in methamphetamine use than youth
who were recruited in Hollywood. Street-based youth were 2.8 times more likely to engage in
methamphetamine use than non-street youth. Being a traveler was highly significant (OR= 2.8,
p< .0001). More alarmingly, youth who reported engaging in sex under the influence of drugs
were 3.7 times more likely to engage in methamphetamine use than youth who did not. Both
descriptive and injunctive norms were highly significantly associated with methamphetamine use
in the bivariate models. Specifically, youth who perceived that their network alters engaged in
methamphetamine use were 6.4 times more likely to engage in methamphetamine use. Youth
who reported that they had more people in their network who would object to them engaging in
methamphetamine use were 87% less likely to engage in the behavior.
In the multivariate model, sexual orientation, engaging in sex under the influence of drugs,
and both kinds of perceived norms (descriptive and injunctive) were significantly associated with
self-reported methamphetamine use. Specifically, non-heterosexual youth were 2 times more
likely to engage in methamphetamine use relative to heterosexual youth. Youth who reported
engaging in sex under the influence of drugs were 2.5 times more likely to report
methamphetamine use. Both perceived norms retained their significance in the multivariate
model even after controlling for other demographic behaviors demonstrating the significance of
social influence on substance use behaviors. Youth who believed that their alters engaged in
methamphetamine use were 2.5 times more likely to engage in methamphetamine use. Youth
who believed that a greater proportion of people in their networks would object to them engaging
in methamphetamine use were 81% less likely to engage in methamphetamine use.

75

Table 7 : Odds ratios for association between perceived norms and methamphetamine use
Adjusted
OR
Demographic Characteristics
Field site (Santa Monica=1) 2.32 1.46 3.71 ** 1.32 0.63 2.75
Age 1.06 0.95 1.18
Time Homeless 1.06 0.98 1.16 # 1.02 0.92 1.13
Gender (Male=1) 1.63 0.95 2.81 #
Sexual Orientation (Non-heterosexual=1) 1.72 1.02 2.92 * 2.08 1.02 4.25 *
Race (White=1) 1.61 1.01 2.58
*
0.84 0.42 1.71
Current Living Situation (Street-based=1) 2.83 1.75 4.56
***
1.40 0.72 2.75
High School Graduate (Yes=1) 0.71 0.44 1.15
Traveler (Yes=1) 2.81 1.74 4.54 *** 1.75 0.92 3.32
Sex Under Influence of Drugs 3.78 2.34 6.09 *** 2.58 1.37 4.85 **
Perceived Norms
Descriptive Norms 6.42 3.87 10.65 *** 2.51 1.29 4.87 **
Injunctive Norms 0.13 0.08 0.22 *** 0.19 0.10 0.37 ***
Pseudo R-Square 0.28
2 Log Likelihood 267.99

#
p<.10 *=p<0.05; **=p<0.01; ***=p<0.001
Note: Only variables significant in the bivariate analyses at p< .10 were included in the final adjusted analyses
Meth Use Meth use
Unadjusted
OR 95% CI 95% CI

Injection Drug Use
Table 8 presents associations between perceived norms of injection drug use (both
descriptive and injunctive) and actual injection drug use among homeless youth.
On the bivariate level, field site, race, living situation, being a traveler, having sex under the
influence of drugs, and perceived norms regarding injection drug use were all associated with
self-reported injection drug use. White youth were 2.7 times more likely to engage in injection
drug use compared to youth of other ethnicities. Street youth were 1.8 times more likely to
engage in injection drug use than youth who are in temporary housing or couch surfing. Youth
who reported being travelers were 2.9 times more likely to engage in injection drug use. Youth
who reported engaging in sex under the influence of drugs were 6.3 times more likely to engage

76

in injection drug use than youth who did not. Youth who reported having alters who were also
injection drug users were 51.8 times more likely to engage in injection drug use. Additionally,
youth who reported that a greater proportion of people in their networks would object to
injection drug use were 83% less likely to engage in injection drug use themselves.
In the multivariate model, reporting having sex under the influence of drugs, descriptive
norms, and injunctive norms regarding injection drug use were all significantly associated with
self-reported injection drug use. Youth who reported having sex under the influence of drugs
were 4.1 times more likely to engage in injection drug use. Youth who believed that their alters
engaged in injection drug use were 15.5 times more likely to engage in injection drug use.
Additionally, youth who reported that a greater proportion of people in their networks would
object to injection drug use were 79 % less likely to engage in injection drug use themselves.

77

Table 8 : Odds ratios for association between perceived norms and injection drug use
Adjusted
OR
Demographic Characteristics
Field site (Santa Monica=1) 3.02 1.56 5.85 ** 1.95 0.78 4.88
Age 1.05 0.90 1.22
Time Homeless 1.05 0.94 1.18
Gender (Male=1) 1.58 0.73 3.42
Sexual Orientation (Non-heterosexual=1) 1.38 0.69 2.76
Race (White=1) 2.71 1.44 5.12
**
1.27 0.54 3.01
Current Living Situation (Street-based=1) 1.86 0.98 3.53
#
0.52 0.21 1.25
High School Graduate (Yes=1) 0.71 0.37 1.37
Traveler (Yes=1) 2.97 1.50 5.88 ** 2.32 0.96 5.58
Sex Under Influence of Drugs 6.30 2.93 13.53 *** 4.16 1.71 10.10 **
Perceived Norms
Descriptive Norms 51.83 14.20 189.24 *** 15.57 3.41 71.08 **
Injunctive Norms 0.17 0.09 0.34 *** 0.21 0.09 0.48 **
Pseudo R-Square 0.21
2 Log Likelihood 179.28

#
p<.10 *=p<0.05; **=p<0.01; ***=p<0.001
Note: Only variables significant in the bivariate analyses at p< .10 were included in the final adjusted analyses
Injection Drug Use Injection Drug Use
Unadjusted
OR 95% CI 95% CI

78

CHAPTER SIX
RESULTS STUDY AIM 2
This chapter highlights the key findings derived from the survey and egocentric data for
Aim 3. The characteristics of the study sample have been presented in the previous chapter (see
Table 2).
Collinearity among Independent Variables
Table 9 presents collinearity diagnostics among key independent variables. A tolerance of
less than 0.20 or 0.10 and/or a VIF of 5 and above indicates a multicollinearity problem
(Mansfield & Helms, 1982). Based on these threshold measures, as seen on Table 9,
multicollinearity can be ruled out.

79

Table 9 : Collinearity among independent variables for Study Aim 2
Variables Tolerance VIF
Field Site 0.67 1.49
Age 0.90 1.11
Time Homeless 0.83 1.20
High School 0.92 1.08
Traveler 0.84 1.20
Male 0.87 1.15
Sexual Orientation 0.88 1.13
Race 0.70 1.43
Street 0.75 1.34
Connect Online with Family 0.76 1.32
Connect Online with Home-Based Peers 0.68 1.46
Connect Online with Caseworkers 0.87 1.15
Connect Online with Street-Based Peers 0.78 1.29

Perceived Norms
Descriptive Norms
Table 10 presents data on the perceived descriptive norms of methamphetamine and
injection drug use behaviors reported by the participants. The perceptions of meth prevalence are
higher among these youth compared to injection drug use. As shown in this table, 70% of
participants indicated that they had no one in their egocentric networks that used
methamphetamine. However, 30.9 % of youth indicated that they at least had one person in their
network who engaged in methamphetamine use. Perceptions of injection drug use were lower.
Only 5% of the youth interviewed indicated that they thought that someone in their networks that
engaged in injection drug use.
Injunctive Norms
Table 10 also presents results for injunctive norms for both methamphetamine and
injection drug use behaviors. Nearly 58% of youth indicated that all (i.e. 100%) of their network

80

would object to them using meth. On the other hand, only 14.7% of the youth indicated that
some, few, or all of their network members would encourage them to use meth. Similarly,
concerning injection drug use, nearly 58.2% of youth reported that they believed that all of their
network members would object to them engaging in injection drug use. Encouragingly, only
8.9% of the youth thought that their network members would encourage them to engage in
injection drug use.
Online Social Capital
Table 11 presents information on online social capital among the participants. As noted in
the previous chapter, the email and social media questions were aggregated to assess sources of
Variables n %
Meth Using Alters
None 248 69.08
Any 110 30.92
Alters Encourage Meth
None 306 85.24
Any 53 14.76
Alters Object Meth
All 114 31.75
Less than All 245 68.25
IDU Alters
None 341 94.99
Any 18 5.01
Alters Encourage IDU
None 327 91.09
Any 32 8.91
Alters Object IDU
All 209 58.22
Less than All 150 41.78
Table 10: Perceived Norms for Meth and IDU

81

bridging and bonding social capital among this sample. Almost 64% of the youth reported that
they used email and social media to keep in touch with family members. More than half of the
youth reported that they maintained connections online to home-based peers. A fifth of the youth
reported connecting with caseworkers through email or social media. Less than half of youth said
that they used email or social media to keep in touch with their street peers.
Bivariate and Multivariate analyses
The second aim of the study was to examine the associations between specific
relationships maintained via internet and social media and perception of norms regarding
methamphetamine and injection drug use among homeless youth. Described below are the
findings from the bivariate and the multivariate analyses. The findings are further separated by
outcomes (descriptive vs. injunctive norms) for each type of drug use (methamphetamine vs.
injection drug use).

Variables n %
Bridging Capital
a
Connect online with family 226 63.84
Connect online with home-based peers 186 52.99
Connect online with caseworkers 73 20.80
Bonding Capital
Connect online with street-based peers 170 48.43
a
Family includes parents (foster/step family), siblings, relatives
Table 11: Online Social Capital among Homeless Youth

82

Methamphetamine Use Norms
Descriptive norms
Bivariate and multivariate results for descriptive norms regarding methamphetamine use
are presented in Table 12. On the bivariate level, in terms of socio-demographics, field site,
sexual orientation, living situation, educational background and being a traveler were all
associated with reporting having network members who use methamphetamine. Specifically,
youth who were recruited in Santa Monica were two times more likely to report having
connections to people who use meth compared to youth in Hollywood. Non-heterosexual youth
also reported having greater number of network members who used methamphetamine (OR=
1.96, p < .01). Notably, street youth were 2.5 times more likely to report having network
members who used methamphetamine relative to youth who had access to some form of
temporary housing. Youth who connected to family members were 55% less likely to report
having connections to people who used methamphetamine. On the contrary, youth who
connected to street peers online were 1.8 times more likely to report having connections to other
methamphetamine using people.
In the multivariate model, only race, living situation, and more importantly, connections
to family members online retained their significance. Sexual orientation (i.e. being non-
heterosexual) continued to be a risk factor. Non-heterosexual youth were 1.7 times more likely to
report having connections to alters who used methamphetamine. Youth who reported living
primarily on the street were also 1.7 times more likely to report having alters who used
methamphetamine. However, connections to family members online continued to be protective
even after controlling for other factors. Youth who reported connecting to family members’

83

Table 12 : Odds ratios for association between social capital and descriptive norms of methamphetamine use
Demographic Characteristics
Field site (Santa Monica=1) 2.00 1.27 3.15 ** 1.50 0.88 2.58
Age 0.99 0.89 1.11
Time Homeless 1.08 0.99 1.17
Gender (Male=1) 0.79 0.48 1.29
Sexual Orientation (Non-heterosexual=1) 1.96 1.17 3.27 ** 1.73 1.01 3.04 *
Race (White=1) 1.46 0.92 2.31
Current Living Situation (Street-based=1) 2.50 1.57 3.99
***
1.79 1.05 3.08 *
High School Graduate (Yes=1) 0.50 0.31 0.80 ** 0.67 0.37 1.21
Traveler (Yes=1) 2.18 1.38 3.46 ** 1.56 0.93 2.63
Social Capital
Bridging Capital
Connect online with Family 0.45 0.28 0.73 ** 0.59 0.35 0.98 *
Connect online with Home-based Peers 0.60 0.38 0.96 * 1.10 0.60 2.01
Connect online with Caseworkers 0.89 0.39 2.01
Bonding Capital
Connect online with Street-based Peers 1.84 1.13 2.97 * 1.45 0.81 2.58
Pseudo R-Square 0.14
2 Log Likelihood 378.28

#
p<.10 *=p<0.05; **=p<0.01; ***=p<0.001
Note: Only variables significant in the bivariate analyses at p< .10 were included in the final adjusted analyses
Unadjusted
OR
Adjusted
OR
Alters Use Meth Alters Use Meth
95% CI 95% CI

online (via email and social media) were 31% less likely to report having connections to people
who used methamphetamine.
Injunctive Norms
Bivariate and multivariate results for injunctive norms (separated by encourage and
object use) regarding methamphetamine use are presented in Table 13 and 14.

84

Encourage Methamphetamine Use
Concerning encouragement of methamphetamine use, on the bivariate level, field site,
sexual orientation, living situation, and connecting online to street peers were all significant.
Specifically, youth recruited in Santa Monica were 2 times more likely to report having network
members who would encourage them to use methamphetamine than youth in Hollywood. Non-
heterosexual youth were 1.8 times more likely to report having network members who would
encourage them to use methamphetamine relative to heterosexual youth. Street based youth were
1.4 times more likely to report having network members who would encourage them to use
methamphetamine. Connecting to street peers online was associated with 2.2 times likelihood of
reporting network members who would encourage youth to engage in methamphetamine use.
On the multivariate level, field site and connections to street peers online retained their
significance in the model. Youth who were recruited in Santa Monica were 2.1 times more likely
to report having network members who would encourage them to engage in methamphetamine
use relative to youth recruited in Santa Monica. More importantly, youth who reported
connecting to street peers online were 2.14 times more likely to report having network members
who would encourage them to engage in methamphetamine use.

85

Table 13: Odds ratios for association between social capital and injunctive norms of methamphetamine use
Unadjusted
OR
Adjusted
OR
Demographic Characteristics
Field site (Santa Monica=1) 2.07 1.14 3.74 * 2.10 1.05 4.19 *
Age 1.10 0.95 1.27
Time Homeless 1.05 0.94 1.17
Gender (Male=1) 1.18 0.60 2.32
Sexual Orientation (Non-heterosexual=1) 1.84 0.97 3.51 # 1.68 0.84 3.36
Race (White=1) 0.74 0.39 1.39
Current Living Situation (Street-based=1) 1.90 1.04 3.46 * 1.48 0.75 2.92
High School Graduate (Yes=1) 0.66 0.35 1.22
Traveler (Yes=1) 1.40 0.78 2.53
Social Capital
Bridging Capital
Connect online with Family 0.69 0.37 1.28
Connect online with Home-based Peers 0.80 0.44 1.47
Connect online with Caseworkers 0.59 0.29 1.18
Bonding Capital
Connect online with Street-based Peers 2.24 1.23 4.09 ** 2.14 1.12 4.09 *
Pseudo R-Square 0.05
2 Log Likelihood 259.86

#
p<.10 *=p<0.05; **=p<0.01; ***=p<0.001
Note: Only variables significant in the bivariate analyses at p< .10 were included in the final adjusted analyses
Encourage Meth Encourage Meth
95% CI 95% CI

Object to Methamphetamine Use
Concerning objection to meth use, on the bivariate level, field site, gender, living
situation, and being a traveler were found to be significant. However, none of the online social
capital measures was significantly associated with this outcome even on the bivariate level.
In the multivariate model, only living situation was significantly associated with having
network members who would object to methamphetamine use. Specifically, street based youth
were 49% less likely to report having network members who would object to their engagement
in methamphetamine use.

86

Unadjusted
OR
Adjusted
OR
Demographic Characteristics
Field site (Santa Monica=1) 0.56 0.37 0.86 ** 0.78 0.48 1.26
Age 0.99 0.89 1.09
Time Homeless 0.96 0.88 1.04
Gender (Male=1) 0.66 0.41 1.08 # 0.80 0.48 1.33
Sexual Orientation (Non-heterosexual=1) 0.70 0.43 1.15
Race (White=1) 0.84 0.55 1.30
Current Living Situation (Street-based=1) 0.44 0.28 0.68 *** 0.51 0.31 0.83 **
High School Graduate (Yes=1) 1.30 0.85 2.01
Traveler (Yes=1) 0.66 0.42 1.00 # 0.90 0.57 1.43
Social Capital
Bridging Capital
Connect online with Family 1.13 0.70 1.80
Connect online with Home-based Peers 0.90 0.57 1.40
Connect online with Caseworkers 1.38 0.80 2.40
Bonding Capital
Connect online with Street-based Peers 0.90 0.57 1.44
Pseudo R-Square 0.045
2 Log Likelihood 451.99

#
p<.10 *=p<0.05; **=p<0.01; ***=p<0.001
Note: Only variables significant in the bivariate analyses at p< .10 were included in the final adjusted analyses
Table 14: Odds ratios for association between social capital and injunctive norms of methamphetamine use
Object Meth Object Meth
95% CI 95% CI

Injection Drug Use (IDU) Norms
Descriptive norms
Bivariate and multivariate results for descriptive norms regarding injection drug use
(IDU) are presented in Table 15.
On the bivariate level, field site, race, current living situation, being a traveler, and
connections to family members online was all significantly associated with reporting having
network members who are IDU’s. Specifically, youth recruited in Santa Monica were 2.7 times

87

more likely to report having connections with people who are IDU’s. White youth were 4.8 times
more likely to report having connections with people who are IDU’s compared to youth of other
ethnicities. Living on the streets was also associated with riskier connections. Street-based youth
were 4.5 times more likely to report having connections with people who are IDU’s compared to
youth who are in temporary housing or couch surfing. Youth who are travelers were 3.9 times
more likely to report having connections with people who are IDU’s than youth who are not
travelers. None of the online social capital variables were however associated with reporting
having connections to people who engage in IDU.
In the multivariate model, only race was significantly associated with having connections
to people who engage in IDU. White youth were 5.7 times more likely to report having
connections to people who engage in IDU compared to youth from other ethnicities.

88

Table 15 : Odds ratios for association between online social capital and descriptive norms of injection drug use
Demographic Characteristics
Field site (Santa Monica=1) 2.77 1.02 7.55 * 1.33 0.40 4.46
Age 0.93 0.74 1.17
Time Homeless 0.94 0.77 1.14
Gender (Male=1) 0.74 0.27 2.03
Sexual Orientation (Non-heterosexual=1) 1.26 0.44 3.65
Race (White=1) 4.89 1.70 14.05 ** 5.74 1.39 8.67 *
Current Living Situation (Street-based=1) 4.56 1.45 14.27 ** 2.30 0.66 8.04
High School Graduate (Yes=1) 0.73 0.27 1.99
Traveler (Yes=1) 3.94 1.27 12.23 * 1.98 0.58 6.73
Social Capital
Bridging Capital
Connect online with Family 0.28 0.11 0.73 ** 0.48 0.17 1.34
Connect online with Home-based Peers 0.78 0.29 2.06
Connect online with Caseworkers 0.52 0.18 1.52
Bonding Capital
Connect online with Street-based Peers
1.67 0.63 4.44
Pseudo Rsquare 0.04
2 Log Likelihood 120.01

#
p<.10 *=p<0.05; **=p<0.01; ***=p<0.001
Note: Only variables significant in the bivariate analyses at p< .10 were included in the final adjusted analyses
Unadjusted
OR
Adjusted
OR
Alters IDU Alters IDU
95% CI 95% CI

Injunctive Norms
Bivariate and multivariate results for injunctive norms (separated by encourage and
object use) regarding IDU are presented in Table 16 and 17.
Encourage Injection Drug Use (IDU)
Concerning encouragement of IDU, bivariate statistics revealed that sexual orientation,
living situation, and being a traveler were all associated with reports of having connections to
people who engaged in IDU. Non-heterosexual youth were 2.2 times more likely to report having
connections to people who would encourage them to engage in IDU compared to heterosexual

89

youth. Street based youth were 3 times more likely to report having connections to people who
would encourage them to engage in IDU compared to non-street youth. Travelers were 2 times
more likely to have network members who engaged in IDU relative to non-travelers.
In the multivariate model, however, only living situation was significantly associated
with having connections to people who would encourage IDU. Specifically street based youth
were 2.4 times more likely to report having connections to people who would encourage them to
engage in IDU.

90

Table 16 : Odds ratios for association between online social capital and injunctive norms of injection drug use
Unadjusted
OR
Adjusted
OR
Demographic Characteristics
Field site (Santa Monica=1) 1.55 0.75 3.22
Age 1.13 0.94 1.36
Time Homeless 0.98 0.85 1.13
Gender (Male=1) 1.38 0.58 3.29
Sexual Orientation (Non-heterosexual=1) 2.22 1.03 4.80 * 2.01 0.90 4.48
Race (White=1) 0.95 0.44 2.05
Current Living Situation (Street-based=1) 3.03 1.38 6.64 ** 2.44 1.07 5.57 *
High School Graduate (Yes=1) 1.17 0.56 2.43
Traveler (Yes=1) 2.05 1.05 4.41 * 1.54 0.68 3.47
Social Capital
Bridging Capital
Connect online with Family 0.51 0.24 1.08
Connect online with Home-based Peers 0.54 0.26 1.12
Connect online with Caseworkers 0.90 0.35 2.28
Bonding Capital
Connect online with Street-based Peers 1.60 0.75 3.42
Pseudo Rsquare 0.06
2 Log Likelihood 191.9

#
p<.10 *=p<0.05; **=p<0.01; ***=p<0.001
Note: Only variables significant in the bivariate analyses at p< .10 were included in the final adjusted analyses
Encourage IDU Encourage IDU
95% CI 95% CI

Object to IDU
On the bivariate level, only gender and living situation were associated with reports of
having connections to people who would object to them engaging in IDU. Males were 43% less
likely to report that they had connections to people who would object to them engaging in IDU.
Similarly, street based youth were 49% less likely to report that they had connections to people
who would object to them engaging in IDU.

91

On the multivariate level, only living situation retained its significance after controlling
for other covariates. Street based youth were 46% less likely to report that they had connections
to people who would object to them engaging in IDU.

92

Unadjusted
OR
Adjusted
OR
Demographic Characteristics
Field site (Hollywood=1) 0.78 0.50 1.22
Age 1.00 0.89 1.11
Time Homeless 1.02 0.93 1.11
Gender (Male=1) 0.57 0.34 0.98 * 0.66 0.38 1.13
Sexual Orientation (Non-heterosexual=1) 0.83 0.49 1.40
Race (White=1) 1.14 0.72 1.82
Current Living Situation (Street-based=1) 0.51 0.33 0.81 ** 0.54 0.34 0.86 *
High School Graduate (Yes=1) 0.95 0.60 1.49
Traveler (Yes=1) 0.69 0.44 1.08
Social Capital
Bridging Capital
Connect online with Family 1.06 0.64 1.74
Connect online with Home-based Peers 0.83 0.52 1.34
Connect online with Caseworkers 1.34 0.75 2.36
Bonding Capital
Connect online with Street-based Peers 0.78 0.48 1.27
Pseudo Rsquare 0.04
2 Log Likelihood 422.13

#
p<.10 *=p<0.05; **=p<0.01; ***=p<0.001
Note: Only variables significant in the bivariate analyses at p< .10 were included in the final adjusted analyses
Table 17 : Odds ratios for association between online social capital and injunctive norms of injection drug use
Object IDU Object IDU
95% CI 95% CI

93

CHAPTER SEVEN
RESULTS STUDY AIM 3
SOCIOMETRIC ANALYSES
This chapter presents results from the sociometric analyses specific to Aim 3 of the
dissertation. First, the socio-demographic characteristics of the sociometric sample are presented
for each network separately. Then network visualizations are depicted for each network (i.e.
Hollywood and Santa Monica) for each kind of outcome (descriptive vs. injunctive norm).
Finally, the statistical models describing the associations between the centrality variables and
each of the outcomes are presented for the two networks separately.
Descriptive Statistics
Socio-Demographic Characteristics
Data on socio-demographic characteristics of participants are presented in Table X. Both
networks were predominantly composed of males, however, there were more females in the
Hollywood sample (35.5%) compared to the Santa Monica sample (19.5%). The networks were
similar with regards to sexual orientation. The Santa Monica network was predominantly White
(64.5%), while the Hollywood network was more diverse, with African Americans comprising
the majority of the youth (40.1%). Additionally, more Santa Monica youth reported to be street-
based (66.6%) compared to youth in Hollywood (28%). A large majority of youth in Santa
Monica were travelers (61.7%). Youth in Santa Monica were slightly older (21.7 years) and had
been homeless a little longer (2.9 years).

94

Table 18 : Descriptive statistics of homeless youth in Hollywood and Santa Monica
Hollywood Santa Monica
n=160 n=130
n % n %
Gender
Male 98 64.47 103 80.47
Female 54 35.53 25 19.53
Sexual Orientation
Heterosexual 121 73.83 101 78.91
Non-heterosexual 31 20.39 27 21.09
Race
Native American 4 2.63 3 2.36
Asian 1 0.66 0 0.00
Black 61 40.13 17 13.39
Native Hawaiin or Pacific Islander 1 0.66 1 0.79
White 27 17.76 82 64.57
Latino 34 22.37 6 4.72
Mixed Race 24 15.79 18 14.17
Current Living Situation
Non Street 108 72.00 42 33.33
Street 42 28.00 84 66.67
"Traveler" 58 38.67 79 61.72
Mean Time Homeless (std dev) 2.53 2.48 2.90 2.90

Perceived Network Norms
Table 19 presents perceived norms for the two sociometric networks. More youth in
Santa Monica thought that their peers engaged in methamphetamine use (24.5%) compared to
youth in Hollywood (17.5%). In both networks, very few youth thought that their peers would
encourage them to engage in methamphetamine use. More youth in Santa Monica reported that

95

their peers would object to them engaging in methamphetamine use (56.9%) compared to youth
in Hollywood (57.7%).
In terms of injection drug use, more youth in Hollywood reported that their peers
engaged in injection drug use (18.1%) compared to youth in Santa Monica (11.2%). In both
networks, perceptions of encouragement of both methamphetamine and injection drug use were
very low. Only eight youth in Hollywood and six youth in Santa Monica believed that their peers
would encourage them to engage in injection drug use. More youth in Hollywood (78.1%)
believed that their peers would object to their engagement in injection drug use compared to
youth in Santa Monica (67.2%).

96

Variables n % n %
Proportion of Meth Using Alters
None 105 65.6 83.0 75.5
Any 28 17.5 27.0 24.6
Proportion of Alters Encourage Meth
None 121 75 100.0 90.9
Any 12 7.5 10.0 9.1
Proportion of Alters Object Meth
All 91 56.9 63.0 57.3
Less than All 42 31.6 47.0 42.7
Proportion of IDU Alters
None 120 75 97.0 88.2
Any 13 18.1 13.0 11.8
Proportion of Alters EncourageIDU
None 125 78.1 104.0 94.6
Any 8 5 6.0 5.5
Proportion of Alters Object IDU
All 98 78.1 74.0 67.3
Less than All 35 21.9 36.0 32.7
Table 19 : Perceived Norms for Meth and IDU
Hollywood Santa Monica

Sociometric Structural Characteristics
The sociometric structural properties of the two networks are presented in Table 20. In
terms of the sociometric characteristics, the two networks are very similar. About 60.5% youth in
the Hollywood network and 54.6% of youth in the Santa Monica network had two or less ties in
the network. However, more youth in Hollywood had more influential connections than youth in
Santa Monica. About 76.3% of youth in Hollywood and 71% of youth in Santa Monica were in
the less than two core component.

97

Table 20 : Sociometric Characteristics of Networks
Network Structure n % n %
Degree Centrality
Degrees 1-2 92 60.53 70 54.69
Degrees 3-14 60 39.47 58 45.31
Eigenvector Centrality
< .04/<.03 121 79.61 87 67.97
> .04/>.03 31 20.39 41 32.03
K-core
Kcore 1-2 116 76.32 91 71.09
Kcore 3-4 36 23.68 37 28.91
Hollywood Santa Monica

Network Visualization
As mentioned in Chapter 4, one of the first steps in network analysis is visualization.
These diagrams are an excellent tool for pattern recognition. Netdraw 2.090 graph visualization
software (Analytic Technologies, Irvine, CA) was used to generate network visualizations using
the spring embedder routine. Since these analyses focused on perception of norms, isolates were
excluded. Isolates are people who have no other connections (or did not nominate or were
nominated by anyone in the network), and therefore data on their perceptions of what others in
their network do or expect them to do is not available.
In addition, there were youth who were nominated, but did not nominate anybody else in
the network. Therefore, we knew how other youth who nominated them believed their substance
use behaviors are, but had no data on their perceptions. However, for the sake of the visual
analyses, these youth were retained in the network matrix. Removing these youth would have

98

compromised the structural integrity of the networks, and information on important structural
attributes would have been modified. In all the figures, they have been color-coded as white.
Hollywood Network of Homeless Youth
Structural Characteristics of Descriptive Norms Regarding Methamphetamine Use
Figure 6 represents the sociometric network of 160 homeless youth in Hollywood, Los
Angeles. Small numbers of ties aggregate into larger network structures of homeless youth.
Visual inspection of this network reveals a large interconnected core, some small components,
and triads and pairs of youth who are connected to each other.
Blue nodes are youth who believe that their peers engage in methamphetamine use. On
visually inspecting the network, one can see that youth who report that at least one of their street
peers engages in meth use are a part of the large interconnected central component of the network.
In addition, these youth appear to be highly nominated (based on the direction of the arrows).

99

Figure 6: Perception of Meth Use in Network (Hollywood, n=160, Ties=290). Blue
circles are youth who believe that their peers use methamphetamine. Arrows indicate direction of
nominations between youth.

100

Structural Characteristics of Injunctive Norms Regarding Methamphetamine Use
Figure 7: Perception of Encouragement of Meth Use in Network (Hollywood, n=160, Ties=290).
Blue circles are youth who believe that their peers will encourage methamphetamine use. Arrows
indicate direction of nominations between youth.

Figure 7 presents the structural characteristics of youth who believed that their peers
would encourage them to use methamphetamine. A visual inspection of the network does not
reveal any structural patterns. Very few youth believed that their peers would approve of their
drug use.

101

Figure 8: Perception of Objection to Meth Use in Network (Hollywood, n=160, Ties=290). Blue
circles are youth who believe that their peers will object to methamphetamine use. Arrows
indicate direction of nominations between youth.

Figure 8 presents the structural characteristics of youth who believed that their peers
would not approve of methamphetamine use. A visual inspection of the network does not reveal
any structural patterns. A large number of youth believed that their peers would approve of their
drug use.

102

Structural Characteristics of Descriptive Norms Regarding Injection Drug Use (IDU)
Figure 9: Perception of Injection Drug Use in Network (Hollywood, n=160, Ties=290). Blue
circles are youth who believe that their peers engage in IDU. Arrows indicate direction of
nominations between youth.

Figure 9 presents the perception of injection drug use in the Hollywood network. Blue
nodes are youth who believe that their peers engage in injection drug use. Perceptions of
injection drug use are less prevalent within the network compared to methamphetamine use.
However, similar to perceptions of methamphetamine use, youth who believed that their peers
engaged in injection drug use are concentrated in the interconnected core of the network.
Additionally, youth who were highly nominated were more likely to think that their peers
engaged in injection drug use.

103

Structural Characteristics of Injunctive Norms Regarding Injection Drug Use (IDU)
Figure 10: Perception of Encouragement of Injection Drug Use in Network (Hollywood, n=160,
Ties=290). Blue circles are youth who believe that their peers would object to engagement in
IDU. Arrows indicate direction of nominations between youth.

Figure 10 presents the structural characteristics of youth who believed that their peers
would encourage them to engage in injection drug use. A visual inspection of the network does
not reveal any structural patterns. Very few youth (n=8) believed that their peers would approve
of their injection drug use.

104

Figure 11: Perception of Objection to Injection Drug Use in Network (Hollywood, n=160,
Ties=290). Blue circles are youth who believe that their peers would object to engagement in
IDU. Arrows indicate direction of nominations between youth.

Figure 11 presents the structural characteristics of youth who believed that their peers
would not approve of injection drug use. A visual inspection of the network does not reveal any
structural patterns. A large number of youth believed that their peers would disapprove of them
engaging in injection drug use.

105

Santa Monica Network of Homeless Youth
Structural Characteristics of Descriptive Norms Regarding Methamphetamine Use
Figure 12: Perception of Methamphetamine Use in Network (Santa Monica, n=130, Ties=242).
Blue circles are youth who believe that their peers engage in meth use. Arrows indicate direction
of nominations between youth.

Figure 12 presents the structural characteristics of perceptions of methamphetamine use
in the Hollywood network. Unlike the Hollywood network, youth who perceive that their peers
use methamphetamine in the Santa Monica network were more likely to be peripheral or be in
the lower kcore regions (i.e. kcore lesser than 2).

106

Structural Characteristics of Injunctive Norms Regarding Methamphetamine Use
Figure 13: Perception of Encouragement of Methamphetamine Use in Network (Santa Monica,
n=130, Ties=242). Blue circles are youth who believe that their peers would encourage them to
engage in meth use. Arrows indicate direction of nominations between youth.

Figure 13 presents the structural attributes of encouragement of methamphetamine use in
the network. Similar to the Hollywood network, very few youth thought that their peers would
encourage them to engage in methamphetamine use. No concrete structural pattern could be
discerned from the visual inspection.

107

Figure 14: Perception of Objection to Methamphetamine Use in Network (Santa Monica, n=130,
Ties=242). Blue circles are youth who believe that their peers would object to them to engaging in
meth use. Arrows indicate direction of nominations between youth.

Figure 14 presents the structural attributes of perceptions of objection to
methamphetamine use in the Santa Monica network. Visual inspection of the network reveals that
youth who are in the higher k-core regions (i.e. have a kcore greater than 2) were more likely to
believe that their peers would object to them engaging in methamphetamine use.

108

Structural Characteristics of Descriptive Norms Regarding Injection Drug Use
Figure 15: Perception of Injection Drug Use in Network (Santa Monica, n=130, Ties=242). Blue
circles are youth who believe that their peers would engage in injection drug use. Arrows indicate
direction of nominations between youth.

Figure 15 presents the structural attributes of injection drug use in the Santa Monica
network. Very few youth thought that their peers would engage in injection drug use. No real
structural pattern could be discerned from the visual inspection of the network.

109

Structural Characteristics of Injunctive Norms Regarding Injection Drug Use
Figure 16: Perception of Encouragement of Injection Drug Use in Network (Santa Monica,
n=130, Ties=242). Blue circles are youth who believe that their peers would encourage them to
engage in injection drug use. Arrows indicate direction of nominations between youth.

Figure 16 presents the structural attributes of encouragement of injection drug use in the
network. Similar to the Hollywood network, very few youth thought that their peers would
encourage them to engage in injection drug use. Similarly, no concrete structural pattern could be
discerned from the visual inspection.

110

Figure 17: Perception of Objection to Injection Drug Use in Network (Santa Monica, n=130,
Ties=242). Blue circles are youth who believe that their peers would object to them engaging in
injection drug use. Arrows indicate direction of nominations between youth.

Figure 17 presents the structural attributes of objection to injection drug use in the Santa
Monica network. From a preliminary inspection of the network, no real visible structural patterns
could be discerned.
Statistical Results
Bivariate and Multivariate Analyses
The third and final aim of the study was to assess the association between sociometric
network characteristics and perceived social network norms (descriptive and injunctive) for
methamphetamine and injection drug use among homeless youth. Described below are the
findings from the bivariate and the multivariate analyses specific to this study aim. The findings

111

are further separated by outcomes (methamphetamine and injection drug use) and networks (i.e.
Hollywood vs. Santa Monica). The analyses assessing associations with encouragement of both
methamphetamine and injection drug use were not conducted because of the lack of variation in
the data.
Hollywood Network
Descriptive Norms (Methamphetamine Use)
Table 21 presents bivariate and multivariate statistics assessing associations between
sociometric characteristics and perceived norms of methamphetamine use in the Hollywood
network. Bivariate statistics revealed that sexual orientation, current living situation, traveler
status, Kcore position, and degree centrality were all significantly associated with perceived
norms regarding methamphetamine use in the network.
In the multivariate model, sexual orientation and kcore status remained significant.
Specifically, non-heterosexual youth were 3.6 times more likely to think that their alters used
methamphetamine. Youth who were in more cohesive networks (as defined by their kcore status)
were 4.5 times more likely to report that their alters used methamphetamine compared to youth
who were in less dense sub-networks.

112

Table 21 Odds ratios for association between sociometric properties and descriptive norms of methamphetamine use
(Hollywood, n=133)
Demographic Characteristics
Age 0.98 0.77 1.23
Time Homeless 0.17 0.98 1.39
Gender (Male=1) 0.44 0.18 1.13
Sexual Orientation (Non-heterosexual=1) 4.41 1.66 11.69 ** 3.67 1.22 11.11 *
Race (White=1) 2.78 0.88 7.73
Current Living Situation (Street-based=1) 2.67 1.00 7.14
*
2.13 0.68 6.65
Traveler (Yes=1) 2.74 1.04 7.18 * 1.91 0.62 5.96
Sociometric Structural Properties
Kcore (Kcore 3-4=1) 5.95 2.25 15.70 *** 4.59 1.08 19.59 *
Degree Centrality (Deg 3-21=1) 2.90 1.12 7.51 * 1.03 0.81 1.30
Eigenvector Centrality (> .03=1) 2.23 0.81 6.12
Pseudo R-Square 0.12
2 Log Likelihood 89.27

#
p<.10 *=p<0.05; **=p<0.01; ***=p<0.001
Note: Only variables significant in the bivariate analyses at p< .10 were included in the final adjusted analyses
Alters Use Meth Alters Use Meth
Unadjusted
OR
Adjusted
OR 95% CI 95% CI

Injunctive Norms (Methamphetamine Use)
Table 22 provides bivariate and multivariate statistics of factors associated with
injunctive norms regarding methamphetamine use (specifically objection to methamphetamine
use). As mentioned before, due to the lack of variation in the data, analyses for encouragement of
methamphetamine use was not conducted.
Only gender and sociometric position (as defined by eigenvector centrality) were
associated with perceptions of objection to methamphetamine use. At the multivariate level, both
variables continued to remain significant. Specifically, males were 67% less likely to believe that
their alters would object to them engaging in methamphetamine use compared to females. Youth

113

who were highly central (as defined by their eigenvector centrality) were 63% less likely to
believe that their alters would object to them engaging in methamphetamine use compared to
peripheral youth.

114

Table 22: Odds ratios for association between sociometric properties and injunctive norms of methamphetamine use
(Hollywood, n=133)
Demographic Characteristics
Age 1.02 0.85 1.23
Time Homeless 0.97 0.83 1.13
Gender (Male=1) 0.38 0.16 0.89 * 0.33 0.13 0.80 *
Sexual Orientation (Non-heterosexual=1) 0.73 0.31 1.68
Race (White=1) 0.65 0.26 1.60
Current Living Situation (Street-based=1) 0.27 0.12 0.62
Traveler (Yes=1) 0.69 0.32 1.48
Sociometric Structural Properties
Kcore (Kcore 3-4=1) 0.52 0.23 1.16
Degree Centrality (Deg 3-21=1) 0.67 0.32 1.42
Eigenvector Centrality (> .03=1) 0.45 0.19 1.06 # 0.37 0.15 0.92 *
Pseudo R-Square 0.07
2 Log Likelihood 150.52

#
p<.10 *=p<0.05; **=p<0.01; ***=p<0.001
Note: Only variables significant in the bivariate analyses at p< .10 were included in the final adjusted analyses
Alters Object Meth Alters Object Meth
Unadjusted
OR
Adjusted
OR 95% CI 95% CI

Descriptive Norms (Injection Drug Use)
Bivariate and multivariate statistics describing associations with descriptive norms
regarding injection drug use are presented in Table 23. On the bivariate level, sexual orientation,
race, living situation, being a traveler and sociometric position (as defined by kcore status) were
associated with perceptions of injection drug use in the network.
However, in the multivariate model, sociometric position (as defined by kcore status)
ceased to be significant. However, race and living situation were significantly associated with
descriptive norms regarding injection drug use. Specifically, White youth were 3.8 times more
likely to think that their alters would engage in injection drug use compared to youth of other
races/ethnicities. Similarly, street based youth were 4.3 times more likely to think that their alters

115

would engage in injection drug use compared to youth who were in some form of temporary
housing.

116

Table 23: Odds ratios for association between sociometric properties and descriptive norms of injection drug use
(Hollywood, n=133)
Demographic Characteristics
Age 1.04 0.85 1.27
Time Homeless 1.04 0.88 1.23
Gender (Male=1) 1.66 0.69 4.05
Sexual Orientation (Non-heterosexual=1) 2.27 0.97 5.30 #
Race (White=1) 5.90 2.37 14.68
***
3.82 1.36 10.79 *
Current Living Situation (Street-based=1) 6.00 2.50 14.41
***
4.34 1.67 11.27 **
Traveler (Yes=1) 2.76 1.21 6.34 * 1.42 0.54 3.72
Sociometric Structural Properties
Kcore (Kcore 3-4=1) 2.72 1.16 6.41 * 1.49 0.54 4.12
Degree Centrality (Deg 3-21=1) 2.03 0.90 4.54
Eigenvector Centrality (> .03=1) 1.24 0.48 3.24
Pseudo R-Square 0.16
2 Log Likelihood 118.26

#
p<.10 *=p<0.05; **=p<0.01; ***=p<0.001
Note: Only variables significant in the bivariate analyses at p< .10 were included in the final adjusted analyses
Alters IDU Alters IDU
Unadjusted
OR
Adjusted
OR 95% CI 95% CI

Santa Monica Network
Descriptive Norms Regarding Methamphetamine Use
Bivariate and multivariate statistics are presented in Table 24. On the bivariate level, age,
time since first homeless and sociometric position (as defined by kcore status) were all
associated with descriptive norms regarding methamphetamine use in the Santa Monica network.
In the multivariate model, time since first homeless and sociometric position (as defined
by kcore status) remained significant. Youth who reported being homeless longer were 1.18
times more likely to report that their alters engaged in methamphetamine use. Contrary to the
Hollywood network, youth who were in more cohesive networks in the Santa Monica network
were 71% less likely to believe that their peers engage in methamphetamine use.

117

Table 24: Odds ratios for association between sociometric properties and descriptive norms of methamphetamine use
(Santa Monica, n=109)
Demographic Characteristics
Age 0.88 0.71 1.09 # 0.86 0.67 1.10
Time Homeless 1.14 0.99 1.32 # 1.18 1.01 1.39 *
Gender (Male=1) 0.79 0.29 2.17
Sexual Orientation (Non-heterosexual=1) 1.47 0.53 4.06
Race (White=1) 0.75 0.31 1.84
Current Living Situation (Street-based=1) 2.13 0.77 5.85
Traveler (Yes=1) 1.12 0.46 2.75
Sociometric Structural Properties
Kcore (Kcore 3-4=1) 0.38 0.13 1.11 # 0.29 0.09 0.97 *
Degree Centrality (Deg 3-21=1) 0.64 0.27 1.54
Eigenvector Centrality (> .03=1) 0.74 0.29 1.90
Pseudo R-Square 0.11
2 Log Likelihood 101.73

#
p<.10 *=p<0.05; **=p<0.01; ***=p<0.001
Note: Only variables significant in the bivariate analyses at p< .10 were included in the final adjusted analyses
Alters Use Meth Alters Use Meth
Unadjusted
OR
Adjusted
OR 95% CI 95% CI

Injunctive Norms Regarding Methamphetamine Use
Table 25 provides bivariate and multivariate statistics of factors associated with injunctive
norms regarding methamphetamine use (objection to methamphetamine use). As mentioned
before, due to the lack of variation in the data, analyses for encouragement of methamphetamine
use was not conducted.
On the bivariate level, age, time since first homeless and sociometric position (as defined
by kcore status and degree centrality independently) were significantly associated with
perceptions of objection to methamphetamine use. In the multivariate model, only kcore status
remained significant. Youth who were in more cohesive networks in the Santa Monica network
were 7.2 times more likely to believe that their alters would object to them engaging in
methamphetamine use.

118

Table 25: Odds ratios for association between sociometric properties and injunctive norms of methamphetamine use
(Santa Monica, n=109)
Demographic Characteristics
Age 1.05 0.88 1.27 # 1.03 0.83 1.28
Time Homeless 0.96 0.84 1.10 # 0.96 0.83 1.12
Gender (Male=1) 0.87 0.35 2.14
Sexual Orientation (Non-heterosexual=1) 0.62 0.25 1.55
Race (White=1) 1.09 0.50 2.41
Current Living Situation (Street-based=1) 0.61 0.27 1.37
Traveler (Yes=1) 0.81 0.37 1.76
Sociometric Structural Properties
Kcore (Kcore 3-4=1) 3.17 1.31 7.65 * 7.29 1.74 30.49 **
Degree Centrality (Deg 3-21=1) 2.51 1.15 5.47 * 0.85 0.62 1.17
Eigenvector Centrality (> .03=1) 1.72 0.76 3.89
Pseudo R-Square 0.12
2 Log Likelihood 122.77

#
p<.10 *=p<0.05; **=p<0.01; ***=p<0.001
Note: Only variables significant in the bivariate analyses at p< .10 were included in the final adjusted analyses
Alters Object Meth Alters Object Meth
Unadjusted
OR
Adjusted
OR 95% CI 95% CI

Descriptive Norms Regarding Injection Drug Use
Bivariate statistics are presented in Table 26. On the bivariate level, only race was
significantly associated with descriptive norms regarding injection drug use in the Santa Monica
network. In the light of this finding, further multivariate analyses were not conducted.

119

Table 26: Odds ratios for association between sociometric properties and descriptive norms of injection drug use
(Santa Monica, n=109)
Demographic Characteristics
Age 0.956 0.721 1.268
Time Homeless 0.93 0.74 1.18
Gender (Male=1) 0.98 0.25 3.87
Sexual Orientation (Non-heterosexual=1) 1.16 0.29 4.60
Race (White=1) 7.86 0.98 62.95
#
Current Living Situation (Street-based=1) 0.80 0.24 2.64
Traveler (Yes=1) 1.03 0.31 3.39
Sociometric Structural Properties
Kcore (Kcore 3-4=1) 0.90 0.26 3.16
Degree Centrality (Deg 3-21=1) 1.24 0.39 3.96
Eigenvector Centrality (> .03=1) 0.53 0.14 2.06
Pseudo R-Square
2 Log Likelihood

#
p<.10 *=p<0.05; **=p<0.01; ***=p<0.001
Note: Only variables significant in the bivariate analyses at p< .10 were included in the final adjusted analyses
Multivariate model not conducted since only one variable significant at the bivariate level
Alters IDU
Unadjusted
OR 95% CI

Injunctive Norms Regarding Injection Drug Use
Table 27 provides bivariate and multivariate statistics of factors associated with
injunctive norms regarding injection drug use (objection to injection drug use). As mentioned
before, due to the lack of variation in the data, analyses for encouragement of injection drug use
was not conducted.
On the bivariate level, age, time since first homeless, gender, sexual orientation, and
sociometric position (as defined by kcore status) were all significantly associated with beliefs
that injection drug use would not find approval in these youth’s networks. On the multivariate

120

level, only sociometric status remained significant. Specifically youth who were in more
cohesive networks were 3 times more likely to report that their alters would object to them
engaging in injection drug use.

121

Table 27: Odds ratios for association between sociometric properties and injunctive norms of injection drug use
(Santa Monica, n=109)
Demographic Characteristics
Age 1.20 0.98 1.47 # 1.12 0.89 1.42
Time Homeless 1.15 0.97 1.37 # 1.13 0.95 1.35
Gender (Male=1) 0.96 0.37 2.49
Sexual Orientation (Non-heterosexual=1) 2.76 0.86 8.84
Race (White=1) 0.62 0.26 1.48
Current Living Situation (Street-based=1) 1.38 0.60 3.18
Traveler (Yes=1) 0.83 0.36 1.89
Sociometric Structural Properties
Kcore (Kcore 3-4=1) 2.13 0.85 5.32 # 3.04 1.09 8.53 *
Degree Centrality (Deg 3-21=1) 1.56 0.70 3.49
Eigenvector Centrality (> .03=1) 1.56 0.70 3.49
Pseudo R-Square 0.09
2 Log Likelihood 114.84

#
p<.10 *=p<0.05; **=p<0.01; ***=p<0.001
Note: Only variables significant in the bivariate analyses at p< .10 were included in the final adjusted analyses
Alters Object IDU Alters Object IDU
Unadjusted
OR
Adjusted
OR 95% CI 95% CI

122

CHAPTER EIGHT
DISCUSSION
Little research has been conducted on the relationships between sociometric network
locations or of personal network characteristics to norms of substance use behaviors related to
HIV transmission among homeless youth. The main purpose of this dissertation was to
understand whether social norms of substance use behaviors are clustered within social networks
and the association of these norms to substance use behaviors (specifically methamphetamine
and injection drug use) among homeless youth. Additionally, researchers have suggested that
future studies must address how different relationship roles influence HIV behavioral norms.
Homeless youth are increasingly using new technology to reach out to family and friends from
home, and this has implications for their risk-taking behaviors. Therefore, this study also sought
to understand the role of new forms of communicative technology (such as email and social
media) in connecting homeless youth to relationships beyond their street peers, and whether
connecting to these diverse relationships lead to more protective norms regarding risky
behaviors. Taken together, the egocentric and sociometric data were able to examine multiple
research questions that have not yet been addressed in the HIV risk norms literature. This chapter
will summarize some of the major findings of this study and discuss how it relates to and
expands on previous research. Additionally, the limitations of the study are addressed. Finally,
the implications of the study for future social work practice, policy, and research is identified.
Discussion of Key Findings
Several important empirical findings emerge from these analyses. This study was unique
because it focused on both social influence as well as positional characteristics of homeless

123

youth’s social networks. While social influence measures the social proximity of adolescents
with other substance users, positional characteristics reflect the social standing of youth relative
to their peers (Ennett et al., 2006). These data suggest that both social proximity as well as social
positioning within networks is associated with risk norms. The results therefore overall
supported the general proposition that both egocentric and sociometric network attributes affect
substance use among homeless youth. Other findings elaborated on these overall results. The
findings are discussed in more detail in relation to each study aim and associated hypothesis.
Norms and Behavior
The first aim of the study was to assess the association between perceived social network
norms (descriptive and injunctive) and HIV risk behaviors (methamphetamine and injection drug
use) among homeless youth. The hypothesis posited that stronger the perceived network norms
approving substance use behaviors, the greater the likelihood that respondents will engage in
these behaviors. The results from this study provide evidence that both descriptive and injunctive
norms play an important role in understanding methamphetamine and injection drug use
behaviors among homeless youth. However, consistent with previous research, the findings
indicate that the perception of these norms, especially with regards to how pervasive these
behaviors are (descriptive norms) and how their network contacts would react to each kind of
substance use (injunctive norm) differed based on the type of substance use (i.e.
methamphetamine vs. injection drug use).
In terms of descriptive norms, overall, more youth believed that their network members
engaged in methamphetamine use than injection drug use. While 30% of the youth indicated that
they knew at least one person in their network that used methamphetamine, only 5% of youth

124

believed that anybody in their network were injection drug users. These results suggest variation
in the social nature of substance use across substances (Ennett et al., 2006). Studies have found
that certain drugs are more stigmatized than others are (Flom et al., 2001). It is possible that
methamphetamine is more acceptable among homeless youth compared to injection drug use.
Alternatively, this may also be related to local drug availability (Gleghorn et al., 1998). For
example, a recent study found that methamphetamine abuse has outpaced both heroin and
cocaine use on the West Coast because of its local production (Weisheit & Wells, 2010).
It is however important to note that perceived norms do not reflect actual norms and
might not always be accurate. Research on college students has shown that they often
underestimate or overestimate their peers’ use (Perkins et al., 1999; Perkins et al., 2005). These
same misconceptions may occur among homeless youth. However, these perceptions, even
though inaccurate, have been found to be as or in many cases more important at influencing
behavior than peers’ actual behavior and therefore need to be assessed (Iannotti & Bush, 1992).
However, further research is also needed that can examine the congruency between actual and
perceived behavior.
This study also investigated injunctive norms, or perceptions of what is considered
acceptable behavior. Concerning injunctive norms regarding substance use, the results suggest
that there is little encouragement of and substantial objection to both methamphetamine and
injection drug use. Specifically, 57.38% of youth indicated that they had a greater proportion of
people in their network who would object to their engagement in methamphetamine use and
almost two-thirds of youth (67%) indicated that a greater proportion of people in their network
who would object to their engagement in injection drug use than people who encourage it. This

125

parallels past research, which has reported very similar findings (Flom, Friedman, Jose, & Curtis,
2001). Flom and colleagues (2001) found in their study of low-income minority non-homeless
youth that 98% of youth indicated that none of their peers would encourage them to engage in
injection drug use, and almost all (97%) of their peers would object to them engaging in injection
drug use.
Given that this is a population of youth, among whom drug use is much higher than
similarly aged housed youth, the lack of encouragement of drug use is somewhat surprising and
definitely encouraging. Since egocentric techniques are able to capture network contacts beyond
just similarly aged friends or similarly situated street peers, it is possible that these injunctive
norms are representative of family members and other adult figures in these youth’s lives.
However, in analyzing the data for the sociometric analyses, which effectively reduced the
network members to only street peers, the findings did not change. Even after accounting for
only the norms of street peers, the findings still indicated that encouragement for either kind of
substance was sparse.
However, these findings also need to be interpreted with caution. Absence of
encouragement does not always imply the lack of influence. One needs to note here that an
important distinction needs to be made between peer pressure and peer influence (McIntosh,
MacDonald, & McKeganey, 2003). Apart from direct coercion, peers also influence individual’s
behavior in other ways (McIntosh et al., 2003). For example, a study of 235 teenagers aged 11
through 19 (Hart & Hunt, 1997) found that direct pressure from peers to use drugs was rare.
Instead, the selection of like-minded friends and the process of socialization by which
individuals internalize the group’s attitudes and values were more significant in understanding

126

the mechanisms through which peer mechanisms influenced drug use (McIntosh et al., 2003).
This reinforces the need to study both descriptive and injunctive norms.
The hypothesis that perceived network norms for substance use behaviors would be
associated with actual drug use behaviors among homeless youth was supported by the results of
the study. Strong statistical associations were found between self-reported methamphetamine and
injection drug use and perceptions of network members’ engaging in and objecting to engaging
in methamphetamine and injection drug use.
As hypothesized, having network members who also engage in drug use, more
specifically methamphetamine and injection drug use, was associated with a greater likelihood
that youth themselves would be engaging in both types of drug use (Wenzel et al., 2010; Tyler,
2008; Rice et al., 2005; Memories et al., 2002). This confirms previous egocentric studies and
reinforces the validity of this study. Conceptually, it provides support for both the principle of
homophily- the notion that people tend to be similar to their peers (McPherson, Smith-Lovin, &
Cook, 2001) and Social Learning Theory- the notion that people learn within a social context
primarily through observational learning (Bandura, 1977). Studies have repeatedly demonstrated
the inclination of youth to associate with people who share their behavioral patterns, such as drug
use (McPherson et al., 2001; Mercken, Candel, Willems, & de Vries, 2007).
However, it is also important to remember that youth select their relationships within the
confines of their immediate environment, the features of which limits the pool of peers available,
the kinds of relationships that develop, and consequently the kind of role models that one has
access to (Epstein & Karweit, 1983). Homeless youth in particular find themselves surrounded
disproportionately by other street youth who engage in various risky behaviors such as illicit

127

drug use, and therefore are more likely to be socialized into these risky practices (Rice et al.,
2005). Disentangling the effects of peer selection vs. influence is beyond the scope of this study;
however previous studies examining peer influence vs. selection effects on youth substance use
has found evidence for both processes (e.g., Bohnert, Bradshaw, & Latkin, 2009; Go, Green,
Kennedy, Pollard, & Tucker, 2010; Hall & Valente, 2007).
Furthermore, injunctive norms continued to be significantly associated with both
methamphetamine and injection drug use even after controlling for descriptive norms. This is
also consistent with previous research, which has found that that descriptive and injunctive
norms have independent effects on behavior across a wide range of behaviors’ including drug
use (McMillan & Conner, 2003), safe sex behavior (White, Terry, & Hogg, 1994), physical
exercise (Rhodes & Courneya, 2003), and aggressive behavior (Norman, Clark, & Walker,
2005). Specifically, the results from this study indicate that having a greater number of alters
who object to drug use compared to those who encourage it is associated with the reduced
likelihood of engaging in both methamphetamine and injection drug use. Therefore, it is
important to understand that networks could provide both risk and protective influences.
However, as long as the protective factors outweigh the detrimental ones, the outcome tends to
be positive.
Internet and social media use among homeless youth- composition of networks and
associations with norms
Most of the research on norms has focused on face-to-face interaction and influences,
limiting our understanding of how norms may be promoted through social networking
technology. This study expanded on the previous research on norms by assessing the kind of
relationships that homeless youth in this study maintain online and whether these connections are

128

associated with more protective substance use norms. While preliminary in nature, these results
support previous research that have found that different reference groups promote different
behavioral norms (Borsari & Carey, 2003; Carey, Borsari, Carey, & Maisto, 2006; Cho, 2006).
Like other recent studies (Bender, Ferguson, Pollio, Thompson, & McClendon, 2009;
Karabanow & Naylor, 2010; Rice et al., 2010), these data suggest that homeless youth are
actively using the internet to connect to relationships outside of street life. More than two-thirds
of the youth in the study reported connecting online to family, and more than half reported
connecting to home-based peers. More notably, about one-fifth of youth reported that they
connected to caseworkers online, which is not only encouraging, but has implications for how
social workers can reach out to these transient youth, and the format through which interventions
can be delivered. More importantly, these findings reinforce previous studies which have found
that e-mail and social networking websites provide new opportunities for homeless youth in
having easier means to engage in processes of bridging to non-street networks (Barman-Adhikari
& Rice, 2011; Rice & Barman-Adhikari, in press).
Perhaps the most compelling aspects of these data are the associations among
connections with family and street-peers and perceptions of methamphetamine use and
encouragement of methamphetamine use among youth in this study. While communicating with
family online was associated with a decreased likelihood of believing that network members
used methamphetamine, connecting to street peers online was associated with a greater
likelihood that these youth would believe that they would be encouraged to use
methamphetamine. These findings support the “Social Identity Theory” (Tajfel & Turner, 1986)
which proposes that norms are linked to specific social groups, and identification with these
groups determines what will be regarded as normative. It can be assumed that homeless youth

129

who are connected to family members online have more role models to rely on in understanding
what can be considered normative behavior. In addition, it is also conceivable that the
conversations that youth are having over email or social media with their family members might
act as sources of informational support, which is known, to influence the expression of norms.
However, it is important to mention that this study did not collect any data on the nature of these
online interactions, more specifically, what was discussed. However, previous studies have
found that communication regarding behaviors inform perceptions of the ‘‘normal’’ or most
prevalent behaviors among network members (Latkin & Knowlton, 2000; Barrington et al.,
2009). A future study that collects detailed network level information on ties and the content of
interactions (both face-to-face and online) across those ties would do much to increase our
understanding of these processes, which could be critical in informing clinical interventions with
youth and families.
It is important to note that none of the online social capital items were significantly
associated with any of the injection drug use norms outcomes. The non-significance of these
relationships may be due to the small number of people who inject drugs, the small number who
have any friends who encourage such use, and the large number of subjects for whom all friends
object to the use of IDU as compared to methamphetamine. Additionally, it is important to
reiterate, that the low self-reports of injection drug use and others’ engagement in it could reflect
local drug availability rather than the norms of acceptable drug use (Gleghorn et al., 1998).
Sociometric Positions and Norms
The third and final aim of the study was to understand how norms of substance use
behaviors linked to HIV transmission are clustered within social networks. The results from this

130

study provide us with some preliminary evidence that substance use norms are associated with
certain structural sociometric characteristics. However, there were variations in the kind of
network characteristics associated with each kind of norm (i.e. descriptive vs. injunctive) and
more notably, the direction of the associations between the network characteristics and norms for
each network (i.e. Santa Monica vs. Hollywood).
Overall, in terms of sociometric network characteristics, the study found that perception
of substance use behaviors (both methamphetamine and injection drug use) and objections to
these behaviors are largely shaped by the cohesiveness in sub-regions of the sociometric network
(defined as Kcore) rather than one’s centrality within these networks (as defined by the number
of connections or number of influential connections one has). From a theoretical as well as
measurement standpoint, a distinction needs to be made between network centrality and network
cohesion (or coreness) within sociometric networks especially since this has implications for not
only the interpretation of the findings but also what kind of intervention programs should be
implemented to change norms about substance use in networks.
Network cohesion in a sociometric or bounded network is identified by focusing on how
an individual is embedded in the structure of groups within a network (Borgatti & Everett, 2000).
Cohesive subgroups are therefore identified based on “dense interconnections within sets of
individuals” (operationalized as the Kcore) (Borgatti & Everett, 2000). Centrality is measured
and operationalized in many ways. For the purposes of this study, degree centrality and
eigenvector centrality were used to arrive at a definition of centrality. While degree centrality is
in most simple terms is defined as the number of connections an individual has within a
sociometric network, eigenvector centrality assumes that connections to high-scoring nodes

131

contribute more to the score of the node in question than equal connections to low-scoring nodes.
The following example best exemplifies the difference between centrality and cohesion.
For instance, it is possible to take the most central actors in a network, according to some
measure of centrality (closeness or degree) and find that the subgraph induced by the set contains
no ties whatsoever - an empty core (Borgatti & Everett, 2000). This is because each actor may
have high centrality by being strongly connected to different cohesive regions of the graph and
need not have any ties to each other (Borgatti & Everett, 2000). Thus, every person in a dense
core is likely to be highly central, but not every central person need be a part of the core.
Consequently, a highly central person might have many connections within a network, but not be
part of an influential group or clique. Therefore, prominence cannot always be equated with
influence (Mizruchi, 1993).
It is therefore not surprising to find that cohesion was more significant in understanding
the perception of norms relative to an individual’s centrality in a network. Because of the dense
nature of ties, cohesion is more salient in understanding how people are socialized into a social
circle and the internalization of group norms, because of the frequency of people’s
communication and the likely sanctions associated with non-conformity within such a closely
connected group (Carpentier & White, 2002; Tyler & Melander, 2012). In the context of
adolescent substance use, cohesion implies that adolescents acquire information about their
friends' attitudes or behaviors through direct communication and reinforcement, which then
becomes a frame of reference leading to behavioral similarity (Fujimoto & Valente, 2012).
However, what may seem paradoxical is the finding that social network cohesion is
associated with both increased and reduced perceptions of substance use in two different

132

networks. More specifically, in the Santa Monica network, being in a cohesive network was
associated with reduced perceptions of methamphetamine use. In Hollywood, on the contrary,
cohesiveness is associated with norms supportive of more risky behaviors. This paradoxical
effect of network cohesion on health outcomes have also been illustrated in other studies.
Previous research confirms that network cohesion can be both beneficial and detrimental for
individuals within the network depending on the social context (Barrington, 2008).
For example, among people with mental illness, it has been found that being a part of a
cohesive network is associated with both increased and reduced use of mental health care
(Kadushin, 1966; Pescosolido, Gardner, & Lubell, 1998). Similarly, among injection drug users;
cohesiveness has been associated with norms supportive of both risk enforcement and risk
avoidance (Barrington, 2008; Latkin et al., 2003). For example, Barrington (2008) found that
being a part of a cohesive network was associated with reduced perceptions of condom use, but
with increased communication about safe sex behavior within the same network. Therefore, it is
important to note that social networks operate differently in diverse contexts.
More significantly, one needs to understand that social network structure and content are
interlinked (Pescosolido, 2006). Specifically, while social network structures play a vital role in
facilitating interaction and interdependencies, the outcomes of these interactions depend on the
nature of these social exchanges (whether they are positive or negative) (Lin, 1999). For
example, cohesive spaces with greater connection to other street youth who also engage in more
risky practices will likely be characterized by more high-risk norms. It is also possible that
engaging in high-risk behavior will help shape individuals' sociometric risk network locations
(Fisher, 1988; Friedman et al., 1997). On the other hand, if these network spaces are associated

133

with people who are not risk-takers and condone more protective substance use norms, one is
more likely to endorse these safer norms as well (Fisher, 1988). It would be useful for future
studies to not only measure the existence but also the nature of these ties (e.g. positive or
negative, strong or weak) to gain better insight into the context of these networks to further
clarify these incongruent findings (Barrington, 2008).
However, this begs the larger question as to why youth who are in cohesive networks in
Santa Monica might be less risk-taking than Hollywood. The Santa Monica network has a
greater number of travelers. Almost 64% of youth in Santa Monica consider themselves travelers
compared to only 37% in Hollywood. Travelers are a migratory subgroup of homeless youth
(Martino et al., 2011) who travel along common routes in the United States, and are less likely
than non-traveling homeless youth to stay in one geographic area for too long. Previous studies
have also found that travelers engage in more risky drug and sexual behaviors relative to non-
traveling homeless youth (Lankenau et al., 2008; Martino et al., 2011; Sanders, Lankenau,
Jackson-Bloom, & Hathazi, 2008). What might be particularly salient in explaining the greater
risk taking in less cohesive networks in Santa Monica, is however the composition of these
traveler’s social network characteristics’. Martino and colleagues (2011) compared traveler and
non-traveler personal network characteristics and found that compared to non-travelers, travelers
were less likely to have conventional social ties (such as family) and more likely to associate
with other risky peers.
However, with respect to a sociometric network, it is possible that since travelers do not
live in one geographic area for too long, even though they associate with similarly situated peers,
these connections are not very stable and likely not very cohesive. It is plausible that travelers

134

are part of these less cohesive network spaces and since research has shown that travelers are
more likely to engage in risky practices, it might explain why risk is distributed differently in the
Santa Monica and Hollywood networks. While not the focus of this study, future studies could
explore the positioning of travelers in both networks and their risk taking behaviors in order to
understand whether this explanation can be confirmed.
It is also important to note here that similar to the online social capital findings,
sociometric associations with injection drug use norms were largely insignificant and might be
explained by the same reasons that were cited for the non-significant social capital findings.
Associations with socio-demographic characteristics
In addition to egocentric and sociometric network characteristics, this study also found
associations between other socio-demographic and behavioral characteristics and self-reported
methamphetamine and injection drug use and associated norms. Sexual orientation in particular
was associated with greater methamphetamine use. Non-heterosexual youth in particular were
more likely to engage in methamphetamine use compared to heterosexual youth and additionally,
believe that their peers were more likely to engage in methamphetamine use. This is consistent
with two previous studies that have assessed methamphetamine use among homeless youth
(Douglas, Colfax, Moss, Bansberg, & Hahn, 2008; Salomonsen-Sautel et al., 2008), and perhaps
specify the need for more targeted intervention programs. These findings suggest that even
among homeless youth who are by nature at-risk; lesbian, gay, bisexual, transgender, questioning
(LGBTQ) youth bear a greater burden of negative outcomes compared to their heterosexual
peers, including a greater risk of drug use. The reasons associated with such use range from

135

coping with their sexual identity, general stigmatization, disparities in health and access to care,
and the availability of drugs at the club scenes (Cochran, Stewart, Ginzler, & Cauce, 2002).
Youth who indicated that they lived on the streets (compared to youth who had temporary
housing) were more likely to believe that their peers and other network contacts practiced risky
behaviors. This finding might be reflective of the inherent perilous nature of street life. Homeless
youth’s experiences are entrenched within the complexity of the challenges that they face in their
day-to-day lives (Karabanow & Taylor, 2007). Street youth have fewer resources and less stable
social networks (Davey-Rothwell, 2006). Street youth often have no ties to conventional support
systems (i.e. to non-street peers) and feel highly alienated and marginalized (Rice et al., 2005). In
addition, street youth use drugs because it helps them numb the pain of and endure the
extremities and stressors of street life (MacNeil & Pauly, 2011). Therefore, drug use among
street youth might be more of a structural dysfunction than a personal pathology (Karabanow &
Clement, 2004).
Another noteworthy but expected finding that emerged out of this study is the association
between engaging in sex under the influence and self-reporting as methamphetamine and
injection drug users. Methamphetamine use has been in previous studies implicated with HIV
transmission because of its association with high-risk sexual behaviors (Clements, Gleghorn,
Garcia, Katz, & Marx, 1997; Huba et al., 2001; Kipke et al., 1997; Martinez et al., 1998;
Whitbeck, Hoyt, Yoder, Cauce, & Paradise, 2001). Methamphetamine is known to increase
sexual arousal while reducing inhibitions. The findings of this study provide further evidence
that methamphetamine use is not just a substance use issue, but also a broader public health
problem, because of the implications it has for HIV and other STD transmission.

136

Study Limitations
As is the case with any study, the findings need to be interpreted in light of the limitations
of the study. First, this study was based on cross-sectional data, and therefore does not allow for
causal interpretations; it is possible that youth who are substance users are more likely to assume
that their peers are also substance users, or/and would approve of them engaging in drug use.
Longitudinal research is needed to examine how network dynamics operate over time in order to
understand the causal pathway through which network attributes affect the development of
norms among this population. Second, all the behavioral variables are based on self-reports,
which could be subject to bias. As with any self-reported data, there is the possibility that
behaviors are under- or over-reported due to social desirability biases. Participants were
reminded that the data was confidential and were encouraged to ask questions while completing
the questionnaire to minimize such invalid data. Additionally, the use of computer-assisted self-
interviews, which was employed, have been documented to mitigate issues of social desirability
and impression management (Schroeder, Carey, & Vanable, 2003), and lead to more honest and
unbiased responses. Additionally, the norms variables were all based on the youth’s perceptions.
As mentioned before, these also might not be accurate. Further research on norms should include
independent confirmation of the norms, which were available for these data, but is the focus of a
complementary study.
The way in which social capital was measured in this study was relatively crude.
However, there has been no consensus on the best way in which social capital should be
measured (MacGillivray & Walker, 2000; Popay, 2000). It has been suggested that measures of
social capital should be thoroughly based on, and tied to, the conceptual framework for the

137

specific study (Cavaye, 2004). Lin (1990, 1999) sees social capital as resources that are activated
by individuals through their social ties. The measures of social capital in this study follow Lin’s
(1999) conceptualization of social capital. Furthermore, there is no data available on the timing
of any of the online social capital connections and all of the online behaviors are based on
lifetime self-reports; and could therefore tap into internet use that occurred before the youth
became homeless. However, the questions immediately followed questions on current internet
access and, therefore, it is likely youth were framing their responses to their current situation. In
addition, it was assumed that non-street ties are a source of positive influence. However, no data
was collected regarding the nature of these ties and the content of these social exchanges.
However, previous studies have found that relationships from home are generally more
protective (Milburn et al., 2006).
Finally, while EBA (Freeman & Webster, 1994) is the one of the only viable methods to
collect sociometric data from homeless youth, it is possible that even this approach could be
improved on. The study attempted to enroll as many repeat visitors as possible, and had a low
refusal rate, but it is possible that important members of the homeless youth community were
excluded.
Implications for Policy, Practice, and Future Research
Although this study is exploratory, the results from this study provide specific directions
for interventions aimed to change risky HIV norms among homeless youth. Evidence is growing
that homeless youth networks are not only an important determinant of their risk as well as
protective behaviors, but that they can also be successfully used for prevention (Friedman et al.,
1997). The findings of this study further reinforce these previous results and suggest that

138

interventions seeking to change social norms among homeless youth might be effective in
reducing HIV risk behaviors among homeless youth.
However, in order to further develop and refine network-based interventions, several
issues need to be addressed. One issue is to determine which kind of network intervention is
appropriate and customize it depending on the characteristics of the network. For example, one
of the most popular and widely used network intervention techniques is the use of popular
opinion leaders (Valente, 2012) or more commonly known as the Popular Opinion Leader (POL)
model (Kelly et al., 1994). However, findings of this and previous studies suggests that leaders
within a community might not necessarily be the best change agents. First, just because one is
central does not mean that they have the best reach within the community. For example, Borgatti
observed that the most centrally located nodes sometimes be linked to the same people, and
therefore not be the best people to disseminate information (Valente, 2012). Second, leaders
within a network might be vested in the status quo because it helps them preserve their status
within the network, and therefore might be resistant to change (Rice & Rhoades, 2013). On the
other hand bridging individuals (who span different networks) might be more appropriate for
these interventions (such as the Hollywood network in this study) because they can reach more
individuals and might be more amenable to change, as it does not affect their status within any
groups.
However, this study also found that network cohesion was more significantly associated
with substance use norms relative to centrality. The significance of network cohesion within
sociometric network implies the presence of close and dense ties within members of a sub-
network within a larger bounded network (Seidman, 1983). The presence of these tightly knit

139

sub-groups suggests that instead of a leader centric technique, network interventions should be
designed to capitalize on the reciprocity and social influence naturally occurring in these sub-
networks (Neaigus, 1998). Valente (2012) suggests that “segmentation” might be the most
effective approach in such a situation. Segmentation involves identifying groups of people that
can be persuaded to change at the same time. When there is a lot of interdependence among
people, they consider any change only when the whole group changes. Therefore, group level
network interventions work best for these individuals.
Another issue that needs to be kept in mind while designing network interventions is the
kind of behavior that the network intervention is trying to change and the prevalence of that
behavior. In designing interventions, most HIV prevention programs for youth do not
differentiate between drugs the youth are currently using. The findings of this study suggest that
methamphetamine use is more prevalent among this group of youth compared to injection drug
use. Because of the differences in risk behaviors between users of different drugs, prevention
programs should assess recent drug use among youth clients in order to more precisely tailor
their interventions (Gleghorn et al., 1998).
These findings also underscore the importance of assessing both egocentric and
sociometric networks. In particular, the use of egocentric and sociometric data in this study
allowed for an examination of both structural aspects (such as network position), as well as the
diversity of social ties (network composition). The influence of each of these network
dimensions could not have been captured if only one kind of data source was used. Additionally,
if one had relied alone on sociometric data, the protective nature of family ties would not have
been detected. This study found that connections with family online were associated with less

140

risky norms. While peers become increasingly more important during adolescence, family
processes are able to buffer risks and act as protective factors for adolescents who are in
environments where their peers are engaging in risky behaviors. Therefore, it might also be
prudent for future studies to examine the feasibility of designing network interventions that can
incorporate not only peers, but promote connections with non-street relationships.
In addition, as this study shows, the options for network interventions have been radically
enhanced by communication and information technologies. This study demonstrates that the
internet and social media can serve as powerful bridging resources for homeless youth, and need
to be leveraged in future policy, interventions, and day-to-day social work practice. The ubiquity
of internet use among homeless youth also opens up the possibility of delivering interventions
online. In previous studies, prevention and intervention programs that are delivered online have
proven to be both viable and effective (Moskowitz, Melton, & Owczarzak, 2009; Ybarra & Bull,
2007). Internet-based interventions are especially useful because the messages can be tailored to
the risk profiles of the participants (Ybarra, Kiwanuka, Emenyonu, & Bangsberg, 2006), youth
can proceed through these interventions at their own pace, and access them at their own
convenience, all of which can be particularly significant for an unstable population like homeless
youth (Alemagno & Kenne, 2012).
On a policy level, given the critical role that internet and social media play in homeless
youth’s lives, it becomes important that internet accessibility is made more convenient for these
youth. Policy-makers should contemplate specialized funding streams for youth-serving agencies
to set up computer labs in order to mitigate this existing “digital divide’. Furthermore, given that
20% of youth indicated using email and social media to communicate with their caseworkers, the

141

Internet would appear to offer an alternative means for social workers to reach these transient
clients.
It is also important to remember that network interventions cannot be effective unless it is
guided by rigorous scientific theory. While diffusion and social influence provide the underlying
mechanism through which change can be perpetuated (Valente, 2012) they do not provide insight
into the ways, in which the different dimensions of social influence can be utilized in order for
the intervention to be most effective. For example, with respect to injunctive norms, this study
found that very few youth perceived that their friends would actively encourage them to use
drugs. A great deal of preventive work, however, has been based on this theory and has sought
to provide the young person with the social and interpersonal skills that will enable them to resist
such proactive pressure (Coggans & McKellar, 1994; Denscombe, 2001). More studies need to
be conducted to understand whether the absence of such peer pressure is a matter of perception,
or pressure works in other subtle ways not captured by survey data (Denscombe, 2001). It is
possible that observational studies or qualitative studies could better capture the nuances of how
such influence is exerted.
Finally, attention needs to be paid to the groups that seem to be at most-risk of negative
outcomes. Therefore, at the individual level, it is important to tailor interventions based on
youth’s unique characteristics and histories. LGBT youth are particularly vulnerable during
periods of homelessness. However, even with such documented vulnerabilities, inclusive and
culturally responsive programs targeted at LGBTQ homeless youth continue to be rare.
Therefore, it is urgent that prevention programs continue to focus on these youth, and providers
are educated and trained to provide them with non-discriminatory support services.

142

References
Ajzen, I. (1991). The theory of planned behavior. Organizational Behavior and Human Decision
Processes, 50(2), 179-211.
Ajzen, I., & Fishbein, M. (1980). Understanding attitudes and predicting social behaviour.
Alemagno, S. A., & Kenne, D. R. (2012). Personal Computer, Mobile Phone and Internet
Technologies to Increase HIV Testing and Prevention.
Alexander, C., Piazza, M., Mekos, D., & Valente, T. (2001). Peers, schools, and adolescent
cigarette smoking. Journal of Adolescent Health, 29(1), 22-30.
Arnold, E. M., & Rotheram-Borus, M. J. (2009). Comparisons of prevention programs for
homeless youth. Prevention Science, 10(1), 76-86.
Asch, S. E. (1951). ‘Effects of group pressure upon the modification and distortion of judgments.
Groups, Leadership, and Men.S, , 222-236.
Bandura, A. (1977). Self-efficacy: Toward a unifying theory of behavioral change.
Psychological Review, 84(2), 191.
Bandura, A. (1997). Self-efficacy: The exercise of control Worth Publishers.
Barman-Adhikari, A., & Rice, E. (2011). Sexual health information seeking online among
runaway and homeless youth. Journal of the Society for Social Work and Research, 2(2),
88.

143

Barman-Adhikari, A., & Rice, E. (2011). Sexual health information seeking online among
runaway and homeless youth. Journal of the Society for Social Work and Research, 2(2),
88.
Barrington, C. (2008). Social networks, norms, and HIV-related behaviors among the male
partners of female sex workers in la romana, dominican republic. ProQuest.
Barrington, C., Latkin, C., Sweat, M. D., Moreno, L., Ellen, J., & Kerrigan, D. (2009). Talking
the talk, walking the walk: Social network norms, communication patterns, and condom use
among the male partners of female sex workers in la romana, dominican republic. Social
Science & Medicine, 68(11), 2037-2044.
Bauman, K. E., & Ennett, S. T. (1996). On the importance of peer influence for adolescent drug
use: Commonly neglected considerations. Addiction, 91(2), 185-198.
Beech, B. M., Myers, L., Beech, D. J., & Kernick, N. S. (2003). Human immunodeficiency
syndrome and hepatitis B and C infections among homeless adolescents. Paper presented at
the Seminars in Pediatric Infectious Diseases, , 14(1) 12-19.
Berkman, L. F., Glass, T., Brissette, I., & Seeman, T. E. (2000). From social integration to
health: Durkheim in the new millennium. Social Science & Medicine, 51(6), 843-857.
Berkowitz, A. D. (2004). The social norms approach: Theory, research and annotated
bibliography. Higher Education Center for Alcohol and Other Drug Abuse and Violence
Prevention.US Department of Education,

144

Blake, J., & Davis, K. (1964). Norms, values, and sanctions.
Bohnert, A. S., Bradshaw, C. P., & Latkin, C. A. (2009). A social network perspective on heroin
and cocaine use among adults: Evidence of bidirectional influences. Addiction, 104(7),
1210-1218.
Booth, R. E., & Zhang, Y. (1996). Severe aggression and related conduct problems among
runaway and homeless adolescents. Psychiatric Services, 47(1), 75.
Borgatti, S. P., & Halgin, D. S. (2011). Network theorizing. Organization Science,
Borgatti, S. P., & Everett, M. G. (2000). Models of core/periphery structures. Social Networks,
21(4), 375-395.
Borgatti, S. P., Everett, M. G., & Linton, C. (2002). Freeman. 2002. Ucinet for Windows:
Software for Social Network Analysis.Boston, MA: Analytic Technologies,
Borsari, B., & Carey, K. B. (2003). Descriptive and injunctive norms in college drinking: A
meta-analytic integration. Journal of Studies on Alcohol, 64(3), 331.
Bourdieu, P. (1986). 15 the forms of capital. Readings in Economic Sociology, , 280.
Brown, J. D. (1997). Skewness and kurtosis. Shiken: JALT Testing and Evaluation SIG
Newsletter, 1(1), 16-18.

145

Buchacz, K., McFarland, W., Kellogg, T. A., Loeb, L., Holmberg, S. D., Dilley, J., & Klausner,
J. D. (2005). Amphetamine use is associated with increased HIV incidence among men who
have sex with men in san francisco. Aids, 19(13), 1423-1424.
Bungay, V., Malchy, L., Buxton, J. A., Johnson, J., MacPherson, D., & Rosenfeld, T. (2006).
Life with jib: A snapshot of street youth's use of crystal methamphetamine. Addiction
Research & Theory, 14(3), 235-251.
Burt, R. S. (1992). Structural holes Harvard Univ. Press.
Buunk, B. P., Bakker, A. B., Siero, F. W., Eijinden, R. J., & Yzer, M. C. (1998). Predictors of
AIDS-preventive behavioral intentions among adult heterosexuals at risk for HIV-infection:
Extending current models and measures. AIDS Education and Prevention,
Campbell, D. T. (1975). On the conflicts between biological and social evolution and between
psychology and moral tradition. American Psychologist, 30(12), 1103.
Carey, K. B., Borsari, B., Carey, M. P., & Maisto, S. A. (2006). Patterns and importance of self-
other differences in college drinking norms. Psychology of Addictive Behaviors: Journal of
the Society of Psychologists in Addictive Behaviors, 20(4), 385.
Carpentier, N., & White, D. (2002). Cohesion of the primary social network and sustained
service use before the first psychiatric hospitalization. The Journal of Behavioral Health
Services and Research, 29(4), 404-418.

146

Carpentier, N., & White, D. (2002). Cohesion of the primary social network and sustained
service use before the first psychiatric hospitalization. Journal of Behavioral Health
Services and Research, 29(4), 404-418.
Celentano, D. D., Valleroy, L. A., Sifakis, F., MacKellar, D. A., Hylton, J., Thiede, H., . . .
LaLota, M. (2006). Associations between substance use and sexual risk among very young
men who have sex with men. Sexually Transmitted Diseases, 33(4), 265-271.
Cho, H. (2006). Influences of norm proximity and norm types on binge and non-binge drinkers:
Examining the under-examined aspects of social norms interventions on college campuses.
Journal of Substance use, 11(6), 417-429.
Christakis, N. A., & Fowler, J. H. (2007). The spread of obesity in a large social network over 32
years. New England Journal of Medicine, 357(4), 370-379.
Christensen, P. N., Rothgerber, H., Wood, W., & Matz, D. C. (2004). Social norms and identity
relevance: A motivational approach to normative behavior. Personality and Social
Psychology Bulletin, 30(10), 1295-1309.
Cialdini, R. B., Reno, R. R., & Kallgren, C. A. (1990). A focus theory of normative conduct:
Recycling the concept of norms to reduce littering in public places. Journal of Personality
and Social Psychology, 58(6), 1015-1026.
Cialdini, R. B., & Trost, M. R. (1998). Social influence: Social norms, conformity and
compliance.

147

Clatts, M. C., Davis, W. R., Sotheran, J. L., & Atillasoy, A. (1998). Correlates and distribution of
HIV risk behaviors among homeless youths in New York city: Implications for prevention
and policy. Child Welfare: Journal of Policy, Practice, and Program,
Clements, K., Gleghorn, A., Garcia, D., Katz, M., & Marx, R. (1997). A risk profile of street
youth in northern California: Implications for gender-specific human immunodeficiency
virus prevention. Journal of Adolescent Health, 20(5), 343-353.
Cochran, B. N., Stewart, A. J., Ginzler, J. A., & Cauce, A. M. (2002). Challenges faced by
homeless sexual minorities: Comparison of gay, lesbian, bisexual, and transgender
homeless adolescents with their heterosexual counterparts. American Journal of Public
Health, 92(5)
Coggans, N., & McKellar, S. (1994). Drug use amongst peers: Peer pressure or peer preference?
Drugs: Education, Prevention, and Policy, 1(1), 15-26.
Coleman, J. S. (1990). The emergence of norms. Social Institutions: Their Emergence,
Maintenance, and Effects, , 35-39.
Cox, J. M. (2010). A social norms approach to college alcohol use: Drinking in a low-use
environment. Proquest Dissertation International.
Das-Douglas, M., Colfax, G., Moss, A. R., Bangsberg, D. R., & Hahn, J. A. (2008). Tripling of
methamphetamine/amphetamine use among homeless and marginally housed persons,
1996–2003. Journal of Urban Health, 85(2), 239-249.

148

Davey-Rothwell, M. A., & Latkin, C. A. (2007). HIV-related communication and perceived
norms: An analysis of the connection among injection drug users. AIDS Education &
Prevention, 19(4), 298-309.
De, P., Cox, J., Boivin, J., Platt, R. W., & Jolly, A. M. (2007). The importance of social networks
in their association to drug equipment sharing among injection drug users: A review.
Addiction, 102(11), 1730-1739.
Denscombe, M. (2001). Peer group pressure, young people and smoking: New developments and
policy implications. Drugs: Education, Prevention, and Policy, 8(1), 7-32.
Deutsch, M., & Gerard, H. B. (1955). A study of normative and informational social influences
upon individual judgment. Journal of Abnormal and Social Psychology, 51(3), 629-636.
DiClemente, R. J., & Wingood, G. M. (2000). Expanding the scope of HIV prevention for
adolescents: Beyond individual-level interventions. J Adolesc Health, 26(6), 377.
DiClemente, R. J., Wingood, G. M., Harrington, K. F., Lang, D. L., Davies, S. L., Hook III, E.
W., . . . Gordon, A. B. (2004). Efficacy of an HIV prevention intervention for African
American adolescent girls. JAMA: The Journal of the American Medical Association,
292(2), 171-179.
Dinwiddie, S. H., Reich, T., & Cloninger, C. R. (1992). Patterns of lifetime drug use among
intravenous drug users. Journal of Substance Abuse, 4(1), 1-11.

149

Doherty, I. A., Padian, N. S., Marlow, C., & Aral, S. O. (2005). Determinants and consequences
of sexual networks as they affect the spread of sexually transmitted infections. Journal of
Infectious Diseases, 191(Supplement 1), S42-S54.
Eaton, D. K., Kann, L., Kinchen, S., Shanklin, S., Flint, K. H., Hawkins, J., . . . Chyen, D.
(2012). Youth risk behavior surveillance-united states, 2011. MMWR Surveill Summ, 61(4),
1-162.
Ellison, N. B. (2007). Social network sites: Definition, history, and scholarship. Journal of
Computer ‐Mediated Communication, 13(1), 210-230.
Ellison, N. B., Steinfield, C., & Lampe, C. (2007). The benefits of facebook “friends:” social
capital and college students’ use of online social network sites. Journal of
Computer ‐Mediated Communication, 12(4), 1143-1168.
Empelen, P. v., Kok, G., Jansen, M., & Hoebe, C. (2001). The additional value of anticipated
regret and psychopathology in explaining intended condom use among drug users. AIDS
Care, 13(3), 309-318.
Ennett, S. T., Bailey, S. L., & Federman, E. B. (1999). Social network characteristics associated
with risky behaviors among runaway and homeless youth. Journal of Health and Social
Behavior, 40(1), 63-78.
Ennett, S. T., Bailey, S. L., & Federman, E. B. (1999). Social network characteristics associated
with risky behaviors among runaway and homeless youth. Journal of Health and Social
Behavior, , 63-78.

150

Ennett, S. T., & Bauman, K. E. (1993). Peer group structure and adolescent cigarette smoking: A
social network analysis. Journal of Health and Social Behavior, , 226-236.
Ennett, S. T., & Bauman, K. E. (1994). The contribution of influence and selection to adolescent
peer group homogeneity: The case of adolescent cigarette smoking. Journal of Personality
and Social Psychology, 67(4), 653.
Ennett, S. T., Bauman, K. E., Hussong, A., Faris, R., Foshee, V. A., Cai, L., & DuRant, R. H.
(2006). The peer context of adolescent substance use: Findings from social network
analysis. Journal of Research on Adolescence, 16(2), 159-186.
Ennett, S. T., Foshee, V. A., Bauman, K. E., Hussong, A., Cai, L., Reyes, H. L. M., . . . DuRant,
R. (2008). The social ecology of adolescent alcohol misuse. Child Development, 79(6),
1777-1791.
Epstein, J. L., & Karweit, N. (1983). Friends in school. New York: Academic, 56.
Festinger, L. (1954). A theory of social comparison processes. Human Relations, 7(2), 117-140.
Fishbein, M., & Ajzen, I. (1975). Belief, attitude, intention and behavior: An introduction to
theory and research. Reading, Addison-Wesley.
Fishbein, M., Middlestadt, S. E., & Hitchcock, P. J. (1994). Using information to change
sexually transmitted disease-related behaviors. Preventing AIDS: Theories and Methods of
Behavioral Interventions, , 61-78.

151

Fisher, J. D., & Fisher, W. A. (1992). Changing AIDS-risk behavior. Psychological Bulletin,
111(3), 455.
Fisher, J. D., & Misovich, S. (1990). Social influence and AIDS-preventive behavior.
Flom, P. L., Friedman, S. R., Kottiri, B. J., Neaigus, A., Curtis, R., Des Jarlais, D. C., . . .
Zenilman, J. M. (2001). Stigmatized drug use, sexual partner concurrency, and other sex
risk network and behavior characteristics of 18-to 24-year-old youth in a high-risk
neighborhood. Sexually Transmitted Diseases, 28(10), 598-607.
Freeman, L. C., & Webster, C. M. (1994). Interpersonal proximity in social and cognitive space.
Social Cognition, 12(3), 223-247.
Frey, B. S., & Meier, S. (2004). Social comparisons and pro-social behavior: Testing"
conditional cooperation" in a field experiment. The American Economic Review, 94(5),
1717-1722.
Friedkin, N. E. (2001). Norm formation in social influence networks. Social Networks, 23(3),
167-189.
Friedman, S. R., Neaigus, A., Jose, B., Curtis, R., Goldstein, M., Ildefonso, G., . . . Des Jarlais,
D. C. (1997). Sociometric risk networks and risk for HIV infection. American Journal of
Public Health, 87(8), 1289-1296.

152

Frosch, D., Shoptaw, S., Huber, A., Rawson, R. A., & Ling, W. (1996). Sexual HIV risk among
gay and bisexual male methamphetamine abusers. Journal of Substance Abuse Treatment,
13(6), 483-486.
Ghanem, K. G., Hutton, H. E., Zenilman, J. M., Zimba, R., & Erbelding, E. J. (2005). Audio
computer assisted self interview and face to face interview modes in assessing response
bias among STD clinic patients. Sexually Transmitted Infections, 81(5), 421-425.
Go, M., Green, H. D., Kennedy, D. P., Pollard, M., & Tucker, J. S. (2010). Peer influence and
selection effects on adolescent smoking. Drug and Alcohol Dependence, 109(1), 239-242.
Granovetter, M. (1985). Economic action and social structure: The problem of embeddedness.
American Journal of Sociology, , 481-510.
Green, H. D., Haye, K., Tucker, J. S., & Golinelli, D. (2013). Shared risk: Who engages in
substance use with American homeless youth? Addiction,
Greene, J. M., Ennett, S. T., & Ringwalt, C. L. (1999). Prevalence and correlates of survival sex
among runaway and homeless youth. American Journal of Public Health, 89(9), 1406.
Greene, J. M., & Ringwalt, C. L. (1996). Youth and familial substance use's association with
suicide attempts among runaway and homeless youth. Substance use & Misuse, 31(8),
1041-1058.
Greenfield, P., & Yan, Z. (2006). Children, adolescents, and the internet: A new field of inquiry
in developmental psychology. Developmental Psychology, 42(3), 391.

153

Greenhow, C., Robelia, B., & Hughes, J. E. (2009). Learning, teaching, and scholarship in a
digital age web 2.0 and classroom research: What path should we take now? Educational
Researcher, 38(4), 246-259.
Guadagno, R. E., Muscanell, N. L., & Pollio, D. E. (2012). The homeless use facebook?!
similarities of social network use between college students and homeless young adults.
Computers in Human Behavior, 29, 86-89.
Gudelunas, D. (2012). There’s an app for that: The uses and gratifications of online social
networks for gay men. Sexuality & Culture, 16(4), 347-365.
Hagan, J., & McCarthy, B. (1998). Mean streets: Youth crime and homelessness Cambridge
Univ Pr.
Halcón, L. L., & Lifson, A. R. (2004). Prevalence and predictors of sexual risks among homeless
youth. Journal of Youth and Adolescence, 33(1), 71-80.
Haley, N., Roy, E., Leclerc, P., Boudreau, J., & Boivin, J. (2004). HIV risk profile of male street
youth involved in survival sex. Sexually Transmitted Infections, 80(6), 526-530.
Halkitis, P. N., Palamar, J. J., & Mukherjee, P. P. (2007). Poly-club-drug use among gay &
bisexual men: A longitudinal analysis. Drug and Alcohol Dependence, 89(2-3), 153.
Hall, H. I., Song, R., Rhodes, P., Prejean, J., An, Q., Lee, L. M., . . . McKenna, M. T. (2008).
Estimation of HIV incidence in the United States. JAMA: The Journal of the American
Medical Association, 300(5), 520-529.

154

Hall, J. A., & Valente, T. W. (2007). Adolescent smoking networks: The effects of influence and
selection on future smoking. Addictive Behaviors, 32(12), 3054-3059.
Hart, L., & Hunt, N. (1997). Choosers not losers? drug offers, peer influences and drug decisions
amongst 11–16 year olds in West Kent. Maidstone, Kent: Invicta Community Care NHS
Trust.
Hechter, M., & Opp, K. (2001). Social norms Russell Sage Foundation Publications.
Hechter, M., & Opp, K. (2001). What have we learned about the emergence of social norms.
Social Norms, , 394-415.
Hogg, M. A., Abrams, D., Otten, S., & Hinkle, S. (2004). The social identity perspective
intergroup relations, self-conception, and small groups. Small Group Research, 35(3), 246-
276.
Hosmer, D. W., & Lemeshow, S. (2000). Applied logistic regression Wiley-Interscience.
Huba, G., Melchior, L. A., Smereck, G. A., Brown, V. B., Jean-Louis, E., German, V. F., . . .
Hughes, C. (2001). Perceived barriers to receiving HIV services in groups of traditionally
underserved individuals: Empirical models. Home Health Care Services Quarterly, 19(1-2),
53-75.
Hyman, H. H. (1980). The psychology of status Arno Press.

155

Iannotti, R. J., & Bush, P. J. (1992). Perceived vs. actual friends' use of alcohol, cigarettes,
marijuana, and cocaine: Which has the most influence? Journal of Youth and Adolescence,
21(3), 375-389.
Institute Inc, S. (2008). SAS/STAT 9.2 User’s guide.
Irwin, J., LaGory, M., Ritchey, F., & Fitzpatrick, K. (2008). Social assets and mental distress
among the homeless: Exploring the roles of social support and other forms of social capital
on depression. Social Science & Medicine, 67(12), 1935-1943.
Johnson, K. D., Whitbeck, L. B., & Hoyt, D. R. (2005). Predictors of social network composition
among homeless and runaway adolescents. Journal of Adolescence, 28(2), 231-248.
Jones, S., & Fox, S. (2009). Generations online in 2009 Pew Internet & American Life Project
Washington, DC.
Kadushin, C. (1966). The friends and supporters of psychotherapy: On social circles in urban
life. American Sociological Review, , 786-802.
Karabanow, J., & Clement, P. (2004). Interventions with street youth: A commentary on the
practice-based research literature. Brief Treatment and Crisis Intervention, 4(1), 93.
Karabanow, J., & Naylor, T. D. (2010). “Being hooked up”: Exploring the experiences of street
youth and information technologies. Digital Diversity, , 161-178.
Kelly, J. A., St Lawrence, J. S., Stevenson, L. Y., Hauth, A. C., Kalichman, S. C., Diaz, Y. E., . .
. Morgan, M. G. (1992). Community AIDS/HIV risk reduction: The effects of

156

endorsements by popular people in three cities. American Journal of Public Health, 82(11),
1483-1489.
Kerr, T., Stoltz, J., Marshall, B. D., Lai, C., Strathdee, S. A., & Wood, E. (2009). Childhood
trauma and injection drug use among high-risk youth. Journal of Adolescent Health, 45(3),
300-302.
Kidder, D. P., Wolitski, R. J., Campsmith, M. L., & Nakamura, G. V. (2007). Health status,
health care use, medication use, and medication adherence among homeless and housed
people living with HIV/AIDS. American Journal of Public Health, 97(12), 2238-2245.
Kincaid, D. L. (2004). From innovation to social norm: Bounded normative influence. Journal of
Health Communication, 9(S1), 37-57.
Kipke, M. D., Unger, J. B., O'Connor, S., Palmer, R. F., & Lafrance, S. R. (1997). Street youth,
their peer group affiliation and differences according to residential status, subsistence
patterns, and use of services. Adolescence, 32(127)
Kipke, M. D., Montgomery, S., & MacKenzie, R. G. (1993). Substance use among youth seen at
a community-based health clinic. Journal of Adolescent Health, 14(4), 289-294.
Kipke, M. D., O'Connor, S., Palmer, R., & MacKenzie, R. G. (1995). Street youth in los angeles:
Profile of a group at high risk for human immunodeficiency virus infection. Archives of
Pediatrics & Adolescent Medicine, 149(5), 513.

157

Knowlton, A., Hua, W., & Latkin, C. (2005). Social support networks and medical service use
among HIV-positive injection drug users: Implications to intervention. AIDS Care, 17 (4),
479-492.
Kottiri, B. J., Friedman, S. R., Neaigus, A., Curtis, R., & Des Jarlais, D. C. (2002). Risk
networks and racial/ethnic differences in the prevalence of HIV infection among injection
drug users. Journal of Acquired Immune Deficiency Syndromes (1999), 30(1), 95.
Kral, A. H., Molnar, B. E., Booth, R. E., & Watters, J. K. (1997). Prevalence of sexual risk
behaviour and substance use among runaway and homeless adolescents in San Francisco,
Denver and New York city. International Journal of STD & AIDS, 8(2), 109-117.
L. Flom, Samuel R. Friedman, Benny Jose, Richard Curtis, Peter. (2001). Peer norms regarding
drug use and drug selling among household youth in a low-income'drug supermarket'urban
neighborhood. Drugs: Education, Prevention, and Policy, 8(3), 219-232.
LaBrie, J. W., & Cail, J. (2011). Parental interaction with college students: The moderating
effect of parental contact on the influence of perceived peer norms on drinking during the
transition to college. Journal of College Student Development, 52(5), 610-621.
Lakon, C. M. (2004). Relationships between Drug, Hang Out, and Sex Partner Network
Characteristics and Needle use among High Risk Youth and Young Adults, Proquest
Dissertation International.

158

Lankenau, S. E., Sanders, B., Bloom, J. J., Hathazi, D., Alarcon, E., Tortu, S., & Clatts, M. C.
(2008). Migration patterns and substance use among young homeless travelers. Geography
and drug addiction (pp. 65-83) Springer.
Lansford, J. E., Killeya-Jones, L. A., Miller, S., & Costanzo, P. R. (2009). Early adolescents’
social standing in peer groups: Behavioral correlates of stability and change. Journal of
Youth and Adolescence, 38(8), 1084-1095.
Lapinski, M. K., & Rimal, R. N. (2005). An explication of social norms. Communication Theory,
15(2), 127-147.
Larimer, M. E., Kaysen, D. L., Lee, C. M., Kilmer, J. R., Lewis, M. A., Dillworth, T., . . .
Neighbors, C. (2009). Evaluating level of specificity of normative referents in relation to
personal drinking behavior. Journal of Studies on Alcohol and Drugs. Supplement,
(Supplement no. 16), 115.
Latane, B. (1981). The psychology of social impact. American Psychologist, 36(4), 343-356.
Latkin, C. A., Forman, V., Knowlton, A., & Sherman, S. (2003). Norms, social networks, and
HIV-related risk behaviors among urban disadvantaged drug users. Social Science &
Medicine, 56(3), 465-476.
Latkin, C. A., Mandell, W., Vlahov, D., Oziemkowska, M., & Celentano, D. D. (1996). The
long-term outcome of a personal network-oriented HIV prevention intervention for
injection drug users: The SAFE study. American Journal of Community Psychology, 24(3),
341-364.

159

Latkin, C., Donnell, D., Celentano, D. D., Aramrattna, A., Liu, T., Vongchak, T., . . . Metzger,
D. (2009). Relationships between social norms, social network characteristics, and HIV risk
behaviors in Thailand and the US. Health Psychology: Official Journal of the Division of
Health Psychology, American Psychological Association, 28(3), 323.
Lenhart, A., Rainie, L., & Lewis, O. (2001). Teenage life online: The rise of the instant-message
generation and the Internet’s impact on friendships and family relationships. Washington,
DC: Pew Internet & American life project, 2001. NEXT: The Future just Happened,
Lin, N. (1999). Building a network theory of social capital. Connections, 22(1), 28-51.
Lovell, A. M. (2002). Risking risk: The influence of types of capital and social networks on the
injection practices of drug users. Social Science & Medicine, 55(5), 803-821.
Luke, D. A., & Harris, J. K. (2007). Network analysis in public health: History, methods, and
applications. Annu.Rev.Public Health, 28, 69-93.
Macalino, G. E., Celentano, D. D., Latkin, C., Strathdee, S. A., & Vlahov, D. (2002). Risk
behaviors by audio computer-assisted self-interviews among HIV-seropositive and HIV-
seronegative injection drug users. AIDS Education and Prevention, 14(5), 367-378.
MacGillivray, A., & Walker, P. (2000). Local social capital: Making it work on the ground.
Social Capital: Critical Perspectives, , 197-211.
MacNeil, J., & Pauly, B. (2011). Needle exchange as a safe haven in an unsafe world. Drug and
Alcohol Review, 30(1), 26-32.

160

Major, A. M. (2000). Norm origin and development in cyberspace: Models of cyber norm
evolution. Wash.ULQ, 78, 59.
Malow, R. M., Kershaw, T., Sipsma, H., Rosenberg, R., & Dévieux, J. G. (2007). HIV
preventive interventions for adolescents: A look back and ahead. Current HIV/AIDS
Reports, 4(4), 173-180.
Mansfield, E. R., & Helms, B. P. (1982). Detecting multicollinearity. The American Statistician,
36(3a), 158-160.
Marsden, P. V., & Friedkin, N. E. (1994). Network studies of social influence. Sage Focus
Editions, 171, 3-3.
Marshall, B. D., Kerr, T., Qi, J., Montaner, J. S., & Wood, E. (2010). Public injecting and HIV
risk behaviour among street-involved youth. Drug and Alcohol Dependence, 110(3), 254-
258.
Marshall, B. D., Kerr, T., Shoveller, J. A., Patterson, T. L., Buxton, J. A., & Wood, E. (2009).
Homelessness and unstable housing associated with an increased risk of HIV and STI
transmission among street-involved youth. Health & Place, 15(3), 753.
Marshall, B. D., Milloy, M., Wood, E., Montaner, J. S., & Kerr, T. (2011). Reduction in
overdose mortality after the opening of North America’s first medically supervised safer
injecting facility: A retrospective population-based study. The Lancet, 377(9775), 1429-
1437.

161

Martinez, T. E., Gleghorn, A., Marx, R., Clements, K., Boman, M., & Katz, M. H. (1998).
Psychosocial histories, social environment, and the HIV risk behaviors of injection and
noninjection drug using homeless youths. Journal of Psychoactive Drugs, 30(1), 1-10.
Martino, S. C., Tucker, J. S., Ryan, G., Wenzel, S. L., Golinelli, D., & Munjas, B. (2011).
Increased substance use and risky sexual behavior among migratory homeless youth:
Exploring the role of social network composition. Journal of Youth and Adolescence,
40(12), 1634-1648.
Mason, W. A., Conrey, F. R., & Smith, E. R. (2007). Situating social influence processes:
Dynamic, multidirectional flows of influence within social networks. Personality and
Social Psychology Review, 11(3), 279-300.
McAdams, R. H. (1997). The origin, development, and regulation of norms. Michigan Law
Review, 96(2), 338-433.
McIntosh, J., MacDonald, F., & McKeganey, N. (2003). The initial use of drugs in a sample of
pre-teenage schoolchildren: The role of choice, pressure and influence. Drugs: Education,
Prevention, and Policy, 10(2), 147-158.
McMillan, B., & Conner, M. (2003). Applying an extended version of the theory of planned
behavior to illicit drug use among Students1. Journal of Applied Social Psychology, 33(8),
1662-1683.

162

McMorris, B. J., Tyler, K. A., Whitbeck, L. B., & Hoyt, D. R. (2002). Familial and on-the-street
risk factors associated with alcohol use among homeless and runaway adolescents. Journal
of Studies on Alcohol, 63(1), 34-43.
McMorris, B. J., Tyler, K. A., Whitbeck, L. B., & Hoyt, D. R. (2002). Familial and “on-the-
street” risk factors associated with alcohol use among homeless and runaway adolescents.
Journal of Studies on Alcohol and Drugs, 63(1), 34.
McPherson, M., Smith-Lovin, L., & Cook, J. M. (2001). Birds of a feather: Homophily in social
networks. Annual Review of Sociology, , 415-444.
Mercken, L., Candel, M., Willems, P., & De Vries, H. (2007). Disentangling social selection and
social influence effects on adolescent smoking: The importance of reciprocity in
friendships. Addiction, 102(9), 1483-1492.
Milburn, N. G., Rosenthal, D., Rotheram-Borus, M. J., Mallett, S., Batterham, P., Rice, E., &
Solorio, R. (2007). Newly homeless youth typically return home. Journal of Adolescent
Health, 40(6), 574-576.
Milburn, N. G., Rice, E., Rotheram-Borus, M. J., Mallett, S., Rosenthal, D., Batterham, P., . . .
Duan, N. (2009). Adolescents exiting homelessness over two years: The risk amplification
and abatement model. Journal of Research on Adolescence, 19(4), 762-785.
doi:10.1111/j.1532-7795.2009.00610.x

163

Milburn, N. G., Rosenthal, D., Rotheram-Borus, M. J., Mallett, S., Batterham, P., Rice, E., &
Solorio, R. (2007). Newly homeless youth typically return home. Journal of Adolescent
Health, 40(6), 574-576. doi:10.1016/j.jadohealth.2006.12.017.
Mitchell, C. U., & LaGory, M. (2002). Social capital and mental distress in an impoverished
community. City & Community, 1(2), 199-222.
Mizruchi, M. S. (1993). Cohesion, equivalence, and similarity of behavior: A theoretical and
empirical assessment. Social Networks, 15(3), 275-307.
Morrison-Beedy, D., Carey, M. P., & Tu, X. (2006). Accuracy of audio computer-assisted self-
interviewing (ACASI) and self-administered questionnaires for the assessment of sexual
behavior. AIDS and Behavior, 10(5), 541-552.
Moskowitz, D. A., Melton, D., & Owczarzak, J. (2009). PowerON: The use of instant message
counseling and the internet to facilitate HIV/STD education and prevention. Patient
Education and Counseling, 77(1), 20-26.
Neaigus, A. (1998). The network approach and interventions to prevent HIV among injection
drug users. Public Health Reports, 113(Suppl 1), 140.
Neighbors, C., LaBrie, J. W., Hummer, J. F., Lewis, M. A., Lee, C. M., Desai, S., . . . Larimer,
M. E. (2010). Group identification as a moderator of the relationship between perceived
social norms and alcohol consumption. Psychology of Addictive Behaviors: Journal of the
Society of Psychologists in Addictive Behaviors, 24(3), 522.

164

Neighbors, C., Lee, C. M., Lewis, M. A., Fossos, N., & Walter, T. (2009). Internet-based
personalized feedback to reduce 21st-birthday drinking: A randomized controlled trial of an
event-specific prevention intervention. Journal of Consulting and Clinical Psychology,
77(1), 51.
Noell, J., Rohde, P., Ochs, L., Yovanoff, P., Alter, M. J., Schmid, S., . . . Black, C. (2001).
Incidence and prevalence of Chlamydia, herpes, and viral hepatitis in a homeless adolescent
population. Sexually Transmitted Diseases, 28(1), 4-10.
Norman, P., Clark, T., & Walker, G. (2005). The theory of planned behavior, descriptive norms,
and the moderating role of group identification. Journal of Applied Social Psychology,
35(5), 1008-1029.
Nyamathi, A., Hudson, A., Greengold, B., & Leake, B. (2012). Characteristics of homeless youth
who use cocaine and methamphetamine. The American Journal on Addictions, 21(3), 243-
249.
Okun, M. A., Ruehlman, L., Karoly, P., Lutz, R., Fairholme, C., & Schaub, R. (2003). Social
support and social norms: Do both contribute to predicting leisure-time exercise? American
Journal of Health Behavior, 27(5), 493-507.
Paluck, E. L., & Shepherd, H. (2012). The salience of social referents: A field experiment on
collective norms and harassment behavior in a school social network.
Papacharissi, Z., & Rubin, A. M. (2000). Predictors of internet use. Journal of Broadcasting &
Electronic Media, 44(2), 175-196.

165

Parker, W. (2004). Rethinking conceptual approaches to behaviour change: The importance of
context. Centre for AIDS Development, Research and Evaluation (CADRE),
Parriott, A. M., & Auerswald, C. L. (2009). Incidence and predictors of onset of injection drug
use in a san francisco cohort of homeless youth. Substance use & Misuse, 44(13), 1958-
1970.
Pascoe, C. (2012). Studying young people's new media use: Methodological shifts and
educational innovations. Theory into Practice, 51(2), 76-82.
Pearson, M., & West, P. (2003). Drifting smoke rings. Connections, 25(2), 59-76.
Pénard, T., & Poussing, N. (2010). Internet use and social capital: The strength of virtual ties.
Journal of Economic Issues, 44(3), 569-595.
Pepitone, A. (1976). Toward a normative and comparative biocultural social psychology.
Journal of Personality and Social Psychology, 34(4), 641.
Pescosolido, B. A. (2006). Of pride and prejudice: The role of sociology and social networks in
integrating the health sciences. Journal of Health and Social Behavior, 47(3), 189-208.
Pescosolido, B. A., Gardner, C. B., & Lubell, K. M. (1998). How people get into mental health
services: Stories of choice, coercion and “muddling through” from “first-timers”. Social
Science & Medicine, 46(2), 275-286.
Pfeifer, R. W., & Oliver, J. (1997). A study of HIV seroprevalence in a group of homeless youth
in Hollywood, California. Journal of Adolescent Health, 20(5), 339-342.

166

Plankey, M. W., Ostrow, D. G., Stall, R., Cox, C., Li, X., Peck, J. A., & Jacobson, L. P. (2007).
The relationship between methamphetamine and popper use and risk of HIV
seroconversion in the multicenter AIDS cohort study. Journal of Acquired Immune
Deficiency Syndromes (1999), 45(1), 85.
Pollio, D. E., Batey, D. S., Bender, K., Ferguson, K., & Thompson, S. (2013). Technology use
among emerging adult homeless in two US cities. Social Work, 58(2), 173-175.
Popay, J. (2000). Social capital: The role of narrative and historical research. Journal of
Epidemiology and Community Health, 54(6), 401-401.
Portes, A. (1998). Social capital: Its origins and applications in modern sociology. Annual
Review of Sociology, 24(1), 1-24.
Putnam, R. D. (2001). Bowling alone: The collapse and revival of American community. Simon
& Schuster.
Quan-Haase, A., & Wellman, B. (2004). How does the internet affect social capital. Social
Capital and Information Technology, 113, 135-113.
Rawson, R. A., Gonzales, R., Obert, J. L., McCann, M. J., & Brethen, P. (2005).
Methamphetamine use among treatment-seeking adolescents in southern California:
Participant characteristics and treatment response. Journal of Substance Abuse Treatment,
29(2), 67-74.

167

Ray, N. (2006). Lesbian, Gay, Bisexual and Transgender Youth: An Epidemic of Homelessness.
New York: National Gay and Lesbian Task Force Policy Institute and the National
Coalition for the Homeless.
Reback, C. J., Larkins, S., & Shoptaw, S. (2004). Changes in the meaning of sexual risk
behaviors among gay and bisexual male methamphetamine abusers before and after drug
treatment. AIDS and Behavior, 8(1), 87-98.
Rhodes, R. E., & Courneya, K. S. (2003). Investigating multiple components of attitude,
subjective norm, and perceived control: An examination of the theory of planned behaviour
in the exercise domain. British Journal of Social Psychology, 42(1), 129-146.
Rice, E. (2010). The positive role of social networks and social networking technology in the
condom-using behaviors of homeless young people.
Rice, E. & Barman-Adhikari, A. (in press). Resources for stability among homeless youth.
JCMC.
Rice, E., Milburn, N. G., & Monro, W. (2011). Social networking technology, social network
composition, and reductions in substance use among homeless adolescents. Prevention
Science, 12(1), 80-88.
Rice, E., Milburn, N., & Rotheram-Borus, M. (2007). Pro-social and problematic social network
influences on HIV/AIDS risk behaviours among newly homeless youth in Los Angeles.
AIDS Care, 19(5), 697-704.

168

Rice, E., Stein, J. A., & Milburn, N. (2008). Countervailing social network influences on
problem behaviors among homeless youth. Journal of Adolescence, 31(5), 625-639.
Rice, E., Kurzban, S., & Ray, D. (2012). Homeless but connected: The role of heterogeneous
social network ties and social networking technology in the mental health outcomes of
street-living adolescents. Community Mental Health Journal, 48(6), 692-698.
Rice, E., Monro, W., Barman-Adhikari, A., & Young, S. D. (2010). Internet use, social
networking, and HIV/AIDS risk for homeless adolescents. Journal of Adolescent Health,
47(6), 610-613.
Rice & Rhoades (2013). How should network-based prevention for homeless youth be
implemented? Addictions.
Rimal, R. N., Lapinski, M. K., Cook, R. J., & Real, K. (2005). Moving toward a theory of
normative influences: How perceived benefits and similarity moderate the impact of
descriptive norms on behaviors. Journal of Health Communication, 10(5), 433-450.
Rimal, R. N., & Real, K. (2003). Understanding the influence of perceived norms on behaviors.
Communication Theory, 13(2), 184-203.
Roberson, J., & Nardi, B. (2010). Survival needs and social inclusion: Technology use among
the homeless. Paper presented at the Proceedings of the 2010 ACM Conference on
Computer Supported Cooperative Work, 445-448.
Robertson, M. (1996). Homeless youth on their own. Berkeley CA: Alcohol Research Group.

169

Roccas, S., & Brewer, M. B. (2002). Social identity complexity. Personality and Social
Psychology Review, 6(2), 88-106.
Roman, N. P., & Wolfe, P. (1995). Web of failure: The relationship between foster care and
homelessness. National Alliance to End Homelessness,
Rosa, C. J. D., Montgomery, S. B., Hyde, J., Iverson, E., & Kipke, M. D. (2001). HIV risk
behavior and HIV testing: A comparison of rates and associated factors among homeless
and runaway adolescents in two cities. AIDS Education and Prevention, 13(2), 131-148.
Russell, K., Dryden, D. M., Liang, Y., Friesen, C., O'Gorman, K., Durec, T., . . . Klassen, T. P.
(2008). Risk factors for methamphetamine use in youth: A systematic review. BMC
Pediatrics, 8(1), 48.
Sanders, B., Lankenau, S. E., Jackson-Bloom, J., & Hathazi, D. (2008). Multiple drug use and
polydrug use amongst homeless traveling youth. Journal of Ethnicity in Substance Abuse,
7(1), 23-40.
Schroder, K. E., Carey, M. P., & Vanable, P. A. (2003). Methodological challenges in research
on sexual risk behavior: II. accuracy of self-reports. Annals of Behavioral Medicine, 26(2),
104-123.
Schwartz, S. H. (1977). Normative influences on Altruism1. Advances in Experimental Social
Psychology, 10, 221-279.

170

Scott, C., & Hofmeyer, A. (2007). Networks and social capital: A relational approach to primary
healthcare reform. Health Research Policy and Systems, 5(1), 9.
Scott, J. (2012). Social network analysis SAGE Publications Limited.
Seidman, S. B. (1983). Network structure and minimum degree. Social Networks, 5(3), 269-287.
Sherif, M. (1966). In common predicament: Social psychology of intergroup conflict and
cooperation Houghton Mifflin Boston.
Shoptaw, S. (2006). Methamphetamine use in urban gay and bisexual populations. Top HIV Med,
14(2), 84-87.
Stablein, T. (2011). Helping friends and the homeless milieu: Social capital and the utility of
street peers. Journal of Contemporary Ethnography, 40(3), 290.
Stricof, R. L., Kennedy, J. T., Nattell, T. C., Weisfuse, I. B., & Novick, L. F. (1991). HIV
seroprevalence in a facility for runaway and homeless adolescents. American Journal of
Public Health, 81 Suppl, 50-53.
Subrahmanyam, K., Reich, S. M., Waechter, N., & Espinoza, G. (2008). Online and offline
social networks: Use of social networking sites by emerging adults. Journal of Applied
Developmental Psychology, 29(6), 420-433.
Sumartojo, E. (2000). Structural factors in HIV prevention: Concepts, examples, and
implications for research. Aids, 14, S3.

171

Tajfel, H. (81). Turner. JC (1986)‘The social identity theory of intergroup behavior’. Psychology
of Intergroup Relations, , 7-24.
Terry, D. J., & Hogg, M. A. (1996). Group norms and the attitude-behavior relationship: A role
for group identification. Personality and Social Psychology Bulletin, 22(8), 776-793.
Terry, D. J., Hogg, M. A., & McKimmie, B. M. (2000). Attitude ‐behaviour relations: The role of
in ‐group norms and mode of behavioural decision ‐making. British Journal of Social
Psychology, 39(3), 337-361.
Tevendale, H. D., Lightfoot, M., & Slocum, S. L. (2009). Individual and environmental
protective factors for risky sexual behavior among homeless youth: An exploration of
gender differences. AIDS and Behavior, 13(1), 154-164.
Thompson, S. J., Rew, L., Barczyk, A., McCoy, P., & Mi-Sedhi, A. (2009). Social estrangement:
Factors associated with alcohol or drug dependency among homeless, street-involved
young adults. Journal of Drug Issues, 39(4), 905-929.
Toro, P. A., Dworsky, A., & Fowler, P. J. (2007). Homeless youth in the united states: Recent
research findings and intervention approaches. Paper presented at the 2007 Symposium on
Homelessness Research, US Department of Health and Human Services. One Study found
that Youth Involved with the Correctional System were More Likely to be Homeless Or
Precariously Housed. this Study Compared, , 209-220.

172

Tsemberis, S., McHugo, G., Williams, V., Hanrahan, P., & Stefancic, A. (2007). Measuring
homelessness and residential stability: The residential time-line follow-back inventory.
Journal of Community Psychology, 35(1), 29.
Tucker, J. S., Edelen, M. O., Ellickson, P. L., & Klein, D. J. (2011). Running away from home:
A longitudinal study of adolescent risk factors and young adult outcomes. Journal of Youth
and Adolescence, 40(5), 507-518.
Tucker, J. S., Hu, J., Golinelli, D., Kennedy, D. P., Green Jr, H. D., & Wenzel, S. L. (2012).
Social network and individual correlates of sexual risk behavior among homeless young
men who have sex with men. Journal of Adolescent Health, 51(4), 386-392.
Turner, C. F., Ku, L., Rogers, S. M., Lindberg, L. D., Pleck, J. H., & Sonenstein, F. L. (1998).
Adolescent sexual behavior, drug use, and violence: Increased reporting with computer
survey technology. Science, 280(5365), 867-873.
Tyler, K. A. (2008). Social network characteristics and risky sexual and drug related behaviors
among homeless young adults. Social Science Research, 37(2), 673-685.
Tyler, K. A., & Johnson, K. A. (2006). Trading sex: Voluntary or coerced? the experiences of
homeless youth. Journal of Sex Research, 43(3), 208-216.
Tyler, K. A. (2013). Homeless youths’ HIV risk behaviors with strangers: Investigating the
importance of social networks. Archives of Sexual Behavior, 1-9.

173

Ueno, K. (2005). The effects of friendship networks on adolescent depressive symptoms. Social
Science Research, 34(3), 484-510.
Valente, T. W., Gallaher, P., & Mouttapa, M. (2004). Using social networks to understand and
prevent substance use: A transdisciplinary perspective. Substance use & Misuse, 39(10-12),
1685-1712.
Valente, T. W. (2012). Network interventions. Science, 337(6090), 49-53.
van Empelen, P., Schaalma, H. P., Kok, G., & Jansen, M. W. (2001). Predicting condom use
with casual and steady sex partners among drug users. Health Education Research, 16(3),
293-305.
Vlahov, D., Fuller, C. M., Ompad, D. C., Galea, S., & Des Jarlais, D. C. (2004). Updating the
infection risk reduction hierarchy: Preventing transition into injection. Journal of Urban
Health, 81(1), 14-19.
Wang, L., Fan, X., & Willson, V. L. (1996). Effects of non-normal data on parameter estimates
and fit indices for a model with latent and manifest variables: An empirical study.
Structural Equation Modeling: A Multidisciplinary Journal, 3(3), 228-247.
Warschauer, M. (2004). Technology and social inclusion: Rethinking the digital divide the MIT
Press.
Wasserman, S., & Galaskiewicz, J. (1994). Advances in social network analysis: Research in the
social and behavioral sciences. Sage Publications, Inc.

174

Weber, A. E., Roy, E., & Haley, N. (2002). Female street youth and prostitution. Focus, 17(4), 5-
6.
Weisheit, R. A., & Wells, L. E. (2010). Methamphetamine laboratories: The geography of drug
production. W.Criminology Rev., 11, 9-18.
Wenzel, S. L., Tucker, J. S., Golinelli, D., Green, H. D., & Zhou, A. (2010). Personal network
correlates of alcohol, cigarette, and marijuana use among homeless youth. Drug and
Alcohol Dependence, 112(1), 140-149.
Wenzel, S., Holloway, I., Golinelli, D., Ewing, B., Bowman, R., & Tucker, J. (2012). Social
networks of homeless youth in emerging adulthood. Journal of Youth and Adolescence,
41(5), 561-571.
Whitbeck, L. B., & Hoyt, D. R. (1999). Nowhere to grow: Homeless and runaway adolescents
and their families Aldine de Gruyter.
Whitbeck, L. B., Hoyt, D. R., Yoder, K. A., Cauce, A. M., & Paradise, M. (2001). Deviant
behavior and victimization among homeless and runaway adolescents. Journal of
Interpersonal Violence, 16(11), 1175-1204.
White, K. M., Terry, D. J., & Hogg, M. A. (1994). Safer sex behavior: The role of attitudes,
norms, and control factors. Journal of Applied Social Psychology, 24(24), 2164-2192.

175

Wood, E., Wood, E., Stoltz, J., Wood, E., Stoltz, J., Zhang, R., . . . Strathdee, S. A. (2008).
Circumstances of first crystal methamphetamine use and initiation of injection drug use
among high-risk youth. Drug and Alcohol Review, 27(3), 270-276.
Woolcock, M. (1998). Social capital and economic development: Toward a theoretical synthesis
and policy framework. Theory and Society, 27(2), 151-208.
Ybarra, M. L., & Bull, S. S. (2007). Current trends in internet-and cell phone-based HIV
prevention and intervention programs. Current HIV/AIDS Reports, 4(4), 201-207.
Ybarra, M. L., Kiwanuka, J., Emenyonu, N., & Bangsberg, D. R. (2006). Internet use among
Ugandan adolescents: Implications for HIV intervention. PLoS Med, 3(11), e433.
Young, S. D., & Rice, E. (2011). Online social networking technologies, HIV knowledge, and
sexual risk and testing behaviors among homeless youth. AIDS and Behavior, 15(2), 253-
260.

Abstract (if available)

Abstract Homeless youth engage in substance use risk behaviors that place them at an increased risk of HIV infection. Studies on this population have shown that their social networks are consistently linked with their HIV risk behaviors. However, recent reviews suggest that substance use prevention programs have had an individualistic focus. This has led leading experts in the field to call for novel network-level prevention interventions to be developed to reduce HIV incidence among homeless youth. Social networks influence behavior through several mechanisms, one of which is through the establishment, and maintenance of social norms. While several theories suggest that norms offer a potent channel of initiating and sustaining behavioral change, intervention efforts have been hampered because of the paucity of research examining clustering of norms within specific risk-taking social network structures. For example, although much of the theoretical research on norm distribution emphasizes structural elements such as network position or the cohesiveness or size of the network in the shaping of perceived norms, research in this area is very limited. The purpose of this proposed study was to utilize sociometric analyses to understand whether social norms of HIV risk behaviors are clustered with social network structures, and whether the norms of these network members are associated with these youth’s risk taking behaviors. Researchers have also suggested that future studies must address how different relationship roles influence HIV behavioral norms. Therefore, this study also utilized egocentric social network analyses to understand which referent groups other than street peers that these youth connect to (especially via new forms of communicative technology), and whether these differential forms of influences contribute to different types of norms regarding these behaviors. ❧ Data for this research came from a larger NIMH funded “Youthnet” study (MH R01 903336). Rice and colleagues are collecting multiple panels of egocentric and sociometric network data over time from homeless youth ages 13-25 in two drop-in centers in Los Angeles, CA. This proposed study used only the baseline data to accomplish its research objectives. Taken together, the egocentric and sociometric data was able examine multiple research questions that have not yet been addressed in the HIV risk norms literature. These data suggest that both social proximity as well as social positioning within networks is associated with risk norms. The results therefore overall supported the general proposition that both egocentric and sociometric network attributes affect substance use among homeless youth. Other findings elaborated on these overall results. The findings can be used to inform new directions in HIV prevention interventions, specifically what network-level interventions could be adapted in the context of the homeless youth population, and the feasibility of online technology as a potential mechanism through which network interventions can be delivered.

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

Social network engagement and HIV risk among homeless former foster youth

PDF

Engaging homeless men and shelter providers to adapt an existing evidence-based HIV prevention intervention

PDF

Couch-surfing among youth experiencing homelessness: an examination of HIV risk

PDF

Youth homelessness and gang involvement: ties, trajectories, and timelines

PDF

A closer look at intention to use condoms and risky sexual behaviors among homeless people: application of the theories of reasoned action and planned behavior

PDF

Sex talks: an examination of young Black women's social networks, sexual health communication, and HIV prevention behaviors

PDF

Social network and contextual influences on substance use and HIV risk behavior among young men who have sex with men

PDF

The algorithm of Black homelessness: a classification tree analysis

PDF

Tobacco use change among formerly homeless supportive housing residents: socioecological barriers and facilitators to cessation

PDF

Contextualizing social network influences on substance use among high risk adolescents

PDF

Main, moderating, and mediating effects of social network compositions on suicidal ideation among North Korean refugees in South Korea

PDF

Using social networks to recruit an HIV vaccine preparedness cohort

PDF

Stage of readiness and learning styles -- a tailored HIV/AIDS prevention intervention for youth detained in the Los Angeles County Juvenile Justice system

PDF

Homeless youth: Reaching the Hard-To-Reach

PDF

Factors and correlates of sexual behaviors among young adults from continuation high schools

PDF

U.S. Latinx youth development and substance use risk: adversity and strengths

PDF

Gendered issues in alcohol abuse and dependence among HIV-positive African Americans

PDF

Homelessness and substance use treatment: using multiple methods to understand risks, consequences, and unmet treatment needs among young adults

PDF

Social network influences on depressive symptoms among Chinese adolescents

PDF

Contagious: social norms about health in work group networks

Asset Metadata

Creator Barman-Adhikari, Anamika (author)

Core Title Social network norms and HIV risk behaviors among homeless youth in Los Angeles, California

School School of Social Work

Degree Doctor of Philosophy

Degree Program Social Work

Publication Date 11/26/2013

Defense Date 07/24/2013

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

Tag HIV risk,homeless youth,OAI-PMH Harvest,social network,social norms

Format application/pdf (imt)

Language English

Contributor Electronically uploaded by the author (provenance)

Advisor Rice, Eric R. (committee chair), Esparza, Nicole (committee member), Wenzel, Suzanne L. (committee member)

Creator Email abarmanadhikari@csufresno.edu,barmanad@usc.edu

Permanent Link (DOI) https://doi.org/10.25549/usctheses-c3-350915

Unique identifier UC11295324

Identifier etd-BarmanAdhi-2183.pdf (filename),usctheses-c3-350915 (legacy record id)

Legacy Identifier etd-BarmanAdhi-2183.pdf

Dmrecord 350915

Document Type Dissertation

Format application/pdf (imt)

Rights Barman-Adhikari, Anamika

Type texts

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

Access Conditions The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the a...

Repository Name University of Southern California Digital Library

Repository Location USC Digital Library, University of Southern California, University Park Campus MC 2810, 3434 South Grand Avenue, 2nd Floor, Los Angeles, California 90089-2810, USA

Social network norms and HIV risk behaviors among homeless youth in Los Angeles, California​

Social network norms and HIV risk behaviors among homeless youth in Los Angeles, California​

Social network norms and HIV risk behaviors among homeless youth in Los Angeles, California

Social network norms and HIV risk behaviors among homeless youth in Los Angeles, California