Close
About
FAQ
Home
Collections
Login
USC Login
Register
0
Selected
Invert selection
Deselect all
Deselect all
Click here to refresh results
Click here to refresh results
USC
/
Digital Library
/
University of Southern California Dissertations and Theses
/
The social groups approach to quitting smoking: An examination of smoking cessation online and offline through the influence of social norms, social identification, social capital and social support
(USC Thesis Other)
The social groups approach to quitting smoking: An examination of smoking cessation online and offline through the influence of social norms, social identification, social capital and social support
PDF
Download
Share
Open document
Flip pages
Contact Us
Contact Us
Copy asset link
Request this asset
Transcript (if available)
Content
THE SOCIAL GROUPS APPROACH TO QUITTING SMOKING:
AN EXAMINATION OF SMOKING CESSATION ONLINE AND OFFLINE
THROUGH THE INFLUENCE OF SOCIAL NORMS, SOCIAL IDENTIFICATION,
SOCIAL CAPITAL AND SOCIAL SUPPORT
by
Joe Jin Phua
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
(COMMUNICATION)
May 2011
Copyright 2011 Joe Jin Phua
ii
TABLE OF CONTENTS
List of Tables iv
List of Figures vi
List of Abbreviations vii
Abstract viii
CHAPTER ONE: INTRODUCTION 1
Overview 1
Health-Based Social Networking Sites 5
Contributions of Dissertation 11
Chapter Summaries 12
CHAPTER TWO: THEORIES ADAPTED AND TESTED 14
Description of Dissertation Studies 14
Social Norms Approach 16
Differences between Social Norms and Social Influence 20
Social Identity Theory 22
Online Social Capital 24
Social Networking Sites and Online Social Capital 28
Online Social Support Groups 29
Lurkers in Online Communities 32
Online Smoking Cessation Programs 35
Smoking Cessation Self-Efficacy 41
Research Questions and Hypotheses (Study I) 44
Research Questions and Hypotheses (Study II) 47
CHAPTER THREE: METHOD 56
Data Collection 56
Study Sites 56
Recruitment of Subjects 59
Explanation of Reliability and Validity 61
Exploratory Factor Analysis 62
Harman’s One-Factor Test (Study I) 64
Measures for Study I 65
Harman’s One-Factor Test (Study II) 68
Measures for Study II 73
Data Analysis Plan 78
iii
CHAPTER FOUR: RESULTS 80
Descriptive Statistics 80
Smoking Behavior 81
Correlation Analyses (Study I) 83
Multivariate Regression (Study I) 85
Internet and Site Usage 100
Correlation Analyses (Study II) 101
Multivariate Regression (Study II) 101
MANOVA and ANOVA Tests (Study II) 109
Structural Equation Modeling (Study II) 118
Mediation Analysis (Study II) 123
CHAPTER FIVE: DISCUSSION 126
Discussion of Results (Study I) 126
Discussion of Results (Study II) 133
Limitations 147
Conclusion 152
BIBLIOGRAPHY 157
APPENDIX 174
iv
LIST OF TABLES
Table 1: List of Smoking Cessation Sites Contacted 58
Table 2: List of Smoking Cessation Sites Studied 59
Table 3: Summary of Exploratory Factor Analysis Results for Study I (N=208) 62
Table 4: Reliability and Validity Statistics for Measures in Study I 64
Table 5: Summary of Exploratory Factor Analysis Results for Study II (N=208) 69
Table 6: Reliability and Validity Statistics for Measures in Study II 73
Table 7: Demographic Characteristics of Questionnaire Respondents (N=252) 80
Table 8: Smoking Behavior of Study I Respondents (N=252) 81
Table 9: Zero-order Correlations of independent variables (Best friends) on
Smoking cessation self-efficacy (N=208) 83
Table 10: Zero-order Correlations of independent variables (Colleagues) on
Smoking cessation self-efficacy (N=208) 84
Table 11: Zero-order Correlations of independent variables (Family members)
on Smoking cessation self-efficacy (N=208) 84
Table 12: Standardized Regression Coefficients (β) of Best Friends’ injunctive
and descriptive norms on Smoking Cessation Self-Efficacy (N=208) 86
Table 13: Collinearity Statistics for Independent Variables (Best Friends) on
Smoking Cessation Self-Efficacy 87
Table 14: Standardized Regression Coefficients (β) of Colleagues’ injunctive
and descriptive norms on Smoking Cessation Self-Efficacy (N=208) 91
Table 15: Collinearity Statistics of Independent Variables (Colleagues) on
Smoking Cessation Self-Efficacy 92
Table 16: Standardized Regression Coefficients (β) of Family members’
injunctive and descriptive norms on Smoking cessation self-efficacy (N=208) 95
v
Table 17: Collinearity Statistics of Independent Variables (Family Members)
on Smoking Cessation Self-Efficacy 96
Table 18: Average Internet and Site Usage by Questionnaire Respondents
(N=252) 100
Table 19: Zero-order Correlations of independent variables on Smoking
cessation self-efficacy (Study II) (N=208) 101
Table 20: Collinearity Statistics of Independent Variables (Study II) on
Smoking Cessation Self-Efficacy 103
Table 21: Standardized Regression Coefficients (OLS Regression) of Selected
Independent Variables on Smoking Cessation Self-Efficacy (N=208) 103
Table 22: Tests of Significance Difference between Active Participants and
Lurkers on Social Variables (N=208) 114
Table 23: Tests of Significance Difference between Active Participants and
Lurkers on Intimacy Levels (N=208) 117
Table 24: Goodness-of-Fit indicators of Measurement Models in Confirmatory
Factor Analysis (N=208) 118
Table 25: Parameter Estimates and Significance Levels for Structural Model
(Standard Errors in Parentheses) (N=208) 121
Table 26: Mediation Analyses (Baron & Kenny, 1986) of Social Variables on
Relationship between Participation Level and Attitude towards Smoking 123
Table 27: Mediation Analyses (Baron & Kenny, 1986) of Social Variables on
Relationship between Participation Level and Smoking Cessation Self-Efficacy 124
vi
LIST OF FIGURES
Figure 1: Reference Group Perspective for Smoking Cessation (Study I) 19
Figure 2: Theoretical Model Specifying Relationships among Variables
(Study II) 55
Figure 3: Structural Model for the Social Groups Approach to Smoking
Cessation (N=208) 122
vii
LIST OF ABBREVIATIONS
ANOVA Analysis of Variance
CMC Computer-Mediated Communication
EPPM Extended Parallel Processing Model
ICT Information and Communication Technology
IRB Institutional Review Board
ISCS Internet Social Capital Scales
MUD Multi-User Dungeon
RSS Really Simple Syndication
SEM Structural Equation Model
SIDE Social Identification Model of Deindividuated Effects
SIP Social Information Processing
SIT Social Identity Theory
SCT Self-Categorization Theory
SNS Social Networking Site
TNSB Theory of Normative Social Behavior
WATI Web-Based Tobacco Interventions
WHO World Health Organization
viii
ABSTRACT
The goal of this dissertation was to investigate the ways in which social groups
influence the individual’s attitudes towards smoking and smoking cessation self-efficacy,
in both offline and online contexts. Researchers have previously found that the injunctive
and descriptive norms of one’s reference groups toward smoking have a profound
influence on one’s own smoking behavior. Additionally, smokers who use online support
groups for smoking cessation have also been found to increase their levels of smoking
cessation self-efficacy. This dissertation builds on this work. I propose a Social Groups
Approach to smoking cessation, which applies and extends traditional theories of peer
influence, including Social Identity Theory (Tajfel & Turner, 1986), Social Norms
Approach (Perkins & Berkowitz, 1986), Social Influence (Rogers, 2003; Valente, 1995),
Social Capital (Putnam, 2000; Williams, 2006) and Social Support (Sarason & Sarason,
2006). The Social Groups Approach is explored here in two studies, one offline and one
online.
Study I assessed participants’ (N=208) identification with three offline reference
groups, hypothesizing that identification would moderate the relationship between
injunctive and descriptive norms and smoking cessation self-efficacy. Study II examined
members (N=208) of online health social networking sites for smoking cessation,
hypothesizing that four social variables: social identification, bridging and bonding social
capital, social norms and social support, would impact the relationship between
participation level and smoking cessation self-efficacy. Study II also hypothesized
ix
significant differences between active participants and lurkers for the four social
variables, and four types of intimacy levels.
For Study One, injunctive and descriptive norms of reference groups significantly
affected smoking cessation self-efficacy, and this relationship was moderated by
identification. Injunctive norms were stronger in predicting higher smoking cessation
self-efficacy than descriptive norms, with injunctive norms of family members and
descriptive norms of best friends having the most significant effect. For Study II,
participation significantly impacted the four social variables, which in turned influenced
smoking cessation self-efficacy. Active participants and lurkers also differed
significantly for intimacy levels on the site, social identification, social capital, social
norms, social support, and smoking cessation self-efficacy.
The dissertation ultimately proposed a model for the application of online social
media to smoking cessation, the “Social Groups Approach to Smoking Cessation,” which
was tested and supported. Implications for future research on theory-based interventions
of smoking cessation using online and face to face influences are discussed.
Keywords: Social networking sites, Smoking cessation, Social identity theory, Social
capital, Social support, Injunctive norms, Descriptive norms, Reference groups, Lurking.
1
CHAPTER ONE: INTRODUCTION
Overview
Friends, family and social networks are important aspects of a person’s life.
There is little doubt that the number of friends one maintains in his or her social network
has a significant impact on a number of aspects of a person’s life, including happiness,
good health and satisfaction (see Christakis & Fowler, 2009; Lehrer, 2009). No single
study demonstrated this better than the Framingham Heart Study (Christakis & Fowler,
2009). This study found that two types of health behavior, obesity and smoking, were
distributed throughout interconnected individuals within social networks, and changed in
systematic ways over time. For example, obesity was largely scattered throughout the
community in 1985, but over time the number of obese individuals increased outward
from the clusters of individuals who were obese. By 2000, over 40% of the population
was obese, and these obese individuals were in clusters of social networks. In fact,
having an obese spouse or friend increased one’s risk of obesity 37% and 171%
respectively. Similarly, quitting smoking was another behavior which diffused
throughout connected individuals in one’s social network. In 1971, smoking was evenly
distributed throughout the Framingham community, but over time smokers and non-
smokers began to drift apart. By 2001, most smokers who were once at the center of
their social network were pushed to the margins, connected to other smokers, or
emerging as socially isolated individuals (Szabo, 2008). When smokers quit, their
friends and family were also 36% more likely to quit as well.
2
The Christakis and Fowler (2009) study attests to the ability of social networks to
influence individual health behavior. Humans are social animals, and in today’s society,
are often intricately interconnected through social networks and communities. They also
engage in interpersonal conversations, which serve to transmit information, social norms
and identities. As Habermas (1984) states in Theory of Communicative Action,
communication between individuals serve to transmit and renew cultural knowledge,
such that mutual understanding is achieved. Action is then coordinated towards social
integration and solidarity, with people forming their personal and social identities as the
end process. Because social networks exert a great amount of influence on the ways in
which people think and act, therefore, in order to understand how to change individual
behavior, it is often very useful to consider the behavior of other people within one’s
social circle. This dissertation reports the results of two studies which examined the
influence of one’s social networks, both offline and online, on smoking cessation.
Despite tough anti-smoking legislation and widespread public awareness of
negative health consequences, including cardiovascular diseases and lung cancer,
smoking currently remains one of the top public health problems around the world. It has
been estimated that 4.9 million people around the world die of smoking-related diseases
each year, and this figure is expected to increase to 10 million per year by 2030
(Balmford, Borland & Benda, 2008). Perhaps more troubling is the fact that half of all
current smokers today, 650 million people, will die as a result of smoking, and that 80%
of all deaths from smoking are projected to occur in the developing world, where tobacco
companies have now focused their marketing efforts (Norman, McIntosh, Selby &
3
Eysenbach, 2008). Additionally, between 80 to 90% of adult smokers begin smoking by
age 18 (Alexander, Piazza, Mekos & Valente, 2001). Research has shown that people’s
social networks exert a great amount of influence on their behavior (Christakis & Fowler,
2009; Perkins & Berkowitz, 1986). Social norms are powerful agents of control whereby
behavior is framed by norms, and the course of behavior commonly taken is typically in
accordance with normative directives of reference groups most important to the
individual (Perkins, 2002). Group norms are reflected in the dominant attitudes,
expectations and behaviors characterizing social groups, and they serve to regulate group
members’ actions so as to perpetuate and follow the collective norm. As such, normative
information exerts an enormous influence on social behavior like smoking. Further, peer
influence has been identified in many studies as a leading correlate and probable cause of
smoking (Hoffman, Sussman, Unger & Valente, 2006; Valente, Unger & Johnson, 2005).
Valente et al. (2005), for example, found that popular students are more likely to have a
strong influence in setting peer group norms and providing cues to the acceptability or
unacceptability of smoking. In communities with high smoking prevalence, peer
influence promotes smoking, while in communities with low smoking prevalence, peer
influence is more likely to discourage smoking. These popular students set trends for
their peer groups, and are associated with increased susceptibility to smoke in schools
with high smoking prevalence.
In recent years, the Internet has become a highly useful and effective avenue for
the promotion of smoking cessation, due to the wide availability of literature online about
the health risks associated with smoking, and information on how to quit and maintain
4
smoke-free behaviors. Cobb and Graham (2006) report that approximately 7%, or 10.2
million, of adult American Internet users search for information about smoking cessation
online each year. Internet-based health information can provide patients with timely
support and materials to achieve positive health outcomes (Bessell et al., 2002).
Additionally, people looking to quit smoking can also join online health support groups,
the majority of which now incorporate forums, message boards and social networking
features to provide social support for their members (Schneider & Tooley, 1986;
Schneider, Walter & O’Donnell, 1990; Stoddard et al., 2005; Stretcher, Shiffman &
West, 2005). Health-based social networking sites are becoming increasingly popular for
people with chronic health conditions because of the many features which increase their
connections to others with similar problems. They are also the focus of this dissertation.
Social networking sites are web-based services in which people create a profile
which may be public or semi-public, connect with a list of other users, and view and
traverse these connections and those made by others within the site (boyd & Ellison,
2007). An online profile acts as a user’s home page on the site, where he or she can
articulate his or her identity by providing personal information, including hobbies,
interests and favorite music and television shows, and also post pictures. Members of the
sites can also add other members to their networks through the process of “friend
requests.” After two members are connected, they can then view each other’s profiles,
post comments, receive updates and more. The popularity of Facebook and Myspace has
led to the emergence of many health-related social networking sites. As Boulos and
Wheelert (2007) point out, Web 2.0, or the Social Web, has enabled more widespread
5
collaboration and interactivity, such that users can simultaneously act as readers and
writers, providing quicker access to relevant information and knowledge. By
incorporating social networking features, health support sites enable patients to better
manage their own health issues and also “gather, learn from, and give support to others in
need” (p. 20).
Health-Based Social Networking Sites
Health-based social networking sites represent a new niche in the age of Web 2.0.
More and more of these sites are emerging every day, and changing the ways in which
health care is conducted, and also the interaction between health professionals, patients
and caregivers. With medical care costs currently on the rise in the United States, the
physician-patient relationship becomes particularly important to consider. This is where
health-based social networking comes in, according to Bottles (2009), who commented
on its ability to maximize the quality of health care, while minimizing associated costs.
Following on Web 2.0, the term “Health 2.0” was coined to describe the use of Web 2.0
tools, including social networking sites, blogs, RSS (Really Simple Syndication) feeds,
podcasts, Wikis, and other new Internet technologies to create user-generated content,
thereby enabling an “architecture of participation” within health care (Bottles, 2009).
According to a report by the Pew Internet and American Life Project released in March
2010, patients with chronic illnesses are less likely to have Internet access, but once they
are online, they are also more likely than other people to blog about their illnesses, and
also participate in online discussions or join social networks devoted to their health
problem (Miller, 2010). Further, they are also more likely to look for health information
6
pertaining to their conditions online and also to use the Internet to alleviate depression,
anxiety and stress. As a 2008 poll by Edelman Trust Barometer found, the most credible
source of health information is from “a person like me”, surpassing public trust in health
professionals and academic experts (Aids.Gov, 2008).
Social networking is fundamentally transforming the relationships between health
patients, their health care providers and caregivers, as well as their interactions with other
patients. Members of health-based social networking sites, particularly, are able to talk
about illness experiences with others with the same conditions, compare treatment
options, provide and receive emotional support, and even aggregate the diseases’
community experience (Bottles, 2009). A New York Times article by Miller (2008)
introduced some of the most popular sites, including “Disaboom,” a site for the disabled
in which they can search for others with the same conditions, “Diabetic Connect,” a site
connecting diabetes patients with their doctors and caregivers, “IMedix,” a general health
social network with a wide variety of health topics, “Inspire,” which has mini social
networks for different health conditions, “Patients Like Me,” “Health Central” and
various patient-started sites on “Ning” and “Wet Paint.” In terms of social networking
features, “Diabetic Connect,” for example, includes doctor-approved reference content,
interactive tools and Facebook-style friend activity feeds, the ability to embed “favorite”
videos and articles from external sources on profiles, commenting and user ratings.
These sites are more useful to patients because, compared to general SNSs like Facebook
and MySpace, they attract people with the same health conditions and experiences, and
therefore connecting with others through the sites becomes much easier.
7
For the people who feel alone in their health struggles, especially those who
cannot leave their homes due to illness, social networking sites often become the center
of their social life (Miller, 2008). For example, Miller (2008) interviewed several
members in “Disaboom” who made plans to watch the same movie at home, while
simultaneously chatting with one another over the Internet about the movie. Miller
(2010) also talked about how social networking sites have become the “lifeline” for many
patients she interviewed, including a woman with multiple sclerosis and Lyme disease,
and a wheelchair-bound quadriplegic man, who find social networking sites particularly
useful for making close friendships with others who share the same experiences.
Compared to traditional online communities, social networking sites are
characterized as being a more supportive environment, with less flaming, negative
attitudes and behavior. Because they are not anonymous, social networking sites also
allow more meaningful and deeper friendships to build, as members can browse others’
profiles and know exactly to whom they are talking. Miller (2010) contends that social
networking sites provide members with a “real-life perspective” on their health
conditions, since they are able to interact with other patients who have lived through the
same conditions and are able to provide practical tips, unlike doctors and other health
professionals whose knowledge may not be as practical. Especially for people with
stigmatizing diseases, social networking sites give them a forum whereby they will not be
judged or criticized harshly (Miller, 2010). This sentiment is echoed by Levine (2009),
who described how HIV awareness agencies like Aids.Gov have used social networking
8
sites, as well as associated widgets and apps (e.g. Nine and a half Minutes, Sexpert, etc.)
to educate teens about safe sex.
Further, Health 2.0 tools are also transforming the physician-patient relationship.
For example, “American Well” and “HelloHealth” are two sites created specifically with
features to enhance physicians’ contact and interaction with their patients online, and
have already been implemented in Hawaii and Minnesota through patients’ Blue
Cross/Blue Shield medical plans (Bottles, 2009). Additionally, physicians are also
currently testing the use of Twitter, combined with sensors which would inform them
when their homebound patients’ vital signs indicate a health emergency. At the Henry
Ford Hospital in Detroit, Twitter has also been used by surgeons during operations as a
way to document and inform the medical community (Bottles, 2009). Finally, the use of
mobile phones, particularly iPhone apps, is another major Health 2.0 application. These
apps, like the “Medzio”, integrate into patients’ everyday activities, and makes accessing
health information and social networking much easier and more fun (Bottles, 2009). As
seen from the above examples, Health 2.0 has the potential to transform health care in
many ways.
However, health-based social networking sites also present several challenges.
One of the most significant problems has to do with the fact that medical advice often
comes from other members rather than health professionals, and as such the quality of the
advice cannot be verified. However, a Pew Internet and American Life Project survey
found that only 2% of patients reported being “harmed” by following medical advice
from the sites. Most of the sites also clearly label information from health professionals
9
and that from other members, and have a disclaimer stating that the information found on
the sites should not be a substitute for actual medical advice (Miller, 2010). Another
major problem is that many of the sites rely on advertising revenue, which can be
bothersome for some members due to the number of pop-up ads. Finally, many of the
sites also have pharmaceutical companies and clinical trial researchers as sponsors, which
can drastically affect the objectivity of the information found on the sites (Miller, 2008).
These challenges, according to an article on Aids.Gov (2008), can be solved through
feedback from patients, doctors and caregivers, all of whom, due to their direct
involvement in the sites, can give the best advice and suggestions to create a sense of
community.
In previous research, participation in online health support groups has been
found to be positively associated with health benefits (Bessell et al., 2002; Hoybye,
Johansen & Tjornhoj-Thomsen, 2005; Shaw, Hawkins, McTavish, Pingree & Gustafson,
2006). For people interested in quitting smoking, self-efficacy has been found in
numerous studies to be an important predictor of smoking cessation and subsequent
abstinence (Stretcher, Becker, Kirscht, Eraker & Graham, 1985; Stretcher, DeVellis,
Becker & Rosenstock, 1986), and Internet-based smoking cessation programs are
particularly useful for increasing self-efficacy (Stretcher et al., 2005; Takahashi et al.,
1999). However, as noted by Burnett (2000) and Demiris (2005), a large number of
members of online health support groups are lurkers who do not participate in
discussions. Lurking in online health social support groups can be both an advantage and
a disadvantage. The lack of activity in many online communities presents a problem
10
because they may not be able to provide members with needed services. This is
especially worrisome for online health groups because important treatments and problems
may not be discussed, thereby depriving members of much-needed informational and
emotional support. Lurkers may also have a negative influence in online communities
because they are regarded by active members as “free riders” (Kollock & Smith, 1996).
However, there is also an upside to lurking. For example, Takahashi, Fujimoto
and Yamasaki (2003) found that lurkers propagated topics discussed in online
communities to non-members, and also used the knowledge gained from online
communities in their own organizational activities, thereby playing an important and
necessary role in the information-dissemination process. Preece, Nonnecke and Andrews
(2004) suggested that lurking benefited online groups because it allowed new members to
learn the group’s norms, see if the concerns addressed were relevant to them, and find out
more about other members before posting any messages. Further, in groups with heavy
traffic, lurking can be good because it reduces the “noise-to-signal” ratio. By posting
only when necessary, lurkers can help online groups maintain the quality of the
information shared and received. Even though in general, lurkers received significantly
fewer benefits than posters from the online support groups, they still felt a sense of
community within the groups (Millen & Patterson, 2003; Preece et al., 2004). In their
systematic review of previous studies about online health groups, Eysenbach, Powell,
Englesakis, Rizo and Stern (2004) concluded that there has been no robust evidence for
health benefits, or negative effects from participation, and that more research is needed to
11
quantify the conditions and target audience for which social support can be effectively
delivered in these online groups.
Contributions of Dissertation
This dissertation addressed the concerns mentioned in the previous section by
utilizing a two-study, mixed-method approach. Theories applied include the Social
Norms (Perkins & Berkowitz, 1986) and Social Influence (Rogers, 2003) approaches for
addictive behavior, Social Identity Theory (Tajfel, 1978), Social Capital (Bourdieu, 1986;
Burt, 1992; Coleman, 1988; Lin, 1999; Putnam, 2000; Williams, 2006) and Social
Support (Tanis, 2008). The dissertation sought to advance scholarly knowledge about
smoking cessation offline and online in the following ways:
Study I proposed a Reference Group Perspective on smoking cessation. It
examined how participants’ social identification with three reference groups (best friends,
family members and colleagues), as well as perceived injunctive and descriptive norms
among these three reference groups, influenced their attitudes towards smoking and
smoking cessation self efficacy, applying the Social Norms Approach (Perkins &
Berkowitz, 1986). The results of Study I added to existing research by positing social
identification (Tajfel & Turner, 1986) as a potential moderator in the relationship
between norms and smoking cessation self efficacy. More precisely, it established social
identity as an important predictor of smoking cessation for people in offline social groups
and relationships.
Study II examined participants’ use of online health social networking sites for
smoking cessation using Social Identity Theory (Tajfel, 1978) and Social Capital
12
(Putnam, 2000). Most current research on health social support sites has mainly focused
on content analysis of message boards and discussion forums. However, because health-
based social networking sites, specifically those which allow members to have their own
profile pages and to connect to other members publicly on friend lists, are becoming
increasingly popular among health patients, the increased health benefits, if any, that can
be received through participation on these sites deserves scholarly research. Study II
contributed to existing research by testing whether social identification, online bridging
and bonding social capital, and social support received from the sites could affect the
relationship between participation level and perceived positive health benefits, such as
smoking cessation self-efficacy. Additionally, it also tested whether there was a
significant difference between active participants and lurkers on the sites in terms of
social identification, bridging and bonding social capital, perceived social support, and
smoking cessation self-efficacy. Antecedents to participation online, such as
commitment, interdependence, communication, network convergence, code change and
predictability, were also predicted to significantly vary between active participants and
lurkers. Finally, an overall model of influence, the Social Groups Approach to Smoking
Cessation, was proposed and tested. Chapter summaries are presented below.
Chapter Summaries
Chapter 2 focuses on theories adapted and tested in the dissertation. It begins by
introducing the proposed Social Groups Approach to Smoking Cessation, both offline
and online. Second, it discusses how the following theories: Social Norms Approach
(Perkins & Berkowitz, 1986), Social Identity Theory (SIT) (Tajfel & Turner, 1986),
13
Social Capital (Bourdieu, 1986; Putnam, 2000), and Social Support (Sarason & Sarason,
2006; Tanis, 2008), are adapted and tested in the two studies within the dissertation.
Third, it reviews the existing literature on Social Networking Sites (SNSs), Online Social
Support Groups, lurkers in online communities, Online smoking cessation programs and
Smoking cessation self-efficacy, and highlights the ways in which the dissertation
contributes to the current field of research on smoking cessation. Chapter 3 begins with a
detailed discussion of the methods employed in the two studies. First, hypotheses and
research questions for both studies are summarized, and then data collection, survey
measures, participation recruitment, site selection, as well as data analysis methods for
each study are described. Chapter 4, meanwhile, reports the results of the data analyses,
including descriptive statistics, multivariate regression, correlation analyses, MANOVA
and ANOVA tests, mediation analyses, and the structural equation model. Finally,
Chapter 5 includes a discussion of the findings of the two studies, and how these findings
contribute to the existing research on smoking cessation. Further, it also discusses the
limitations associated with the two studies, as well as implications for future research on
smoking cessation and health-based social networking sites.
14
CHAPTER TWO: THEORIES ADAPTED AND TESTED
The first part of Chapter two will introduce the two studies in the dissertation, and
set up the rationale for the Social Groups Approach to Smoking Cessation, both offline
and online. Second, it will discuss theories adapted and tested in the dissertation: Social
Norms Approach (Perkins & Berkowitz, 1986), Social Identity Theory (SIT) (Tajfel &
Turner, 1986), Social Capital (Bourdieu, 1986; Putnam, 2000), and Social Support
(Sarason & Sarason, 2006; Tanis, 2008). Third, it will review the literature on Social
Networking Sites (SNSs), Online Social Support Groups, Lurkers in online communities,
Online smoking cessation programs and Smoking cessation self-efficacy, and address the
contributions of the two studies to the current field of research on smoking cessation.
Description of Dissertation Studies
The purpose of this dissertation is to examine how people’s social groups can
be utilized to help one quit smoking in both offline (Study I) and online (Study II)
contexts. Study I assessed whether the injunctive and descriptive norms of one’s
reference groups, including best friends, family members, and colleagues, and social
identification with these groups influence one’s own smoking. The theories used in
Study I include SIT (Tajfel & Turner, 1986) and the Social norms approach (Perkins &
Berkowitz, 1986). Particularly, social identification was tested as a moderator of the
relationship between reference group norms and one’s own smoking.
Study II assesses SNSs for smoking cessation. It is important to note that
health-based SNSs differ from traditional online social support groups in several ways.
First, the main purpose of SNSs is to help members develop their social networks and
15
establish an online presence, as opposed to online communities which mainly allow
people to improve their understanding on a particular topic (Ahn, Han, Kwak, Moon &
Jeong, 2007; Rau, Gao & Ding, 2008). Second, relationships on SNSs are more direct,
visible and interpersonal than traditional online communities. This is because users are
interconnected with other members through “friend requests,” and therefore explicitly
state their relationship with others. As such, “connections come before content,” and not
the other way round, according to Rau et al. (2008). Also, because of the less anonymous
nature of SNSs, users are more likely to define their own identities in an authentic way
than in text-based online communities. Third, users of SNSs are connected in “bottom-
up” networks rather than the hierarchical, “top down” structure inherent in online
communities. According to Rau et al. (2008), SNSs are “user-centric, context-driven,
user-controlled and self-organizing,” whereas online communities are “moderator-
controlled, topic-driven, centralized and architected” (p. 2759).
SNSs are therefore seen as being more utilitarian, and the nature of the
relationships formed and sustained on the sites can lead to more social bridging and
bonding, and act as more of a substitute or supplement to people’s offline social networks
than traditional online communities. These differences mean that SNSs for smoking
cessation may have different implications for their users, compared to traditional
information-based online communities, since the sites fulfill different support needs
(social and emotional, rather than informational), and hence, their effect on smoking
cessation and subsequent abstinence may be different. Study II posited that these features
of SNSs allowed for the development of four social variables which contributed to
16
smoking cessation for active members: social identification, social norms, social capital
and social support.
Social Norms Approach
The Social Norms Approach (Perkins & Berkowitz, 1986) dates to research on
alcohol use by college students. Perkins and Berkowitz (1986) found that students tended
to have more moderate alcohol consumption habits, but misperceived how much alcohol
they thought others consumed. It is this degree of discrepancy between students’ own
consumption and the perceived campus norm which strongly predicts drinking. Perkins
and Berkowitz (1986) argued that a correction of misconceptions regarding group norms
will result in decreased problem behavior and increased healthy behavior. There is strong
and growing body of literature supporting a positive outcome on healthy behavior if
discrepancies are reduced (Borsari & Carey, 2003; Cialdini, Reno & Kallgren, 1990;
Elek, Miller-Day & Hecht, 2006; Gunther & Storey, 2003; Iannotti & Bush, 1992;
Iannotti, Bush & Weinfurt, 1996; Kallgren, Reno & Cialdini, 2000). Particularly, two
types of norms have been identified as predictors of behavioral change: injunctive and
descriptive norms (Perkins & Berkowitz, 1986). Injunctive norms are group members’
feelings about what ought to be done, based on morals and beliefs, and characterize
approval or disapproval of a behavior. Meanwhile, descriptive norms are group
members’ beliefs about what other group members actually do.
Over the years, researchers have developed theories concerning the impact of
Social Norms through the addition of new variables which affect the relationship between
peer norms and behavior. These include: peer proximity, social distance, peer
17
conversations, identification and sensation seeking. For peer proximity, Yanovitzky,
Stewart and Lederman (2006) found that perceived alcohol use by best friends (socially-
equivalent proximal peers) was highly predictive of own alcohol use. Similarly, Lewis
and Neighbors (2006) found that good friends (proximal peers) have a stronger influence
on behavior than typical students (distal peers). Paek and Gunther (2007) also found that
the perceived influence of anti-smoking messages on proximal peers had a stronger
positive effect on students’ own smoking than perceived influence on distal peers.
For social distance, Borsari and Carey (2003) found that misperceptions increase
as social distance increases, as adolescents misperceived the norms of the typical student
more than those of their close friends. The importance of peer conversation was
highlighted by Rimal and Real (2003), who found that peer communication about alcohol
use occurred among students who perceived positive benefits to alcohol use, and this
directly predicted alcohol use. More recently, Real and Rimal (2007) found that the more
students talked to their peers about alcohol use, the greater their estimation of prevalence
of normative alcohol consumption in their peer groups, hence increasing their own use.
Social identification with peer groups is another variable which influences the
relationship between social norms and behavior. Social Identity Theory (SIT) (Tajfel &
Turner, 1986), which will be discussed in greater detail in the next section, states that
one’s self-concept derives from knowledge of membership in social groups, and the
membership has value and emotional significance. Young people identify with their peer
groups, often looking to these groups for normative information regarding appropriate
social behaviors, and therefore, their attitudes and behaviors can be shaped by social
18
identification with, and perceived social norms of their peers. Individual group members
inculcate the values and emotions of the social group, defining themselves in terms of
membership in the social group, and creating a shared social identity. As group identity
becomes salient, group members experience positive distinctiveness by conforming to
group norms. Reed, Lange, Ketchie and Clapp (2007) found that stronger identification
with reference groups (friends, peers and Greek members) was associated with heavy
drinking, and among students who identified strongly with a reference group, perceptions
of heavy drinking acceptability was associated with greater heavy drinking.
Identification with peer groups also had a strong moderating influence on injunctive
norms and alcohol use.
Study I proposes the reference group perspective to smoking cessation in an
offline context by applying the Social Norms Approach (Perkins & Berkowitz, 1986),
and Social Identity Theory (SIT) (Tajfel & Turner, 1986), hypothesizing that the
injunctive and descriptive norms of one’s reference groups have a direct impact on one’s
attitude towards smoking and smoking cessation self-efficacy. Additionally, the effect of
reference groups’ descriptive and injunctive norms on attitude towards smoking and
smoking cessation self-efficacy is also moderated by one’s social identification with the
members of these reference groups.
Figure 1 presents the Reference Group Perspective to Smoking Cessation. As can
be seen, in conventional, offline smoking cessation interventions, the health care
professional (doctor) provides a treatment to the patient (smoker), resulting in a direct
health outcome (smoking cessation). Study one proposes an indirect health outcome that
19
can be harnessed through the patient’s (smoker) social ties to his or her reference groups.
Approval or disapproval (injunctive norm) of smoking, and actual smoking (descriptive
norm) among members of the reference group, as well as the patient’s social
identification with the reference group, will have an indirect health outcome on the
patient, by impacting his or her attitude towards smoking, and smoking cessation self-
efficacy. More negative attitude towards smoking, and higher smoking cessation self-
efficacy, will in turn lead back to the direct health outcome, smoking cessation.
Figure 1: Reference Group Perspective for Smoking Cessation
20
Differences between Social Norms and Social Influence
The Social Norms approach to smoking cessation used in Study I differs from the
Social Influence approach, which looks at a person’s social network, focusing on patterns
of friendships among members of the social system, inferring how diffusion and behavior
uptake occur based on such concepts as social position, friendship links, and exposure
(Rogers, 2003; Valente, 1995). By emphasizing patterns of relationships, the social
influence approach asserts that social groups encourage both positive and negative health
habits in reference to people’s positions within social networks (Wellman & Berkowitz,
1988). Members of a social group may pick up smoking at different times, and as the
number of smoking peers a person has increases, so does his or her likelihood of picking
up the habit. In their study of the Framingham, Massachusetts community, Christakis
and Fowler (2008) reported that people quit smoking together over time. Those who
were connected to others who had quit smoking were also significantly more likely to
quit. They, in turn, influenced other people they were connected to, until the behavior
spread throughout the network (Szabo, 2008). According to Christakis and Fowler
(2008), within the Framingham community, smoking cessation by a spouse decreased a
person’s chance of smoking by 67%, smoking cessation by a sibling decreased the chance
by 25%, smoking cessation by a friend decreased the chance by 34%, and smoking
cessation by a co-worker decreased the chance by 34%. This study suggests that groups
of interconnected people can stop smoking together by influencing each other, so
interventions can make use of this phenomenon to reduce and prevent smoking.
21
Many research studies applying the Social Influence approach to adolescent
smoking behavior found that popular students, i.e. those who occupy central positions
within their social networks, are significantly more likely to set trends regarding the
acceptability or unacceptability of smoking, and therefore can be harnessed to improve
health outcomes (Alexander, Piazza, Mekos & Valente, 2001; Ennett & Bauman, 1993;
Valente, Chou & Pentz, 2007; Valente, Hoffman, Ritt-Olson, Lichtman & Johnson, 2003;
Valente, Unger & Johnson, 2005). Many studies also found that adolescents associate
with others like themselves, and this tendency for homophily leads to the uptake of
smoking within adolescent social groups (Ennett, Bauman & Koch, 1994; Hoffman,
Sussman, Unger & Valente, 2006; Phua, 2010). A major source of debate in the
literature on peer influence is whether adolescents are influenced by their peers, or select
friends based on similar interests and behaviors. Peer influence occurs when one’s
connections to peers influence one’s own attitudes and behavior, while peer selection
occurs when one’s attitudes and behavior influences one’s selection of friends (Hoffman,
Monge, Chou & Valente, 2007). Several studies have looked at peer influence and peer
selection, and found that they work concurrently in affecting adolescents’ health
behaviors (Aloise-Young, Graham & Hansen, 1994; Hall & Valente, 2007; Hoffman et
al., 2007; Jones, Schroeder & Moolchan, 2004; Kirke, 2004; Ritt-Olson et al., 2005),
suggesting that interventions should take both peer influence and peer selection into
account, encouraging teens to select friends on healthier interests.
22
Social Identity Theory
Study I hypothesizes that injunctive and descriptive norms of reference groups
will significantly impact smoking cessation self-efficacy, while social identification with
the groups moderates the relationships between injunctive/descriptive norms and
smoking cessation self-efficacy. Study II hypothesizes that participation in SNSs for
smoking cessation significantly increases social identification, which in turn predicts
smoking cessation self-efficacy. Social Identity is defined as the part of an individual’s
self-concept which derives from the knowledge of his or her membership in a social
group, together with the emotional and value significance attached to this membership
(Tajfel & Turner, 1986). Social Identity Theory (SIT) states that human behavior ranges
on a spectrum from purely interpersonal to purely intergroup (Tajfel, Billig, Bundy &
Flament, 1971; Tajfel & Turner, 1986). In an interpersonal interaction, people relate to
others as individuals, with no concept of social categories, while in an intergroup
interaction, people relate to others as representatives of the group, with no individualizing
characteristics. When social identity is salient, the psychological separation between the
self and the group as a whole generally disappears, resulting in individuals seeing
themselves as interchangeable exemplars of their social group.
The three main principles of SIT are: people strive to achieve and maintain a
positive social identity, the positive social identity is achieved by favorable comparisons
with and positive differentiation from a relevant out-group, and when one’s social
identity is unsatisfactory, he or she will employ self-enhancement strategies which serve
to help him or her achieve a positive self-identity. The central motivation for intergroup
23
behavior is one’s desire for a positive and secure self-concept, and this is formed and
maintained through evaluations of the in-group with reference to relevant out-groups
(Tajfel & Turner, 1986). Individuals are motivated to act and think in ways that achieve
positive distinctiveness for the in-group, leading to high levels of intergroup
differentiation and outgroup derogation. The three conditions necessary for intergroup
differentiation are: group membership must be internalized, the social situation must
allow for intergroup comparison, and the out-group is relevant for intergroup comparison
(Hogg, 2003).
SIT conceptualized group members’ motivation for self-enhancement as the need
for positive self-esteem, which leads to the achievement and maintenance of a positive
social identity (Abrams & Hogg, 1988). Additionally, three self-enhancement strategies
are used by group members: social mobility, social change and social creativity, in
reaction to a threatened social identity. Social mobility involves dis-identifying with and
leaving one’s low-status group, and joining the higher-status group, social change
involves direct collective action or intergroup competition to overturn the existing
intergroup hierarchy, while social creativity involves changing the dimension of
comparison, so as to improve the evaluative consequence for one’s current group. In
1987, Turner et al. proposed the Self-Categorization Theory (SCT), adding cognitive
elements, and elaborating on intra-group processes in addition to intergroup processes.
SCT differs from SIT by characterizing identity as operating on three levels of
inclusiveness (super-ordinate, intermediate and subordinate) which are determined by
accessibility and fit, and also elaborates on the depersonalization process, stating that
24
people cognitively represent members of social groups as prototypes (Hogg, 2003; Turner
et al., 1987). In this dissertation, social identification with reference groups is predicted
to significantly influence one’s attitude towards smoking, and smoking cessation self-
efficacy (Study one). Social identification, along with social capital, social support and
social norms, is also predicted to be one of the outcomes of participation on social
networking sites for smoking cessation (Study II). The following sections review
theories used for Study II: Social capital (Putnam, 2000; Williams, 2006) and Social
support (Tanis, 2008), as well as lurkers in online communities, and previous studies
about online smoking cessation programs.
Online Social Capital
Study II of the dissertation measures online social capital, hypothesizing that
participation on SNSs for smoking cessation significantly increases social capital, which
in turn predicts smoking cessation self-efficacy. Social capital is derived from people’s
interactions within social relationships (Bourdieu, 1986; Burt, 1992; Coleman, 1988;
Jones, 1996; Lin, 1999; Putnam, 2000; Valenzuela et al., 2009; Williams, 2006,
Woolcock, 1998). An early definition was offered by Bourdieu (1986), who
conceptualized it as resources made possible by formal and informal networks and social
relationships. Similarly, Burt (1992) characterizes social capital as contact points
through which one receives opportunities to use their human and financial capital. 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)
25
conceptualizes social capital as a set of relationships among people that facilitate
productive activity.
Meanwhile, Putnam (2000) defines social capital as a system of trust and
reciprocity within social relationships. The advent of new technologies such as the
Internet has resulted in further debate about social capital formation and maintenance in
the information age. The utopian perspective argues that the Internet allows people to
extend their social lives, giving them more opportunities to socialize with existing offline
networks, and also meet new friends online, while the dystopian perspective argues that
the Internet takes time away from offline social activities, causing people to become
socially isolated. For example, Kraut et al. (1996) found that greater internet use
displaced offline social activities and substituted strong offline ties for weak Internet ties.
In a follow-up study, Kraut et al. (2002) found that Internet use significantly decreased
depression, and increased social circles and face-to-face communication. Similarly,
Wellman, Quan-Haase, Witte and Hampton (2001) found that the internet increased
social capital, allowing people with same interests to meet and contact geographically
disperse friends, but also decreased social capital because online ties were weaker, time
was taken away from offline social activities, and participants reported increased
depression and loneliness. Resnick (2001) argued the Internet increases social capital in
the civic arena by reducing costs of coordinating and publicizing activities, and acting as
a new space for public deliberation, thereby producing new forms of socio-technical
capital, or productive combinations of social relations and ICTs. The Internet also
26
enabled people to seek and receive help from diverse social networks and resources, a
process known as networked individualism (Boase, Horrigan, Wellman & Rainie, 2006).
Two further research paradigms for online social capital are the “Rich get richer”
hypothesis (Kavanaugh et al., 2005; Kraut et al., 2002) and the displacement hypothesis
(Nie & Hillygus, 2002). The former states that extroverts and those with greater offline
social support reap more benefits from the Internet than introverts, while the latter states
that the Internet takes away time spent on face-to-face social activities. The location and
timing of Internet use can also determine how social relations are affected (Nie &
Hillygusm 2002; Shklovski, Kiesler & Kraut, 2006). Cyber-balkanization, the formation
of homogeneous online groups based on interests or specialization, can also affect online
social capital (Papacharissi, 2002; Van Alstyne & Brynjolffson, 2005; Williams, 2007).
Two main types of social capital, bridging and bonding, have been identified
(Putnam, 2000; Williams, 2006). Bridging social capital refers to loose, informal ties
which may not provide emotional support, but increase opportunities for new ideas and
information. Bonding social capital refers to close relationships which provide high
levels of trust, social and emotional support. Closely related to bridging and bonding
social capital are the concepts of weak ties and strong ties. Weak ties are heterogeneous,
diverse and have breadth, allowing new information to diffuse into social networks (Burt,
1992; Granovetter, 1983), while strong ties are homogeneous, similar and have depth,
facilitating local cohesion, high trust and support. Weak ties are important because they
act as bridges allowing for diffusion to take place across group boundaries, providing
new information, diversity and resources critical for survival (Lin, 1990).
27
Online groups allow for formation of both bridging and bonding social capital.
Bridging social capital is created through the development of weak ties in online message
boards, forums and social networking sites, while bonding social capital is formed
through the maintenance of strong ties by email and IM (Haythornthwaite, 2002; Norris,
2002). The Internet Social Capital Scale (ISCS) by Williams (2006) included criteria for
bridging and bonding social capital. For bridging social capital, questions accessed
outward-looking, contact with a broad range of people, view of oneself as part of a larger
group, and diffuse reciprocity with the surrounding community. For bonding social
capital, questions accessed emotional support, links to scarce and limited resources,
ability to mobilize solidarity, and outgroup antagonism. The Internet currently provides a
wide variety of sites and applications for different purposes. For example, online forums,
message boards and chat-rooms may be based around common interests and hobbies, and
therefore will attract people from diverse backgrounds and locations. Users of these
types of sites will be more likely to build weak ties, hence increasing their bridging social
capital. On the other hand, other features of the Internet, including email, IM and social
networking sites, are used by people to keep in contact with existing offline friends and
family, and thereby they may help build strong ties, hence increasing their bonding social
capital. Therefore, depending on features selected (chat, forum, email, social
networking), there are different implications for bridging and bonding, as well as online
and off line social capital (Haythornthwaite, 2002; Williams, 2007).
28
Social Networking Sites and Online Social Capital
Social networking sites (SNSs) for smoking cessation are the focus of Study II.
boyd and Ellison (2007) define SNSs as “web-based services that allow individuals to
construct a public or semi-public profile within a bounded system, articulate a list of
other users with whom they share a connection, and view and traverse their list of
connections and those made by others within the system” (p. 211). Online profiles act as
a user’s home page on the site, where he or she is able to articulate his or her identity
through pictures, personal information, hobbies, favorite music and other information.
Members can also link up with friends and acquaintances and add them to their networks.
They may then view each other’s profiles, post comments and receive updates about
people in their network. Because SNSs allow people to connect with new people online,
as well as close friends and family, they lead to formation and maintenance of bridging
and bonding social capital. It is estimated that in the United States, over 88.1 million
adults over the age of 18 regularly use SNSs, and this number will increase to 115 million
by 2013 (Walsh, 2009).
Previous research has found that SNSs allow for the formation and maintenance
of online bridging and bonding social capital (Ellison, Steinfield & Lampe, 2007; Lampe,
Ellison & Steinfield, 2006; Zywica & Danowski, 2009). In their survey of MSU
students, Ellison et al. (2007) found that intensive Facebook use was significantly
associated with bonding, bridging and high school (maintained) social capital. Students
used the site mainly to contact offline friends, connect with classmates, and maintain
relationships with family and high school friends back home, with those who used
29
Facebook more intensively experiencing greater bridging social capital and engagement
with the college. Students with lower self-esteem and satisfaction with life also gained in
bridging social capital through Facebook use. SNSs can also foster political and civic
participation, with bridging social capital facilitating mobilization of citizens, and
bonding social capital encouraging traditional political engagement through strengthening
of offline relationships (Skoric, Ying & Ng, 2009; Valenzuela, Park & Kee, 2009)
Study II of the dissertation assesses online bridging and bonding social capital
formed through participation in SNSs for smoking cessation, hypothesizing that social
capital would be positively associated with higher smoking cessation self-efficacy. SNSs
can be a potentially useful medium in the health care arena as they enable patients and
caregivers to connect with friends, relatives and other people from their offline lives, so
as to receive social support, thereby helping to build bonding social capital. At the same
time, SNSs may also enable patients to connect to others with the same health conditions,
but from diverse geographic locations and backgrounds, leading to the formation of weak
ties, thereby facilitating diffusion of new information, and the formation of bridging
social capital. SNSs extend the efficacy of traditional online social support groups due to
greater interactivity, and features which encourage the formation of relationships online,
as well as its eventual movement offline.
Online Social Support Groups
Study II of the dissertation measures social support received from SNSs for
smoking cessation, hypothesizing that participation on the sites significantly increases
social support, which in turn predicts smoking cessation self-efficacy. The Internet has
30
become an indispensible medium for people dealing with health issues, who seek
information about their health conditions and social support from others with the same
health conditions. In a Pew Internet and American Life Project report, Jones and Fox
(2009) report that in 2008, over 40 million adults in the United States use the Internet as
their primary source of information about health, and that over 84% of all Americans
over the age of 18 have searched for health and medical information online or
participated in online health support groups. These online groups have also been
identified in previous studies as one of the primary methods for online health information
seeking as well as for social support (Cotton & Gupta, 2004; Kummervold, Gammon,
Bergvik, Johnsen, Hasvold & Rosenvinge, 2002; Preece, 20000). Social support is “the
communication between recipients and providers that reduces uncertainty about the
situation, the self, the other or the relationships and functions to enhance a perception of
personal control in one’s life experience” (Tanis, 2008, p. 292). There are three main
types of social support: instrumental support, informational support and emotional
support. Instrumental support refers to the providing of goods and services, and giving
practical assistance with daily living. Informational support is the exchange of practical
information which expands people’s knowledge about their health conditions, and
reduces uncertainty. Emotional support is the display of understanding about what other
people are going through, including compassion, sympathy and solidarity.
Online groups can foster patients’ empowerment through providing support which
increases their ability to act efficaciously by accessing relevant medical sources, exercise
critical thinking regarding their health conditions, and exert control over their illnesses so
31
as to bring about desired results (Barak, Bonel-Nissim & Suler, 2008; Barnett & Hwang,
2006; Lasker, Sogolow & Sharim, 2005; Sharf, 1997; White & Dorman, 2001; Wright,
2002). Additionally, they also facilitate online friendships with high levels of intimacy,
self-disclosure, and psychological well-being, which eventually move offline (Hoybye,
Johansen & Tjornhoj-Thomsen, 2005; Rodgers & Chen, 2005; Schweizer, Leimeister &
Krcmar, 2006; Shaw et al., 2006). Online social support groups are also particularly
helpful for those who are physically immobile, as they can access the group from
anywhere (Braithwaite & Waldron, 1999; Cummings, Sproull & Kiesler, 2002; Finn,
1999). For people with stigmatized illnesses, online support groups also enable them to
get in touch with people with similar conditions online so they can talk about subjects
and topics which may be considered taboo to others offline (Houston, Cooper & Ford,
2002; Kral, 2006; Lapinski, 2006; Salem, Bogar & Reid, 1997).
Many difficulties associated with traditional, face-to-face social support are
mitigated by online social support groups, including mismatching of caregivers with
recipients (Turner, Grube & Meyers, 2001; Walther & Boyd, 2002), high levels of stress
and emotional strain experienced by caregivers (Sarason & Sarason, 2006), caregivers’
withholding of honest feedback so as not to hurt the patient (Turner et al., 2001), and
threat to patient self-esteem through seeking help (Walter & Boyd, 2002; Sarason &
Sarason, 2006). For patients with no direct access to their family members, fictive kin, or
family friends, may be called upon to become caregivers (Jordan-Marsh & Harden,
2005). Offline social support can backfire because well-intentioned supportive efforts
32
have unexpected negative effects on both givers and receivers, and may end up
worsening patients’ health conditions (Sarason & Sarason, 2006).
Online social support groups offer many advantages over face-to-face groups,
including the ability for people to read and reply to messages asynchronously (Wright &
Bell, 2003), transcendence of temporal and geographic constraints (Papacharissi, 2002),
the valuing of people based on contributions rather than salient physical characteristics
(Tanis, 2008), therapeutic value of disclosing traumatic experiences (Braithwaite et al.,
1999), fostering of greater self-disclosure and high levels of intimacy (Drentea & Moren-
Cross, 2005; Walther & Boyd, 2002), allowance for people with shy personalities to
participate (Rains & Young, 2007), low costs and high storage capacity (Papacharissi,
2002). There are also several disadvantages. For example, patients may find it difficult
to sort through inaccurate, misleading information (Rains & Young, 2007). Also, a
digital divide continues, with participants of online support groups being of higher SES,
younger and more educated (Cotton & Gupta, 2004; Wright, 2002). There is also no
feeling of immediacy (Barak et al., 2008), and many privacy and security concerns
(Preece & Maloney-Krichmar, 2005). The disinhibiting nature of online groups
encourages hostile behaviors like flaming, trolling, and lurking (Tanis, 2008), while the
lack of cues fosters deceptive identities (Wright & Bell, 2003).
Lurkers in Online Communities
Study II of the dissertation hypothesizes that active participants and lurkers on
SNSs for smoking cessation derive significantly different levels of social identification,
social support, social norms and social capital from the sites, and therefore, there will be
33
a significant difference between these two groups for smoking cessation self-efficacy.
Lurkers are members of online communities who only read messages without posting or
participating in any discussions, and do not actively give support to other members
(Nonnecke & Preece, 2000; White & Dorman, 2001). Studies have found that lurking is
one of the most commonly cited problems in online support groups, with the majority of
users in online communities being lurkers (Katz, 1998; Kollock & Smith, 1996;
Nonnecke & Preece, 2000). The lack of activity in many online communities presents a
major problem because without participation, they are not able to provide their members
with needed services, including the giving and receiving of social support, and many
groups, therefore, cannot be sustained over a long period of time (Butler, Sproull, Kiesler
& Kraut, 2002; Ramirez, Zhang, McGrew & Lin, 2007). This is especially worrisome for
online health communities because important treatments and problems may not be
discussed, depriving members of much-needed help (Bishop, 2007).
Encouraging lurkers to actively participate and post messages is a major
challenge for administrators of online communities due to the difficulty of reaching them.
Even if an online community has the latest interactive technology available, the
community will not flourish without participation. Research has indicated that lurking is
a post-joining adaptive strategy, with many lurkers being “shy” and “introverted”
individuals (Nonnecke, Preece & Andrews, 2004). Even though lurkers felt like they
were part of the community, active posters had a significantly greater feeling of solidarity
with the online group (Preece et al., 2004). Lurking, according to Katz (1998), allows
people to have the best of both worlds because they can access information and other
34
people’s insights and experiences while bypassing flaming, insults and abuse. In order to
attract lurkers to post, researchers suggest strategies including smaller group activities,
stronger moderation and privacy protection (Nonnecke & Preece, 2003; Nonnecke et al.,
2004), as well as trustworthiness ratings, increasing member credibility, and positive
comments (Bishop, 2007)
Research has also found that lurking can be advantageous for online groups
because indiscriminate posting, including spam, flaming and trolling, can cause groups to
be less efficient, taking more time for members to sift through unwanted material
(Kollock & Smith, 1996). Lurking also benefited online groups as it allowed new
members to learn group norms before posting messages (Preece et al., 2004). Studies
also found that for most lurkers, reading other people’s posts and identifying with their
experiences can lead to positive health outcomes; therefore, there would be need for them
to post to benefit from the online groups (Grohl, 2008; Van Udern-Kraan et al., 2008). In
line with these findings, Study II of the dissertation sought to assess whether active
participants and lurkers in SNSs for smoking cessation derived significantly different
social benefits from the sites.
Rau et al. (2008) postulated that factors influencing SNS users’ public postings
were different from those influencing participation in traditional online communities.
Because SNSs are more focused on socio-emotional than informational support, the
authors suggested that the level of interpersonal intimacy (including verbal and affective
components) is related to users’ posting behavior. The study found that both verbal
intimacy and affective intimacy were positively correlated with posting frequency.
35
Significant differences in both verbal and affective intimacy were also found between
posters and lurkers, suggesting that people become lurkers in SNSs because they feel that
their socio-emotional needs cannot be fulfilled even when they post regularly. Rau et al.
(2008) hence concluded that while in traditional online communities, lurkers did not post
due to informational support reasons, in SNSs, socio-emotional support was more
important and hence influenced SNS users’ posting decisions more. Low affective
intimacy increases relationship uncertainty and lowers the possibility of being understood
correctly, while low verbal intimacy causes people to be uncomfortable discussing
various topics and self-disclosing. In traditional online communities, people can read
messages to find the information they need, and only post when they have specific
questions on a topic, knowing that a stranger with more expertise will be able to respond.
In SNSs however, posting messages is akin to talking to friends and acquaintances, and
therefore, socio-emotional needs, including intimacy level, will influence people’s
decision to post. As such, steps should be taken to foster a more verbally open and
affective environment in SNSs, so as to encourage lurkers to participate more actively in
the community.
Online Smoking Cessation Programs
This section summarizes current research on online smoking cessation
programs. Smoking is highly addictive, and smokers who want to quit, or have already
quit, often end up relapsing within a year (Cobb & An, 2007). The main reason for this is
that there has been no smoking cessation technique which has proven to be truly effective
for long-term abstinence (Stoddard et al., 2005). Hospitals and health agencies typically
36
use two methods for smoking cessation: face-to-face programs; and self-help printed
guides and manuals. Face-to-face programs have the advantage of making available the
services of health professionals and social support from other smokers. However, they
are often very time-consuming, require extensive travel time, especially for people living
further away from hospitals, and can also be very difficult to schedule. Self-help guides,
meanwhile, are expensive due to the high costs of printing and mailing, the constant need
to update information, and lack of tailoring for individuals (Stoddard et al., 2005).
The Internet represents a viable avenue for smoking cessation programs for
many reasons. The cost associated with travel and scheduling is virtually eliminated
since the Internet can be accessed at any time of the day from anywhere. Also, self-help
guides that are posted on the Internet can be delivered for little to no cost, and can be
easily updated. In addition, information on smoking cessation can be tailored to each
individual smoker. Further, social support is readily available in the form of other
Internet users, even after the program has ended. This is especially useful for ex-smokers
who wish to maintain their abstinence from cigarettes and need some quick
encouragement from other individuals in similar situations. Usage data and cessation
rates can also be easily accessed by health agencies and hospitals for evaluation purposes.
Online smoking cessation programs can also reach segments of the population who are
unresponsive to traditional offline programs.
For example, Graham et al. (2005) found that the majority of people who
enrolled in an online smoking cessation trial (QuitNet.com) were female (60.5%), white
(86.4%) and college-educated (48.4%). Cobb and Graham (2006) also found among
37
people who sought information about smoking cessation online, 59% of the 2265
respondents were female, and younger than the general Internet population. Current
smokers preferred information on how to quit and medications available, while former
smokers preferred information on how to cope with withdrawal, with individually-
tailored information and withdrawal information rated the most useful. Online smoking
cessation programs may therefore appeal more to younger, higher-educated and female
populations, compared to face-to-face interventions.
Online smoking cessation programs have become increasingly popular due to
higher levels of Internet access, cell phone use, and mobile web devices in most countries
around the world, and the ability of the Internet to reach out to large numbers of people at
any one time. The rise of Web 2.0 technologies, including SNSs like Facebook and
media-sharing sites like YouTube, has also allowed for user-created content which
engage the public, resulting in opportunities to discuss smoking cessation and influence
public health policies.
Many research studies conducted on online smoking cessation programs have
found positive health benefits for participants, including increased smoking cessation
self-efficacy and longer periods of abstinence (Cobb et al., 2005; Etter, 2005; Lenert et
al., 2003; Stoddard et al., 2005; Takahashi et al., 1999). An early study by Schneider and
Tooley (1986) assessed an electronic bulletin board for smoking cessation, and found that
over a 3-month period, participants increased the frequency of self-disclosure and
supportive comments, with quit rates higher than face-to-face programs. In a follow-up
study, Schneider, Walter and O’Donnell (1990) assigned 1158 smokers on a CompuServe
38
network to four groups in a 2X2 design, with half the groups receiving access to a full
cessation program or a control version, and half participating in a forum or not. Results
indicated that participants in the forum reported higher quit rates (17.7%), with the
highest cessation rate reported by the group with access to the program plus the forum.
Additionally, Takahashi et al. (1999) examined an online 60-day cessation
program (Quit Smoking Marathon) which consisted of health information, daily emails,
and a mailing list linking participants to doctors and other supporters, and found that 52%
of light smokers, and 43.4% of heavy smokers successfully quit and abstained from
cigarettes 12 months after completion. Lenert et al. (2003) evaluated an 8-week online
cessation program which included self-monitoring tools, and tailored emails timed to
participants’ stage of quitting, and found that 34% successfully quit smoking, while 50%
reported smoking less. Stoddard et al. (2005) tested a 30-day online program in which
participants answered questionnaires measuring nicotine dependence, depressive
symptomatology and menstrual distress, and were given individually tailored information
based on their responses. Results indicated that at the 30-day follow-up, 8.3% of
participants reported abstinence for 7 days or more, while 40% made a serious attempt to
quit smoking. Cobb et al. (2005) studied QuitNet.com, one of the first online multi-
modal cessation programs centered on a support community, and found that social
support was significantly associated with higher abstinence. Etter and Perneger (2001)
had their participants complete a baseline questionnaire, after which they received
personalized counseling letters, followed by monthly email reminders. Results indicated
that 10.9% of baseline current smokers, and 25.2% of baseline former smokers
39
successfully quit. The studies described show that individually tailored information,
combined with encouragement and social support from other members, contributed to
high quit rates for current and former smokers (Cobb et al., 2005; Etter & Perneger, 2001;
Takahashi et al., 1999).
Many recent studies also found that integrated online & mobile phone-delivered
programs can be highly efficacious. Brendryen, Drozd and Kraft (2008) evaluated Happy
Ending (HE), a 1-year cessation program delivered via email, text messages, and
interactive voice responses, finding that the treatment group reported significantly higher
abstinence and self-efficacy. Finkelstein, Lapshin and Cha (2008) tested a tablet PC-
delivered interactive cessation educational module on a group of 35 former methadone
users, and found that module use significantly increased scores on the Hazards of
Smoking Knowledge Survey (HSKS). Houston et al. (2008) tested an online intervention
to increase implementation of provider advice for cessation in dental practices, and found
that practices which used the intervention improved their performance on asking patients
about smoking (4%) and advising patients about quitting (11%). Rabius, Pike, Wiatrek
and McAlister (2008) conducted a study in which smokers were assigned to 5 tailored
interactive cessation sites or one minimally interactive site, and found that the more
interactive sites were associated with higher quit rates at a 13-month follow-up. Saul et
al. (2007) evaluated Quitplan.com, which included social support features, chat,
counseling, expert information and tailored emails, and found a 13.2% quit rate at the 6-
month follow-up. Stretcher et al. (2008) accessed engagement (web sections opened) in
an online cessation site, finding that higher engagement significantly predicted smoking
40
cessation, with each section opened predicting 18% higher likelihood of quitting.
Whittaker et al. (2008) pilot-tested a multimedia cessation mobile phone-delivered
platform among teen smokers in New Zealand, with features like interactive forums,
blogs and animations, and found that out of fifteen teens studied, nine stopped smoking
during the program. Zbikowski, Hapgood, Barnwell and McAfee (2008) examined the
Free and Clear Quit for Life Program, in which 11, 143 participants accessed an
interactive site, phone counseling and tailored emails, and found that utilization of online
plus phone components was most significantly associated with quitting.
Most researchers agree, however, that several issues concerning online smoking
cessation programs need to be resolved before they can substitute for face-to-face
cessation programs (Balmford et al., 2008; Cunningham, 2008; McKay et al., 2008;
Stoddard, Augustson & Moser, 2008). A telephone survey of 8467 Canadian adults by
Cunningham (2008) found that smokers (74%) were less likely than non-smokers (81%)
to be Internet users, with only 46% of Internet-using smokers interested in online
cessation programs. Access and interest are hence issues to consider regarding online
programs. Balmford et al. (2008) assessed an Australian online program, QuitCoach,
and found that despite individually tailored advice and email reminders, only 27% of
users returned for a second visit. McKay et al. (2008) compared participants in two web-
based cessation programs: QSN (Quit Smoking Network), which included engaging
multimedia components, and Active Lives, which was less interactive, and found that
while QSN users spent more time on the site than Active Lives users, no difference in
user engagement was found. Both studies concluded that in order for online cessation
41
programs to be more effective, measures should be taken to improve participant
engagement on the sites.
Another issue with online cessation programs is that more research should be
done to determine which features of the sites, such as tailored emails, message boards,
forums, self-help tools, informational guides, interactive elements and social networking
features, are most effective for smoking cessation, and what target population each of
them most appeals to (Bock, Graham, Whiteley & Stoddard, 2008; Etter, 2006). Because
Internet use is so prevalent, many new online cessation support groups are formed, but
also abandoned every day. These groups offer similar content and services to users, and
most have large numbers of lurkers and low participation rates. Bock, Graham, Whiteley
and Stoddard (2008) conducted a review of 2008 versus 2004 online cessation programs,
and found that 2008 sites included higher-quality advice and more personally-relevant
counseling, with interactive sites increasing from 39% in 2004 to 56% in 2008. In their
study of Quitplan.com, An et al. (2008) found that the most widely-used resources among
607 participants were interactive tools and community features, which significantly
predicted increased abstinence.
Smoking Cessation Self-Efficacy
Self-efficacy, with respect to addictive behavior, refers to one’s ability to resist
the temptation to engage in a certain behavior, rather than one’s ability to perform a task
(Young, Oei & Crook, 1991). According to Bandura (2002), self-efficacy is the
foundation of human agency, because unless people believe they can produce a desired
effect and inhibit an undesired one by their actions, they have little incentive to act.
42
Three distinct dimensions of self-efficacy have been identified, based on an individual’s
confidence in performing a set of graded behaviors (Young et al., 1991). Magnitude of
self-efficacy refers to the ordering of tasks by difficulty level (Bandura, 2002). People
with low-magnitude self-efficacy believe they can only perform the simpler of a series of
tasks, while people with high-magnitude self-efficacy believe they can perform the more
difficult tasks (Stretcher, DeVillis, Becker & Rosenstock, 1986). Strength of self-
efficacy refers to the confidence level of an individual in performing a specific task,
while generality of self-efficacy refers to the extent to which self-efficacy can be
generalized across situations (Bandura, 2002). An individual’s self-efficacy is dependent
on the particular task or context which he or she is currently confronting. It influences all
aspects of behavior, including the acquisition of new behaviors, inhibition of existing
behaviors, as well as disinhibition of behaviors (Stretcher et al., 1986). Over the years,
Bandura’s (2002) construct of self-efficacy has been used in the development of various
health behavioral interventions, including smoking cessation (Stretcher et al., 1985),
weight control (Chambliss & Murray, 1979), alcohol (Young et al., 1991) and exercise
(Weinberg, Gould & Jackson, 1979). In most cases, level of self-efficacy has been found
to correlate strongly with one’s ability to engage in health behavior change.
Previous research has demonstrated a positive association between self-efficacy
and smoking cessation behaviors. DiClemente (1981) analyzed the relationship between
self-efficacy and subjects’ ability to maintain post-treatment abstinence from smoking at
a 5-month follow-up, and found that subjects who maintained nonsmoking behaviors
received higher self-efficacy scores than recidivists. Godding and Glasgow (1985) found
43
that efficacy beliefs to resist temptation to smoke predicted reduction in the number of
cigarettes smoked, the amount of tobacco per smoke, and the nicotine content. Nicki,
Remington and MacDonald (1985) found that self-efficacy training, in combination with
nicotine fading, was significantly more effectively than nicotine fading alone in
predicting smoking cessation. Mudde, Kok and Stretcher (1989) found that self-efficacy
beliefs increased after a smoking cessation treatment, and participants who acquired the
highest levels of self-efficacy were successful in maintaining their non-smoking behavior
for a longer period of time. Stacy, Sussman, Dent, Burton and Flay (1992) found that
perceived self-efficacy mediates peer social influence on smoking in teenagers. Conrad,
Flay and Hill (1992) found that the combination of peer pressure and low self-efficacy
predicted the onset of smoking and substance use in adolescents. Kowalski (1997) found
that smokers who received self-help smoking cessation materials scored higher in self-
efficacy, and made plans to quit smoking more often than those who did not. Dijkstra,
De Vries and Roijackers (1998) found that self-efficacy enhancing information had a
significant effect on predicting quitting behavior. Shiffman et al. (2000) found that low
smoking cessation self-efficacy was highly predictive of smoking relapse after a quit
attempt. Ockene et al. (2000) also found that low self-efficacy was a strong predictor of
smoking relapse in recent quitters. Etter and Perninger (2001) found that participation in
a computer-tailored smoking cessation program was positively associated with self-
efficacy and quitting rates. Stretcher et al. (2005) found that participants in web-based
smoking cessation programs reported higher self-efficacy than face-to-face programs, and
this was associated with higher continuous abstinence rates. All of these studies suggest
44
that self-efficacy is a strong predictor of smoking cessation, and maintenance of a smoke-
free lifestyle.
Study II of the dissertation examines SNSs for smoking cessation,
hypothesizing that participation in SNSs significantly influences social identification,
social support, bridging and bonding social capital, and social norms, which in turn lead
to more negative attitude towards smoking, and higher smoking cessation self-efficacy.
This indirect health outcome then contributes to the direct health outcome of quitting
smoking.
Research Questions and Hypotheses
Study I
Study I elaborated upon Cobb’s (2007) expanded perspective on medical care,
which posited that patients’ social ties can bring about collateral health outcomes which
add value to direct health outcomes that are provided by the health care given by doctors.
The Reference Group Perspective on smoking cessation proposed in Study I hypothesizes
that perceived injunctive and descriptive norms of one’s reference groups (best friends,
colleagues, family members) would predict respondents’ smoking cessation self-efficacy,
especially among respondents who identified strongly with the groups constituting
sources of social norms.
Reference Group One: Best Friends
H1a: Best friends’ injunctive norms towards smoking will be positively associated
with respondents’ smoking cessation self-efficacy.
45
H1b: Best friends’ descriptive norms towards smoking will be positively
associated with respondents’ smoking cessation self-efficacy.
H1c: Social identification with best friends will be positively associated with
respondents’ smoking cessation self-efficacy.
H1d: Respondents’ attitude towards smoking will be positively associated with
respondents’ smoking cessation self-efficacy.
H1e: Social identification with best friends will moderate the effect of best
friends’ injunctive norms towards smoking and best friends’ descriptive norms
towards smoking on respondents’ smoking cessation self-efficacy.
Reference Group Two: Colleagues
H2a: Colleagues’ injunctive norms towards smoking will be positively associated
with respondents’ smoking cessation self-efficacy.
H2b: Colleagues’ descriptive norms towards smoking will be positively
associated with respondents’ smoking cessation self-efficacy.
H2c: Social identification with colleagues will be positively associated with
respondents’ smoking cessation self-efficacy.
H2d: Respondents’ attitude towards smoking will be positively associated with
respondents’ smoking cessation self-efficacy.
H2e: Social identification with colleagues will moderate the effect of colleagues’
injunctive norms towards smoking and colleagues’ descriptive norms towards
smoking on respondents’ smoking cessation self-efficacy.
46
Reference Group Three: Family Members
H3a: Family members’ injunctive norms towards smoking will be positively
associated with respondents’ smoking cessation self-efficacy.
H3b: Family members’ descriptive norms towards smoking will be positively
associated with respondents’ smoking cessation self-efficacy.
H3c: Social identification with family members will be positively associated with
respondents’ smoking cessation self-efficacy.
H3d: Respondents’ attitude towards smoking will be positively associated with
respondents’ smoking cessation self-efficacy.
H3e: Social identification with family members will moderate the effect of family
members’ injunctive norms towards smoking and family members’ descriptive
norms towards smoking on respondents’ smoking cessation self-efficacy.
Also, social distance of the reference group has been found to be a strong variable
predicting college students’ personal behavior (Lewis & Neighbors, 2006; Paek &
Gunther, 2007; Yanovitzky, Stewart & Lederman, 2006). Proximate peers, including
close friends and best friends, have been found to be more influential in affecting
behavior than more distant peers like the “typical” student. As such, Study I also
hypothesized that best friends’ injunctive and descriptive norms towards smoking would
be stronger predictors of respondents’ smoking cessation self-efficacy than those of distal
groups: colleagues and family members.
47
H4: Best friends’ injunctive norms towards smoking will be a stronger predictor
of smoking cessation self-efficacy than colleagues’ and family members’
injunctive norms.
H5: Best friends’ descriptive norms towards smoking will be a stronger predictor
of smoking cessation self-efficacy than colleagues’ and family members’
descriptive norms.
Additionally, Study I also sought to find out whether injunctive or descriptive
norms had a greater impact on one’s smoking cessation self-efficacy. Rimal and Real
(2003), for example, found that injunctive norms had a greater impact on reducing
alcohol consumption among students than descriptive norms. Park and Smith (2007) also
found that injunctive and descriptive norms were distinct in their impact on enrollment as
an organ donor, and engagement in family discussion about organ donation.
RQ1: Which type of norm: injunctive or descriptive, will have a stronger effect on
respondents’ smoking cessation self-efficacy (controlling for identification with
each reference group)?
Study II
Study II assessed members of SNSs for smoking cessation, hypothesizing that
social identification with other members, perceived bridging and bonding social capital,
and social support would influence the relationship between participation level on the
sites and smoking cessation self-efficacy, which is one’s ability to abstain from smoking.
Self-efficacy has been demonstrated in previous research to be significantly associated
with smoking cessation, as well as longer periods of abstinence from cigarettes
48
(DiClemente, 1981; Godding & Glasgow, 1985; Kowalski, 1997; Mudde, Kok &
Stretcher, 1989; Nicki, Remington & MacDonald, 1985). Many studies have also found
that peer influence has a significant impact on smoking cessation self-efficacy among
adolescents, with peer pressure and low self-efficacy predicting the onset of smoking and
substance use (Conrad, Flay & Hill, 1992; Stacy, Sussman, Dent, Burton & Flay, 1992).
Further, self-efficacy enhancing information, including tailored emails, pamphlets,
forums and message boards, can have a significant effect on predicting quitting behavior
(Dijkstra, De Vries & Roijackers, 1998; Ockene et al., 2000; Shiffman et al., 2000).
More recent studies have also found that participation in computer-tailored smoking
cessation programs is positively associated with smoking cessation self-efficacy, higher
quit rates and abstinence, and lower risk of relapse (Etter & Perninger, 2001; Stretcher et
al., 2005). All of these studies suggest that self-efficacy is a strong predictor of smoking
cessation, and maintenance of a smoke-free lifestyle.
H6: Participation level on the site will be positively associated with smoking
cessation self-efficacy.
Also, according to SIT (Tajfel, 1978), when people’s social identities are salient,
they experience depersonalization as group members, defining themselves in terms of a
shared identity, thereby conforming to perceived group norms. Reed et al. (2006), for
example, examined college students’ drinking behavior through their social identification
with reference groups, and found that respondents who identified strongly with peer
groups where there was high drinking acceptability were significantly more likely to
consume more alcohol. It is possible that on smoking cessation SNSs, members who
49
identified more with other members, and also conformed more to the injunctive and
descriptive norms towards smoking of other members, would have greater smoking
cessation self-efficacy.
H7: Social identification with other members will be positively associated with
smoking cessation self-efficacy.
H8: Conforming to social norms regarding smoking will be positively associated
with smoking cessation self-efficacy.
Next, utilizing an online social capital lens, this dissertation sought to find out
whether members of SNSs for smoking cessation derive social capital from participation.
The dissertation also tested whether the social capital derived from participation in these
SNSs leads to greater smoking cessation self-efficacy. Health SNSs generally include
features such as forums, message boards and chat rooms. Because they are based on
health problems, they may attract people from diverse backgrounds and geographical
locations, and therefore are more likely to build weak ties, and hence bridging social
capital. However, many of these groups are now also incorporating social networking
features, such as the ability to create profiles, connect with other members, and articulate
friendship links. According to Rau et al. (2008), people join SNSs mainly to fulfill socio-
emotional needs, including to find friendships and establish intimate interactions with
others, rather than to seek information. Because people are connected to other members
directly through friendship links, the social ties and interpersonal relationships are more
direct on SNSs than on traditional health support groups. Therefore, they can motivate
people to seek closer, more intimate and validating relationships, use the social
50
networking features to connect with existing friends and family, and even health
professionals, thereby building strong ties, and bonding social capital. This would, in
turn, bring about positive health benefits, in this case, smoking cessation self-efficacy.
H9a: Bridging social capital on the site will be positively associated with smoking
cessation self-efficacy.
H9b: Bonding social capital on the site will be positively associated with smoking
cessation self-efficacy.
Following on the social support literature, Study II also sought to find out whether social
support received from the site would be associated with smoking cessation self-efficacy.
Cobb et al. (2005) found that in QuitNet.com, a large-scale online smoking cessation
program, perceived social support received by registered members of the site was
positively associated with more than three times greater abstinence, and more than four
times greater continuous abstinence. They thereby concluded that social support is an
important variable to consider when discussing the efficacy of online cessation programs
to help people quit smoking. Because it is difficult to sustain such support in face-to-face
and telephone smoking cessation interventions, especially after the treatment period, the
unique properties of the Internet provides a promising way for patients to receive social
support whenever they need it.
H10: Social support received on the site will be positively associated with
smoking cessation self-efficacy.
51
Next, Study II also proposed that the social group variables tested: social identification,
social capital, social norms and social support, would interact with participation level on
the site, to influence smoking cessation self-efficacy.
H11: Social identification with other members will moderate the association
between participation level on the site and smoking cessation self-efficacy.
H12a: Bridging social capital on the site will moderate the association between
participation level on the site and smoking cessation self-efficacy.
H12b: Bonding social capital on the site will moderate the association between
participation level on the site and smoking cessation self-efficacy.
H13: Conforming to social norms will moderate the association between
participation level on the site and smoking cessation self-efficacy.
H14: Social support from the site will moderate the association between
participation level on the site and smoking cessation self-efficacy.
Also, previous research has found that the majority of online groups comprise
of many lurkers who do not actively participate (Eysenbach et al., 2004; Preece et al.,
2004, Van Uden-Kraan et al., 2008). Despite not directly posting messages, the impact
of online messages posted by other members in the groups on lurkers’ feelings towards
the groups and, more importantly, on their own health conditions, is important to
understand. Study II thereby assessed active participants versus lurkers on the health
SNSs for smoking cessation, hypothesizing that there would be a significant difference
between active participants and lurkers in terms of social identification, bridging and
52
bonding social capital, conforming to social norms, social support, attitude towards
smoking and smoking cessation self-efficacy.
H15: Active participants and lurkers will differ significantly in social
identification with other members on the site.
H16a: Active participants and lurkers will differ significantly in bridging social
capital on the site.
H16b: Active participants and lurkers will differ significantly in bonding social
capital on the site.
H17: Active participants and lurkers will differ significantly in conforming to
social norms of other members towards smoking on the site.
H18: Active participants and lurkers will differ significantly in social support on
the site.
H19: Active participants and lurkers will differ significantly on attitude towards
smoking.
H20: Active participants and lurkers will differ significantly in smoking cessation
self-efficacy.
A study by Parks and Floyd (1999) examined relationship development in Internet
newsgroups by conceptualizing the relationships formed online along seven specific
dimensions: increases in interdependence, in breadth and depth of interaction, in
interpersonal predictability and understanding, in code change towards more personalized
communication, in convergence of social networks, and in commitment. In their
examination of lurking on SNSs, Rau et al. (2008) incorporated the seven relational
53
dimensions as part of their measures of verbal and affective intimacy. Verbal intimacy
refers to self-disclosure. When a SNS user feels a high level of verbal intimacy from the
site, he or she is more likely to disclose various aspects of him or herself, including
uploading photos, links, blog messages and profile updates. The study measured verbal
intimacy through: breadth and depth of online interactions. Affective intimacy refers to
the emotional closeness, affective feelings and reciprocal support (Rau et al., 2008).
Users who perceive higher affective intimacy will post more because they are more likely
to seek emotional support from their SNS connections, and the expectation that they will
receive reciprocal support is also high. Affective intimacy was measured in the study
through: interdependence, interpersonal predictability and understanding, and
commitment. Rau et al. (2008) found that intimacy levels on SNSs had a significant
effect on participation level. As such, it is possible that on SNSs for smoking cessation,
lurkers do not participate because they feel a low level of relational development with
other users, and therefore are not comfortable to reveal themselves. In line with these
findings, Study II sought to find out if there was a significant difference between active
participants and lurkers along four levels of intimacy: verbal, affective, cognitive and
physical intimacy. Cognitive intimacy was measured by the code change dimension
proposed by Parks and Floyd (1996), while physical intimacy was measured by the
network convergence dimension.
RQ2: Is there a significant difference in verbal intimacy between active
participants and lurkers on the site?
54
RQ3: Is there a significant difference in affective intimacy between active
participants and lurkers on the site?
RQ4: Is there a significant difference in cognitive intimacy between active
participants and lurkers on the site?
RQ5: Is there a significant difference in physical intimacy between active
participants and lurkers on the site?
All the variables described were then tested as part of a structural model in Study
II. The model was partly based on Cobb’s (2007) Post-Hoc model for the Social network
approach to smoking cessation, which proposed that social facilitation, social learning
and social comparison would have an effect on social support, which would in turn
facilitate smoking cessation behavior. Study II included four “social” components: social
identification, social capital, social norms, and social support, which were hypothesized
to have a significant impact on two dependent variables: attitude towards smoking, and
smoking cessation self-efficacy. Additionally, the four types of “intimacy”, which were
modified from studies by Parks and Floyd (1996) and Rau et al. (2009), were
hypothesized to have a significant impact on participation level. Participation level was,
in turn, hypothesized to have a significant effect on the four social variables. Figure 2
presents the theoretical model for Study II, specifying the relationships among the
variables tested. (Figure 2)
55
Figure 2: Theoretical Model specifying Relationships among Variables in Study II
56
CHAPTER THREE: METHOD
Data Collection
Data for both Studies I and II were collected using an online questionnaire created
on Qualtrics.com. The first page of the questionnaire contained an Information Sheet,
which described the purpose of the study, procedures, potential risks and discomforts,
potential benefits to subjects and/or society, payment of subjects, confidentiality of data,
participation and withdrawal, rights of research participants, and identity and contact
address and email of the primary investigator, as well as approval of the study by the
Institutional Review Board at the University of Southern California. Potential subjects
were told that their participation was strictly voluntary, and that they may choose not to
answer any questions, and withdraw from the questionnaire at any time. They also had to
be 18 years of age and above, and all data collected would be strictly anonymous. They
were also reminded that they would be entered in a raffle to win $10 gift certificates from
Amazon.com. After participants read the Information sheet, and clicked on a link to the
questionnaire, they have indicated their agreement to participate, in lieu of providing a
signed consent form. The online questionnaires were also set up in Qualtrics.com to
prevent anyone from taking them more than once.
Study Sites
Study sites for the dissertation were selected through a comprehensive web search
through Google, Yahoo and other major search engines. Key terms used included
“Smoking Cessation”, “Social Networking”, “Quit Smoking”, “Moderator”,
“Membership”, “Post Profile” and “Friend Connections”. Other sources, including
57
newspaper and magazine reports, were also used to generate a list of possible sites. The
sites on the list fulfilled the following three criteria: (1) the site should focus on smoking
cessation, or include a section for smoking cessation, (2) the site should include social
networking features, including the ability to create individual member profile pages, and
friend connections, (3) the site should not be sponsored by pharmaceutical companies
(e.g. Nicorette), and (4) the site should be moderated. Websites which only provided
health information about quitting, those catering specifically to health professionals, and
several sites which provided social support for smokers to continue smoking, were also
excluded. A total of 25 sites were generated. The number of sites on the list was then
further reduced through a process of screening for active sites. Those sites with
discussion forums that had at least one posting during the most recent week were
considered to be active. Several smoking cessation sites that were hosted by Google,
Ning, and Yahoo, were deemed to be inactive, and therefore excluded from the sample.
A total of 18 active sites were finally selected. Six of these sites focused solely on
smoking cessation, while the other twelve were general health SNSs with smoking
cessation communities within the site. These sites are presented in Table 1.
58
Table 1: List of Smoking Cessation Sites Contacted
Study Site Type of Site Study Period
Care Pages General Health No
Daily Strength General Health Yes
Healia General Health No
Hope Cube General Health No
IC You General Health No
iMedix General Health No
iVillage Health General Health Yes
Inspire Smoking Cessation Requested additional info
MD Junction General Health No
Med Help General Health No
Patients Like Me General Health Requested additional info
Quit Net Smoking Cessation No
Quit Smoking Journals Smoking Cessation No
Quit Smoking Support Smoking Cessation Requested additional info
We Go Health General Health No
Well Sphere General Health Yes
Why Quit Smoking Cessation Requested additional info
Woofmang Smoking Cessation No
Following selection, moderators of each of these 18 sites were contacted for
permission to post a link for the questionnaire. The emails to the moderators included a
detailed research protocol, information sheet, explanation of study objectives, as well as
of the measures on the questionnaires. Among those contacted, three moderators granted
permission to post the survey links on their site. Four others requested additional
information, while the other moderators did not reply. A letter from IRB (Institutional
Review Board) with the letter head of the Annenberg School for Communication and
Journalism, which stated that the study protocol was approved and that the research
conducted would be supervised by faculty members of the University of Southern
California, was sent to the four moderators who requested additional information. Three
of these moderators finally granted permission to recruit participants from their sites,
59
while the fourth denied permission, stating that they wanted to maintain privacy for the
members of their group. As such, six sites were finally chosen for the study. These sites
are presented in Table 2.
Table 2: List of Smoking Cessation Sites Studied
Study Site Type of Site Study Period
iVillage General Health Dec 2009-Jan 2010
Daily Strength General Health Dec 2009-Jan 2010
Inspire Smoking Cessation Dec 2009-Jan 2010
Quit Smoking Support Smoking Cessation Feb 2010-March 2010
Why Quit Smoking Cessation May 2010-June 2010
Well Sphere General Health May 2010-June 2010
Recruitment of Subjects
Subjects for Studies I and II were recruited from the same smoking cessation
sites. A recruitment message, as well as a URL link to the online questionnaire, was sent
to the moderators of the six smoking cessation sites, who then posted it on the discussion
forums of the respective sites. The moderators of two of the sites also sent the
recruitment message and URL link as email attachments to individual members of the
sites. These moderators also included an endorsement message along with the email,
stating that the study has been approved for distribution by the site administrators.
The recruitment message that was sent out in all cases contained a description of
the primary investigator’s qualifications as a Ph.D. candidate at Annenberg School for
Communication and Journalism, as well as the objectives and goals of the study, to find
out more about the use of social networking sites for smoking cessation. Additionally, it
specified that subjects’ participation in the study was strictly voluntary, and that if they
participate, they could choose not to answer any questions they did not wish to answer,
60
could also opt out of the survey at any time, and that the questionnaires would take
approximately ten to fifteen minutes to complete. Additionally, they had to be 18 years
of age and above, and be registered members of the study sites. Also, they were
informed that personally identifying information would not be collected in the
questionnaire. They were also told specifically that they would not receive any
immediate incentives for participation, but upon completion, they should email the
primary investigator individually requesting to be entered in a raffle which would entitle
them to win $10 gift certificates from the online retailer Amazon.com. This also ensured
that the data from the questionnaires remained anonymous.
The entire study period for the six sites lasted for a period of seven months, from
December 2009 to June 2010. For each site, the online questionnaire was posted for one
month. For sites 1, 2 and 3, the questionnaire was posted from December 2009 to
January 2010. For site 4, the questionnaire was posted from February to March 2010,
and for sites 5 and 6, the questionnaire was posted from May to June 2010. A total of
318 subjects completed the questionnaire. After subtracting ineligible subjects and
incomplete questionnaires, 252 completed questionnaires remained. The raffle for the
$10 gift certificates was conducted at the end of the study period in July 2010. The
primary investigator randomly selected 5 email addresses from each site to win the gift
certificates, making a total of 30 winners for the six sites combined. These 30 winners
were contacted through email, and asked to send the primary investigator their mailing
address, after which the gift certificates were mailed physically to them.
61
The online questionnaire, created using Qualtrics.com, included all the measures
used for Studies I and II. Measures used for Study I were: respondents’ attitude towards
smoking, identification with reference groups (best friends, colleagues and family
members), perceived smoking norms (injunctive and descriptive norms), smoking
cessation self-efficacy, and demographics. Measures used for Study II were:
participation level on the site, social identification with other members, online bridging
and bonding social capital, perceived social support on the site, smoking cessation self-
efficacy, intimacy levels (verbal, affective, cognitive, physical) and demographics.
Explanation of Reliability and Validity
Reliability establishes the external and internal consistency of the measures
(Williams & Monge, 2001). It is the degree to which each instrument measures the same
way each time it is used under the same conditions with the same subjects (Carmines &
Zeller, 1979). In other words, it is the lack of distortion or precision of a measuring
instrument (Kerlinger, & Lee, 2000). Validity, meanwhile, establishes the degree to
which each instrument measures what it is supposed to measure, or the accuracy of the
measurements (Carmines & Zeller, 1979; Williams & Monge, 2001). There are three
types of validity. Face validity refers to the extent to which the items in an instrument
actually address the important aspects of the domain the instrument is intended to assess
(Kerlinger & Lee, 2000). Content validity refers to the extent to which the items in an
instrument address the full range of the important aspects of the domain being addressed;
the representativeness or sampling adequacy of the measuring instrument (Kerlinger &
Lee, 2000). Construct validity refers to the extent to which the items in an instrument
62
address the underlying latent factors within a domain (DeVillis, 2003).
Exploratory Factor Analysis
Principal component analysis (PCA) with varimax rotation was used to test for
reliability and validity of the measures in both studies. Results indicated that all
measures loaded significantly onto their intended latent factors, establishing construct
validity. Percentages of variances also supported construct validity, as a substantial
amount of the variances in the measures were captured by the latent constructs.
Additionally, acceptable Cronbach’s α values were obtained for each of the respective
factors, providing for internal consistency. The factors in Study I with eigenvalues
greater than 1.00 are reported, along with their Cronbach’s α coefficients and factor
loading for each question in the measures, in Table 3. Meanwhile, Table 4 reports the
eigenvalues and % of variance explained for each of the factors.
Table 3: Summary of Exploratory Factor Analysis Results for Study (N=208)
Factor Factor Loadings
1 2 3 4 5
Factor 1: Injunctive norms (α=.87)
Respond to knowing about smoking 0.80
Respond to smoking around them 0.82
Factor 2: Descriptive norms (α=.87)
How common is smoking 0.93
How socially acceptable is smoking 0.90
Factor 3: Identification (α=.91)
Identify with reference group 0.95
Similarity of attitudes and beliefs 0.88
Feel strong bonds to reference group 0.83
Importance of group to sense of self 0.87
Factor 4: Attitude towards smoking (α=.85)
All forms of smoking are dangerous 0.80
63
Table 3: Continued
Factor Factor Loadings
1 2 3 4 5
Smoking during pregnancy is harmful 0.76
Smoking should be avoided 0.81
Smoking seriously damages health 0.90
Smoking shortens a person’s life 0.69
Smoking is purposeless activity 0.59
Quitting can totally reverse damage 0.67
Smoking kills 0.87
Second-hand smoke is harmful 0.74
Smokers die younger 0.73
Smoking is revolting 0.79
No one should be allowed to smoke 0.56
Lung cancer can develop from smoking 0.93
Damage done by smoking is irreversible 0.77
Lung cancer rate is higher for smokers 0.61
Easy to develop tobacco-related disorders 0.85
Smokers more exposed to heart disease 0.86
One of life’s basic pleasures* 0.66
Nothing like a good smoke* 0.67
Only heavy smoking is dangerous* 0.73
Smoking not as harmful as drugs* 0.71
Smoking less dangerous than other risks* 0.72
Health hazards of smoking are misleading* 0.68
Anti-smoking ads exaggerate ill effects* 0.61
Smoking is relatively harmless* 0.63
Low-tar cigarettes reduce risk* 0.51
Second-hand smoke is not dangerous* 0.53
Life is too short to worry about smoking* 0.56
Must smoke a long time to develop risks* 0.63
No difference in mortality rate for smokers* 0.59
Old people who smoke show smoking is OK* 0.66
Factor 5: Smoking cessation self-efficacy (α=.87)
When nervous 0.86
When depressed 0.87
When angry 0.85
When anxious 0.83
When thinking about difficult problem 0.81
When I feel the urge to smoke 0.79
When drinking with friends 0.87
When celebrating 0.81
When drinking alcohol 0.90
64
Table 3: Continued
Factor Factor Loadings
1 2 3 4 5
When I am with other smokers 0.83
After a meal 0.74
When having coffee or tea 0.73
________________________________________________________________________
*Statement reversed during analysis
Table 4: Reliability and Validity Statistics for Measures in Study One
Factor Eigenvalue Cronbach’s
α
%
Variance
Explained
Injunctive Norms 2.89 0.87 10.21
Descriptive Norms 2.93 0.87 9.90
Social Identification 3.75 0.91 12.73
Attitude Towards
Smoking
5.18 0.85 13.68
Smoking Cessation
Self-efficacy
6.71 0.87 15.05
Harman’s One-Factor Test (Study I)
A potential threat to internal validity is Common Method Variance (CMV), which
can occur when questionnaires are used to collect responses in a single setting (Malhotra,
Kim & Patil, 2006). The threat of common method bias is high when a single factor
accounts for the majority of the covariance among the variables, and when a single factor
emerges from factor analysis (Yu, Lu & Liu, 2010). Common method variance can
inflate or deflate observed relationships between constructs, leading to both Type I and
Type II errors.
65
Harman’s one-factor test was used to test for common method variance. In the
exploratory factor analysis, all 5 variables in Study I were entered, using principal
component analysis with varimax rotation, to determine the number of factors necessary
to account for the variance in the variables. The results of the principal component
analysis with varimax rotation, revealed the presence of 5 distinct factors with
eigenvalues > 1.0, rather than one single factor. The 5 factors together accounted for
61.57% of the total variance, and the largest factor (smoking cessation self-efficacy) only
accounted for 15.05% of the variance. Therefore, no general factor is detected in study
one, and common method variance is not present.
Measures for Study I
Attitude towards smoking: Attitude toward smoking was measured by using
Haddad and Malak’s (2002) positive and negative attitudes towards smoking subscales,
which have fourteen items and seventeen items respectively, on seven-point Likert scales,
ranging from “Strongly Disagree” to “Strongly Agree”. Haddad and Malak (2002) pilot-
tested the instrument on a group of university students at Jordan University, and found
high reliability and internal consistency, with Cronbach’s α = 0.83. Additionally, they
also assessed content validity by having three public health experts and nursing educators
in Jordan evaluate the instrument, and found the total content validity index to be 1.00.
In the current study, the fourteen positive attitude questions were reverse-coded, and
scaled with the seventeen negative attitude questions. A higher score indicated a more
negative attitude towards smoking. Cronbach’s α for the scale combining the 31
66
questions was 0.85 (M=4.55, SD=1.58), indicating moderately high reliability. The
factor analysis yielded an eigenvalue of 5.18, demonstrating high construct validity.
Identification with reference groups: Identification questions were modified from
a study by Reed et al. (2007), and measured respondents’ identification with three
reference groups: best friends, colleagues and family members. There are four questions
for each reference group, making a total of twelve questions, measured on seven-point
Likert scales, ranging from “Do not identify” to “Strongly identify”. Reed et al. (2007)
assessed college students’ identification with their reference groups, and found high
internal consistency for the scale, with Cronbach’s α for their identification with close
friends = 0.87, Cronbach’s α for their identification with university peers = 0.82, and
Cronbach’s α for their identification with Greek members = 0.80. In the present study,
Cronbach’s α for identification with best friends was 0.92 (M=4.38, SD=1.36),
Cronbach’s α for identification with colleagues was 0.88 (M=3.76, SD=2.21), and
Cronbach’s α for identification with family members was 0.89 (M=4.21, SD=1.86),
thereby indicating reliability for the three scales. In the exploratory factor analysis, the
measure had a Cronbach’s α=0.91, and eigenvalue=3.75, indicating construct validity.
Perceived smoking norms: Perceived smoking norms questions were adapted
from Reed et al. (2007), and measured acceptance of smoking (injunctive norm) and
actual smoking (descriptive norm) by respondents’ best friends, colleagues and family
members, on seven-point Likert scales. There are two questions each for injunctive
norms and descriptive norms for each reference group, making a total of twelve
questions. Reed et al. (2007) found high internal consistency for the instrument, with the
67
following Cronbach’s α values for injunctive norms of close friends =0.78, university
peers =0.80, and Greek members =0.85, and Cronbach’s α values for descriptive norms
of close friends =0.81, university peers =0.83, and Greek members=0.85. In this study,
injunctive norms of best friends achieved a Cronbach’s α=0.81 (M=3.87, SD=1.38),
while descriptive norms of best friends achieved a Cronbach’s α=0.82 (M=4.17,
SD=2.11); injunctive norms of colleagues achieved a Cronbach’s α=0.83 (M=5.66,
SD=2.37), while descriptive norms of colleagues achieved a Cronbach’s α=0.82
(M=4.56, SD=2.38); and injunctive norms of family members achieved a Cronbach’s
α=0.87 (M=4.89, SD=2.55), and descriptive norms of family members achieved a
Cronbach’s α=0.91 (M=4.37, SD=2.08). In the exploratory factor analysis, injunctive
norms had a Cronbach’s α=0.87 and eigenvalue=2.89, while descriptive norms had a
Cronbach’s α=0.87, and eigenvalue=2.93.
Smoking measure: The smoking measure consisted of eight questions from the
American Legacy Foundation survey, which accessed respondents’ smoking behavior.
The first question asked whether they have ever tried cigarettes. For smokers, four
questions asked them: (1) the number of cigarettes smoked in their entire life, (2) whether
they have smoked everyday for the last 30 days, (3) the number of days smoked in the
last 30 days, and (4) the number of cigarettes smoked in the last 30 days. For non-
smokers, three questions asked them: (1) whether they would smoke a cigarette anytime
during the next year, (2) whether they will be smoking cigarettes 5 years from now, and
(3) whether they will smoke if a friend offered them a cigarette.
68
Demographic measure: This included questions accessing respondents’ gender,
age, ethnicity, marital status, educational level, nationality and annual household income.
All the questions used in the measures for Study I are included in Appendix A.
Principal component analysis (PCA) with varimax rotation was used to test for
reliability and validity of the Study II measures. All measures were found to load
significantly onto their intended latent factors, establishing construct validity, with a
substantial amount of the variances captured by the latent constructs, and acceptable
Cronbach’s α values obtained for each of the respective factors, providing for internal
consistency. Factors in Study II with eigenvalues greater than 1.00 are reported in Table
5, while table 6 reports eigenvalues and % of variance explained for each of the factors.
Harman’s One-Factor Test (Study II)
Harman’s one-factor test was also used to test for common method variance in
Study II. All 12 variables in the study were entered in the exploratory factor analysis,
using principal components analysis with varimax rotation, to determine the number of
factors necessary to account for the variance in the variables. The results of the principal
components analysis with varimax rotation revealed the presence of 12 distinct factors,
each with eigenvalues > 1.0, rather than a single factor. The 12 factors accounted for
75.74% of the total variance, and the largest factor (smoking cessation self-efficacy), at
9.81%, did not account for the majority of the variance. As such, no general factor was
detected, and therefore, the threat of common method variance is not a problem.
69
Table 5: Factor Loadings and Reliability Statistics for Study II Measures (N=208)
Construct Factor Loadings
1 2 3 4 5 6
7 8 9 10 11 12
Factor 1: Participation level (α=.87)
Contribute information 0.87
Receive information 0.85
Post questions 0.77
Answer questions 0.79
Give advice 0.76
Receive advice 0.77
Give encouragement 0.81
Receive encouragement 0.86
Correspond with others 0.93
Update profile and status 0.91
Start new threads 0.79
Post links 0.67
Part of daily activity 0.63
Proud to tell people I use the site 0.65
Part of my routine 0.68
Feel out of touch if not logged in 0.67
Feel part of the community 0.63
Sorry if the site shut down 0.58
Factor 2: Social identification (α=.91)
Identify with other members 0.95
Similarity of attitudes and beliefs 0.88
Feel strong bonds to other members 0.83
Importance to sense of self 0.87
Factor 3: Bridging social capital (α=.89)
Sense of belonging 0.83
Interested in what goes on 0.85
Community is good place to be 0.76
Willing to offer help to other 0.73
Want to try new things 0.66
Feel part of larger community 0.80
Spend time to support others 0.68
Come into contact with new people 0.67
Reminds me everyone is connected 0.65
Factor 4: Bonding social capital (α=.87)
Several people I trust to solve my problems 0.73
Someone for emergency loan 0.71
Turn to for important advice 0.90
70
Table 5: Continued
Construct Factor Loadings
1 2 3 4 5 6
7 8 9 10 11 12
Provide me with emotional support 0.83
Do not know well enough for important things* 0.80
Factor 5: Social norms (α=.87)
Respond same way to other member smoking 0.80
Respond same way to smoking around us 0.82
Same view regarding own smoking 0.93
Same view regarding smoking acceptability 0.90
Factor 6: Social support (α=.91)
Find others to share experiences 0.93
Find others to give encouragement 0.91
Find other who value my opinions 0.88
Find others who care about me 0.89
Find others who listen to me 0.93
Find others who give me advice 0.87
Find others who understand my problems 0.88
Find others to give me helpful information 0.76
Find others to give emotional support 0.83
Find others to give companionship 0.81
No one cares about me* 0.71
I feel all alone* 0.73
No one I can turn to* 0.72
Factor 7: Self-efficacy (α=.87)
When nervous 0.86
When depressed 0.87
When angry 0.85
When anxious 0.83
When thinking about difficult problem 0.81
When I feel the urge to smoke 0.79
When drinking with friends 0.87
When celebrating 0.81
When drinking alcohol 0.90
When I am with other smokers 0.83
After a meal 0.74
When having coffee or tea 0.73
Factor 8: Attitude towards smoking (α=.85)
All forms of smoking are dangerous 0.80
Smoking during pregnancy is harmful 0.76
Smoking should be avoided 0.81
Smoking seriously damages health 0.90
71
Table 5: Continued
Construct Factor Loadings
1 2 3 4 5 6
7 8 9 10 11 12
Smoking shortens a person’s life 0.69
Smoking is purposeless activity 0.59
Quitting can totally reverse damage 0.67
Smoking kills 0.87
Second-hand smoke is harmful 0.74
Smokers die younger 0.73
Smoking is revolting 0.79
No one should be allowed to smoke 0.56
Lung cancer can develop from smoking 0.93
Damage done by smoking is irreversible 0.77
Lung cancer rate is higher for smokers 0.61
Easy to develop tobacco-related disorders 0.85
Smokers more exposed to heart disease 0.86
One of life’s basic pleasures* 0.66
Nothing like a good smoke* 0.67
Only heavy smoking is dangerous* 0.73
Smoking not as harmful as drugs* 0.71
Smoking less dangerous than other risks* 0.72
Health hazards of smoking are misleading* 0.68
Anti-smoking ads exaggerate ill effects* 0.61
Smoking is relatively harmless* 0.63
Low-tar cigarettes reduce risk* 0.51
Second-hand smoke is not dangerous* 0.53
Life is too short to worry about smoking* 0.56
Must smoke a long time to develop risks* 0.63
No difference in mortality rate for smokers* 0.59
Old people who smoke show smoking is OK* 0.66
Factor 9: Verbal intimacy (α=.87)
Communication covers wide range of topics 0.87
Move easily from one topic to another 0.87
Contact each other in variety of ways 0.89
Usually tell exactly how I feel 0.79
Told what I like about him or her 0.78
Can confide in this person 0.90
Told this person secrets about me 0.81
Communication stays on surface of topics* 0.76
Communication limited to few topics* 0.72
Keep personal judgments to myself* 0.73
Never tell this person anything intimate* 0.71
72
Table 5: Continued
Construct Factor Loadings
1 2 3 4 5 6
7 8 9 10 11 12
Factor 10: Affective intimacy (α=.89)
Two of us depend on each other 0.81
Wait to consult other on important matters 0.83
Great deal of effect on each other 0.73
Influence each other’s feelings 0.77
Accurately predict this person’s response 0.65
Can tell what this person feels inside 0.83
Accurately predict this person’s attitude 0.71
Committed to maintaining this relationship 0.55
Relationship is big part of who I am 0.73
Make great effort to maintain relationship 0.72
Feel close to this person most of the time 0.70
Feel like being encouraging 0.68
Important to listen to his/her disclosures 0.73
Important to understand my feelings 0.84
Important to be encouraging 0.81
Satisfied with this relationship 0.80
Neither set aside time to communicate* 0.72
Little influence on each other’s thoughts* 0.68
Uncertain about what this person is like* 0.70
Do not know this person well* 0.81
Relationship not very important* 0.73
Do not expect relationship to last long* 0.61
Factor 11: Cognitive intimacy (α=.83)
Read between the lines of each other 0.71
Use private signals 0.69
Use special nicknames 0.72
Can get idea across with simple message 0.71
Share special language jargon 0.70
No difference in way I talk with this person* 0.61
Factor 12: Physical intimacy (α=.82)
Introduced each other to family and friends 0.69
Contact same people on the Internet 0.73
Involved in same groups and sites 0.75
Overlapping social circles on the Internet 0.74
Overlapping social circles offline 0.67
Do not know any of the same people* 0.71
*Statement reversed during analysis
73
Table 6: Reliability and Validity Statistics for Measures in Study II
Measures for Study II
Participation level on the site: Participation level on the site was measured by a
series of eighteen questions, modified from Ellison et al.’s (2007) Facebook Intensity
Scale. In their study, Ellison et al. (2007) found a Cronbach’s α value of 0.83 for the
instrument, establishing internal consistency. In the present study, the questions, utilizing
seven-point Likert scales, ranging from “Strongly Disagree” to “Strongly Agree”,
assessed respondents’ use of various features of the sites, including posting messages,
Factor Eigenvalue Cronbach’s
α
%
Variance
Explained
Smoking Cessation Self-Efficacy 6.71 0.87 9.81
Attitude Towards Smoking 5.18 0.85 8.79
Perceived Social Support 6.41 0.91 6.33
Social Norms 5.21 0,87 5.11
Bridging Social Capital 2.19 0.89 6.12
Bonding Social Capital 2.89 0.87 5.93
Identification with other members 6.37 0.91 6.26
Participation Level on the site 5.19 0.87 5.08
Verbal Intimacy 3.15 0.87 5.21
Affective Intimacy 2.31 0.89 5.18
Cognitive Intimacy 1.91 0.83 5.93
Physical Intimacy 2.89 0.82 5.99
74
videos, links, replying to others’ messages and more. These eighteen questions were
summed to create a single Participation scale. Cronbach’s α was 0.87 (M=4.91,
SD=1.48), indicating high reliability. In the exploratory factor analysis, the instrument
yielded an eigenvalue= 5.19, establishing construct validity. Six additional open-ended
questions asked respondents to estimate: (1) the average number of hours spent on the
site per week, (2) the average number of hours spent surfing the Internet per week, (3) the
average number of messages they post to the site per week, (4) the number of connections
to other members they have on the site, (5) the number of friends they have offline who
are also members of the site, and (6) the number of members of the site whom they have
since developed offline friendships with.
Social Identification with other members: Identification questions were modified
from a study by Reed et al. (2007), and measured respondents’ identification with other
members. There were four questions, measured on seven-point Likert scales, ranging
from “Do not identify” to “Strongly identify”. These four questions were identical to the
questions for identification with the three reference groups. Reed et al. (2007) assessed
college students’ identification with their reference groups, and found high internal
consistency for the scale, with Cronbach’s α for their identification with close friends =
0.87, Cronbach’s α for their identification with university peers = 0.82, and Cronbach’s α
for their identification with Greek members = 0.80. In the present study, the
identification with other members scale achieved Cronbach’s α=0.91 (M=3.87,
SD=2.11), indicating high reliability. The factor analysis yielded an eigenvalue=6.37.
75
Online Bridging Social Capital: Online bridging social capital was measured
using a modified version of Williams’ (2006) bridging social capital subscale. Williams
(2006) found high reliability for the subscale, with Cronbach’s α=0.84. Ellison et al.
(2007), who used the subscale to assess MSU students’ engagement with Facebook, also
found a Crobach’s α=0.87 for the scale. There are nine questions, measured on seven-
point Likert scales, accessing subjects’ interactions with heterogeneous and weak ties
online. The present study found high reliability for the subscale, with Cronbach’s α=0.89
(M=4.51, SD=1.38), with the factor analysis yielding an eigenvalue=2.19.
Online Bonding Social Capital: Online bonding social capital was measured using
a modified version of Williams’ (2006) bonding social capital subscale. Williams (2006)
found a Cronbach’s α=0.89 for the subscale, while Ellison et al. (2007) also found
moderately high reliability, with Cronbach’s α=0.75. The subscale included six
questions, measured on seven-point Likert scales, accessing subjects’ interactions with
homogeneous and strong ties online. In the present study, the bonding social capital
subscale had Cronbach’s α = 0.87 (M=4.28, SD=1.68), and eigenvalue=2.89.
Social Norms: Social norms questions measured respondents’ similarity in, and
conformation to perceived injunctive and descriptive norms towards smoking of other
members of the site. Questions were modified from Reed et al. (2007), and were
measured on seven-point Likert scales. There were two questions each for injunctive
norms and descriptive norms. These questions were identical to those in Study I which
measured social norms of reference groups. Reed et al. (2007) found high internal
consistency for the instrument, with the following Cronbach’s α values for injunctive
76
norms of close friends =0.78, university peers =0.80, and Greek members =0.85, and
Cronbach’s α values for descriptive norms of close friends =0.81, university peers =0.83,
and Greek members=0.85. In the present study, Cronbach’s α = 0.87 (M=4.33,
SD=1.68), and eigenvalue=5.21 was obtained, indicating reliability and validity.
Smoking Cessation Self-Efficacy: Smoking cessation self-efficacy was measured
using the Smoking Self-Efficacy Questionnaire (SEQ-12), a twelve-item measure of self-
efficacy (Etter, Bergman, Humair, & Perenger, 2000). The scale assessed current and
former smokers in their ability to abstain from smoking in situations where they might be
tempted to smoke. The scale was divided into two subscales: internal stimuli (e.g. feeling
depressed), and external stimuli (e.g. facing peer pressure to smoke). Etter et al. (2000)
found high internal consistency for the instrument, with Cronbach’s α = 0.95 for internal
stimuli, and Cronbach’s α = 0.94 for external stimuli). In the present study, the smoking
cessation self-efficacy scale found Cronbach’s α=0.87 (M=5.01, SD=3.83), and in the
factor analysis, eigenvalue=6.71.
Perceived Social Support: Perceived social support from the site was measured
using a modified version of the “Multidimensional Scale of Perceived Social Support”
(Canty-Mitchell & Zimet, 2000; Zimet, Dahlem, Zimet & Farley, 1988). The scale
included thirteen questions assessing respondents’ giving and receiving of informational,
emotional and tangible support on the site. Seven-point Likert scales, ranging from
“Strongly Disagree” to “Strongly Agree”, were used. Canty-Mitchell and Zimet (2000)
tested the scale on urban adolescents, and found internal consistency, with Cronbach’s
77
α=0.93. In the present study, the summed social support scale achieved Cronbach’s
α=0.91 (M=3.88, SD=2.09), and the factor analysis yielded eigenvalue=6.41.
Verbal Intimacy: Verbal intimacy consisted of eleven items, modified from Rau et
al. (2008), which asked respondents to answer the questions when thinking about the
most intimate relationships they had experienced through the site. Four questions
measured breadth of interaction, while seven questions measured depth of interaction.
Rau et al. (2008) measured internal consistency for the scale, and found Cronbach’s
α=0.73. The present study tested reliability, and found Cronbach’s α=0.87 (M=4.17,
SD=1.78), and in the factor analysis, the eigenvalue obtained=3.15.
Affective Intimacy: Affective intimacy consisted of 22 questions modified from
Rau et al. (2008), with six items measuring interdependence, five items measuring
interpersonal predictability and understanding, five items measuring commitment, and six
items from the Miller Social Intimacy Scale (MSIS) (Miller & Lefcourt, 1982) which
measured frequency of affective intimacy. Rau et al. (2008) established internal
consistency, with Cronbach’s α=0.89 for the combined scale. In the present study,
Cronbach’s α=0.88 (M=3.87, SD=1.39), and eigenvalue=2.31.
Cognitive Intimacy: The cognitive intimacy scale was modified from the code
change subscale by Parks and Floyd (1996), which assessed the degree to which online
communication partners develop specialized ways of communicating, including personal
idioms, that allow them to express themselves in more efficient ways and that reinforce
their relational identity. Parks and Floyd (1996) found internal consistency for the
78
measure, with Cronbach’s α=0.81. In the present study, the six items in the instrument
achieved Cronbach’s α=0.83 (M=3.95, SD=1.97), with eigenvalue=1.91.
Physical Intimacy: The Physical intimacy scale was modified from Parks and
Floyd’s (1996) network convergence subscale, and consisted of six questions which
measured the degree to which online relationships moved offline. Parks and Floyd
(1996) found Cronbach’s α=0.79 for the scale. In the present study, the instrument
demonstrated high reliability, with Cronbach’s α=0.82 (M=4.13, SD=1.92), as well as
construct validity, with eigenvalue=2.89.
Data Analysis Plan
Data analysis was conducted in the following ways. SPSS version 17.0 was used
for descriptive statistics, correlation analyses, multivariate regression analyses and factor
analyses. LISREL version 8.8 was used to test the structural model for Study II.
The following data analysis was done for Study I. To answer H1, H2, and H3,
multivariate regression analyses were conducted, in which respondents’ attitude towards
smoking was entered as the dependent variable, with demographic attributes (Model 1),
injunctive norms and descriptive norms (Model 2), social identification (Model 3), and
attitude towards smoking (Model 4) entered as independent variables. Interaction terms
“Injunctive norms * Social identification” and “Descriptive norms * Social
identification” were also created for each of the reference groups, and entered as Model 5
in the regression. Three separate regression tables were created altogether, one for each
reference group (best friends, colleagues, and family members. To answer H4, the
standardized regression coefficients (β) for injunctive norms across the three reference
79
groups were compared. Similarly, to answer H5, the standardized regression coefficients
(β) for descriptive norms across the three reference groups were also compared. To
answer RQ1, the standardized regression coefficients (β) for descriptive norms and
injunctive norms were compared, for each of the three reference groups.
Meanwhile, the following data analysis was carried out for Study II. To answer
H6, H7, H8, H9 and H10, multivariate regression analyses were conducted, in which
smoking cessation self-efficacy was entered as the dependent variable, with demographic
attributes (Model 1), participation level on the site (Model 2), social identification with
other members (Model 3), perceived bridging and bonding social capital (Model 4),
social norms (Model 5), social support (Model 6) and attitude towards smoking (Model
7) as independent variables. To answer H11, H12, H13 and H14, the interaction terms
“Participation level * Social identification”, “Participation level * Bridging social
capital”, “Participation level * Bonding social capital”, “Participation level * Social
norms” and “Participation level * Social support” were created, and entered in the
regression as Models 7, 8, 9 and 10 respectively. To answer H15, H16, H17, H18, H19
and H20, as well as RQ2, RQ3, RQ4 and RQ5, omnibus MANOVA tests, plus univariate
ANOVA tests were conducted. Finally, confirmatory factor analysis (CFA) and
structural equation modeling (SEM) was applied to test the structural model. Baron and
Kenny’s (1986) mediation analysis technique was then used to test for potential
mediators between participation level on the site with attitude towards smoking and
smoking cessation self-efficacy.
80
CHAPTER FOUR: RESULTS
Descriptive Statistics
Table 7 presents the descriptive statistics for the respondents in Study I (N=252).
Mean age of respondents was 40.4 years.
Table 7: Demographic Characteristics of Questionnaire Respondents (N=252)
Variable Category Number Percentage
Gender Male 118 46.8%
Female 134 53.2%
Ethnicity Asian 17 6.7%
Black 15 6.0%
Latino 24 9.5%
White 175 69.4%
Native American 4 1.6%
Mixed/Other 17 6.7%
Marital Status Single, Never Married 62 24.6%
Married 153 60.7%
Divorced 21 8.3%
Separated 11 4.4%
Widowed 5 2.0%
Annual Income Less than $25,000 3 1.2%
$25,001 to $40,000 14 5.6%
$40,001 to $60,000 64 25.4%
$60,001 to $80,000 84 33.3%
$80,001 to $100,000 66 26.2%
More than $100,000 21 8.3%
Educational Level Junior High 2 0.8%
Some High School 10 4.0%
High School Graduate 65 25.8%
Vocational/Trade School 10 4.0%
Some College 6 2.4%
Associate Degree 55 21.8%
Bachelor Degree 90 35.7%
Masters Degree 10 4.0%
Professional Degree 3 1.2%
Doctorate Degree 1 0.4%
81
Table 7: Continued
Variable Category Number Percentage
Country United States 164 65.1%
United Kingdom 31 12.3%
Canada 28 11.1%
Australia 5 2.0%
Germany 4 1.6%
Mexico 4 1.6%
Puerto Rico 2 0.8%
South Africa 2 0.8%
Spain 2 0.8%
Belgium 1 0.4%
Greece 1 0.4%
Hong Kong 1 0.4%
Ireland 1 0.4%
Japan 1 0.4%
New Zealand 1 0.4%
Scotland 1 0.4%
South Korea 1 0.4%
Sweden 1 0.4%
Switzerland 1 0.4%
Smoking Behavior
Smoking behavior of respondents in Study I is summarized in Table 4. In the data
analysis, the 37 non-smokers, and the 7 respondents who were unsure of whether they
have ever smoked, were excluded, leaving the 208 respondents who smoked.
Table 8: Smoking Behavior of Study I Respondents (N=252)
Variable Category Number Percentage
Ever Tried Smoking Yes 208 82.5%
No 37 14.7%
Not Sure 7 2.8%
For Smokers (N=208)
Cigarettes Smoked in Entire Life 1 or more puffs 8 3.2%
1 cigarette 17 6.7%
2 to 5 cigarettes 18 7.1%
82
Table 8 (Continued)
Variable Category Number Percentage
6 to 15 cigarettes 23 9.1%
16 to 25 cigarettes 28 11.1%
26 to 99 cigarettes 25 9.9%
5 packs or more 89 35.3%
Ever Smoked Everyday for 30 days Yes 98 38.9%
No 110 43.7%
Cigarettes Smoked per day <1 per day 31 12.3%
1 per day 28 11.1%
2 to 5 per day 28 11.1%
6 to 10 per day 41 16.3%
11 to 20 per day 49 19.4%
>20 per day 31 12.3%
For Non-Smokers (N=44)
Likely to smoke in next year Definitely yes 1 0.4%
Probably yes 3 1.2%
Not sure 3 1.2%
Probably not 8 3.2%
Definitely not 29 11.5%
Likely to smoke in next 5 years Definitely yes 2 0.8%
Probably yes 3 1.2%
Not sure 1 0.4%
Probably not 14 5.6%
Definitely not 24 9.5%
Will smoke if offered by friend Definitely yes 2 0.8%
Probably yes 3 1.2%
Not sure 2 0.8%
Probably not 4 1.6%
Definitely not 33 13.1%
83
STUDY I
Correlation Analysis
For Study one, Zero-order correlation analyses were first conducted on the
independent and dependent variables in the study for each of the three reference groups.
The correlation matrices are presented in Tables 9, 10, and 11.Since the correlation
coefficient values were lower than the threshold of .70 as recommended by Campbell
(1998), there was no multicollinearity of the variables. This was further verified in the
regression analyses by the collinearity statistics. There was no multicollinearity problem
if tolerance did not approach 0, and VIF did not approach 10.
Table 9: Zero-order Correlations of independent variables (Best friends) on Smoking
cessation self-efficacy (N=208)
Smoking
Cessation
Self-
Efficacy
Attitude
towards
Smoking
Social
identificatio
n (Best
friends)
Descriptive
norms (Best
friends)
Injunctive
norms
(Best
friends)
Smoking
Cessation
Self-Efficacy
-- 0.39** 0.36** 0.37** 0.38**
Attitude
towards
Smoking
0.39** -- 0.29** 0.21* 0.24**
Social
identification
(Best
friends)
0.36** 0.29** -- 0.26** 0.23**
Descriptive
norms (Best
friends)
0.37** 0.21* 0.26** -- 0.31**
Injunctive
norms (Best
friends)
0.38** 0.24** 0.23** 0.31** --
p<.05*, p<.01**, p<.001***
84
Table 10: Zero-order Correlations of independent variables (Colleagues) on Smoking
cessation self-efficacy (N=208)
p<.05*, p<.01**, p<.001***
Table 11: Zero-order correlations of Independent variables (Family members) on
Smoking Cessation Self-Efficacy (N=208)
p<.05*, p<.01**, p<.001***
Smoking
Cessation
Self-
Efficacy
Attitude
towards
Smoking
Social
identification
(Colleagues)
Descriptive
norms
(Colleagues)
Injunctive
norms
(Colleagues)
Smoking
Cessation
Self-Efficacy
-- 0.37** 0.35** 0.36** 0.43**
Attitude
towards
Smoking
0.37** -- 0.26** 0.24** 0.25**
Social
identification
(Colleagues)
0.35** 0.26** -- 0.36** 0.31**
Descriptive
norms
(Colleagues)
0.36** 0.24** 0.36** -- 0.39**
Injunctive
norms
(Colleagues)
0.43** 0.25** 0.31** 0.39** --
Smoking
Cessation
Self-
Efficacy
Attitude
towards
Smoking
Social
identification
(Family
members)
Descriptive
norms
(Family
members)
Injunctive
norms
(Family
members)
Smoking
Cessation
Self-Efficacy
-- 0.30** 0.31** 0.27** 0.32**
Attitude
towards
Smoking
0.30** -- 0.26** 0.23** 0.25**
Social
identification
(Family
members)
0.31** 0.26** -- 0.21* 0.25**
Descriptive
norms
(Family
members)
0.27** 0.23** 0.21* -- 0.29**
Injunctive
norms
(Family
members)
0.32** 0.25** 0.25** 0.29** --
85
Multivariate Regression
During the data analysis, multivariate regression analyses were performed.
Respondents’ smoking cessation self-efficacy was entered as the dependent variable.
Three separate regression tables were constructed altogether, one for each reference
group. For each set of regressions, Model 1 included the demographic variables, Model 2
included injunctive or descriptive norms, and Model 3 included the social identification
with the reference group. Model 4 included respondents’ attitude towards smoking.
Model 5 included the interaction terms: “injunctive norms * social identification” and
“descriptive norms * social identification”. Dummy variables were created for each of
the demographic variables. Gender was re-coded for “Female”, ethnicity was re-coded
for “Asian”, “Black”, “Latino”, “Native American” and “Other”, and marital status was
re-coded for “Married”, “Divorced”, “Widowed” and “Separated”. Meanwhile,
education level was re-categorized into five groups: less than high school, high school
graduate, associate degree, bachelor degree and masters/professional degree and above.
Finally, income was re-coded as a quantitative numeric variable. Results of the
multivariate regression analyses are presented below in Tables 12, 14, and 16.
86
Table 12: Standardized Regression Coefficients (β) of Best Friends’ injunctive and
descriptive norms on Smoking cessation self-efficacy (N=208)
Independent Variable
Model 1 Model 2 Model 3 Model 4 Model 5
Gender
Female
0.26* 0.21* 0.10 (ns) 0.09 0.09
Ethnicity
Asian
0.03 0.03 0.04 0.03 0.03
African American
-0.10 -0.09 -0.06 -0.05 -0.05
Latino
-0.09 -0.09 -0.06 -0.04 -0.04
Native American
-0.05 -0.05 -0.05 -0.02 -0.01
Mixed/Other
0.08 0.08 0.03 0.03 0.03
Age
0.23* 0.09 (ns) 0.08 0.08 0.09
Income
0.21* 0.17* 0.09 (ns) 0.08 0.07
Education
High School
0.19* 0.06 (ns) 0.06 0.03 0.02
Associate Degree
0.21* 0.07 (ns) 0.06 0.04 0.04
Bachelor Degree
0.22* 0.07(ns) 0.06 0.05 0.05
Masters/Professional
Degree
0.22* 0.08 (ns) 0.06 0.05 0.05
Marital Status
Married
0.25* 0.21* 0.09 (ns) 0.06 0.06
Divorced
-0.09 -0.07 -0.05 -0.05 -0.05
Widowed
-0.12 -0.08 -0.07 -0.07 -0.07
Separated
-0.08 -0.06 -0.06 -0.02 -0.02
Injunctive Norms (Best
Friends)
0.42** 0.31** 0.21* 0.19*
Descriptive Norms (Best
Friends)
0.28* 0.26* 0.18* 0.17*
Identification (Best
Friends)
0.17* 0.15* 0.14*
Attitude Towards Smoking
0.47*** 0.21*
Injunctive
Norms*Identification
0.20*
Descriptive
Norms*Identification
0.19*
DF
21 23 24 25 27
R²
0.08 0.15 0.21 0.28 0.30
F
4.68 8.23*** 12.37*** 15.51*** 18.89***
p<.05*, p<.01**, p<.001***
87
Table 13: Collinearity Statistics for Independent Variables (Best Friends) on Smoking
Cessation Self-Efficacy
IV Tolerance VIF
Injunctive
Norms (Best
Friends)
.815 1.227
Descriptive
Norms (Best
Friends)
.826 1.211
Identification
(Best Friends)
.792 1.263
Attitude
Towards
Smoking
.682 1.466
As can be seen in Table 13, no multicollinearity problem was detected, as the
collinearity statistics indicated that tolerance for all variables were >.50, and VIF were <
and did not approach 10. For best friends, in Model 1, controlling for other demographic
variables, ethnicity was found to have no effect on smoking cessation self-efficacy.
Among the other demographic variables, female respondents (β=0.26, p<.05) reported
significantly higher smoking cessation self-efficacy than male respondents. Those who
were older (β=0.23, p<.05) and had higher income (β=0.21, p<.05) also reported
significantly greater smoking cessation self-efficacy. For education level, those who had
high school education (β=0.19, p<.05), associate degrees (β=0.21, p<.05), bachelor
degrees (β=0.22, p<.01) and masters/professional degrees (β=0.22, p<.01) reported
significantly higher smoking cessation self-efficacy than those with less than high school
education. For marital status, only respondents who are married (β=0.25, p<.01) reported
significantly higher smoking cessation self-efficacy than single respondents. Model 1
explained 8% of the variance in smoking cessation self-efficacy scores.
88
In Model 2, both injunctive norms and descriptive norms of best friends were
found to be significantly associated with smoking cessation self-efficacy, controlling for
demographic variables. Respondents who perceived their best friends to be less
approving of smoking reported significantly higher smoking cessation self-efficacy
(β=0.42, p<.01). Those whose best friends are non-smokers are also likely to have
significantly higher smoking cessation self-efficacy (β=0.28, p<.01). Therefore, both
H1a and H1b were supported. Also, the effect of education level and age on smoking
cessation self-efficacy disappeared after controlling for injunctive and descriptive norms
of best friends. In terms of the other demographic variables, respondents who are female
(β=0.21, p<.05), with higher income (β=0.17, p<.05) and are married (β=0.21, p<.05),
continued to have a significantly greater smoking cessation self-efficacy. Model 2
explained 15% of variance in smoking cessation self-efficacy, which represented a
significant improvement from the previous model (F (23,184) =8.23, p<.001).
Further, in Model 3, identification with best friends was found to have a
significant positive effect on smoking cessation self-efficacy, controlling for injunctive
and descriptive norms. Those who were more highly-identified with their best friends
(β=0.17, p<.05) reported significantly greater smoking cessation self-efficacy. As such,
H1c was supported. Meanwhile, controlling for identification with best friends,
injunctive norms and descriptive norms of best friends continued to have a significant
impact on smoking cessation self-efficacy. Respondents who perceived their best friends
to be less approving of smoking continued to report significantly higher smoking
cessation self-efficacy (β=0.31, p<.01), and those who perceived smoking to be less
89
common among their best friends were also significantly more likely to report higher
smoking cessation self-efficacy (β=0.26, p<.05), controlling for identification with best
friends. The effect of higher income, being female and married, on smoking cessation
self-efficacy also disappeared. Model 3 explained 21% of variance in smoking cessation
self-efficacy, a significant improvement from Model 2 (F (24, 183) =12.37, p<.001).
In Model 4, respondents’ attitude towards smoking (β=0.47, p<.001) was found to
be positively associated with smoking cessation self-efficacy. Those who had more
negative attitudes towards smoking were also significantly more likely to report higher
smoking cessation self-efficacy levels. As such, H1d was supported. Injunctive norms of
best friends (β=0.21, p<.05), descriptive norms of best friends (β=0.18, p<.05) and
identification with best friends (β=0.15, p<.05) also continued to have a significant
impact on smoking cessation self-efficacy, holding attitude towards smoking constant.
Model 4 explained 28% of variance in smoking cessation self-efficacy scores, a
significant improvement from the previous model (F=15.51 (25, 182), p<.001).
Finally, in Model 5, the two-way interaction terms “Injunctive norms of best
friends * Identification with best friends” and “Descriptive norms of best friends *
Identification with best friends” were added to the regression, and were found to be
significantly associated with smoking cessation self-efficacy. Those respondents who
reported that their best friends were less approving of smoking, and who identified more
with their best friends were more likely to have greater smoking cessation self-efficacy
(β=0.20, p<.05), and those who reported their best friends were non-smokers and who
identified more with their best friends were also more likely to have greater smoking
90
cessation self-efficacy (β=0.19, p<.05), holding attitude towards smoking constant.
Model 5 (F (27, 180) =18.89, p<.001) explained 30% of the variance in smoking
cessation self-efficacy, a significant improvement in variance from Model 4. As can be
seen, identification with best friends moderated the effect of injunctive norms and
descriptive norms of best friends on respondents’ smoking cessation self-efficacy. As
such, H1e was also supported.
91
Table 14: Standardized Regression Coefficients (β) of Colleagues’ injunctive and
descriptive norms on Smoking cessation self-efficacy (N=208)
p<.05*, p<.01**, p<.001***
Independent Variable
Model 1 Model 2 Model 3 Model 4 Model 5
Gender
Female
0.26* 0.20* 0.09 0.07 0.07
Ethnicity
Asian
0.03 0.03 0.02 0.02 0.02
African American
-0.10 -0.08 -0.06 -0.06 -0.05
Latino
-0.09 -0.08 -0.06 -0.03 -0.03
Native American
-0.05 -0.05 -0.05 -0.05 -0.03
Mixed/Other
0.08 0.08 0.06 0.05 0.05
Age
0.23* 0.11 0.08 0.07 0.07
Income
0.21* 0.18* 0.09
(ns)
0.08 0.07
Education
High School
0.19* 0.06 (ns) 0.05 0.03 0.02
Associate Degree
0.21* 0.07 (ns) 0.06 0.04 0.03
Bachelor Degree
0.22* 0.07 (ns) 0.06 0.05 0.03
Masters/Professional
Degree
0.22* 0.08 (ns) 0.06 0.06 0.04
Marital Status
Married
0.25* 0.15* 0.09
(ns)
0.08 0.06
Divorced
-0.09 -0.05 -0.04 -0.04 -0.03
Widowed
-0.12 -0.08 -0.08 -0.08 -0.06
Separated
-0.08 -0.03 -0.02 -0.02 -0.02
Injunctive Norms
(Colleagues)
0.29* 0.20* 0.19* 0.18*
Descriptive Norms
(Colleagues)
0.17* 0.16* 0.16* 0.16*
Identification
(Colleagues)
0.16* 0.15* 0.15*
Attitude towards smoking
0.33** 0.22*
Injunctive
Norms*Identification
0.18*
Descriptive
Norms*Identification
0.19*
DF
21 23 24 25 27
R²
0.08 0.15 0.21 0.28 0.30
F
4.68 8.11*** 12.24**
*
15.57*** 18.85***
92
Table 15: Collinearity Statistics of Independent Variables (Colleagues) on Smoking
Cessation Self-Efficacy
IV Tolerance VIF
Injunctive Norms
(Colleagues)
.836 1.196
Descriptive Norms
(Colleagues)
.812 1.232
Identification
(Colleagues)
.763 1.311
Attitude Towards
Smoking
.686 1.458
As can be seen in Table 15, no multicollinearity problem was detected, as the
collinearity statistics indicated that tolerance for all variables were >.50, and VIF were <
and did not approach 10. For colleagues, in Model 2, controlling for demographic
variables, both injunctive and descriptive norms were also found to be significantly
associated with smoking cessation self-efficacy. Respondents who perceived their
colleagues to be less approving of smoking were found to report significantly higher
smoking cessation self-efficacy (β=0.29, p<.05), thereby supporting H2a. Meanwhile,
respondents who perceived smoking to be less common among their colleagues also
reported significantly higher smoking cessation self-efficacy (β=0.17, p<.05), thereby
supporting H2b. Additionally, controlling for injunctive and descriptive norms of college
peers, being female (β=0.20, p<.05), married (β=0.15, p<.05) and having higher income
(β=0.18, p<.05) continued to significantly predict higher smoking cessation self-efficacy,
while the effect of age and educational level disappeared. Model 2 explained 15% of the
variance in respondents’ smoking cessation self-efficacy scores, which was a significant
improvement from the previous model (F (23, 184)=8.11, p<.001).
93
In Model 3, identification with colleagues was found to have a significant effect
on smoking cessation self-efficacy, holding injunctive and descriptive norms of
colleagues constant. Respondents who identified more closely with their colleagues
reported significantly higher smoking cessation self-efficacy (β=0.16, p<.05), thereby
supporting H2c. Also, holding identification with colleagues constant, both injunctive
and descriptive norms of colleagues also continued to have a significant effect on
smoking cessation self-efficacy. Respondents who perceived their colleagues to be less
approving of smoking reported significantly higher smoking cessation self-efficacy
(β=0.20, p<.05). Similarly, those who perceived smoking to be less common among their
colleagues also reported higher smoking cessation self-efficacy scores (β=0.16, p<.05).
Also, the effect of being female, married and of higher income on smoking cessation self-
efficacy disappeared. Model 3 explained 21% of the variance in smoking cessation self-
efficacy, a significant change from Model 2 (F(24, 183)=12.24, p<.001).
In Model 4, attitude towards smoking was added to the regression, and also found
to be significantly associated with smoking cessation self-efficacy. Respondents who had
more negative attitudes towards smoking (β=0.33, p<.01) were also more likely to have
higher smoking cessation self-efficacy. H2d was hence supported. Meanwhile,
injunctive norms of colleagues (β=0.19, p<.05), descriptive norms of colleagues (β=0.16,
p<.05) and identification with colleagues (β=0.15, p<.05) also continued to significantly
affect smoking cessation self-efficacy. Model 4 explained 28% of variance in smoking
cessation self-efficacy, a significant improvement from Model 3 (F (25, 182) =15.57,
p<.001).
94
For Model 5, two-way interaction terms “Injunctive norms of colleagues *
Identification with colleagues”, and “Descriptive norms of colleagues * Identification
with colleagues” were added to the regression. Identification with colleagues was found
to interact with both injunctive norms and descriptive norms to significantly affect
smoking cessation self-efficacy. Those respondents who perceived their colleagues to be
less approving of smoking and identified more strongly with their colleagues (β=0.18,
p<.05), and those who perceived their colleagues to smoke less, and identified more
strongly with their colleagues (β=0.19, p<.05), were found to have significantly higher
smoking cessation self-efficacy. Model 5 (F (27, 180) =18.85, p<.001) explained 30% of
the variance in smoking cessation self-efficacy, a significant improvement of 1% from
Model 4. As can be seen, social identification with colleagues moderated the
associations between both injunctive and descriptive norms on respondents’ smoking
cessation self-efficacy. Thereby, H2e was supported.
95
Table 16: Standardized Regression Coefficients (β) of Family members’ injunctive and
descriptive norms on Smoking cessation self-efficacy (N=208)
Independent
Variable
Model 1 Model 2 Model 3 Model 4 Model 5
Gender
Female
0.26* 0.10 (ns) 0.08 0.08 0.06
Ethnicity
Asian
0.03 0.03 0.03 0.02 0.01
African American
-0.10 -0.07 -0.06 -0.05 -0.05
Latino
-0.09 -0.08 -0.06 -0.05 -0.05
Native American
-0.05 -0.05 -0.05 -0.05 -0.04
Mixed/Other
0.08 0.08 0.07 0.06 0.06
Age
0.23* 0.19* 0.08 (ns) 0.07 0.05
Income
0.21* 0.17* 0.09 (ns) 0.08 0.06
Education
High School
0.19* 0.04 (ns) 0.04 0.03 0.03
Associate Degree 0.21* 0.05 (ns) 0.04 0.04 0.03
Bachelor Degree
0.22* 0.07 (ns) 0.05 0.05 0.04
Masters/Professional
Degree
0.22* 0.08 (ns) 0.06 0.06 0.04
Marital Status
Married
0.25* 0.18* 0.09 (ns) 0.08 0.08
Divorced
-0.09 -0.06 -0.04 -0.04 -0.05
Widowed
-0.12 -0.08 -0.08 -0.08 -0.08
Separated
-0.08 -0.05 -0.04 -0.04 -0.02
Injunctive Norms
(Family members)
0.33** 0.22* 0.19* 0.18*
Descriptive Norms
(Family members)
0.31** 0.16* 0.15* 0.15*
Identification
(Family members)
0.17* 0.16* 0.15*
Attitude towards
smoking
0.38** 0.22*
Injunctive
Norms*Identification
0.16*
Descriptive
Norms*Identification
0.16*
DF
21 23 24 25 27
R²
0.08 0.15 0.21 0.28 0.30
F
4.68 8.18*** 12.23*** 16.49*** 18.57***
p<.05*, p<.01**, p<.001***
96
Table 17: Collinearity Statistics of Independent Variables (Family Members) on Smoking
Cessation Self-Efficacy
IV Tolerance VIF
Injunctive Norms
(Family Members)
.832 1.202
Descriptive Norms
(Family Members)
.818 1.222
Identification
(Family Members)
.782 1.279
Attitude Towards
Smoking
.680 1.471
As can be seen in Table 17, no multicollinearity problem was detected, as the
collinearity statistics indicated that tolerance for all variables were >.50, and VIF were <
and did not approach 10. For family members, in Model 2, both injunctive and
descriptive norms were found to have a significant positive effect on respondents’
smoking cessation self-efficacy, controlling for demographic variables. Respondents
who perceived their family members to be less approving of smoking reported
significantly greater smoking cessation self-efficacy (β=0.33, p<.01). As such, H3a was
supported. Similarly, respondents who perceived actual smoking to be less common
among their family members also reported significantly higher smoking cessation self-
efficacy (β=0.31, p<.01). As such, H3b was supported. For demographic variables, the
gender effect became non-significant. Controlling for injunctive and descriptive norms
of family members, females no longer differed significantly from males for smoking
cessation self-efficacy. However, those respondents who are older (β=0.19, p<.05), have
higher income (β=0.17, p<.05) and are married (β=0.18, p<.05), continue to report higher
smoking cessation self-efficacy, holding injunctive and descriptive norms of family
members constant. Model 2 explained 15% of the variance in smoking cessation self-
97
efficacy scores among respondents, a significant improvement from Model 1 (F (23, 184)
=8.18, p<001).
In Model 3, identification with family members was entered into the regression,
and found to have a significant effect on smoking cessation self-efficacy (β=0.17, p<.05).
Therefore, H3c was supported. Holding identification with family members constant,
injunctive and descriptive norms of family members also continued to have a significant
effect on respondents’ smoking cessation self-efficacy. Those who perceived their family
members to be less approving of smoking were more likely to have significantly higher
smoking cessation self-efficacy (β=0.22, p<.05). Those who perceived smoking to be
less common among their family members also reported significantly higher smoking
cessation self-efficacy (β=0.16, p<.05). The effect of age, income and married status
meanwhile, disappeared. Model 3 accounted for 21% of variance in smoking cessation
self-efficacy, a significant improvement from Model 2 (F (24, 183) =12.23, p<.001).
In Model 4, attitude towards smoking was added to the regression, and found to
be significantly associated with smoking cessation self-efficacy. Respondents’ who
reported more negative attitudes towards smoking (β=0.38, p<.01) were also significantly
more likely to have higher smoking cessation self-efficacy. Therefore, H3d was
supported. Meanwhile, injunctive norms of family members, descriptive norms of family
members, and social identification with family members, all continued to have a
significant positive effect on smoking cessation self-efficacy, holding attitude towards
smoking constant. Model 4 explained 28% of variance in smoking cessation self-
efficacy, a significant change from Model 3 (F (25, 182) =16.49, p<.001).
98
Finally, in Model 5, the 2-way interaction terms “Injunctive norms of family
members * Identification with family members”, and “Descriptive norms of family
members * Identification with family members” were added to the regression, and also
found to be significant at the p<.05 level, with β=0.16 and 0.16 respectively. Likewise,
Model 5 (F (27, 180) =18.57, p<.001) explained 30% of the variance in smoking
cessation self-efficacy, a significant increase from Model 4. Those whose family
members are less approving of smoking and who identified more with their family
members (β=0.16, p<.05) were significantly more likely to have higher smoking
cessation self-efficacy. Those whose family members smoked less and who identified
more with their family members (β=0.16, p<.05) were also significantly more likely to
have higher smoking cessation self-efficacy, controlling for attitude toward smoking.
Identification with family members can be said to moderate the associations between
injunctive and descriptive norms of family members on respondents’ reported smoking
cessation self-efficacy, thereby supporting H3e.
To answer H4 and H5, the standardized regression coefficients for the effect of
injunctive and descriptive norms of each reference group on smoking cessation self-
efficacy were compared. The effect of best friends’ injunctive norms on respondents’
smoking cessation self-efficacy was β=0.42, p<.01, while the effect of colleagues’
injunctive norms was β=0.29, p<.05, and the effect of family members’ injunctive norms
was β=0.33, p<.01. As expected, injunctive norms of best friends had the greatest effect
on respondents’ smoking cessation self-efficacy, followed by family members, and
finally colleagues. For respondents in the sample, their smoking cessation self-efficacy
99
was most influenced by the injunctive norms of their best friends, or their most proximal
reference group, therefore H4 was supported. Meanwhile, the effect of best friends’
descriptive norms on respondents’ smoking cessation self-efficacy was β=0.28, p<.05,
while the effect of colleagues’ and family members’ descriptive norms were β=0.17,
p<.05, and β=0.31, p<.01 respectively. Contrary to expectations, family members’
descriptive norms had the greatest effect among the three reference groups on
respondents’ smoking cessation self-efficacy, followed by best friends and finally
colleagues, therefore H5 was not supported. This can be explained by the fact that the
average age of the sample in the current study is 40.4 years old, and the majority of
respondents (60.7%) are married. As such, the actual smoking behavior, or descriptive
norms, of their family members, including spouses, children and other immediate
relatives, may be the most predictive of their own smoking cessation self-efficacy.
To answer RQ1, the standardized regression coefficients for the effect of
injunctive and descriptive norms on smoking cessation self-efficacy for each reference
group were compared. For best friends, injunctive norms had an effect of β=0.42, p<.01,
while descriptive norms had an effect of β=0.28, p<.05. For colleagues, injunctive norms
had an effect of β=0.29, p<.05, while descriptive norms had an effect of β=0.17, p<.05.
For family members, injunctive norms had an effect of β=0.33, p<.01, while descriptive
norms had an effect of β=0.31, p<.01. As can be seen, injunctive norms of all three
reference groups had a greater effect on respondents’ smoking cessation self-efficacy
than descriptive norms. As such, in answer to RQ1, for the respondents in the sample,
their reference groups’ approval or disapproval of smoking played a more important role
100
in influencing their smoking cessation self-efficacy than members of their reference
groups’ actual smoking behavior.
STUDY II
Internet and Site Usage
The respondents for Study II were identical to those for those for Study one,
therefore the descriptive statistics for gender, ethnicity, age, annual household income,
highest educational level attained, and marital status, as well as smoking behavior,
remained the same. For Study II, additional descriptive statistics were obtained for the
following. Respondents spent an average of 2.7 hours on the site per week. They spent
an average of 10.1 hours surfing the Internet per week. Additionally, the average number
of postings per week per respondents was 3.1. The average number of connections that
respondents had on the site was 43. The average number of friends from offline who
were also members of the site was 12. Finally, the average number of offline friendships
developed from people initially met through the site was 3. (Table 18)
Table 18: Average Internet and Site Usage by Questionnaire Respondents (N=252)
Mean SD
Number of hours spent on the
site per week
2.7 3.9
Number of hours spent surfing
the Internet per week
10.1 8.5
Number of postings on the site
per week
3.1 5.7
Number of friend connections
on the site
43 12
Number of offline friends who
are also members of the site
8 5
Number of offline friendships
developed with people initially
met on the site
3 4
101
Correlation Analysis
For Study II, zero-order correlation analyses were conducted on the independent
and dependent variables in the study. The complete correlation matrix is presented in
Table 19. No multicollinearity issue was detected, as the correlation coefficients were
lower than the threshold of 0.70 recommended by Campbell (1998).
Table 19: Zero-order Correlations of independent variables on Smoking cessation self-
efficacy (Study II) (N=208)
p<.05*, p<.01**, p<.001
Multivariate Regression
Following the correlation analyses, multivariate regression was performed using
the variables tested. Smoking cessation self-efficacy was entered as the dependent
Smoking
Cessation
Self-
efficacy
Attitude
towards
Smoking
Social
Support
Social
Norms
Bonding
Social
Capital
Bridging
Social
Capital
Social
Identifi
cation
Participa
tion
Level
Smoking
Cessation
Self-
efficacy
-- 0.31** 0.36** 0.36** 0.28** 0.23* 0.22* 0.37**
Attitude
towards
Smoking
0.31** -- 0.37** 0.31** 0.21* 0.22* 0.26** 0.28**
Social
Support
0.36** 0.37** -- 0.26** 0.21* 0.22* 0.21* 0.35**
Social
Norms
0.36** 0.31** 0.26** -- 0.25* 0.21* 0.24* 0.26**
Bonding
Social
Capital
0.28** 0.21* 0.21* 0.25* -- 0.41** 0.25* 0.29**
Bridging
Social
Capital
0.23* 0.22* 0.22* 0.21* 0.41** -- 0.24* 0.28**
Social
Identificat
ion
0.22* 0.26** 0.21* 0.24* 0.25* 0.24* -- 0.26**
Participati
on Level
0.37** 0.28* 0.35** 0.26** 0.29** 0.28** 0.26** --
102
variable, with demographic attributes (Model 1), participation level on the site (Model 2),
social identification with other members on the site (Model 3), perceived online bridging
and bonding social capital (Model 4), social norms (Model 5), social support (Model 6)
and attitude towards smoking (Model 7) entered as independent variables. Additionally,
the interaction terms “Participation level * Social identification”, “Participation level *
Bridging social capital”, “Participation level * Bonding social capital”, “Participation
level * Social norms” and “Participation level * Social support” were created, and entered
in the regression as Models 8, 9, 10 and 11 respectively. The dummy variables for the
demographic attributes, which were created earlier for Study I, were also used for the
regression analyses in Study II. Gender was re-coded for “Female”, ethnicity was re-
coded for “Asian”, “Black”, “Latino”, “Native American” and “Other”, and marital status
was re-coded for “Married”, “Divorced”, “Widowed” and “Separated”. Education level
was re-categorized into five groups: less than high school, high school graduate, associate
degree, bachelor degree and masters/professional degree and above. Income was re-
coded as a quantitative numeric variable. Only smokers (N=208) were included in the
data analysis. As can be seen from Table 20, no multicollinearity problem was detected,
as the tolerance for the independent variables did not approach 0, and VIF did not
approach 10.
103
Table 20: Collinearity Statistics of Independent Variables (Study II) on Smoking
Cessation Self-Efficacy
IV Tolerance VIF
Participation Level .891 1.122
Social Identification .737 1.357
Bridging Social
Capital
.802 1.247
Bonding Social
Capital
.783 1.277
Social Norms .816 1.225
Social Support .787 1.271
Attitude Towards
Smoking
.702 1.425
Table 21: Standardized Regression Coefficients (OLS Regression) of Selected
Independent Variables on Smoking Cessation Self-Efficacy (N=208)
p<.05*, p<.01**, p<.001***
104
In Model 1, controlling for other demographic variables, there was a significant
effect for gender, age, income, educational level, and marital status. Female respondents
(β=0.26, p<.05) reported significantly higher smoking cessation self-efficacy than male
respondents. Additionally, those who were older (β=0.23, p<.05) and had higher income
(β=0.21, p<.05) were also more likely to have higher smoking cessation self-efficacy. As
for educational level, relative to respondents with less than high school education, those
with high school education (β=0.19, p<.05), associate degrees (β=0.21, p<.05), bachelor
degrees (β=0.22, p<.05), and masters/professional degrees (β=0.22, p<.05) also reported
significantly higher smoking cessation self-efficacy. In terms of marital status, only
those who are married (β=0.25, p<.05) were more likely to have higher smoking
cessation self-efficacy, compared to those who are single. Model 1 accounted for 8% of
variance in smoking cessation self-efficacy.
For Model 2, participation level on the site was added to the regression, and found
to be significantly related to smoking cessation self-efficacy, controlling for demographic
variables. Those who participated more on the site (β=0.49, p<.001) also scored
significantly higher on smoking cessation self-efficacy. As for the demographic
variables, the effect of educational level and marital status on smoking cessation self-
efficacy disappeared, while females (β=0.19, p<.05), whose who were older (β=0.17,
p<.05), and those with higher income (β=0.19, p<.05) continued to report significantly
higher smoking cessation self-efficacy, holding participation level on the site constant.
Model 2 explained 15% of variance in smoking cessation self-efficacy, representing a
significant increase from Model 1 (F (22, 185)= 9.37, p<.001).
105
Following this, social identification with other members on the site was added to
the regression in Model 3, and also found to be significantly associated with smoking
cessation self-efficacy. Those who identified more highly with other members of the site
(β=0.29, p<.05), who had either quit smoking, or were in the process of quitting, scored
significantly higher in smoking cessation self-efficacy. Participation level on the site
(β=0.39, p<.01) also continued to significantly predict smoking cessation self-efficacy.
Additionally, the effect of gender, age and income on smoking cessation self-efficacy
also disappeared in Model 3, holding social identification and participation level
constant. Model 3 explained 22% of variance in smoking cessation self-efficacy, a
significant increase from Model 2 (F (23, 184)= 13.91, p<.001).
In Model 4, online bridging and bonding social capital were added, and found to
be significantly associated with smoking cessation self-efficacy. Respondents who
perceived greater bridging social capital, or weak ties, from the site (β=0.33, p<.01)
scored significantly higher on smoking cessation self-efficacy. Those who perceived
greater bonding social capital, or strong ties, from the site (β=0.28, p<.05) also reported
significantly higher smoking cessation self-efficacy, although the effect of bridging social
capital is higher. Meanwhile, participation level (β=0.31, p<.05), and social
identification (β=0.26, p<.05) continued to significantly predict smoking cessation self-
efficacy. Model 4 explained 28% of variance in smoking cessation self-efficacy, a
significant increase from Model 3 (F (25, 182)= 16.57, p<.001).
For Model 5, conforming to social norms was added as an independent variable.
Those respondents who conformed more to the norms of the site with regards to smoking
106
(β=0.19, p<.05) were found to have significantly higher smoking cessation self-efficacy.
This can be explained by the fact that members of the site are trying to quit smoking, and
therefore the norm would be to not smoke. Further, participation level (β=0.29, p<.05),
social identification (β=0.24, p<.05) and online bridging (β=0.27, p<.05) and bonding
social capital (β=0.25, p<.05) also continued to significantly predict smoking cessation
self-efficacy. Model 5 explained 31% of variance in smoking cessation self-efficacy,
representing a significant increase from Model 4 (F (26, 181)= 18.93, p<.001).
For Model 6, social support from the site was added to the regression, and found
to be significantly associated with smoking cessation self-efficacy. Respondents who
perceived greater social support (β=0.37, p<.01) from other members of the site, were
also more likely to have higher smoking cessation self-efficacy. Additionally,
participation level (β=0.27, p<.05), social identification (β=0.22, p<.05), bridging
(β=0.19, p<.05) and bonding social capital (β=0.18, p<.05), and social norms (β=0.18,
p<.05) also significantly predicted smoking cessation self-efficacy. Model 6 accounted
for 33% of variance in smoking cessation self-efficacy, a significant increase from Model
5 (F (27, 180)= 21.23, p<.001).
In Model 7, attitude towards smoking was added to the regression, and found to
be significantly associated with smoking cessation self-efficacy, holding the other
independent variables constant. Those respondents who had more negative attitudes
towards smoking (β=0.22, p<.05) also reported significantly higher smoking cessation
self-efficacy. Holding attitude towards smoking constant, participation level (β=0.26,
p<.05), social identification (β=0.22, p<.05), bridging (β=0.18, p<.05) and bonding social
107
capital (β=0.18, p<.05), social norms (β=0.18, p<.05), and social support (β=0.38, p<.01)
also continued to significantly predict smoking cessation self-efficacy. Overall, Model 7
explained 35% of variance in smoking cessation self-efficacy, a significant increase from
Model 6 (F (28, 179) = 25.61, p<.001).
For Model 8, the interaction term “Participation level * Social identification” was
added to the regression, and found to be significantly associated (β=0.27, p<.05) with
smoking cessation self-efficacy. Respondents who did not identify as highly with other
members of the site, but who participated more on the site, were able to gain in smoking
cessation self-efficacy. As can be seen, social identification moderated the relationship
between participation level and smoking cessation self-efficacy. As such, H11 was
supported. Meanwhile, participation level (β=0.26, p<.05), social identification (β=0.18,
p<.05), bridging (β=0.18, p<.05) and bonding social capital (β=0.16, p<.05), social norms
(β=0.17, p<.05), social support (β=0.33, p<.01), and attitude towards smoking (β=0.20,
p<.05) also continued to significantly predict smoking cessation self-efficacy. Model 8 (F
(29, 178)=30.23, p<.001) explained 38% of variance in smoking cessation self-efficacy, a
significant increase of 3% from Model 7.
For Model 9, the interaction terms “Participation level * bridging social capital”
(β=0.20, p<.05), and “Participation level * bonding social capital” (β=0.18, p<.05) were
added to the regression as independent variables, and found to exert a significant effect
on smoking cessation self-efficacy. Respondents with lower levels of bridging social
capital on the site, were able to increase their smoking cessation self-efficacy through
greater participation on the site. Similarly, those who had lower levels of bonding social
108
capital on the site also scored significantly higher in smoking cessation self-efficacy as a
result of greater participation on the site. Bridging and bonding social capital therefore
moderated the association between participation level and smoking cessation self-
efficacy, thereby supporting H12a and H12b. Among the other independent variables,
participation level (β=0.25, p<.05), social identification (β=0.20, p<.05), bridging social
capital (β=0.17, p<.05), bonding social capital (β=0.16, p<.05), social norms (β=0.16,
p<.05), and social support (β=0.32, p<.01) and attitude towards smoking (β=0.19, p<.05)
continued to significantly predict smoking cessation self-efficacy. Model 9 (F (30, 177)
=30.18, p<.001) accounted for 38% of variance in smoking cessation self-efficacy, a
significant increase of 3% from Model 7.
In Model 10, the interaction term “Participation level * Social norms” (β=0.18,
P<.05) was added, and found to significantly predict smoking cessation self-efficacy.
Those respondents who did not conform as much to the social norms of other members of
the site with regards to smoking, were able to increase their levels of smoking cessation
self-efficacy, through more intensive participation on the site. Social norms therefore
moderated the association between participation level and smoking cessation self-
efficacy. Hence H13 was supported. As for the other independent variables,
participation level (β=0.23, p<.05), social identification (β=0.20, p<.05), bridging
(β=0.17, p<.05) and bonding social capital (β=0.15, p<.05), social norms (β=0.15, p<.05),
social support (β=0.32, p<.01) and attitude towards smoking (β=0.19, p<.05) significantly
predicted smoking cessation self-efficacy. Model 10 (F (29, 178)=30.21, p<.001)
109
explained 38% of variance in smoking cessation self-efficacy, a significant increase of
3% from Model 7.
Finally, in Model 11, the interaction term “Participation level * Social support”
(β=0.31, p<.01) was added to the regression, and found to have a significant impact on
smoking cessation self-efficacy. Respondents that felt lower levels of social support
from the site were able to increase their smoking cessation self-efficacy through greater
participation on the site. Social support therefore moderated the relationship between
participation level and smoking cessation self-efficacy. As such, H14 was supported.
Further, participation level (β=0.22, p<.05), social identification (β=0.20, p<.05),
bridging (β=0.16, p<.05) and bonding social capital (β=0.15, p<.05), social norms
(β=0.15, p<.05), social support (β=0.31, p<.01) and attitude towards smoking (β=0.18,
p<.05) also significantly predict smoking cessation self-efficacy. Model 11 (F (29,
178)=30.29, p<.001) accounted for 38% of variance in smoking cessation self-efficacy, a
significant increase of 3% from Model 7.
MANOVA and ANOVA Tests
One-way MANOVA and ANOVA tests were then carried out to find out whether
there was a significant difference between active participants and lurkers on the sites in
levels of social identification with other members (H15), perceived bridging and bonding
social capital on the site (H16), conforming to social norms of other members (H17),
perceived social support on the site (H18), attitude towards smoking (H19), and smoking
cessation self-efficacy (H20). Before the MANOVA and ANOVA tests could be
conducted, participation level on the site was re-coded into a categorical variable with
110
four categories: lurker, low participation, medium participation and high participation.
“Lurker” was defined as those with scores in between 18 (the minimum score after
answering 18 questions, denoting no participation) and 45, “Low participation” was
defined as those with scores in between 46 and 72, “Medium participation” was defined
as those with scores in between 73 and 99, and “High participation was defined as those
with scores in between 100 and 126. The Lurking index was entered in the one-way
MANOVA and ANOVA tests as the “Factor” variable.
First, the hypotheses (H15, H16, H17, H18, H19 and H20) were tested using a
overall, omnibus MANOVA test. The MANOVA test was done so as to reduce the
experiment-wise level of Type I error. The one-way MANOVA test revealed a
significant multivariate main effect for participation level on the site. Wilks’ λ (or % of
variance in the DVs not explained by differences in the level of participation) = .431,
F(12, 128) = 69.87, p<.001, partial eta square =.245, and observed power to detect the
effect was .985. Box’s M = 18.91, p=.23, which was non-significant, thus there were no
significant differences in the covariance matrices of the dependent variables across
participation levels, and as such, an assumption of MANOVA was not violated. The
omnibus MANOVA test revealed that there are significant differences in social
identification with other members (H15), perceived bridging and bonding social capital
on the site (H16), conforming to social norms (H17), perceived social support (H18),
attitude towards smoking (H19) and smoking cessation self-efficacy (H20) across the
participation categories.
111
Given the significance of the overall MANOVA test, the univariate main effects
were examined. Since 6 univariate significance tests were done, a smaller confidence
interval of .008 was set, so as to protect against inflated alpha error. For H15, the
univariate test revealed a significant main effect for participation on social identification,
F(3, 204) = 248.31, p<.001, partial eta square = .221, power =.837. The Levene test for
equality of variances indicated variances were not equal (F=6.85, p<.001); therefore an
alternative post-hoc test for pair-wise differences of means was used. Mean
identification score was 8.21 (SD=1.86) for lurkers, 10.8 (SD=2.09) for low participation,
13.7 (SD=3.1) for medium participation, and 20.86 (SD=4.21) for high participation.
The Tamhane test of post-hoc differences indicated significant differences at the p<.05
level in mean identification scores between lurkers and high participation, and between
low participation with high participation. As such, H15 was supported. There was a
significant difference between lurkers and active participants for social identification with
other members.
For H16a, the univariate test revealed a significant main effect for participation
level on bridging social capital, F (3, 204) = 108.6, p<.001, partial eta square =0.239,
power =0.922. The Levene test for equality of variances indicated variances were not
equal (F=3.09, p<.05); therefore an alternative post-hoc test for pair-wise differences of
means was used. Mean bridging social capital was 21.35 (SD=6.89) for lurkers, 27.66
(SD=1.23) for low participation, 34.8 (SD=8.16) for medium participation, and 37.65
(SD=5.62) for high participation. The Tamhane test of post-hoc differences indicated
significant differences at the p<.05 level in mean bridging social capital between lurkers
112
with low and high participation, and between low participation with high participation.
As such, H16a was supported. A significant difference between lurkers and active
participants was found in perceived bridging social capital.
For H16b, the univariate test revealed a significant main effect for participation
on level of bonding social capital, F (3, 204) = 68.91, p<.001, partial eta square = .189,
power =.943. The Levene test for equality of variances indicated variances were equal
(F=2.48, p=0.09); hence no alternative post-hoc test for pair-wise differences of means
was needed. Mean bonding social capital was 11.36 (SD=4.34) for lurkers, 16.1
(SD=1.36) for low participation, 17.8 (SD=3.13) for medium participation, and 19.26
(SD=3.61) for high participation. Therefore, H16b was supported. There was a
significant difference between lurkers and active participants and lurkers for bonding
social capital on the site.
For H17, the univariate test revealed a significant main effect for participation on
conforming to social norms on the site, F (3, 204) = 128.49, p<.001, partial eta
square=.193, power =.887. The Levene test for equality of variances was non-significant
(F=0.39, p=0.83), indicating variances were equal; hence an alternative post-hoc test for
pair-wise differences of means was not used. Mean social norms scores were as follows:
Lurkers (M=10.28, SD=3.36), Low participation (M=14.13, SD=4.24), Medium
participation (M=18.12, SD=2.82), and High participation (M=22.55, SD=3.56). A
significant difference was found between lurkers and active participants in conforming to
social norms regarding smoking on the site.
113
Similarly, for H18, the univariate test revealed a significant main effect for
participation on perceived social support from the site, F (3, 204) = 782.61, p<.001,
partial eta square = .135, power =.932. The Levene test for equality of variances
indicated variances were not equal (F=9.51, p<.001); hence an alternative post-hoc test
for pair-wise differences of means was used. Mean social support score was 16.89
(SD=3.85) for lurkers, 23.71 (SD=2.86) for low participation, 42.9 (SD=3.95) for
medium participation, and 57.76 (SD=5.87) for high participation. The Tamhane test of
post-hoc differences indicated significant differences at the p<.05 level in mean social
support scores between lurkers and high participation, and between low participation and
high participation. As such, H18 was supported. A significant difference was found
between lurkers and active participants for social support on the site.
For H19, the univariate test did not reveal a significant main effect for
participation on attitude towards smoking, F (3, 204) = 2.96, p=.109, partial eta square =
.465, power =.771. The Levene test for equality of variances was non-significant
(F=3.25, p=0.12), indicating variances were equal; hence an alternative post-hoc test for
pair-wise differences of means was not used. Mean attitude towards smoking scores
were as follows: Lurkers (M=124.73, SD=33.86), Low participation (M=137.21,
SD=31.67), Medium participation (M=114.6, SD=8.78), and High participation
(M=135.61, SD=27.88). As such, H19 was not supported.
For H20, the univariate test revealed a significant main effect for participation on
smoking cessation self-efficacy, F (3, 204) = 179.12, p<.001, partial eta square = .183,
power =.905. The Levene test for equality of variances was insignificant, indicating
114
variances were equal (F=3.18, p=.16); hence an alternative post-hoc test for pair-wise
differences of means was not used. Mean smoking cessation self-efficacy was 22.68
(SD= 9.23) for lurkers, 26.78 (SD=4.23) for low participation, 48.66 (SD=6.88) for
medium participation, and 52.69 (SD=10.12) for high participation. As such, H20 was
supported. A significant difference was found between lurkers and active participants for
smoking cessation self-efficacy. Results of the MANOVA and ANOVA tests are
summarized in Table 22.
Table 22: Tests of Significance Difference between Active Participants and Lurkers on
Social Variables (N=208)
Next, to answer RQ2, RQ3, RQ4 and RQ5, an overall, omnibus MANOVA test
was conducted. The one-way MANOVA test revealed a significant multivariate main
Variable F Significance Partial Eta
Square
Observed
Power
Overall
(MANOVA)
69.87 p<.001 .245 .985
Social
Identification
(H15)
248.31 p<.001 .221 .837
Bridging Social
Capital (H16a)
108.6 p<.001 .239 .922
Bonding Social
Capital (H16b)
68.91 p<.001 .189 .943
Social Norms
(H17)
128.49 p<.001 .193 .887
Social Support
(H18)
782.61 p<.001 .135 .932
Attitude
Towards
Smoking (H19)
2.96
p=.109
(ns)
.465 .771
Smoking
Cessation
Self-Efficacy
(H20)
179.12 p<.001 .183 .905
115
effect for participation level on the site. Wilks’ λ = .589, F (10, 126) = 69.31, p<.001,
partial eta square =.195, and observed power to detect the effect was .935. Box’s M =
12.91, p=.31, which was non-significant, thus there were no significant differences in the
covariance matrices of the dependent variables across participation levels. The omnibus
MANOVA test revealed that there are significant differences in perceived verbal
intimacy (RQ2), affective intimacy (RQ3), cognitive intimacy (RQ4) and physical
intimacy (RQ5) levels across the participation categories.
Given the significance of the overall MANOVA test, the univariate main effects
were examined. Since 4 univariate significance tests were done, a smaller confidence
interval of .013 was set, so as to protect against inflated alpha error. For RQ2, the
univariate ANOVA test revealed a significant main effect for participation on perceived
verbal intimacy on the site, F (3, 204) = 155.61, p<.001, partial eta square = .258, power
=.837. The Levene test for equality of variances indicated variances were not equal
(F=6.96, p<.001); therefore an alternative post-hoc test for pair-wise differences of means
was used. Mean verbal intimacy score was 18.91 (SD=5.68) for lurkers, 25.81
(SD=6.81) for low participation, 37.10 (SD=3.25) for medium participation, and 44.55
(SD=10.10) for high participation. The Tamhane test of post-hoc differences indicated
significant differences at the p<.05 level in mean verbal intimacy scores between lurkers
and high participation. To answer RQ2, a significant difference was found between
lurkers and active participants for verbal intimacy on the site.
For RQ3, the univariate ANOVA test revealed a significant main effect for
participation on perceived affective intimacy on the site, F (3, 204) = 100.81, p<.001,
116
partial eta square = .314, power =.866. The Levene test for equality of variances
indicated variances were not equal (F=4.35, p<.01); therefore an alternative post-hoc test
for pair-wise differences of means was used. Mean affective intimacy score was 40.86
(SD=18.24) for lurkers, 62.26 (SD=10.91) for low participation, 69.10 (SD=3.90) for
medium participation, and 86.80 (SD=16.55) for high participation. The Tamhane test of
post-hoc differences indicated significant differences at the p<.05 level in mean affective
intimacy scores between lurkers with medium and high participation. Therefore, to
answer RQ3, a significant difference was found between lurkers and active participants
for affective intimacy on the site.
For RQ4, the univariate ANOVA test revealed a significant main effect for
participation on perceived cognitive intimacy on the site, F (3, 204) = 76.12, p<.001,
partial eta square = .198, power =.923. The Levene test for equality of variances
indicated variances were not equal (F=4.12, p<.05); hence an alternative post-hoc test for
pair-wise differences of means was used. Mean cognitive intimacy score was 11.38
(SD=5.80) for lurkers, 16.85 (SD=3.13) for low participation, 16.75 (SD=1.48) for
medium participation, and 23.10 (SD=3.86) for high participation. The Tamhane test of
post-hoc differences indicated significant differences at the p<.05 level in mean cognitive
intimacy scores between lurkers with medium and high participation, and between
medium participation and high participation. To answer RQ3, a significant difference
was therefore found between lurkers and active participants for cognitive intimacy.
For RQ5, the univariate ANOVA test revealed a significant main effect for
participation on perceived physical intimacy on the site, F (3, 204) = 94.18, p<.001,
117
partial eta square = .173, power =.933. The Levene test for equality of variances
indicated variances were not equal (F=3.61, p<.05); hence an alternative post-hoc test for
pair-wise differences of means was used. Mean physical intimacy score was 11.12
(SD=5.86) for lurkers, 16.50 (SD=4.45) for low participation, 17.05 (SD=1.35) for
medium participation, and 23.50 (SD=6.00) for high participation. The Tamhane test of
post-hoc differences indicated significant differences at the p<.05 level in mean physical
intimacy scores between lurkers with medium and high participation, and between
medium and high participation. To answer RQ5, a significant difference was found
between lurkers and active participants for physical intimacy on the site. Results of the
MANOVA and ANOVA tests are summarized in Table 23.
Table 23: Tests of Significance Difference between Active Participants and Lurkers on
Intimacy Levels (N=208)
Variable F Significance
Level
Partial Eta
Squared
Observed
Power
Overall
(MANOVA)
69.31 p<.001 .195 .935
Verbal
Intimacy
(RQ2)
155.61 p<.001 .258 .837
Affective
Intimacy
(RQ3)
100.81 p<.001 .314 .866
Cognitive
Intimacy
(RQ4)
76.12 p<.001 .198 .923
Physical
Intimacy
(RQ5)
94.18 p<.001 .173 .933
118
Structural Equation Modeling
In the confirmatory factor analysis, all 12 variables in the study were first loaded
as one single factor to examine the fit of the one-factor measurement model versus the
12-factor measurement model. As according to Harman’s one-factor test, if common
method variance is present, the one-factor measurement model should fit the data well.
The results of the confirmatory factor analysis showed that the one-factor model did not
fit the data well (χ
2
(29) = 68.31, p<.001, GFI=0.72, AGFI=0.70, RMSEA=0.18), when
compared to the full 12-factor model. Hence, one single factor did not account for the
variance in the data, and consequently, common method variance was not considered a
threat in the data, and is unlikely to confound the interpretation of the results.
Table 24: Goodness-of-Fit indicators of Measurement Models in Confirmatory Factor
Analysis (N=208)
Model χ
2
DF χ
2
/DF GFI AGFI RMSEA
Single Factor 68.31*** 29 2.36 0.72 0.71 0.18
Twelve Factor 13.38 (ns) 16 0.84 0.98 0.96 0.01
p<.05*, p<.01**, p<.001***
Following the confirmatory factor analysis, structural equation modeling (SEM)
analysis was applied to test the proposed structural model, which included H6, H7, H8,
H9a, H9b and H10. According to Hoe (2008), a sample size of ten for each free
parameter estimated, and a critical sample size of 200, is needed to provide sufficient
statistical power for SEM data analysis. The sample size for this study was N=208, and
twelve free parameters were estimated; thereby fulfilling SEM requirements.
119
The model addressed the following: (1) four types of intimacy (verbal, affective,
cognitive, and physical) which influence participation level on the site, (2) the effect of
participation on the site on four types of social variables (social identification, bridging
and bonding social capital, social norms, and social support) and the dependent variables:
attitude towards smoking, and smoking cessation self-efficacy, and (3) the effect of each
of the four social variables on attitude towards smoking, and smoking cessation self-
efficacy.
The hypothesized structural model was tested in LISREL 8.8, using the maximum
likelihood estimation procedure. The error term for each endogenous variable was set at
a mean of 0, and a variance of 1. The model fit was assessed through the following
criteria: (1) a non-significant χ
2
statistic, (2) a goodness-of-fit index (GFI), and adjusted
goodness-of-fit index (AGFI) of 0.90 and above, and (3) a root mean square error of
approximation (RMSEA) of equal to or less than 0.06 (Byrne, 1998).
Results showed that the model was a good fit to the data. The chi-square
goodness-of-fit test was not significant (χ
2
(16) = 13.38, p=0.68), the RMSEA was < 0.01,
goodness-of-fit index (GFI) = 0.98, and the adjusted goodness-of-fit index (AGFI) =
0.96. Therefore, the model showed support for the predicted mechanisms put forth in H6
to H10. Parameter estimates and significance levels for the structural model are
summarized in Table 25, and the complete model is shown in Figure 3.
The standardized coefficients obtained in the structural model are different than
those from the multivariate regression earlier in the chapter, due to the inclusion of the
four types of intimacy, as well as attitude towards smoking as an endogenous variable.
120
Verbal intimacy (β=0.26, p<.05), affective intimacy (β=0.28, p<.05), cognitive intimacy
(β=0.20, p<.05), and physical intimacy (β=0.18, p<.05) were all positively associated
with participation level.
Participation level was also positively associated with higher levels of social
identification with other members (β=0.33, p<.01), perceived bridging social capital
(β=0.28, p<.05) and bonding social capital (β=0.22, p<.05), conforming to social norms
of other members regarding smoking (β=0.26, p<.05) and perceived social support
(β=0.38, p<.01). There was also a significant direct effect of participation level on
attitude towards smoking (β=0.37, p<.01) and smoking cessation self-efficacy (β=0.29,
p<.05). Those who participated more on the site had significantly more negative attitudes
towards smoking, and higher smoking cessation self-efficacy. Hence, H6 was supported.
Social identification with other members on the site was positively associated
with attitude towards smoking (β=0.19, p<.05) and smoking cessation self-efficacy
(β=0.21, p<.05), thereby supporting H7. Conforming to social norms was positively
associated with attitude towards smoking (β=0.19, p<.05) and smoking cessation self-
efficacy (β=0.18, p<.05), thereby supporting H8.
Bridging social capital was positively associated with attitude towards smoking
(β=0.21, p<.05) and smoking cessation self-efficacy (β=0.19, p<.05), hence supporting
H9a. Bonding social capital was positively associated with attitude towards smoking
(β=0.18, p<.05) and smoking cessation self-efficacy (β=0.17, p<.05), hence supporting
H9b. Social support was positively associated with attitude towards smoking (β=0.25,
p<.05) and smoking cessation self-efficacy (β=0.30, p<.01), hence supporting H10.
121
Further, the significant direct and indirect paths leading from participation to
attitude towards smoking and smoking cessation self-efficacy also suggest that the four
social variables (social identification, bridging and bonding social capital, social norms
and social support) are potential mediators.
Table 25: Parameter Estimates and Significance Levels for Structural Model (Standard
Errors in Parentheses) (N=208)
122
Figure 3: Structural Model for the Social Groups Approach to Smoking Cessation
(N=208)
123
Mediation Analysis
Mediation analysis (Baron & Kenny, 1986) tested whether social identification,
bridging and bonding social capital, social norms, and social support, were mediators
between participation level and attitude towards smoking, and participation level and
smoking cessation self-efficacy. The results are summarized in Tables 26 and 27.
Table 26: Mediation Analyses (Baron & Kenny, 1986) of Social Variables on
Relationship between Participation Level and Attitude towards Smoking
IV DV Mediator Pearson’s R
Participation Level Attitude Towards
Smoking
0.28**
Participation Level Social
Identification
0.26**
Social Identification Attitude Towards
Smoking
0.26**
Participation Level Attitude Towards
Smoking
Social
Identification
0.08 (ns)
Participation Level Bridging
Social Capital
0.28**
Bridging
Social Capital
Attitude Towards
Smoking
0.22*
Participation Level Attitude Towards
Smoking
Bridging
Social Capital
0.21*
Participation Level Bonding
Social Capital
0.29**
Bonding
Social Capital
Attitude Towards
Smoking
0.21*
Participation Level Attitude Towards
Smoking
Bonding
Social Capital
0.19*
Participation Level Social Norms 0.26**
Social Norms Attitude Towards
Smoking
0.31**
Participation Level Attitude Towards
Smoking
Social Norms 0.21*
Participation Level Social Support 0.35**
Social Support Attitude Towards
Smoking
0.37**
Participation Level Attitude Towards
Smoking
Social Support 0.05(ns)
p<.05*, p<.01**, p<.001***
124
Table 27: Mediation Analyses (Baron & Kenny, 1986) of Social Variables on
Relationship between Participation Level and Smoking Cessation Self-Efficacy
IV DV Mediator Pearson’s R
Participation Level Smoking Cessation
Self-Efficacy
0.37**
Participation Level Social Identification 0.26**
Social
Identification
Smoking Cessation
Self-Efficacy
0.22*
Participation Level Smoking Cessation
Self-Efficacy
Social
Identification
0.09(ns)
Participation Level Bridging
Social Capital
0.28**
Bridging
Social Capital
Smoking Cessation
Self-Efficacy
0.23*
Participation Level Smoking Cessation
Self-Efficacy
Bridging
Social Capital
0.18*
Participation Level Bonding
Social Capital
0.29**
Bonding
Social Capital
Smoking Cessation
Self-Efficacy
0.28**
Participation Level Smoking Cessation
Self-Efficacy
Bonding
Social Capital
0.19*
Participation Level Social Norms 0.26**
Social Norms Smoking Cessation
Self-Efficacy
0.36**
Participation Level Smoking Cessation
Self-Efficacy
Social Norms 0.21*
Participation Level Social Support 0.35**
Social Support Smoking Cessation
Self-Efficacy
0.36**
Participation Level Smoking Cessation
Self-Efficacy
Social Support 0.06 (ns)
p<.05*, p<.01**, p<.001***
The mediation analyses (Baron & Kenny, 1986) show that among the social
variables tested, social identification and social support both mediate the relationship
between participation level on the site and attitude towards smoking. The partial
correlation between participation level and attitude towards smoking, when social
125
identification was held constant, was non-significant (Pearson’s R=0.08, p=0.21, one-
tailed). The partial correlation between participation level and attitude towards smoking,
when social support was held constant, was non-significant (Pearson’s R=0.05, p=0.28,
one-tailed). Both social identification and social support act as mediators influencing the
relationship between participation level and attitude towards smoking. When social
identification is held constant, participation level no longer has a significant effect on
influencing attitude towards smoking. After controlling for social support, participation
level no longer has an influence on attitude towards smoking.
The mediation analyses conducted using the five social variables on the
relationship between participation level and smoking cessation self-efficacy show a
similar result with social identification and social support both mediating this
relationship. The partial correlation between participation level and smoking cessation
self-efficacy, when social identification was held constant, was non-significant
(Pearson’s R=0.09, p=0.20, one-tailed). Likewise, the partial correlation between
participation level and smoking cessation self-efficacy, controlling for social support, was
also non-significant (Pearson’s R=0.06, p=0.26, one-tailed). Therefore, social
identification and social support also act as mediators influencing the relationship
between participation level and smoking cessation self-efficacy. When social
identification on the site is held constant, participation level no longer has a significant
effect on smoking cessation self-efficacy. When social support is held constant,
participation level also no longer has an influence on smoking cessation self-efficacy.
126
CHAPTER FIVE: DISCUSSION
Study I
Study I examined smoking cessation in an offline context. It assessed the role
that people’s reference groups played in influencing their attitude towards smoking and
smoking cessation self-efficacy. Specifically, the study looked at injunctive norms
(approval or disapproval of smoking) and descriptive norms (actual smoking behavior) of
three reference groups: best friends, colleagues, and family members, as well as
respondent’s social identification with each of these groups, hypothesizing that social
identification would affect the relationships between injunctive and descriptive norms,
and smoking cessation self-efficacy. Among the findings in this study, the most
important is that social identification was not only significantly associated with smoking
cessation self-efficacy for all three reference groups: best friends (H1c), colleagues
(H2c), and family members (H3c), but was also found to moderate the effect of injunctive
and descriptive norms of all three groups (H1e, H2e, H3e) best friends (H1e), on smoking
cessation self-efficacy. This finding is in line with Hogg and Reid (2006), who
concluded that social identification plays a role in reinforcing and modifying people’s
perceptions of group norms, as well as Reed et al.’s (2006) contention that people often
use their social groups as sources of normative information regarding unhealthy behavior
like drinking and smoking. Particularly, those who identified more strongly with social
groups would conform more strongly to perceived in-group norms.
Another major finding is that both the injunctive and descriptive norms of best
friends (H1a, H1b), colleagues (H2a, H2b) and family members (H3a, H3b) had a
127
significant impact on respondents’ smoking cessation self-efficacy. Respondents in the
sample, therefore, looked at their reference groups’ approval or disapproval of smoking,
as well as their actual smoking behavior, when searching for clues as to the
appropriateness of smoking in social settings. The greater the disapproval from the
reference groups, and the less people in the reference groups smoked, the greater one’s
ability to refrain from smoking in social situations, as measured in this study by smoking
cessation self-efficacy (SEQ-12). Also, for all three reference groups, respondents’
attitude towards smoking was positively associated with their smoking cessation self-
efficacy (H1d, H2d and H3d). People who had more negative attitude towards smoking
were also more likely to have higher smoking cessation self-efficacy. This finding
replicates results from previous studies on smoking cessation (Iannotti & Bush, 1992;
Iannotti et al., 1996), whereby in order to change smoking behavior, it was important to
first target behavioral intention, which would then lead to actual behavior change. People
who are exposed to reference groups that disapprove of smoking, and in which smoking
is uncommon, would tend to develop a more negative attitude towards smoking, in line
with the injunctive and descriptive norms of these groups, and this would in turn lead to
greater ability to refrain from smoking in social situations.
Another major finding of Study I is that the injunctive norms of all three reference
groups (best friends, colleagues and family members) had a stronger effect on
respondents’ smoking cessation self-efficacy than did their descriptive norms (RQ1).
This lends further support to previous studies (Park & Smith, 2006; Rimal & Real, 2003)
which found that perceived injunctive and descriptive norms of social groups had a
128
different impact on changing people’s behavior. In the present study, injunctive norms,
or the perceived approval of smoking by members of a reference group, were found to be
more important than descriptive norms, or the actual smoking behavior of the group
members, in predicting respondents’ smoking cessation self-efficacy. Further, the
present study also found that descriptive norms of family members had the strongest
effect on respondents’ smoking cessation self-efficacy, followed by that of best friends,
and then colleagues (H5). Family members’ actual smoking behavior, therefore, had a
greater influence than that of best friends and colleagues on people’s ability to abstain
from smoking. This finding is congruent with the idea that more proximate peers are
more influential than more distal peers in effecting attitude and behavior change
(Yanovitzky et al., 2006).
In the present study, the mean age of respondents was 40.4 years, and the majority
of them were married (60.7%), and as such, it makes intuitive sense that their most
proximal reference group would be their family members, specifically their immediate
family members, like spouses and children, whom they live with and probably see every
day. In earlier studies which surveyed college students, it was found that close friends at
college were the most influential reference group for smoking (Paek & Gunther, 2007;
Perkins, 2002; Yanovitzky et al., 2006). Because these college students lived on campus,
and away from their family, their most proximal social group was their peer group, and
therefore college peers became most important in terms of setting group norms regarding
smoking and other social behavior. In the present study, family members were the most
proximal social group due to respondents being older, and married. Descriptive norms
129
of best friends and colleagues therefore become less important since respondents may not
spend as much time with these groups on a day-to-day basis. Social distance, therefore,
plays an important role in influencing people’s attitudes and behavior regarding smoking,
as they model and conform to in-group norms of those closest to them, both perceived
and actual (Lewis & Neighbors, 2006; Paek & Gunther, 2007).
Another interesting finding of the present study is that in contrast to descriptive
norms, injunctive norms of best friends had the strongest effect on respondents’ smoking
cessation self-efficacy, followed by family members, and colleagues (H4). Earlier, it was
argued that because the majority of respondents were in their 40s, and married, their most
proximal social group would be their family members. This finding suggests that for
injunctive norms (or the approval or disapproval of smoking), best friends play a more
important role than family members in effecting attitude and behavioral change regarding
smoking. A possible explanation for this is that respondents may be more intimately
connected to, or have already formed such strong bonds with their family members, that
they may not care as much about whether these family members approved or disapproved
of smoking when they smoked, as compared to best friends and colleagues, whom they
do not live with. They may therefore be more respectful of their best friends’ approval or
disapproval of smoking when deciding whether or not to light up a cigarette. A second
possible explanation is that respondents may already share similar smoking behavior as
their family members (e.g. spouses smoke together, or quit together), and therefore, their
injunctive norms become less important in influencing their smoking cessation self-
130
efficacy than those of best friends. These will of course, need to be tested in a future
study.
For all three reference groups, gender, age, income, educational level and marital
status, had a significant impact on smoking cessation self-efficacy. Female respondents
in the sample reported significantly higher smoking cessation self-efficacy than male
respondents. This gender effect remained when controlling for injunctive and descriptive
norms of the three reference groups, and only disappeared after social identification was
held constant in Model 3. A possible explanation for this is that female respondents, in
general, identified with their reference groups more strongly, and therefore when social
identification is controlled for, their smoking cessation self-efficacy scores did not differ
significantly from male respondents anymore. Another possible explanation is that
female respondents in the sample smoked less, and therefore had higher smoking
cessation self-efficacy. Descriptive statistics appear to support this explanation, since
67.6% of the 208 smokers in this study were male.
As for age, it was found that for all three reference groups, those respondents who
were older reported significantly higher smoking cessation self-efficacy, compared to
younger respondents. This positive effect was no longer significant, however, after
controlling for injunctive and descriptive norms in Model 2. Similarly, for educational
level, respondents with higher education were found to have significantly higher smoking
cessation self-efficacy. Judging from the standardized regression coefficients across the
five education groups (less than high school, high school graduate, associate degree,
bachelor degree, and masters/professional degree), each progressively higher-educated
131
group reported significantly greater smoking cessation self-efficacy than the group below
it. As such, the higher one’s educational level was, the greater the ability to refrain from
smoking cigarettes in social situations. Like age, the significant impact of educational
level also disappeared after controlling for injunctive and descriptive norms, indicating
that those who were older, and more highly educated, may be more heavily influenced by
the approval/disapproval of smoking, and actual smoking behavior of their social groups.
As for income, it was found that respondents with higher income reported higher
smoking cessation self-efficacy. In terms of marital status, respondents who were
married reported significantly higher smoking cessation self-efficacy than respondents
who were single. Those who were divorced, widowed or separated did not report
significantly different smoking cessation self-efficacy than those who were single. The
significant positive effect of having a higher income, and being married, on smoking
cessation self-efficacy remained when controlling for injunctive and descriptive norms
(Model 2), and disappeared after controlling for social identification (Model 3). As such,
it is possible that like female respondents in the sample, those with higher income, or are
married, identify more highly with their reference groups. When identification was held
constant, they no longer reported significantly higher smoking cessation self-efficacy
than those with lower income and of other marital status.
Study I contributes to existing research on smoking cessation because it proposes
that the social groups we are part of can have an important impact on our smoking
behavior. Particularly, social identification with one’s reference groups, during which
one identifies himself or herself as part of the social group, creating a shared group
132
identity (Tajfel & Turner, 1986), can affect behavioral intentions and behavioral change
with respect to addictive activities such as smoking. Because people identify with their
reference groups, often looking to these reference groups for normative information
regarding appropriate social behavior, so as to “fit in” with the social norms of the
groups, it is useful to examine whether their smoking cessation self-efficacy, or ability to
refrain from smoking in social situations, can be shaped by social identification with, and
perceived injunctive and descriptive norms of these groups. By applying the Social
Norms Approach (Perkins & Berkowitz, 1986) and Social Identity Theory (Tajfel &
Turner, 1986), Study I found strong evidence that respondents’ identification with three
reference groups: best friends, colleagues, and family members, and the perceived
injunctive and descriptive norms regarding smoking among these three reference groups,
influenced their attitude towards smoking and smoking cessation self-efficacy. Perkins
(2002) talked about how social norms are powerful agents of control, whereby behavior
is framed by norms, and the course of behavior commonly taken is typically in
accordance with normative directives of reference groups which are most important to the
individual. Group norms are reflected in the dominant attitudes, expectations and
behaviors characterizing social groups, and they serve to regulate group members’
actions so as to perpetuate and follow the collective norm (Perkins & Berkowitz, 1986).
Individual group members also inculcate the values and emotions of the social group,
defining themselves in terms of membership in the social group, thus creating a shared
social identity. As group identity becomes salient, group members experience positive
distinctiveness by conforming to group norms. Through this, normative information
133
exerts a great deal of influence on behavior within social groups. By including a social
identity measure to test the relationship between perceived norms and smoking cessation
self-efficacy, Study I demonstrated, most importantly, that social identification with
reference groups could moderate health behaviors like smoking.
Study II
Study II, meanwhile, assessed smoking cessation in an online context. It
examined the role of participation in online social networking sites for smoking
cessation, hypothesizing that participation level on the sites would significantly impact
four different social variables: (1) social identification, (2) bridging and bonding social
capital, (3) social norms, and (4) social support. These four social variables would, in
turn, have a positive effect on changing attitudes towards smoking, as well as increasing
smoking cessation self-efficacy. Additionally, each of these variables would also interact
with participation level to exert an influence on attitude towards smoking and smoking
cessations self-efficacy. Next, Study II also sought to find out whether significant
differences existed between active participants and lurkers on the sites for each of these
six variables. Finally, Study II extended previous research by Parks and Floyd (1996)
and Rau et al. (2008) by proposing that active participants and lurkers would differ on
levels of verbal, affective, cognitive and physical intimacy with another member on the
site.
Among the findings in Study II, the most important is that participation level on
the site was significantly associated with all four of the social variables tested: social
identification, social norms, bridging and bonding social capital, and social support.
134
Respondents who participated more actively on the site, including posting of messages on
forums, updating their profiles, sharing links and videos, and corresponding with other
members, were found to have significantly higher levels of social identification with
other members. Through being actively involved with various activities on the site, these
members come to share a group identity with other members whom they interact with,
resulting in a high level of social group identification on the site, thus supporting one of
the tenets set forth by SIT (Tajfel & Turner, 1986), which states that people in social
groups come to see themselves as interchangeable exemplars of the social group, and that
their identity becomes more or less tied to that of the group in question. Further, these
same members were also found to conform more strongly to the social norms of other
members on the site regarding the social acceptability (injunctive norm) and social utility
(descriptive norm) of smoking, providing support for the notion that people most often
look to important social groups in their lives for clues as to the type of behavior they
should exhibit in various situations, and following these implicit rules so as to fit in with
the group (Perkins, 2002). Additionally, more active participants also reported
significantly greater sense of belonging to the community, with the development of more
weak ties with other members of diverse background and geographic locations.
This finding points to the ability of the social networking features on the health
sites in connecting disparate people from all walks of life who may not be exposed to one
another otherwise, thereby increasing their levels of bridging social capital (Williams,
2006). At the same time, active participants were also able to build more trusting,
emotional relationships with other members online, hence increasing their levels of
135
bonding social capital. Also, participation on the site was significantly associated with
perceived social support. Those who frequently posted messages, updated their status,
and shared their experiences on the site, were able to receive more encouragement, as
well as feelings of shared solidarity and trust, thereby empowering them to take more
active steps to improve their health situations (Sarason & Sarason, 2006). Since smoking
is a social activity, it is important for people who are trying to quit to surround
themselves with other people who would guide them in the right direction. Study II
suggests that online health social networking sites provide this very function.
Another important finding in Study II is that all the independent variables tested
in the multivariate regression analysis were significantly associated with smoking
cessation self-efficacy. As previously articulated, research has shown that self-efficacy,
or the ability for someone to refrain from smoking in certain high risk situations (e.g.
when hanging out with smokers, when stressed or angry etc.), is a strong predictor of
subsequent smoking cessation (Etter & Perninger, 2001; Stretcher et al., 2005). In the
current study, participation level on the site was significantly associated with smoking
cessation self-efficacy (H6). Members who participated more on the site reported higher
smoking cessation self-efficacy than those who participated less. This is congruent with
previous studies on online smoking cessation programs (Brendryen et al., 2008;
Finkelstein et al., 2008; Houston et al., 2008; Schneider et al., 1990; Stoddard et al.,
2008), which found that the more online features one used on the site, the greater the
smoking cessation self-efficacy outcome. Social identification with other members also
had a significant impact on smoking cessation self-efficacy (H7). As Hogg (2003)
136
explained, when one’s social identity was salient, he or she would inculcate the values of
the social group as their own. Respondents who identified more highly with other
members on the site reported greater ability to refrain from smoking in social situations,
most likely because the other members on the site are also trying to quit smoking, or have
recently quit; hence social identification with these members suggests that these
respondents would have less of an impetus to smoke. Related to this is the finding that
respondents who conformed more to the social norms of other members on the site
regarding smoking, also reported significantly higher smoking cessation self-efficacy
(H8). According to the Social Norms Approach (Perkins & Berkowitz, 1986), people
look to their reference groups for clues to appropriate behavior in social situations. Study
One showed that the injunctive and descriptive norms of respondents’ offline reference
groups had an impact on their smoking cessation self-efficacy. As seen in Study II, these
respondents’ online reference group, in the form of other members of the site, also
influence their smoking cessation self-efficacy in the same way. Additionally, bridging
social capital (H9a) and bonding social capital (H9b) were also positively associated with
smoking cessation self-efficacy. In previous research, online bridging and bonding social
capital has been found to positively impact Internet users’ offline behavior (Skoric et al.,
2008; Valenzuela et al., 2009; Zywica & Danowski, 2008). In the present context,
respondents who were able to use the social networking features of the site to form more
weak ties, or loose connections to other members from disparate backgrounds and
geographic locations (bridging social capital), as well as to build more strong, trusting
relationships (bonding social capital) with people whom they already know from offline
137
(average number of offline friends who were also site members was 8), reported
significantly higher smoking cessation self-efficacy, showing that the social capital which
was developed online could have an effect on their offline social behavior, thereby
concurring with previous studies. Finally, social support was also significantly
associated with smoking cessation self-efficacy (H10). As the social support literature
has firmly established, online health groups enable people to find others with the same
health conditions so as to share experiences, encouragement and help each other with
informational, emotional and tangible needs (Braithwaite & Waldron, 1999; Sarason &
Sarason, 2006; Tanis, 2008). For respondents in the study, those who received a greater
amount of support from other members on the site with regards to quitting smoking had
significantly higher smoking cessation self-efficacy. As such, Study II shows that online
social networking sites for smoking cessation can play a complementary role to one’s
offline social groups in helping people to quit smoking, especially in circumstances
whereby one may not be able to find the necessary offline support that he or she requires
in his or her quest to stamp out cigarettes for good.
Another major contribution of Study II is that the five social variables: social
identification (H11), bridging social capital (H12a) and bonding social capital (H12b),
social norms (H13) and social support (H14) each influenced the relationship between
participation level and smoking cessation self-efficacy. In other words, participation
level on the site interacted with these four social variables to affect smoking cessation
self-efficacy scores. Respondents who identified less highly with other members on the
site, were able to increase their level of smoking cessation self-efficacy through more
138
active participation on the site. Similarly, those who reported lower levels of bridging
and bonding social capital, but who participated more on the site, were also able to
significantly increase their smoking cessation self-efficacy. Further, those who
conformed less to the social norms of other members regarding smoking, were able to use
greater participation level on the site to increase their smoking cessation self-efficacy
scores. Finally, those who perceived a lower amount of social support from other
members on the site, were also able to significantly impact their smoking cessation self-
efficacy by participating more on the site. In previous research on social networking
sites, intensive SNS usage has been found to help people who reported lower levels of
self-esteem and satisfaction with life become more engaged with their community
(Ellison et al., 2007). In the present context, greater participation on the site was able to
help people who reported lower levels of social identification, social capital, social norms
and social support, increase their ability to refrain from lighting up in social situations.
This has important implications especially for health-based social networking sites
because through more active online participation, health patients can mitigate certain
inhibiting factors in their lives, which may have been caused by their individual health
conditions, to achieve more positive health outcomes.
In terms of demographic variables, the multivariate regression analysis for
Study II showed that gender, age, income, educational level and marital status had a
significant impact on smoking cessation self-efficacy. Controlling for other demographic
attributes, respondents with high school education, associate degrees, bachelor’s degrees,
and master’s/professional degrees reported significantly higher smoking cessation self-
139
efficacy than respondents with less than high school education. Similarly, controlling for
other demographic attributes, respondents who were married also reported significantly
higher smoking cessation self-efficacy than those who were single, divorced, widowed or
separated. However, when participation level on the site was held constant, both the
effect of educational level and being married disappeared. This may suggest that
respondents who were more highly educated and who were married participated more
actively on the site, thereby leading to higher smoking cessation self-efficacy levels.
Additionally, female respondents also reported significantly higher smoking cessation
self-efficacy than male respondents. Those who were older, or had higher income, also
scored significantly higher on smoking cessation self-efficacy. This positive gender, age
and income effect remained when controlling for participation level, and only became
non-significant after holding social identification constant. This may suggest that female
respondents, older respondents or those with higher income have higher levels of
identification with other members on the site, thereby leading to greater smoking
cessation self-efficacy. None of the other remaining demographic variables had any
significant impact on smoking cessation self-efficacy.
Additionally, Study II also found significant differences between active
participants and lurkers for the four social variables tested: social identification (H15),
bridging social capital (H16a) and bonding social capital (H16b), social norms (H17), and
social support (H18), as well as for the two dependent variables: attitude towards
smoking (H19) and smoking cessation self-efficacy (H20). As existing literature on
lurking has established, lurkers can feel as if they are a part of the online community, and
140
receive social support just by reading the messages, without any form of direct
participation (Grohl, 2008; Preece et al., 2004; Van Uden-Kraan et al., 2008). The
current study, however, disputes this idea of “online bibliotherapy” as it found significant
differences between active and non-active members in terms of both the social aspects of
the sites, as well as potential health outcomes. Participants and lurkers differed
significantly in their level of social identification with other members. Those who were
more active participants saw themselves as being more of a part of the social in-group on
the site, and this led to greater smoking cessation self-efficacy. Participants and lurkers
also differed significantly on bridging and bonding social capital levels. People who
were more active participants on the sites managed to establish a greater number of weak
connections on the site, as well as to build a greater number of strong, trusting
relationships through the site. These in turn predicted greater smoking cessation self-
efficacy, because active participants were able to “cash in” their bridging and bonding
social capital during times of need (Williams, 2006), to achieve their smoking cessation
goals. It is possible that in situations when active participants are feeling the temptation
to smoke, they may be able to log on, and immediately connect to their weak or strong
online ties, thereby helping them to refrain from smoking. Additionally, active
participants and lurkers also differed significantly in conforming to social norms
regarding smoking on the site. Because these sites cater to people who are trying to quit
smoking, or have already quit, it is safe to assume that the prevailing social norms on the
site would be inhibitive to smoking. Active participants conformed more to these norms,
and as found in the study, this had a positive impact on smoking cessation self-efficacy.
141
Further, active participants and lurkers also differed significantly in social support.
Those who took part more intensively in site activities, including corresponding with
other members, offering their help and encouragement, updating their profile and posting
helpful links perceived greater social support from other members on the site. Even
though lurkers may feel a certain level of support and identification with others’
experiences by merely reading messages (Grohl, 2008; Preece et al., 2004), the vastly
increased level of social support received by active members was able to bring about a
more tangible health outcome, in the form of significantly higher smoking cessation self-
efficacy. Finally, active participants and lurkers also differed significantly in terms of
attitude towards smoking and smoking cessation self-efficacy. Those who participated
more on the site reported more negative attitudes towards smoking, and also higher
smoking cessation self-efficacy than lurkers. It is possible that through direct
participation on the site with other members, people develop a more negative stance
towards smoking. As such, active participants were also better able to refrain from
smoking when they get tempted to do so in social situations.
Further, following on previous research (Parks & Floyd, 1996; Rau et al.,
2008), Study II also found that active participants and lurkers on the sites differed
significantly in their perceived levels of verbal, affective, cognitive and physical intimacy
with another member with whom they had developed a friendship. The original levels of
development in online relationships scale was by Parks and Floyd (1996), who divided it
into six subscales: (1) interdependence, (2) depth and breadth of communication, (3) code
change, (4) predictability and understanding, (5) commitment, and (6) network
142
convergence. These were later modified by Rau et al. (2008) into two subscales: (1)
verbal intimacy, which included the breadth and depth items, and (2) affective intimacy,
which included the predictability/understanding, commitment and interdependence items.
The current study attempted to extend this line of research further by including two
additional subscales for intimacy. Cognitive intimacy included the code change items,
which measured the ability of members on the site to cognitively “read between the lines”
of each other’s messages. Questions included whether there was sharing of special
nicknames, use of private signals, language or jargon that no one else would be able to
decipher. Meanwhile, physical intimacy included the network convergence subscale, and
measured whether two members shared the same online and offline social groups, and
also introduced each other to their close friends, family members and work associates.
While Rau et al. (2008) reported results of t-tests, and used these to support their
hypotheses, the current study employed MANOVA and ANOVA between-subjects tests
which reduce associated measurement errors. RQ2, RQ3, RQ4 and RQ5 variously asked
whether there was a significant difference between active participants and lurkers on the
sites for the four types of intimacy. Results indicated that significant differences existed
between the two groups for each of the measures. Active participants reported
significantly greater levels of verbal intimacy with another member on the site (RQ2).
The breadth and depth of their communication online included a wider variety of topics,
more in-depth and honest discussions, and greater self-disclosure of personal matters.
Additionally, active participants also reported significantly higher levels of affective
intimacy with another member on the site (RQ3). Through their communication, their
143
relationship became more interdependent, they were able to better understand each other
and accurately predict each other’s reactions to various situations, and also more
committed to maintaining the relationship. Further, active participants also had
significantly greater cognitive intimacy with another member on the site, compared to
lurkers (RQ4). Those who participated more actively saw their relationship with another
member develop to an extent such that they were able to read between the lines of each
other’s communication, and use their own jargon, signals, private language, and
nicknames. Finally, active participants also differed significantly from lurkers for
physical intimacy with another member on the site (RQ5). These members saw their
relationship develop offline, as they introduced close friends, colleagues and family
members to each other, and also had overlapping social circles both online and offline.
As such, Study II lends further support to the existing literature on lurking in online
health sites, showing that active and non-active members derive significantly different
levels of relationship development with other members, and that these social relations
online eventually lead to different health outcomes depending on participation level on
the site.
The final contribution of Study II to existing research on the use of social
networking sites for smoking cessation is that it proposed a model of influence which
included four types of social variables: social identification, social norms, bridging and
bonding social capital, and social support, which were hypothesized to significantly
impact attitude towards smoking and smoking cessation self-efficacy. Participation on
the site meanwhile, was hypothesized to not only have a direct effect on both attitude
144
towards smoking and smoking cessation self-efficacy, but also to exert an indirect effect
via the four social variables. Additionally, the mediation analyses (Baron & Kenny,
1986) found that among the social variables tested, social identification and social
support both mediated the relationship between participation level and attitude towards
smoking, as well as the relationship between participation level and smoking cessation
self-efficacy. Social identification on the site, rather than participation level on the site,
was the reason for respondents developing more negative attitudes towards smoking, and
also greater smoking cessation self-efficacy. Similarly, the effect of social support on
the site, rather than participation level, also increased negative attitudes towards smoking,
and smoking cessation self-efficacy.
The model also hypothesized further that verbal, affective, cognitive and
physical intimacy, significantly affected participation level on the site. Finally, attitude
towards smoking and smoking cessation self-efficacy were hypothesized to significantly
impact each other. This model of influence was named the “Social Groups Approach to
Smoking Cessation”, and as of this writing, is the first to propose that participation level
in online health sites could significantly impact attitude towards smoking (behavioral
intention) and smoking cessation self-efficacy (behavioral change), particularly through
four types of social influences: social identification, social norms, bridging and bonding
social capital, and social support. The support for the model in SEM analysis using
LISREL 8.80, shows that, collectively, these four types of social influence are important
variables affecting health outcomes in online health SNSs. The features of SNSs,
including the ability to upload pictures, create personal profiles, connect to other
145
members through friend requests, write on others’ walls, and post links, videos and other
media, naturally lend them the ability to allow people to engage in more social,
interactive activities with others online. By including four different theoretical
perspectives: Social Identity Theory (SIT) (Tajfel & Turner, 1986), Social Norms
Approach (Perkins & Berkowitz, 1986), Social Capital (Bourdieu, 1986; Burt, 1992;
Coleman, 1988; Putnam, 2000; Williams, 2006), and Social Support (Sarason & Sarason,
2006; Tanis, 2008) to the study of online SNSs for smoking cessation, the current study
attempted to extend the research paradigm for social media in the health communication
arena. Results of the study indicate that this line of research presents some promising
answers on whether online health SNSs can lead to positive health outcomes. As online
health SNSs catering to various health conditions continue to proliferate on the Internet, it
will be up to future studies to further develop the literature, and hopefully validate the
present findings.
Together, the two studies in this dissertation sought to show that people’s social
groups, in both offline and online environments, can have an influence on their smoking
behavior. As previous research has found, one’s friends and family can have a positive
impact on their happiness, health and satisfaction with life (Christakis & Fowler, 2009).
As this dissertation suggests, the social circles that one has can significantly affect his or
her smoking behavior, and play an important role in helping one to quit smoking, or
otherwise impede the smoking cessation process. For Study I, which examined three
offline reference groups (best friends, colleagues, family members) and their impact on
respondents’ smoking cessation self-efficacy, it was found that both the injunctive and
146
descriptive norms of these reference groups were significantly associated with
respondents’ smoking cessation elf-efficacy, and that social identification with the
reference groups acted as a moderating variable in the relationship. For Study II, which
looked at participation in online health SNSs for smoking cessation, and its implications
for smoking cessation self-efficacy, it was found that social variables, including social
identification with other members, social norms of other members, bridging and bonding
social capital from the site, and perceived social support from the site, all significantly
impacted respondents’ smoking cessation self-efficacy. Further, participation was
influenced by four types of perceived intimacy (verbal, affective, cognitive, and physical)
with other members on the sites, and lurkers were found to derive significantly less
smoking cessation self-efficacy than active participants. These results lend support to the
theories studied: Social identity theory (SIT) (Tajfel & Turner, 1986), Social norms
approach (Perkins & Berkowitz, 1986), Social capital (Bourdieu, 1986; Coleman, 1988;
Putnam, 2000) and Social support (Sarason & Sarason, 2006; Tanis, 2008). In social
groups, when one’s social identity is salient, he or she would see himself or herself as
part of the in-group, and inculcate the values and norms of the in-group as their own. It is
likely that for respondents in both studies, their social identification with relevant
reference groups, whether they are best friends, colleagues, family members or other
online SNS users, play an important role in changing their values and norms regarding
smoking. Also, the injunctive and descriptive norms of reference groups are also
important to changing smoking behavior. The approval or disapproval of smoking by
one’s friends and family, as well as the actual smoking of one’s friends and family, exert
147
a heavy influence on one’s own smoking because smoking cigarettes is a social activity.
Further, in online health SNSs, perceived bridging and bonding social capital, as well as
perceived social support, from the sites, have been shown to also significantly impact
one’s smoking. The more one perceives that they have people online, both weak and
strong ties, to encourage, empower and provide social support for quitting smoking, the
easier it is for them to refrain from smoking, as measured by smoking cessation self-
efficacy. The two studies therefore show that one’s offline and online social groups play
important, and sometimes complementary, roles in impacting one’s smoking behavior.
Limitations
There are some limitations to this dissertation. First, the use of an online
questionnaire for both studies meant that all measures were self-reported by respondents.
As such, the data collected may be subject to some self-reporting biases. For Study I, it
may have been difficult for respondents to recall which people among their reference
groups smoked, as well as predict their attitudes towards smoking. For Study II,
respondents may also not remember exactly the amount of time they spent on the site, the
average number of posts per week, their number of friend connections, etc. As such,
future studies should also include system-recorded data, especially that which is provided
by the developers of health SNSs, so that data analysis can be conducted on actual
statistics of SNS usage by members of the sites. Second, both studies were cross-
sectional in design. Future studies should be longitudinal, so as to establish whether the
dependent variables in the studies: attitude towards smoking and smoking cessation self-
efficacy, change over time.
148
A third limitation is that both studies in this dissertation used the Social norms
approach (Perkins & Berkowitz, 1986) to smoking cessation, and not the Social influence
approach (Rogers, 2003; Valente et al., 2005). Particularly, social network analysis
(SNA), or the examination of friendship ties among people in a social network, would be
a particularly useful data analysis method to use, as the network diagrams would be able
to show how smoking behavior spreads throughout a network. Considering the
importance of both one’s online and offline social groups to one’s own smoking cessation
self-efficacy, it is possible that the use of SNA would have given this dissertation a more
nuanced view of behavioral uptake within the social groups studied. As such, future
studies should use SNA in addition to multivariate analysis and structural equation
modeling so as to establish a more holistic picture of smoking cessation behavior in
online and offline social networks. A fourth limitation, which applies specifically to
Study I, is that only three reference groups (best friends, colleagues and family members)
were examined. Since the respondents in the sample were very diverse in terms of age,
income, ethnicity, educational level and marital status, it is possible there may be other
reference groups in their lives whose injunctive and descriptive norms towards smoking
may be more important. These may include groups such as friends from church,
neighbors, relatives, and in some cases, media personalities, such as celebrities. Future
studies should definitely examine more, and other, reference groups based on the target
population being studied.
Fifth, the primary investigator had limited control over the type and number of
sites examined, as permission was needed from moderators to post the URL link to the
149
online questionnaire. As such, Study II included both general health SNSs (albeit those
with a smoking cessation component within the sites), as well as those SNSs specifically
for smoking cessation. Some of the sites which the primary investigator originally
identified as those he wanted to study also did not grant permission for the research.
Additionally, even though a conscious effort was made to examine sites with similar
social networking features, functions and activity, the six sites which were finally
examined in the present research did have some of these differences. As these six sites
varied in terms of both structural (design of site, number of forums and message boards,
profile information, friend connections, wall posts, etc.) and social elements (amount of
activity on site, number of active members, etc.), they may have very different
implications for the measures in the online questionnaire, particularly the social variables
(social identification, social capital, social support and social norms). Sixth, the current
research only included sites for smoking cessation. In order to establish the efficacy of
SNSs for health patients, future studies should include more types of health SNSs,
especially those catering to other health conditions, such as cancer and diabetes, so as to
establish the generalizability of the present findings in order to help build more effective
sites.
Seventh, the two studies in this dissertation did not examine whether country of
residence and amount of Internet use (including time spent on the site per week) had an
impact on the social variables, participation level, as well as attitude towards smoking
and smoking cessation self-efficacy. Future studies should examine whether there are
any cross-cultural differences with regards to how participation level may affect
150
perceived social identification, social support, social norms and social capital. Also, the
respondents in the two studies were, in general, older, more highly-educated, and of a
higher income group, compared to the general population at large. Future studies should
definitely examine lower-income and lower-educated groups, especially those from
under-served and under-privileged communities with limited access to offline resources,
who may thereby benefit more from participating in health SNSs for smoking cessation.
Also, because previous research has established that people tend to start smoking at a
younger age (Alexander et al., 2001; Valente et al., 2005), it might be useful for future
studies to examine SNSs which have a higher proportion of younger users, particularly
adolescents and young adults. Another limitation with the dissertation is that even
though the primary investigator examined smoking cessation in both offline (Study I) and
online (Study II) contexts, the independent and dependent variables in both studies were
not exactly the same. Study I only looked at social norms and social identification, while
Study II included social capital and social support as well. In order to enable comparison
of the offline context to the online context, future research should include the same
variables.
A ninth limitation is that the health SNSs examined in this dissertation were
those which have social networking features, message boards and forums, and did not
include any actual Web-Assisted Tobacco interventions (WATI). Health SNSs are only
able to provide support, both social and informational, to their members to help them quit
smoking, rather than actual tailored programs which have been found to be efficacious
for people looking to quit smoking (Bock et al., 2008; Brendryen et al.; 2008;
151
Cunningham, 2008; Graham et al., 2005). It is possible that a combination of both social
networking features, plus an Internet-delivered smoking cessation intervention, would
have the optimal effect on smoking cessation among members. As such, future studies
should examine sites which include both of these features.
Finally, use of other new types of technology, including mobile phones, may be
highly feasible for smoking cessation, when combined with the use of health SNSs.
Research has shown that the use of tailored mobile phone messages can aid in smoking
cessation (Whittaker et al., 2008; Zbikowski et al., 2008). Because mobile phones are
widely used, and can be easily accessible to health patients, they may even more useful
for delivering smoking cessation interventions than the Internet alone. Future studies
should therefore explore the use of mobile phones and other technology for smoking
cessation support and intervention. Other viable new technologies for smoking cessation
also include virtual reality and video games. For example, a Canadian study conducted
by Girard, Turcotte, Bouchard and Girard (2009) assigned participants to two virtual
reality conditions in a modified 3D-game. During four weekly sessions, one group of
participants simulated crushing virtual cigarettes, while another group grasped virtual
balls. Over a twelve-week period, Girard et al. (2009) found that the cigarette-crushing
group reported a significantly higher abstinence rate, stayed in the treatment program
longer, and reported not having relapsed at the six-month follow-up. Studies like these
point to the potential of new technologies to help smokers quit their addiction to
cigarettes. Future studies should examine the use of these new avenues for smoking
cessation interventions.
152
Conclusion
As can be seen from the two studies in this dissertation, the social groups that
people associate with, both online and offline, exert a considerable influence on their own
smoking behavior. Study I showed that people identify with certain reference groups in
their lives whose own injunctive and descriptive norms towards smoking set trends for
other members of the groups. Social identification moderated the relationship between
injunctive/descriptive norms and smoking cessation self-efficacy. Further, injunctive
norms, or perceived approval of smoking in the group, were a stronger predictor of
smoking cessation self-efficacy than descriptive norms. More proximal reference groups:
family members and best friends, were also more important in influencing smoking
among participants. Study II, on the other hand, showed that participation in online
social networking sites for smoking cessation positively influenced smoking cessation
self-efficacy, with four social variables: social identification, social norms, social capital
and social support, acting as moderators in this relationship. Further, active participants
and lurkers derived significant differences in terms of perceived social identification,
social norms, social capital and social support from the sites, and also smoking cessation
self-efficacy. The more one participated on the sites, the more social benefits one derived
from the sites, and this in turn leads to greater smoking cessation self-efficacy, or the
ability to refrain from smoking in social situations. The model of influence proposed, the
social groups approach to smoking cessation, is a useful framework to consider for
designing and implementing online smoking cessation interventions in the future.
153
Smoking continues to be a major public health problem around the world today.
As previously discussed, smoking is a social activity, with peer influence being a leading
correlate and probable cause of smoking (Alexander et al., 2001; Hoffman et al., 2006;
Valente et al., 2005). Smoking behavior, and smoking cessation, can spread quickly
within social networks because people become influenced by the smoking behavior of
other people in their social groups, and as such, the smoking norms of important
reference groups can have a major impact on one’s own smoking behavior (Christakis &
Fowler, 2009; Cobb & An, 2007). In particular, group norms reflect dominant attitudes,
expectations and behaviors among social groups, often serving to regulate individuals’
actions within the group (Perkins & Berkowitz, 1986). While earlier smoking cessation
interventions have mostly been conducted in face-to-face settings between health
professionals and patients, the Internet has, over the past two decades, become an
important avenue for the promotion of smoking cessation. Health-based social support
groups for smoking cessation have existed for a while, but recently social networking
sites have begun to proliferate as well. By incorporating social networking features,
these sites allow members to interact with others in a manner similar to how they
associate with their offline social groups.
The existing literature on smoking cessation health sites have identified features
including message boards, forums, personal profiles, friend connections, health
information updates and tailored messages as being highly efficacious in enabling people
to quit smoking (Houston et al., 2008; Schneider et al., 1990; Stretcher et al., 2005;
Whittaker et al., 2008; Zbikowski et al., 2008). More than any other medium used for
154
smoking cessation in the past, the Internet, particularly health-based SNSs, has allowed
for a much higher degree of communication and interaction among users, thereby making
available in a shorter period of time relevant information, resources and support (Boulos
& Wheeler, 2007). Members of these sites become both producers and consumers of
valuable tangible, informational and social support made available on the sites, thereby
empowering them to better manage their health issues outside of actual professional
healthcare. The main strength of health-based SNSs lies in their ability to add value to
traditional health care, by maximizing its benefits through online social interactions with
other patients, while minimizing health costs (Bottles, 2009; Miller, 2010). Health-based
SNSs play a complementary role to traditional health care, cutting across geographic and
temporal boundaries to allow people to find others who share their experiences. This
“architecture of participation” (Bottles, 2009) transforms the notion of health care as
simply being a relationship between patients, caregivers and health care professionals.
The inclusion of people’s social ties and associated norms, group behavior and resources,
within health care relationships allows patients to have more avenues whereby they can
seek and receive help. For most patients, including smokers looking to quit smoking, “a
person like me” is the most credible source of health information and provider of support
(Edelman Trust Barometer, 2008), and as such health-based SNSs have great potential to
bring about positive health benefits.
Human beings are highly social animals. As the world population continues to
proliferate, it is next to impossible for anybody to remain unconnected to other people
socially. People’s networks transmit social norms, provide a sense of identity, give rise
155
to social capital and resources, and exchange social support in times of need. The social
groups that one participates in exert a great deal of influence on attitudes and behavior.
The two studies in this dissertation posited a “Social Groups Approach” to smoking
cessation. The underlying principle is that one’s personal relationships, and the attendant
social benefits associated with these relationships, can have a positive impact on health.
By examining smoking cessation in both offline and online social contexts, this
dissertation sought to answer the question of whether the various aspects of interpersonal
communication among people, such as group identification, group norms, exchange of
social resources, and social support, can lead to a greater ability to quit smoking. The
results suggest that for many people, the social norms of important reference groups in
their lives, their identification with these groups, the social resources derived from group
membership, as well as various forms of support given and provided by others in the
social groups, play an important part in increasing one’s smoking cessation self-efficacy,
or the ability to refrain from smoking in social situations. The fact that smoking
continues to claim millions of lives each year cannot be overstated. With 10 million
people projected to die from smoking-related diseases each year by 2030, and 80% of all
deaths estimated to occur in the developing world (Balmford et al., 2008), it is of utmost
importance that public health research concentrate on finding more effective ways to curb
smoking. Health-based SNSs combine the utility of the Internet with the affordances
provided by one’s social ties, and therefore present a promising venue for the fight
against big tobacco. As more people become connected online, the Social Web will
continue to increase in importance for smoking cessation interventions. This idea,
156
combined with the recent spike in public interest about potential health benefits arising
from people’s social groups (Bottles, 2009; Christakis & Fowler, 2009; Miller, 2010),
provided the impetus for the current dissertation. Future research should continue to
investigate the intersection between technology and social ties for helping people to quit
smoking.
157
BIBLIOGRAPHY
Abrams, D. & Hogg, M.A. (1988). Comments on the motivational status of self-esteem
in social identity and intergroup discrimination. European Journal of Social
Psychology, 18(4), 317-334.
Ahn, Y. Y., Han, S., Kwak, H., Moon, S., & Jeong, H. (2007). Analysis of topological
characteristics of huge online social networking services. Proceedings of the 16th
International Conference on the World Wide Web. pp. 835–844.
Aids.Gov (2008, April 9). New media: Health-focused social networking sites.
Retrieved from: http://blog.aids.gov/2008/04/health-focused.html
Alexander, C., Piazza, M., Mekos, D. & Valente, T.W. (2001). Peers, schools and
Adolescent cigarette smoking. Journal of Adolescent Health, 29, 22-30.
Aloise-Young, P., Graham, J.W. & Hansen, W.B. (1994). Peer influence on smoking
initiation during early adolescence: A comparison of group members and group
outsiders. Journal of Applied Psychology, 79, 281-287.
American Legacy Foundation (2002). National Youth Tobacco Survey (NYTS) 2002
Questionnaire. Retrieved from:
http://www.legacyforhealth.org/PDFSurveys/NYTS_2002_4-
30_final_version.pdf
An, L.C., Schillo, B.A., Saul, J.E., Wendling, A.H., Klatt, C.M., Berg, C.J., Ahulwalia,
J.S., Kavanaugh, A.M., Christenson, M. & Luxenberg, M.G. (2008). Utilization
of smoking cessation informational, interactive and online community resources
as predictors of abstinence: Cohort study. Journal of Internet Medical Research,
10(5), e55.
Balmford, J., Borland, R. & Benda, P. (2008). Patterns of use of an automated interactive
personalized coaching program for smoking cessation. Journal of Internet
Medical Research, 10(5), e54.
Bandura, A. (2002). Social Cognitive Theory of Mass Communication. In Bryant, J. &
Zillmann, D. (Eds.). Media Effects: Advances in Theory and Research (2
nd
Edition). Mahwah, NJ: Lawrence Erlbaum.
Barak, A., Bonel-Nissim, M. & Suler, J. (2008). Fostering empowerment in online
support groups, Computers in Human Behavior, 24(5), 1867-1883.
158
Barnett, G. A., & Hwang, J. (2006). The use of the Internet for health information and
social support: A content analysis of online breast cancer discussion groups. In
M. Muero & R.E. Rice (Eds.). pp. 101-132. The Internet and Health Care:
Theory, Research & Practice. Mahwah, NJ: Lawrence Erlbaum.
Baron, R.M. & Kenny, D.A. (1986). The moderator-mediator distinction in social
psychological research: Conceptual, strategic and statistical considerations.
Journal of Personality and Social Psychology, 51(6), 1173-1182.
Bessell, T.L., McDonald, S., Silagy, C.A., Anderson, J.N., Hiller, J.E. & Sansom, L.N.
(2002). Do internet interventions for consumers cause more harm than good?: A
systematic review. Health Expert, 5(1), 28-37.
Bishop, J. (2007). Increasing participation in online communities: A framework for
human-computer interaction. Computers in Human Behavior, 23, 1881-1893.
Boase, J., Horrigan, J.B., Wellman, B. & Rainie, L. (2006). The strength of internet ties.
Pew Internet and American Life Project Report. Retrieved from:
http://www.pewinternet.org/~/media//Files/Reports/2006/PIP_Internet_ties.pdf.pd
f
Bock, B.C., Graham, A.L., Whiteley, J.A., & Stoddard, J.L. (2008). A review of web-
assisted tobacco interventions (WATIs). Journal of Medical Internet Research,
10(5), e39.
Borsari, B., & Carey, K. B. (2003). Descriptive and injunctive norms in college drinking:
A meta-analytic integration. Journal of Studies on Alcohol, 64, 331-341.
Bottles, K. (2009). Patients, doctors and health 2.0 tools. Physician Executive Journal of
Medical Management, July/Aug 2009, 22-25.
Boulos, M.N.K. & Wheelert, S. (2007). The emerging Web 2.0 social software: An
enabling suite of sociable technologies in health and health care education.
Health Information and Libraries Journal, 24, 2-23.
Bourdieu, P. (1986). The forms of capital. In J.G. Richardson (Ed). Handbook of
Theory and Research for the Sociology of Education. pp. 241-258. Westport,
CT: Greenwood.
boyd, d. m., & Ellison, N. B. (2007). Social network sites: Definition, history, and
scholarship. Journal of Computer-Mediated Communication, 13, article 11.
159
Braithwaite, D. O. & Waldron, V.R. (1999). Communication of social support in
computer-mediated groups for people with disabilities. Health Communication,
11(2), 123-151.
Brendryen, H., Drozd, F. & Kraft, P. (2008). A digital smoking cessation program
delivered through internet and cell phone without nicotine replacement (Happy
Ending): Randomized controlled trial. Journal of Internet Medical Research,
10(5), e51.
Burnett, G. (2000). Information exchange in virtual communities: A typology.
Information Research, 5(4). Retrieved from: http://InformationR.net/ir/5-
4/paper82.html
Burt, R.S. (1992). Structural Holes: The Social Structure of Competition. New York,
NY: Cambridge University Press.
Butler, B., Sproull, L., Kiesler, S. & Kraut, R. (2002). Community effort in online
groups: Who does the work and why? In S. Weisband & L. Atwater (Eds).
Leadership at a Distance. Mahwah, NJ: Lawrence Erlbaum.
Byrne, B.M. (1998). Structural Equation Modeling with LISREL, PRELIS, and
SIMPLIS: Basic Concepts, Applications, and Programming. Mahwah, NJ:
Lawrence Erlbaum.
Campbell, D. T. (1998). Recommendations for APA test standards regarding construct,
trait, or discriminant validity. The American Psychologist, 15, 546-553.
Canty-Mitchell, J., & Zimet, G. D. (2000). Psychometric properties of the
Multidimensional Scale of Perceived Social Support in urban adolescents.
American Journal of Community Psychology, 28, 391-400.
Carmines, E. & Zeller, R. (1979). Reliability and Validity Assessment. Newbury Park,
NJ: Sage Publications.
Chambliss, C. & Murray, E.J. (1979). Cognitive procedures for smoking reduction:
Symptom attribution versus efficacy attribution. Cognitive Therapy and Research,
3(1), 91-95.
Christakis, N.A. & Fowler, J.H. (2008). The collective dynamics of smoking in a large
social network. The New England Journal of Medicine, 358, 2249-2258.
Christakis, N.A. & Fowler, J.H. (2009). Connected: The Surprising Power of Our Social
Networks and How They Shape Our Lives. Boston, MA: Little, Brown and
Company.
160
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 & Social Psychology, 58, 1015-1026.
Cobb, N.K. (2007). Social networking: Online support mechanisms for smoking
cessation. Poster presented at National Conference on Tobacco or Health,
Minneapolis, MN, October 2007.
Cobb, N.K. & An, L.C. (2007). Social networking: Effective tools to help people quit
smoking. Oral presentation at National Conference on Tobacco or Health,
Minneapolis, MN, October 2007.
Cobb, N.K. & Graham, A.L. (2006). Characterizing internet searchers of smoking
cessation information. Journal of Medical Internet Research, 8(3), e17.
Cobb, N.K., Graham, A.L., Bock, B.C., Papandonatos, G. & Abrams, D.B. (2004).
Initial evaluation of a real-world internet smoking cessation system. Nicotine &
Tobacco Research, 7(2), 207-216.
Coleman, J. S. (1988). Social capital in the creation of human capital. American Journal
of Sociology, 94, S95-S121.
Conrad, K.M., Flay, B.R. & Hill, D. (1992). Why children start smoking cigarettes:
predictors of onset. Addiction, 87(12), 1711-1724.
Cummings, J. N., Sproull, L. & Kiesler, S.B. (2002). Beyond hearing: Where real world
and online support meet. Group Dynamics, 6(1), 78-88.
Cunningham, J.A. (2008). Access and interest: Two important issues in considering the
feasibility of web-assisted tobacco interventions. Journal of Internet Medical
Research, 10(5), e69.
Demiris, G. (2006). The diffusion of virtual communities in health care: Concepts and
challenges. Patient Education and Counseling, 62, 178-188.
DeVillis, R.F. (2003). Scale Development: Theory and Applications. pp. 94-96.
Thousand Oaks, CA: Sage Publications.
DiClemente, C. (1981). Self-efficacy and smoking cessation maintenance: A preliminary
report. Cognitive Therapy and Research, 5(2), 175-187.
161
Dijkstra, A., DeVries, H. & Roijackers, J. (1998). Long-term effectiveness of computer-
generated tailored feedback in smoking cessation. Health Education Research,
13(2), 207-214.
Drentea, P. & Moren-Cross, J.L. (2005). Social capital and social support on the web:
The case of an internet mother site. Sociology of Health & Illness, 27, 920-943.
Elek, E., Miller-Day, M., & Hecht, M. L. (2006). Influences of personal, injunctive, and
descriptive norms on early adolescent substance use. Journal of Drug Issues, 36,
147-172.
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.
Ennett, S.T. & Bauman, K.E. (1993). Peer group structure and adolescent cigarette
smoking: A social network analysis. Journal of Health and Social Behavior, 34,
226-236.
Ennett, S.T., Bauman, K.E. & Koch, G.G. (1994). Variability in cigarette smoking within
and between adolescent friendship cliques. Addictive Behavior, 19, 295-305.
Etter, J. (2005). Comparing the efficacy of two Internet-based, computer-tailored
smoking cessation programs: A randomized trial. Journal of Medical Internet
Research, 7(1), e2.
Etter, J. (2006). Internet-based smoking cessation programs. International Journal of
Medical Informatics, 75(1), 110-116.
Etter, J., Bergman, M., Humair, J. & Perneger, T.V. (2000). Development and validation
of a scale measuring self-efficacy of current and former smokers. Addiction,
95(6), 901-913.
Etter, J. & Perneger, T. (2001). Effectiveness of a computer-tailored smoking cessation
program: A randomized trial. Archives of Internal Medicine, 161, 2596-2601.
Eysenbach, G., Powell, J., Englesakis, M., Rizo, C. & Stern, A. (2004). Health-related
virtual communities and electronic support groups: Systematic review of the
effects of online peer to peer interactions. British Medical Journal, 328, 1166-
1172.
Finkelstein, J., Lapshin, O. & Cha, E. (2008). Feasibility of promoting smoking
cessation among methadone users using multimedia computer-assisted education.
Journal of Internet Medical Research, 10(5), e33.
162
Girard, B., Turcotte, V., Bouchard, S. & Girard, B. (2009). Crushing virtual cigarettes
reduces tobacco addiction and treatment discontinuation. CyberPsychology &
Behavior, 12(5), 477-483.
Godding, P.R. & Glasgow, R.E. (1985). Self-efficacy and outcome expectations as
predictors of controlled smoking status. Cognitive Therapy and Research, 9(5),
583-590.
Graham, A.L., Bock, B.C., Cobb, N.K., Niaura, R. & Abrams, D.B. (2005).
Characteristics of smokers reached and recruited to an internet smoking cessation
trial: A case of denominators. Nicotine & Tobacco Research, 8(S1), 43-48.
Granovetter, M. (1983). The strength of weak ties. American Journal of Sociology, 78
(6), 1360-1380.
Grohl, J.M. (2008, July 24). Do lurkers benefit from online support groups? World of
Psychology. Retrieved from:
http://psychcentral.com/blog/archives/2008/07/24/do-lurkers-benefit-from-online-
support-groups/
Gunther, A.C. & Storey, J.D. (2003). The influence of presumed influence. Journal of
Communication, 53(2), 199-215.
Habermas, J. (1984). The Theory of Communicative Action, Vol. 1: Reason and the
Rationalization of Society. Boston, MA: MIT Press.
Haddad, L.G. & Malak. M.Z. (2002). Smoking habits and attitudes towards smoking
among university students in Jordan. International Journal of Nursing Studies,
39, 793-802.
Hall, J.A. & Valente, T.W. (2007). Adolescent smoking networks: The effects of
influence and selection on future smoking. Addictive Behavior, 32(12), 3054-
3059.
Haythornthwaite, C. (2002). Strong, weak and latent ties and the impact of new media.
The Information Society, 18(5), 385-401.
Hoe, S.L. (2008). Issues and procedures in adopting structural equation modeling
technique. Journal of Applied Quantitative Methods, 3(1), 76-83.
Hoffman, B.R., Monge, P.R., Chou, C.P. & Valente, T.W. (2007). Perceived peer
influence and peer selection on adolescent smoking. Addictive Behavior, 32(8),
1546-1554.
163
Hoffman, B.R., Sussman, S., Unger, J.B. & Valente, T.W. (2006). Peer influences on
adolescent cigarette smoking: A theoretical review of the literature. Substance
Use & Misuse, 41(1), 103-155.
Hogg, M.A. (2003). Social Identity. In M.R. Leary & J.P. Tangney (Eds). Handbook of
Self and Identity. Chapter 23. New York, NY: Guilford Press.
Hogg, M.A. & Reid, S.A. (2006). Social identity, Self-categorization and the
communication of group norms. Communication Theory, 16, 7-30.
Houston, T.K., Cooper, L.A. and Ford, D.E. (2002). Internet support groups for
depression: A 1-year prospective cohort study. American Journal of Psychiatry,
159, 2062-2068.
Houston T.K., Richman, J.S., Ray, M.N., Allison, J.J., Gilbert, G.H., Shewchuk, R.M,
Kohler, C.L. & Kiefe, C.I. (2008). Internet delivered support for tobacco control
in dental practice: Randomized controlled trial. Journal of Internet Medical
Research, 10(5), e38.
Hoybye, M.T., Johansen, C. & Tjornhoj-Thomsen, T. (2005). Online interaction: Effects
of storytelling in an internet breast cancer support group. Psychooncology, 14(3),
211-20.
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.
Iannotti, R.J., Bush, P.J. & Weinfurt, K.P. (1996). Perceptions of friends’ use of alcohol,
cigarettes and marijuana among urban schoolchildren: A longitudinal analysis.
Addictive Behaviors, 21(5), 615-632.
Jones, K. (1996). Trust as an affective attitude. Ethics, 107, 4-25.
Jones, D.N., Schroeder, J.R. & Moolchan, E.T. (2004). Time spent with friends who
smoke and quit attempts among teen smokers. Addictive Behaviors, 29(4), 723-
729.
Jones, S. & Fox, S. (2009). Generations online in 2009: A Pew Internet & American Life
Project Report. Retrieved from:
http://www.floridatechnet.org/Generations_Online_in_2009.pdf
Jordan-Marsh, M. & Harden, J.T. (2005). Fictive kin: Friends as family supporting older
adults as they age. Journal of Gerontological Nursing, February 2005, 25-59.
164
Kallgren, C.A., Reno, R.R. & Cialdini, R.B. (2000). A focus theory of normative
conduct: When norms do and do not affect behavior. Personality and Social
Psychology Bulletin, 26, 1002-1012.
Katz, J. (1998). Luring the Lurkers. Retrieved from:
http://www.slashdot.org/features/98/12/28/1745252.shtml
Kavanaugh, A., Carroll, J.M., Rosson, M.B., Zin, T.T. & Reese, D.D. (2005).
Community networks: Where offline communities meet online. Journal of
Computer-Mediated Communication, 10(4), article 3.
Kerlinger, F.N. & Lee, H.B. (2000). Foundations of Behavioral Research. pp. 642-645.
Florence, KY: Wadsworth Publishing.
Kirke, D.M. (2004). Chain reactions in adolescents’ cigarette, alcohol and drug use:
Similarity through peer influence or the patterning of ties in peer networks?
Social Networks, 26, 3-28.
Kollock, P. & Smith, M. (1996). Managing the virtual commons: Cooperation and
conflict in computer communities. In S.C. Herring (Ed.). Computer-Mediated
Communication: Linguistic, Social, and Cross-Cultural Perspectives. pp. 109-
128. Amsterdam, Netherlands: John Benjamin.
Kowalski, S.D. (1997). Self-esteem and self-efficacy as predictors of success in smoking
cessation. Journal of Holistic Nursing, 15, 128-142.
Kral, G. (2006). Online communities for mutual help: Fears, fiction and facts. In M.
Murero & R.E. Rice (Eds). The Internet and Health Care: Theory, Research and
Practice. Mahwah, NJ: Lawrence Erlbaum.
Kraut, R., Kiesler, S., Boneva, B., Cummings, J., Helgegson, V., & Crawford, A. (2002).
Internet paradox revisited. Journal of Social Issues, 58(1), 49-74.
Kraut, R., Patterson, M., Lundmark, V., Kiesler, S., Mukhopadhyay, T., & Scherlis, W.
(1996). Internet paradox: A social technology that reduces social involvement and
psychological well-being? American Psychologist, 53, 1011-1031.
Kummervold, P.E., Gammon, D., Bergvik, S., Johnsen, J.K., Hasvold, T. & Rosenvinge,
J.H. (2002). Social support in a wired world: Use of online mental health forums
in Norway. Nordic Journal of Psychiatry, 56(1), 59-65.
165
Lampe, C., Ellison, N. & Steinfield, C. (2006). A Face(book) in the crowd: Social
searching vs. social browsing. Proceedings of CSCW 2006. pp. 167-170. New
York, NY: ACM Press.
Lapinski, M. K. (2006). StarvingforPerfect.com: A theoretically based content analysis
of pro-eating disorder web sites. Health Communication, 20, 243-253.
Lasker, J.N., Sogolow, E.D. & Sharim, R.R. (2005). The role of an online community for
people with a rare disease: Content analysis of messages posted on a primary
Biliary Cirrhosis mailing list. Journal of Medical Internet Research, 7(1), e10.
Lehrer, J. (2009). The buddy system: How medical data revealed secret to health and
happiness. Wired, 17(10). Retrieved from:
http://www.wired.com/medtech/health/magazine/17-
10/ff_christakis?currentPage=all
Lenert, L., Munoz, R.F., Stoddard, J., Delucchi, K., Bansod, A., Skoczen, S. & Perez-
Stable, E.S. (2003). Design and pilot evaluation of an Internet smoking cessation
program. Journal of the American Medical Information Association, 10, 16-20.
Levine, D. (2009, October). Using new media to promote adolescent sexual health:
Examples from the field. prACTice Matters. Ithaca, NY: ACT for Youth Center
of Excellence.
Lewis, M.A. & Neighbors, C. (2006). Social norms approaches using descriptive
drinking norms education: A review of the research on personalized normative
feedback. Journal of American College Health, 54(4), 213-218.
Lin, N. (1990). Social resources and social mobility: A structural theory of status
attainment. In R. L. Breiger (Ed). Social Mobility and Social Structure. pp. 247-
271. New York, NY: Cambridge University Press.
Lin, N. (1999). Social networks and social attainment. Annual Review of Sociology, 25,
467-487.
Malhotra, N.K., Kim, S.S. & Patil, A. (2006). Common method variance in IS research:
A comparison of alternative approaches and a reanalysis of past research.
Management Science, 52(12), 1865-1883.
Marzano, R. J. (2003). What Works in Schools: Translating Research into Action.
Alexandria, VA: Association for Supervision and Curriculum Development.
166
McKay, H.G., Danaher, B.G., Seeley, J.R., Lichtenstein, E. & Gau, J.M. (2008).
Comparing two web-based smoking cessation programs: Randomized controlled
trial. Journal of Medical Internet Research, 10(5), e40.
Millen, D.R. & Patterson, J.F. (2003). Identity disclosure and the creation of social
capital. Conference on Human Factors in Computing Systems (CHI 2003). pp.
720-721. Fort Lauderdale, FL: ACM Press.
Miller, C.C. (2008, October 24). Social networking for patients. The New York Times.
Retrieved from: http://bits.blogs.nytimes.com/2008/10/24/social-networking-for-
patients/
Miller, C.C. (2010, March 24). Social networks a lifeline for the chronically ill. The
New York Times, B3. Retrieved from:
http://www.nytimes.com/2010/03/25/technology/25disable.html
Miller, R.S. & Lefcourt, H.M. (1982). The assessment of social intimacy. Journal of
Personality Assessment, 46, 514-518.
Mudde, A., Kok, G., & Strecher, V. (1989). Self-efficacy and success expectancy as
predictors of the cessation of smoking. Paper presented at the First European
Congress of Psychology, Amsterdam, Netherlands.
Nicki, R.M., Remington, R.E. & MacDonald, G.A. (1985). Self-efficacy, nicotine-fading,
self-efficacy monitoring and cigarette smoking behavior. Patient Education
Counseling, 7, 249-262.
Nie, N. H., & Hillygus, D. S. (2002). The impact of internet use on sociability: Time-
diary findings. IT & Society, 1(1), 1-20.
Nonnecke, B. (2000). Lurking in email-based discussion lists. Unpublished Ph.D.
dissertation, South Bank University, London, England.
Nonnecke, B. & Preece, J. (2003). Lurker demographics: Counting the silent. Paper
presented at the ACM CHI 2000 Conference on Human Factors in Computing
Systems, Hague: The Netherlands, pp. 73-80.
Nonnecke, B., Preece, J. & Andrews, D. (2004). What lurkers and posters think of each
other. Paper presented at the Proceedings of the 37
th
Annual Hawaii International
Conference on System Sciences.
Norman, C.D., McIntosh, S., Selby, P. & Eysenbach, G. (2008). Web-assisted tobacco
interventions: Empowering change in the global fight for the public’s (e)health.
Journal of Internet Medical Research, 10(5), e1.
167
Norris, P. (2002). The bridging and bonding role of online communities (Editorial).
Harvard International Journal of Press/Politics, 7(3).
Ockene, J.K., Emmons, K.M., Mermelstein, R.J., Perkins, J.A., Bonollo, D.S., Voorhees,
C.C. & Hollis, J.F. (2000). Relapse and maintenance issues for smoking
cessation. Health Psychology, 19, 17-31.
Paek, H.J. & Gunther, A.C. (2007). How peer proximity moderates indirect media
influence on adolescent smoking. Communication Research, 34, 407-432.
Papacharissi, Z. (2002). The virtual sphere: The Internet as a public sphere. New Media
& Society, 4(1), 9-27.
Parks, R.M. & Floyd, K. (1996). Making friends in cyberspace. Journal of Computer-
Mediated Communication, 1(4). Retrieved from:
http://jcmc.indiana.edu/vol1/issue4/parks.html
Perkins, H.W. (2002). Social norms and the prevention of alcohol misuse in collegiate
contexts. Journal of Studies on Alcohol, 14, 164-172.
Perkins, H.W. & Berkowitz, A.D. (1986). Perceiving the community norms of alcohol
use among students: Some research implications for campus alcohol education
programming. Substance Use & Misuse, 21(9), 961-976.
Phua, J.J. (2010). Influence of peer norms and popularity on smoking and drinking
behavior among college fraternity members: A social network analysis. Paper
presented at the International Communication Association conference, June 2010
in Singapore, Singapore.
Phua, J.J. (2008). Posters and lurkers: Participation in online smoking cessation
communities and its implications for perceived social support and smoking
cessation self-efficacy. Paper presented at the National Communication
Association conference, November 2008 in San Diego, California.
Portes, A. (1998). Social capital: Its origins and applications in modern sociology.
Annual Review of Sociology, 24, 1-24.
Preece, J. (2000). Online Communities: Designing Usability, Supporting Sociability.
New York, NY: Wiley.
Preece, J. & Maloney-Krichmar, D. (2005). Online communities: Design, theory and
practice. Journal of Computer Mediated Communication, 2005.
168
Preece, J., Nonnecke, B. & Andrews, D. (2004). The top five reasons for lurking:
Improving community experiences for everyone. Computers in Human Behavior,
2, 1.
Putnam, R. D. (2000). Bowling alone: The collapse and revival of American community.
New York, NY: Simon & Schuster.
Rabius, V., Pike, K.J., Wiatrek, D. & McAlister, A.L. (2008). Comparing internet
assistance for smoking cessation: 13-month follow-up of a six-arm randomized
controlled trial. Journal of Medical Internet Research, 10(5), e45.
Rains, S.A. & Young, V. J. (2007). The health outcomes associated with participating in
computer-mediated support groups: A meta-analysis. Tucson, AZ: University of
Arizona Press.
Ramirez, A., Zhang, S., McGrew, C. & Lin, S.F. (2007). Relational communication in
computer-mediated interaction revisited: A comparison of participant-observer
perspectives. Communication Monographs, 74(4), 492-516.
Rau, P.L.P., Gao, Q. & Ding, Y. (2008). Relationship between the level of intimacy and
lurking in online social network services. Computers in Human Behavior, 24,
2757-2770.
Real, K. & Rimal, R.N. (2007). Friends talk to friends about drinking: Exploring the role
of peer communication in the theory of normative social behavior. Health
Communication, 22(2), 169-180.
Reed, M.B., Lange, J.E., Ketchie, J.M. & Clapp. J.D. (2007). The relationship between
social identity, normative information and college student drinking. Social
Influence, 2(4), 269-294.
Resnick, P. (2001). Beyond bowling together: SocioTechnical capital. In J. Carroll (Ed.),
HCI in the New Millennium. pp. 647-672. New York, NY: Addison-Wesley.
Rimal, R.N. & Real, K. (2003). Understanding the influence of perceived norms on
behaviors. Communication Theory, 13(2), 184-203.
Ritt-Olson, A., Unger, J.B., Valente, T.W., Nezami, E., Chou, C.P., Trinidad, D., Milam,
J., Earleywine, M., Tan, S. & Johnson, C.A. (2005). Exploring peers as mediator
of the association between depression and smoking in young adolescents.
Substance Use & Misuse, 40(1), 77-98.
169
Rodgers, S. & Chen, Q. (2005). Internet community group participation: Psychosocial
benefits for women with breast cancer. Journal of Computer Mediated
Communication, 10(4), 1-42.
Rogers, E.M. (2003). Diffusion of Innovations. (5
th
edition). New York, NY: Free Press.
Salem, D.A., Bogar, G.A. & Reid, C. (1997). Mutual help goes online. Journal of
Community Psychology, 25(2), 198-207.
Sarason, B.R. & Sarason, I.G. (2006). Close relationships and social support:
Implications for the measurement of social support. In A.L. Vangelisti & D.
Perlman (Eds.). The Cambridge Handbook of Personal Relationships. New
York, NY: Cambridge University Press.
Saul, J.E., Schillo, B.A., Evered, S., Luxenberg, M.G., Kavanaugh, A., Cobb, N. & An,
L.C. (2007). Impact of a statewide internet-based tobacco cessation intervention.
Journal of Internet Medical Research, 9(3), e28.
Schneider, S.J. & Tooley, J. (1986). Self-help computer conferencing. Computers and
Biomedical Research, 19, 274-281.
Schneider, S.J., Walter, R. & O’Donnell, R. (1990). Computerized communication as a
medium for behavioral smoking cessation treatment: controlled evaluation.
Computers in Human Behavior, 6, 141-151.
Schweizer, K. J., Leimeister, J.M. & Krcmar, H. (2006). Virtual communities & the
social network of cancer patients. Americas Conference on Information Systems,
Acapulco, Mexico.
Sharf, B.F. (1997). Communicating breast cancer online: Support and empowerment on
the Internet. Women & Health, 26(1), 65-84.
Shaw, B. R., Hawkins, R., McTavish, F., Pingree, S., & Gustafson, D. (2006). Effects of
insightful disclosure within computer mediated support groups on women with
breast cancer. Health Communication, 19, 133-142.
Shiffman S., Balabanis M.H., Paty J.A., Engberg J., Gwaltney C.J., Liu K.S., Gnys M.,
Hickcox, M. & Paton, S.M. (2000). Dynamic effects of self-efficacy on smoking
lapse and relapse. Health Psychology, 19(4), 315-323.
Shklovski, I., Kiesler, S. & Kraut, R. E. (2006). The Internet and Social Interaction: A
Meta-Analysis and Critique of Studies. In R. Kraut, M. Brynin & S. Kiesler (Eds).
Computers, Phones and the Internet: The Social Impact of Information
Technology. pp. 251-264. Oxford, England: Oxford University Press.
170
Skoric, M.M., Ying, D. & Ng, Y. (2009). Bowling online, not alone: Online social
capital and political participation in Singapore. Journal of Computer-Mediated
Communication, 14(2), 414-433.
Stacy, A. W., Sussman, S., Dent, C. W., Burton, D., & Flay, B. R. (1992). Moderators of
peer social influence in adolescent smoking. Personality and Social Psychology
Bulletin, 18, 163-172.
Stoddard, J.L., Augustson, E.M. & Moser, R.P. (2008). Effect of adding a virtual
community (bulletin board) to Smokefree.gov: Randomized controlled trial.
Journal of Medical Internet Research,10(5), e53.
Stoddard, J.L., Delucchi, K.L., Munoz, R.F., Collins, N.M., Perez-Stable, E.J. &
Augustson, E. (2005). Smoking cessation research via the Internet: A feasibility
study. Journal of Health Communication, 10, 27-41.
Stretcher, V.J., Becker, M.H., Kirscht, J.P., Eraker, S.A. & Graham, R. (1985).
Psychosocial aspects of changes in cigarette smoking behavior. Patient Education
Counselling, 7, 249-262.
Stretcher, V.J., DeVellis, M., Becker, M. & Rosenstock, I. (1986). The role of self-
efficacy in achieving health behavior change. Health Education Behavior, 13, 73-
92.
Stretcher, V.J., McClure, J., Alexander, G., Chakraborty, B., Nair, V., Konkel, J., Greene,
S., Couper, M., Carlier, C., Wiese, C., Little, R., Pomerleau, C. & Pomerleau, O.
(2008). The role of engagement in a tailored web-based smoking cessation
program: Randomized controlled trial. Journal of Medical Internet Research,
10(5), e36.
Strecher, V.J., Shiffman, S. & West, R. (2005). Randomized controlled trial of a web-
based computer-tailored smoking cessation program as a supplement to nicotine
patch therapy. Addiction, 100(5), 682-688.
Szabo, L. (2008, May 22). Study: Smokers get the signal from social networks and quit.
USA Today. Retrieved from: http://www.usatoday.com/news/health/2008-05-21-
smoking-quitting_N.htm
Tajfel, H. (1978). Social categorization, social identity and social comparison. In H.
Tajfel (Ed.). Differentiation between social groups: Studies in the social
psychology of intergroup relations. pp. 61-76. London, England: Academic
Press.
171
Tajfel, H., Billig, M.G., Bundy, R.P. & Flament, C. (1971). Social categorization and
intergroup behavior. European Journal of Social Psychology, 1(2), 149-178.
Tajfel, H. & Turner, J.C. (1986). The social identity theory of intergroup behavior. In
D.A. Stapel & H. Blanton, H. (Eds). Social Comparison Theories: Key Readings
in Social Psychology. New York, NY: Taylor & Francis.
Takahashi, Y., Satomura, K., Miyagishima, K., Nakahara, T., Higashiyama, A., Iwai, K.
& Nakamura, K. (1999). A new smoking cessation program using the Internet.
Tobacco Control, 8(1), 109-110.
Takahashi, M., Fujimoto, M. & Yamasaki, N. (2003). The active lurker: A new viewpoint
for evaluating the influence of an in-house online community. SIGGROUP
Bulletin, 23(3), 29-33.
Tanis, M. (2008). What makes the internet a place to seek social support. In E.A.
Konjin, S. Utz, M. Tanis & S.B. Barnes (Eds). Mediated Interpersonal
Communication. New York, NY: Routledge.
Turner, J.C., Hogg, M.A., Oakes, P.J., Reicher, S.D. & Wetherell, M.S. (1987).
Rediscovering the Social Group: A Self-Categorization Theory. Oxford, England:
Blackwell.
Turner, J.W., Grube, J.A. & Meyers, J. (2001). Developing an optimal match within
online communities: An exploration of CMC support communities and traditional
support. Journal of Communication, June 2001, 231-251.
Valente, T.W. (1995). Network Models of the Diffusion of Innovations. Cresskill, NJ:
Hampton Press.
Valente, T.W., Chou, C.P. & Pentz, M.A. (2007). Community coalition networks as
systems: Effects of network change on adoption of evidence-based prevention.
American Journal of Public Health, 97(5), 880-886.
Valente, T.W., Hoffman, B.R., Ritt-Olson, A., Lichtman, K. & Johnson, C.A. (2003).
Effects of a social network method for group assignment strategies on peer-led
tobacco prevention programs in schools. American Journal of Public Health, 93,
1837-1843.
Valente, T.W., Unger, J. & Johnson, A.C. (2005). Do popular students smoke? The
association between popularity and smoking among middle school students.
Journal of Adolescent Health, 37, 323-329.
172
Valenzuela, S., Park, N. & Kee, K.F. (2009). Is there social capital in a social network
site?: Facebook use and college students’ life satisfaction, trust and participation.
Journal of Computer-Mediated Communication,14, 875-901.
VanAlstyne, M., & Brynjolffson, E. (2005). Global village or cyber-balkans? Modeling
and measuring the integration of electronic communities. Management Science,
51(6), 851-868.
Van Uden-Kraan, C.F., Drossaert, C.H.C., Taal, E., Seydel, E.R. & Van de Laar,
M.A.F.J. (2008). Self-reported differences in empowerment between lurkers and
posters in online patient support groups. Journal of Medical Internet Research,
10(2), e18.
Walsh, M. (2009). eMarketer: U.S. social network users to grow 44% by 2013. Online
Media Daily, February 17, 2009. Retrieved from:
http://www.mediapost.com/publications/?fa=Articles.showArticle&art_aid=1004
85
Walther, J. B., & Boyd, S. (2002). Attraction to computer-mediated social support. In
C.A. Lin & D. Atkin (Eds). Communication Technology and Society: Audience
Adoption and Uses. pp.153-188. Cresskill, NJ: Hampton Press.
Weinberg, R., Gould, D. & Jackson, A. (1979). Expectations and Performances: An
empirical test of Bandura’s Self-Efficacy theory. Journal of Sport Psychology, 1,
320-331.
Wellman, B. & Berkowitz, S.D. (1988). Social Structures: A Network Approach.
Cambridge, England: Cambridge University Press.
Wellman, B., Quan-Haase, A.Q., Witte, J. & Hampton, K. (2001). Does the Internet
increase, decrease or supplement social capital? Social networks, participation
and community commitment. American Behavioral Scientist, 45, 436-455.
White, M. & Dorman, S.M. (2001). Receiving social support online: Implications for
health education. Health Education Research, 16(6), 693-707.
Whittaker, R., Maddison, R., McRobbie, H., Bullen, C., Denny, S., Dorey, E., Ellis-
Pegler, M., Van Rooyen, J. & Rodgers, A. (2008). A multimedia mobile phone-
based youth smoking cessation intervention: Findings from content development
and piloting studies. Journal of Medical Internet Research, 10(5), e49.
Williams, D. (2006). Measuring bridging and bonding online and off: The development
and validation of a social capital Instrument. Journal of Computer-Mediated
Communication, 11(2).
173
Williams, D. (2007). The impact of time online: Social capital and cyberbalkanization.
CyberPsychology & Behavior. Retrieved from:
http://www.dmitriwilliams.com/Timeonlineprepub.doc
Williams, F. & Monge, P. (2001). Reasoning with Statistics: How to Read Quantitative
Research. pp. 29-30. Florence, KY: Wadsworth Publishing.
Woolcock, M. (1998). Social capital and economic development: Towards a theoretical
synthesis and policy framework. Theory and Society, 27, 151-208.
Wright, K. (2002). Social support within an on-line cancer community: An assessment of
emotional support, perceptions of advantages and disadvantages, and motives
focusing the community from a communication perspective. Journal of Applied
Communication Research, 30(3), 195-209.
Wright, K.B. & Bell, S.B. (2003). Health related support groups on the Internet: Linking
Empirical findings to social support and Computer Mediated Communication
Theory. Journal of Health Psychology, 8(1), 39-54.
Yanovitzky, I., Stewart, L.P. & Lederman, L.C. (2006). Social distance, perceived
drinking by peers and alcohol use by college students. Health Communication,
19(1), 1-10.
Young, R., Oei, T. & Crook, G.M. (1991). Development of a drinking self-efficacy
questionnaire. Journal of Psychopathology and Behavioral Assessment, 13, 1
Yu, T.K., Lu, L.C. & Liu, T.F. (2010). Exploring factors that influence knowledge
sharing behavior via weblogs. Computers in Human Behavior, 26, 32-41.
Zbikowski, S.M., Hapgood, J., Barnwell, S.S. & McAfee, T. (2008). Phone and web-
based tobacco cessation treatment: Real-world utilization patterns and outcomes
for 11,000 tobacco users. Journal of Medical Internet Research, 10(5), e41.
Zimet, G.D., Dahlem, N.W., Zimet, S.G. & Farley, G.K. (1988). The multidimensional
scale of perceived social support. Journal of Personality Assessment, 52, 30-41.
Zywica, J. & Danowski, J. (2008). The faces of Facebookers: Investigating social
enhancement and social compensation hypotheses; predicting Facebook and
offline popularity from sociability and self-esteem, and mapping the meanings of
popularity with semantic networks. Journal of Computer-Mediated
Communication,14(1), 1-34.
174
APPENDIX:
ONLINE QUESTIONNAIRE FOR USERS OF SMOKING CESSATION HEALTH
WEBSITES
Section 1: Social Identity Measure (Offline)
Best Friends
1) How much do you feel you identify with your best friends?
Do not identify 1 2 3 4 5 6 7 Strongly identify
2) How similar do you feel your attitudes and beliefs are to your best friends?
Very dissimilar 1 2 3 4 5 6 7 Very similar
3) To what extent do you feel strong bonds to your best friends?
No strong bond 1 2 3 4 5 6 7 Very strong bond
4) How important are your best friends to your sense of who you are - your self-identity?
Not important 1 2 3 4 5 6 7 Very important
Colleagues
1) How much do you feel you identify with your colleagues?
Do not identify 1 2 3 4 5 6 7 Strongly identify
2) How similar do you feel your attitudes and beliefs are to your colleagues?
Very dissimilar 1 2 3 4 5 6 7 Very similar
3) To what extent do you feel strong bonds to your colleagues?
No strong bond 1 2 3 4 5 6 7 Very strong bond
4) How important are your colleagues to your sense of who you are - your self-identity?
Not important 1 2 3 4 5 6 7 Very important
Family members
1) How much do you feel you identify with your family members?
Do not identify 1 2 3 4 5 6 7 Strongly identify
2) How similar do you feel your attitudes and beliefs are to your family members?
Very dissimilar 1 2 3 4 5 6 7 Very similar
3) To what extent do you feel strong bonds to your family members?
No strong bond 1 2 3 4 5 6 7 Very strong bond
175
4) How important are your family members to your sense of who you are - your self-
identity?
Not important 1 2 3 4 5 6 7 Very important
Section 2: Perceived Social norms measure (Offline)
Best Friends
1) Supposing you are a smoker, how would your best friends respond if they know about
your smoking?
Strong disapproval 1 2 3 4 5 6 7 Strong approval
2) Supposing you are a smoker, how would your best friends respond if you smoke
cigarettes around them?
Strong disapproval 1 2 3 4 5 6 7 Strong approval
3) How common is smoking among your best friends?
Not common 1 2 3 4 5 6 7 Very common
4) How socially acceptable is smoking among your best friends?
Very unacceptable 1 2 3 4 5 6 7 Very acceptable
Colleagues
1) Supposing you are a smoker, how would your colleagues respond if they know about
your smoking?
Strong disapproval 1 2 3 4 5 6 7 Strong approval
2) Supposing you are a smoker, how would your colleagues respond if you smoke
cigarettes around them?
Strong disapproval 1 2 3 4 5 6 7 Strong approval
3) How common is smoking among your colleagues?
Not common 1 2 3 4 5 6 7 Very common
4) How socially acceptable is smoking among your colleagues?
Very unacceptable 1 2 3 4 5 6 7 Very acceptable
Family members
1) Supposing you are a smoker, how would your family members respond if they know
about your smoking?
Strong disapproval 1 2 3 4 5 6 7 Strong approval
2) Supposing you are a smoker, how would your family members respond if you smoke
cigarettes around them?
Strong disapproval 1 2 3 4 5 6 7 Strong approval
176
3) How common is smoking among your family members?
Not common 1 2 3 4 5 6 7 Very common
4) How socially acceptable is smoking among your family members?
Very unacceptable 1 2 3 4 5 6 7 Very acceptable
Section 3: Attitudes towards Smoking
1) All forms of smoking are dangerous as opposed to only heavy smoking.
Strongly disagree 1 2 3 4 5 6 7 Strongly Agree
2) Smoking during pregnancy is harmful to the unborn baby.
Strongly disagree 1 2 3 4 5 6 7 Strongly Agree
3) Smoking should be avoided.
Strongly disagree 1 2 3 4 5 6 7 Strongly Agree
4) Smoking seriously damages health.
Strongly disagree 1 2 3 4 5 6 7 Strongly Agree
5) Smoking shortens a person’s life.
Strongly disagree 1 2 3 4 5 6 7 Strongly Agree
6) Smoking is a purposeless activity.
Strongly disagree 1 2 3 4 5 6 7 Strongly Agree
7) Smokers can totally reverse damage to their health by giving up smoking.
Strongly disagree 1 2 3 4 5 6 7 Strongly Agree
8) Smoking kills.
Strongly disagree 1 2 3 4 5 6 7 Strongly Agree
9) Second-hand smoke is harmful to the health of non-smokers.
Strongly disagree 1 2 3 4 5 6 7 Strongly Agree
10) Smokers die younger than non-smokers.
Strongly disagree 1 2 3 4 5 6 7 Strongly Agree
11) Smoking is a revolting habit.
Strongly disagree 1 2 3 4 5 6 7 Strongly Agree
12) No one should be allowed to smoke.
Strongly disagree 1 2 3 4 5 6 7 Strongly Agree
177
13) The risk of developing lung cancer as a direct result of smoking is very high.
Strongly disagree 1 2 3 4 5 6 7 Strongly Agree
14) The damage done through the inhalation of tobacco smoke is irreversible.
Strongly disagree 1 2 3 4 5 6 7 Strongly Agree
15) The lung cancer rate is significantly higher for smokers than non-smokers.
Strongly disagree 1 2 3 4 5 6 7 Strongly Agree
16) One does not have to smoke for a long time to be in danger of developing tobacco-
related disorders.
Strongly disagree 1 2 3 4 5 6 7 Strongly Agree
17) Smokers are more exposed to heart and arteriosclerosis diseases than non-smokers.
Strongly disagree 1 2 3 4 5 6 7 Strongly Agree
18) Smoking is one of life’s basic pleasures (REVERSED)
Strongly disagree 1 2 3 4 5 6 7 Strongly Agree
19) There is nothing like a good smoke (REVERSED)
Strongly disagree 1 2 3 4 5 6 7 Strongly Agree
20) Only heavy smoking is dangerous (REVERSED)
Strongly disagree 1 2 3 4 5 6 7 Strongly Agree
21) Smoking is not as harmful as taking drugs or drinking alcohol (REVERSED)
Strongly disagree 1 2 3 4 5 6 7 Strongly Agree
22) Smoking is less of a danger than other risks, such as the risk of a car accident
(REVERSED)
Strongly disagree 1 2 3 4 5 6 7 Strongly Agree
23) Statistics that show a relationship between smoking and health hazards are generally
misleading (REVERSED)
Strongly disagree 1 2 3 4 5 6 7 Strongly Agree
24) Anti-smoking advertisements exaggerate the dangers of smoking (REVERSED)
Strongly disagree 1 2 3 4 5 6 7 Strongly Agree
25) Smoking is relatively harmless (REVERSED)
Strongly disagree 1 2 3 4 5 6 7 Strongly Agree
178
26) Smoking low-tar cigarettes reduces the risk of developing serious diseases
(REVERSED)
Strongly disagree 1 2 3 4 5 6 7 Strongly Agree
27) The health of non-smokers is not affected by breathing cigarette smoke
(REVERSED)
Strongly disagree 1 2 3 4 5 6 7 Strongly Agree
28) Life is too short to worry about the harmful effects of smoking (REVERSED)
Strongly disagree 1 2 3 4 5 6 7 Strongly Agree
29) One has to smoke for a long period of time to be in danger of developing serious
disease (REVERSED)
Strongly disagree 1 2 3 4 5 6 7 Strongly Agree
30) There is no significant differences regarding mortality rate between smokers and non-
smokers (REVERSED)
Strongly disagree 1 2 3 4 5 6 7 Strongly Agree
31) Many old people who have smoked for years and have not developed lung cancer is
clear evidence that lung cancer is not caused by smoking (REVESED)
Strongly disagree 1 2 3 4 5 6 7 Strongly Agree
Section 4: Demographic Characteristics
1) What is your gender?
(1) Male, (2) Female
2) What is your age? ____
3) What is your ethnicity?
(1) Asian, (2) Black, (3) Latino, (4) White, (5) Native American, (6) Mixed/Other
4) What is your marital status?
(1) Single, Never Married, (2) Married, (3) Divorced, (4) Widowed, (5) Separated
5) What is your highest educational level?
(1) None, (2) Elementary, (3) Junior high/middle school, (4) Some high school/did not
finish,
(5) high school graduate, (6) Vocational/trade school, (7) some college/no degree, (8)
Associate degree (2 year degree), (9) Bachelor’s degree (4 year degree), (10) Master’s
degree, (11) Professional degree (MD, LLD), (12) Doctorate degree (PhD, EDD)
179
6) What is your annual household income?
(1) $25,000 or less, (2) $25,001 to $40000, (3) $40001 to $60000, (4) $60001 to $80000,
(5) $80001 to $100000, (6) More than $100,000
7) What is your country of citizenship? _____
8) What is your country of residence? ______
Section 5: Smoking Measure
1) Have you ever tried cigarette smoking, even 1 or 2 puffs?
(1) Yes (continue to 2), (2) No (skip to 6), (3) Don’t know (skip to 6)
For Smokers Only (1=Yes)
2) How many cigarettes have you smoked in your entire life?
(1) 1 or more puffs, but never a whole cigarette, (2) 1 cigarette, (3) 2 to 5 cigarettes,
(4) 6 to 15 cigarettes or about half a pack, (5) 16 to 25 cigarettes or about a pack, (6) 16
to 99 cigarettes or between 1 to 5 packs, (7) 5 packs or more, (8) Don’t know
3) Have you ever smoked at least one cigarette every day for 30 days?
(1) Yes, (2) No, (3) Don’t know
4) During the last 30 days, on how many days did you smoke cigarettes, even 1 or 2
puffs?
(1) 0 days (Skip to 6), (2) 1 or 2 days, (3) 3 to 5 days, (4) 6 to 9 days, (5) 10 to 19 days,
(6) 20 to 29 days, (7) All 30 days, (8) Don’t know
5) During the last 30 days, on the days you smoked, on average, how many cigarettes did
you smoke per day?
(1) <1 per day, (2) 1 cigarette per day, (3) 2 to 5 cigarettes per day, (4) 6 to 10 cigarettes
per day, (5) 11 to 20 cigarettes per day, (6) more than 20 cigarettes
For Non-Smokers Only (1=No, Don’t Know)
6) Do you think you will smoke a cigarette, even one of two puffs, at anytime during the
next year?
(1) Definitely yes, (2) Probably yes, (3) I’m not sure, (4) Probably not, (5) Definitely not
7) Do you think you will be smoking cigarettes 5 years from now?
(1) Definitely yes, (2) Probably yes, (3) I’m not sure, (4) Probably not, (5) Definitely not
8) If one of your best friends offered you a cigarette in the next 30 days, would you
smoke it?
(1) Definitely yes, (2) Probably yes, (3) I’m not sure, (4) Probably not, (5) Definitely not
180
Section 6: Social Identification (Online)
1) How much do you feel you identify with other members of the site?
Do not identify 1 2 3 4 5 6 7 Strongly identify
2) How similar do you feel your attitudes and beliefs are to other members of the site?
Very dissimilar 1 2 3 4 5 6 7 Very similar
3) To what extent do you feel strong bonds to other members of the site?
No strong bond 1 2 3 4 5 6 7 Very strong bond
4) How important are other members of the site to your sense of who you are – your self-
identity?
Not important 1 2 3 4 5 6 7 Very important
Section 7: Bridging Social Capital (Online)
1) I feel I am a part of the community on this site.
Strongly disagree 1 2 3 4 5 6 7 Strongly Agree
2) I am interested in what goes on in this online community.
Strongly disagree 1 2 3 4 5 6 7 Strongly Agree
3) This online community is a good place to be.
Strongly disagree 1 2 3 4 5 6 7 Strongly Agree
4) I will be willing to offer my help to other people on the site.
Strongly disagree 1 2 3 4 5 6 7 Strongly Agree
5) Interacting with other people on the site makes me want to try new things
Strongly disagree 1 2 3 4 5 6 7 Strongly Agree
6) Interacting with other people on the site makes me feel like part of a larger
community
Strongly disagree 1 2 3 4 5 6 7 Strongly Agree
7) I am willing to spend time to support others in this online community
Strongly disagree 1 2 3 4 5 6 7 Strongly Agree
8) In this online community, I come into contact with new people all the time
Strongly disagree 1 2 3 4 5 6 7 Strongly Agree
9) Interacting with other people in this online community reminds me that everyone
in the world is connected
Strongly disagree 1 2 3 4 5 6 7 Strongly Agree
181
Section 8: Bonding Social Capital (Online)
1) There are several people on this site that I trust to help me solve my problems
Strongly disagree 1 2 3 4 5 6 7 Strongly Agree
2) If I needed an emergency loan of $100, I know someone on the site I can turn to
Strongly disagree 1 2 3 4 5 6 7 Strongly Agree
3) There is someone on the site I can turn to for advice for making very important
decisions
Strongly disagree 1 2 3 4 5 6 7 Strongly Agree
4) The people I interact with on the site provide me with emotional support when I
feel down
Strongly disagree 1 2 3 4 5 6 7 Strongly Agree
5) I do not know people on the site well enough to get them to do anything important
(REVERSED)
Strongly disagree 1 2 3 4 5 6 7 Strongly Agree
Section 8: Participation Level (Online)
1) I regularly contribute information I find about smoking on this site.
Strongly disagree 1 2 3 4 5 6 7 Strongly Agree
2) I regularly receive information from other posters about smoking on this site.
Strongly disagree 1 2 3 4 5 6 7 Strongly Agree
3) I regularly post questions about smoking on this site.
Strongly disagree 1 2 3 4 5 6 7 Strongly Agree
4) I regularly answer questions about smoking on this site.
Strongly disagree 1 2 3 4 5 6 7 Strongly Agree
5) I regularly give advice to other members of this site.
Strongly disagree 1 2 3 4 5 6 7 Strongly Agree
6) I regularly receive advice from other members on this site.
Strongly disagree 1 2 3 4 5 6 7 Strongly Agree
7) I regularly give encouragement to other members of this site.
Strongly disagree 1 2 3 4 5 6 7 Strongly Agree
8) I regularly receive encouragement from other members on this site.
Strongly disagree 1 2 3 4 5 6 7 Strongly Agree
9) I regularly correspond with other members of this site.
Strongly disagree 1 2 3 4 5 6 7 Strongly Agree
182
10) I regularly update my profile and posts on this site.
Strongly disagree 1 2 3 4 5 6 7 Strongly Agree
11) I regularly start new threads or topics on this site.
Strongly disagree 1 2 3 4 5 6 7 Strongly Agree
12) I regularly post links, images, video clips, etc. that I find useful to other members on
this site.
Strongly disagree 1 2 3 4 5 6 7 Strongly Agree
13) This social networking site is part of my daily activity.
Strongly disagree 1 2 3 4 5 6 7 Strongly Agree
14) I am proud to tell other people that I am using this social networking site.
Strongly disagree 1 2 3 4 5 6 7 Strongly Agree
15) This social networking site has become part of my daily routine.
Strongly disagree 1 2 3 4 5 6 7 Strongly Agree
16) I feel out of touch if I have not logged onto this social networking site for a while.
Strongly disagree 1 2 3 4 5 6 7 Strongly Agree
17) I feel I am part of the community on this social networking site.
Strongly disagree 1 2 3 4 5 6 7 Strongly Agree
18) I would be sorry if this social networking site shut down.
Strongly disagree 1 2 3 4 5 6 7 Strongly Agree
19) I visit this site only to read the posts, as I do not participate.
Strongly disagree 1 2 3 4 5 6 7 Strongly Agree
20) I get encouragement from the posts that I read, even though I do not participate.
Strongly disagree 1 2 3 4 5 6 7 Strongly Agree
21) I get information from the posts I read, even though I do not participate.
Strongly disagree 1 2 3 4 5 6 7 Strongly Agree
22) I get advice from the posts I read, even though I do not participate.
Strongly disagree 1 2 3 4 5 6 7 Strongly Agree
23) On average, I spend approximately ____ hours on this site per week.
24) On average, I spend approximately ____ hour surfing the Internet each week.
25) On average, I post ____ times on this site per week.
26) My profile is connected to ____ other members on this site.
27) I have ____ friends from offline who are also members of this site.
183
28) I have developed offline friendships with ____ people whom I initially met on this
site.
Section 9: Smoking Cessation Self-Efficacy (SEQ-12)
The following are some situations in which certain people might be tempted to smoke.
Please indicate whether you are sure you could refrain from smoking in each situation.
1) When I feel nervous
Not at all sure 1 2 3 4 5 6 7 Absolutely sure
2) When I feel depressed
Not at all sure 1 2 3 4 5 6 7 Absolutely sure
3) When I feel angry
Not at all sure 1 2 3 4 5 6 7 Absolutely sure
4) When I feel very anxious
Not at all sure 1 2 3 4 5 6 7 Absolutely sure
5) When I think about a difficult problem
Not at all sure 1 2 3 4 5 6 7 Absolutely sure
6) When I feel the urge to smoke
Not at all sure 1 2 3 4 5 6 7 Absolutely sure
7) When having a drink with friends
Not at all sure 1 2 3 4 5 6 7 Absolutely sure
8) When celebrating something
Not at all sure 1 2 3 4 5 6 7 Absolutely sure
9) When drinking beer, wine or other spirits
Not at all sure 1 2 3 4 5 6 7 Absolutely sure
10) When I am with smokers
Not at all sure 1 2 3 4 5 6 7 Absolutely sure
11) After a meal
Not at all sure 1 2 3 4 5 6 7 Absolutely sure
12) When having coffee or tea
Not at all sure 1 2 3 4 5 6 7 Absolutely sure
184
Section 10: Perceived Social Support
1) I can find other people to share experiences about quitting smoking on this site.
Strongly disagree 1 2 3 4 5 6 7 Strongly Agree
2) I can find other people to give me encouragement to quit smoking on this site.
Strongly disagree 1 2 3 4 5 6 7 Strongly Agree
3) I can find other people who value my advice and opinions about quitting smoking on
this site.
Strongly disagree 1 2 3 4 5 6 7 Strongly Agree
4) I can find other people who care about me on this site.
Strongly disagree 1 2 3 4 5 6 7 Strongly Agree
5) I can find other people to listen to me when I need to talk about quitting smoking on
this site.
Strongly disagree 1 2 3 4 5 6 7 Strongly Agree
6) I can find other people to give me advice about quitting smoking on this site.
Strongly disagree 1 2 3 4 5 6 7 Strongly Agree
7) I can find other people who understand my problems on this site.
Strongly disagree 1 2 3 4 5 6 7 Strongly Agree
8) I can find other people to give me helpful information about quitting smoking on this
site.
Strongly disagree 1 2 3 4 5 6 7 Strongly Agree
9) I can find other people to give me emotional support to quit smoking on this site.
Strongly disagree 1 2 3 4 5 6 7 Strongly Agree
10) I can find other people to give me companionship for quitting smoking on this site.
Strongly disagree 1 2 3 4 5 6 7 Strongly Agree
11) I feel like no one on this site cares about me (REVERSED)
Strongly disagree 1 2 3 4 5 6 7 Strongly Agree
12) I feel pretty much all alone on this site (REVERSED)
Strongly disagree 1 2 3 4 5 6 7 Strongly Agree
13) There is no one on this site I can turn to for advice (REVERSED)
Strongly disagree 1 2 3 4 5 6 7 Strongly Agree
185
Section 11: Social Norms (Online)
1) I would respond in the same way as other members of the site if I found out that
someone on the site smokes.
Strongly Disagree 1 2 3 4 5 6 7 Strongly Agree
2) I would respond in the same way as other members of the site if someone were to
smoke cigarettes around us.
Strongly Disagree 1 2 3 4 5 6 7 Strongly Agree
3) I share the same view as other members of the site regarding about my own
smoking.
Strongly Disagree 1 2 3 4 5 6 7 Very
4) I share the same view as other members on the site regarding the social
acceptability of smoking cigarettes.
Strongly Disagree 1 2 3 4 5 6 7 Strongly Agree
Section 12: Intimacy Level (Online)
Please answer the following questions in reference to the most intimate relationship with
another person you have experienced on this site.
Verbal Intimacy
1) Our communication is limited to just a few specific topics (REVERSED).
Strongly disagree 1 2 3 4 5 6 7 Strongly Agree
2) Our communication ranges over a wide variety of topics.
Strongly disagree 1 2 3 4 5 6 7 Strongly Agree
3) Once we get started, we move easily from one topic to another.
Strongly disagree 1 2 3 4 5 6 7 Strongly Agree
4) We contact each other in a variety of ways other than on this site.
Strongly disagree 1 2 3 4 5 6 7 Strongly Agree
5) I usually tell this person exactly how I feel.
Strongly disagree 1 2 3 4 5 6 7 Strongly Agree
6) I try to keep my personal judgments to myself when this person says or does
something with which I disagree (REVERSED)
Strongly disagree 1 2 3 4 5 6 7 Strongly Agree
7) I have told this person what I like about him or her.
Strongly disagree 1 2 3 4 5 6 7 Strongly Agree
186
8) I feel I can confide in this person about almost anything.
Strongly disagree 1 2 3 4 5 6 7 Strongly Agree
9) I would never tell this person anything intimate or personal about myself
(REVERSED)
Strongly disagree 1 2 3 4 5 6 7 Strongly Agree
10) I have told this person things about myself that he or she could not get from any
other source.
Strongly disagree 1 2 3 4 5 6 7 Strongly Agree
11) Our communications stays on the surface of most topics (REVERSED)
Strongly disagree 1 2 3 4 5 6 7 Strongly Agree
Affective Intimacy
1) The two of us depend on each other.
Strongly disagree 1 2 3 4 5 6 7 Strongly Agree
2) There have been times when each of us waited to see what the other thought
before making a decision of some kind.
Strongly disagree 1 2 3 4 5 6 7 Strongly Agree
3) Neither of us sets aside time to communicate with the other (REVERSED)
Strongly disagree 1 2 3 4 5 6 7 Strongly Agree
4) We have a great deal of effect on each other.
Strongly disagree 1 2 3 4 5 6 7 Strongly Agree
5) We often influence each other’s feelings towards the issues we are dealing with.
Strongly disagree 1 2 3 4 5 6 7 Strongly Agree
6) The two of us have little influence on each other’s thoughts (REVERSED)
Strongly disagree 1 2 3 4 5 6 7 Strongly Agree
7) I am very uncertain about what this person is really like (REVERSED)
Strongly disagree 1 2 3 4 5 6 7 Strongly Agree
8) I can accurately predict how this person will respond to me in most situations.
Strongly disagree 1 2 3 4 5 6 7 Strongly Agree
9) I can usually tell what this person is feeling inside.
Strongly disagree 1 2 3 4 5 6 7 Strongly Agree
187
10) I can accurately predict what this person’s attitudes are.
Strongly disagree 1 2 3 4 5 6 7 Strongly Agree
11) I do not know this person very well (REVERSED)
Strongly disagree 1 2 3 4 5 6 7 Strongly Agree
12) I am very committed to maintaining this relationship.
Strongly disagree 1 2 3 4 5 6 7 Strongly Agree
13) This relationship is not very important for me (REVERSED)
Strongly disagree 1 2 3 4 5 6 7 Strongly Agree
14) This relationship is a big part of who I am.
Strongly disagree 1 2 3 4 5 6 7 Strongly Agree
15) I would make a great effort to maintain our relationship.
Strongly disagree 1 2 3 4 5 6 7 Strongly Agree
16) I do not expect this relationship to last long (REVERSED)
Strongly disagree 1 2 3 4 5 6 7 Strongly Agree
17) I feel close to this person most of the time.
Strongly disagree 1 2 3 4 5 6 7 Strongly Agree
18) I feel like being encouraging and supportive to this person when he/she is
unhappy.
Strongly disagree 1 2 3 4 5 6 7 Strongly Agree
19) It is important for me to listen to this person’s very personal disclosures.
Strongly disagree 1 2 3 4 5 6 7 Strongly Agree
20) It is important to me that this person understands my feelings.
Strongly disagree 1 2 3 4 5 6 7 Strongly Agree
21) It is important to me that this person be encouraging and supportive to me when I
am unhappy.
Strongly disagree 1 2 3 4 5 6 7 Strongly Agree
22) I am satisfied with our relationship.
Strongly disagree 1 2 3 4 5 6 7 Strongly Agree
188
Cognitive Intimacy
1) There is not much difference between the way I communicate with this person
and the way I generally communicate on the Internet (REVERSED)
Strongly disagree 1 2 3 4 5 6 7 Strongly Agree
2) We have developed the ability to “read between the lines” of each other’s
messages to figure out what is really on each other’s mind.
Strongly disagree 1 2 3 4 5 6 7 Strongly Agree
3) The two of us use private signals that communicate in ways outsiders would not
understand.
Strongly disagree 1 2 3 4 5 6 7 Strongly Agree
4) We have special nicknames that we just use with each other.
Strongly disagree 1 2 3 4 5 6 7 Strongly Agree
5) I can get an idea across to this person with a much simpler message than I would
have to use with other people.
Strongly disagree 1 2 3 4 5 6 7 Strongly Agree
6) We share a special language or jargon that sets our relationship apart.
Strongly disagree 1 2 3 4 5 6 7 Strongly Agree
Physical Intimacy
1) This person and I do not know any of the same people. (REVERSED)
Strongly disagree 1 2 3 4 5 6 7 Strongly Agree
2) We have introduced each other (face-to-face or otherwise) each other to members
of each other’s circle of friends and family.
Strongly disagree 1 2 3 4 5 6 7 Strongly Agree
3) We contact a lot of the same people on the Internet.
Strongly disagree 1 2 3 4 5 6 7 Strongly Agree
4) This person and I are involved with many of the same sites and groups.
Strongly disagree 1 2 3 4 5 6 7 Strongly Agree
5) We have overlapping social circles on the Internet.
Strongly disagree 1 2 3 4 5 6 7 Strongly Agree
6) We have overlapping social circles outside of the Internet.
Strongly disagree 1 2 3 4 5 6 7 Strongly Agree
Abstract (if available)
Linked assets
University of Southern California Dissertations and Theses
Conceptually similar
PDF
Benefits of social networking in online social support groups
PDF
The formation and influence of online health social networks on social support, self-tracking behavior and weight loss outcomes
PDF
A network analysis of online and offline social influence processes in relation to adolescent smoking and alcohol use
PDF
Contagious: social norms about health in work group networks
PDF
Leveraging social normative influence to design online platforms for healthy eating
PDF
The effect of Twitter social support on health outcomes and its mediators: a randomized controlled trial of a social support intervention on Twitter for patients affected by cancer
PDF
Exploring the association of number of cigarettes smoked and confidence to quit smoking in Korean American emerging adults: a multilevel modeling approach
PDF
Motivation for participating in virtual religious communities: developing social capital
PDF
Understanding normative influence of neighborhoods: a multilevel approach to promoting Latinas’ cervical cancer prevention behaviors in urban ethnic communities
PDF
The role of social control in promoting healthy eating behavior among Chinese immigrants: an ecological approach
PDF
The effect of an educational fotonovela on the prevention of secondhand smoke
PDF
Hispanic advertising as social orientation: an examination of the advertising industry as a field of cultural production
PDF
Examining the relationship between Latinx community college STEM students’ self-efficacy, social capital, academic engagement and their academic success
PDF
The Women and Girls Center for Maternal Health intergenerational approaches to address the Black maternal health crisis
PDF
A view from below: the development and role of organizational social capital in neighborhood regeneration in Los Angeles
PDF
TikToking and Instagramming: following high school teacher influencers' roles in supporting and informing teacher practices
PDF
Emotion culture through the lens of PUA (pickup artist): emotions and the social media narratives of structual gendered violence
PDF
Personal and environmental influences on preparing for transition from sport: a social cognitive approach to exploring the collegiate experiences of Black former NFL players
PDF
An examination of a social justice teacher cohort and its capacity to support transformational professional learning
PDF
Personalized normative feedback applied to undergraduates with problem drinking: a comparison with psychoeducation and an examination of cognitive-affective change mechanisms via the articulated ...
Asset Metadata
Creator
Phua, Joe Jin
(author)
Core Title
The social groups approach to quitting smoking: An examination of smoking cessation online and offline through the influence of social norms, social identification, social capital and social support
School
Annenberg School for Communication
Degree
Doctor of Philosophy
Degree Program
Communication
Publication Date
04/12/2011
Defense Date
11/29/2010
Publisher
University of Southern California
(original),
University of Southern California. Libraries
(digital)
Tag
descriptive norms,injunctive norms,lurking,OAI-PMH Harvest,reference groups,smoking cessation,social capital,social identity theory,social networking sites,social support
Language
English
Contributor
Electronically uploaded by the author
(provenance)
Advisor
Cody, Michael J. (
committee chair
), Jordan-Marsh, Maryalice (
committee member
), McLaughlin, Margaret (
committee member
)
Creator Email
jophua@gmail.com,phua@usc.edu
Permanent Link (DOI)
https://doi.org/10.25549/usctheses-m3736
Unique identifier
UC1189325
Identifier
etd-Phua-4241 (filename),usctheses-m40 (legacy collection record id),usctheses-c127-447489 (legacy record id),usctheses-m3736 (legacy record id)
Legacy Identifier
etd-Phua-4241.pdf
Dmrecord
447489
Document Type
Dissertation
Rights
Phua, Joe Jin
Type
texts
Source
University of Southern California
(contributing entity),
University of Southern California Dissertations and Theses
(collection)
Repository Name
Libraries, University of Southern California
Repository Location
Los Angeles, California
Repository Email
cisadmin@lib.usc.edu
Tags
descriptive norms
injunctive norms
lurking
reference groups
smoking cessation
social capital
social identity theory
social networking sites
social support