Au . .
3.2.1.?
‘2
f
s

l
n

{5.1.x
ta ,
53%

.AT.¢ 1..

a. a:

o:
‘ .Me\id»s 81-50:, .
23.3: n.9,!
humour .u a...

i3
xv. 3.; .. 2.7.
.11 .4.

: 53 ..
3393.31.41
9.x. ‘1 -. ,.

. x
it";
‘1‘ ll):
'0 (w.
)i la

101‘
.,.n..v.x&t..:x
1.4:: i .. I.

 

 

r 12:2:

 

f‘

A
200?

This is to certify that the
dissertation entitled

ONLINE SOCIAL SUPPORT AND CANCER: UNDERSTANDING
THE IMPACT OF SURVIVAL RATE, GENDER, AND DIFFERENT
COMMUNITY CHARACTERISTICS

presented by

Lorraine R. Buis

has been accepted towards fulfillment
of the requirements for the

PhD. degree in Mass Media

 

 

... -....
. _ ..

I

.\

I.

I a two-[l i" I I t L L LW’V

Major Profesfsor’s Signature

{cgab/ 5%. c9 <73 ll

Date

MSU Is an affirmative-action, equal-opportunity employer

 

---—— - —-.--.--—-—-—._.--.—.-.-—.—.—a-.-.-.-.-.-.-v_.-.-.-.-._._._.-.-.-A_

-.... —.- —.- —.-

 

LIBRARY
Michigan State
University

 

 

 

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

 

DATE DUE DATE DUE DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

6/07 p:/ClRC/Date0ue.indd-p.1

ONLINE SOCIAL SUPPORT AND CANCER: UNDERSTANDING THE IMPACT
OF SURVIVAL RATE, GENDER, AND DIFFERENT COMMUNITY
CHARACTERISTICS

By

Lorraine R. Buis

A DISSERTATION

Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of

DOCTOR OF PHILOSOPHY

Mass Media

2007

ABSTRACT
ONLINE SOCIAL SUPPORT AND CANCER: UNDERSTANDING THE IMPACT
OF SURVIVAL RATE, GENDER, AND DIFFERENT COMMUNITY
CHARACTERISTICS
By
Lorraine R. Buis

Introduction: Social support is an important part of dealing with cancer.
Traditional support groups where people connect with others experiencing similar
situations have been commonly utilized for several decades. With the advent of
the World Wide Web, online support communities have become a very important
tool for finding social support when experiencing illness. Due to the benefits that
online support communities afford, it is unsurprising that millions of people turn to
online support groups to find emotional and informational support for a variety of
different diseases and conditions.

Objective: This investigation sought to compare social support messages
contained within eight online support communities for cancers with varying five-
year relative survival rates, as well as understand how the five-year relative
survival rate of specific cancers and other community characteristics such as
community host, openness, post volume, and size affected the demographic
makeup of community members in terms of gender and community member role,
as well as social support messages. In addition, this investigation sought to test
Optimal Matching Theory, a theory of social support, in this electronic realm.

Methods: A quantitative content analysis was conducted on 3,717

messages located in eight different health-related online support communities

focused on four different types of cancer with high and low five-year relative
survival rates. Of the eight communities located in this sample, four were hosted
by YahoolGroups (a general-purpose web portal) and four were hosted by ACOR
(a cancer-specific community host). Intercoder reliability was obtained with all
variables receiving a Cohen’s Kappa agreement of at least .85.

Results: Data from this investigation indicate interesting findings as
significant relationships were found between types of social support messages
and community member role, five-year survival rate, as well as different
characteristics of the communities such as host, openness, and post volume.
Although support was found for the first hypothesis indicating women will post
more frequently than men and will provide a greater proportion of emotional
support provisions, no support was found for hypotheses predicting that
emotional support provisions would be more prevalent in communities for low
survival rate cancers, while informational support provisions would be more
prevalent in communities for high survival rate cancers.

Discussion: Results from this investigation point to a number of
interesting conclusions that have significant prescriptive implications for the
design of online support communities and web-based services. In addition, this
investigation raises new and important questions to guide further research in this

field.

To my father Ronald Stewart, my grandparents Stanley and Ruby Stewart, and
my cousin Christopher Parka, who passed away before having the chance to see

me achieve this remarkable accomplishment.

Acknowledgments

I would like to take the opportunity to acknowledge many of the people
who have contributed considerably to the completion of this dissertation, as well
as played a significant role in helping me complete my doctoral degree.

First of all, I would like to acknowledge my advisor Dr. Pamela Whitten,
who has acted as my mentor during my doctoral career. Over the last few years,
she has served as an excellent role model by demonstrating how to have a
productive career as a researcher and an educator, in addition to having a life
and family outside of academia.

I also would like to acknowledge the members of my guidance and
dissertation committees, Dr. Charles Steinfield, Dr. Nicole Ellison, and Dr. Kami
Silk, who along with Dr. Fred Fico have all provided me with invaluable insight
into the academic world and encouraged me in my academic pursuits. During my
doctoral program, I was also privileged to have the opportunity to work Serena
Carpenter in assorted classes and outside projects and would like to thank her
for helping to make this experience more enjoyable.

Outside of MSU, I would like to thank Reema Kadri who assisted with the
data collection in this investigation. Not only was Reema essential in the
completion of this project, but with her husband Josh Hathaway, they have been
two of the most treasured friends a person could ever have. They have been just
as important and influential to me as any of my family members and I look

fonNard to many more years of travel, games, friendship, and laughter.

The completion of my formal education would not have been possible
without the love, support, and encouragement of my parents Susan and Tom
Hart and my grandparents Gladys and David Parka. They raised me to believe
that with a little hard work, I could achieve anything. In addition, I would also like
to thank John and Judie Buis, David L. Parka, Robin Hepfinger, and Katie and
Missy Hepfinger. With my parents and grandparents, they have taught me how
important family is and through the years, they have never stopped supporting
me. I truly appreciate and love them all very much.

Finally, I would like to thank Jeffrey Buis for always being there for me.
Since I was fifteen, he has stood by my side, doing everything in his power to
make me happy. He has encouraged me when I when I felt unsure, made me
laugh when l was down, helped me to relax when l was as tightly wound as one
person could be, and loved me without question. Furthermore, through teaching
me to love knowledge, he has challenged me to become a better researcher and
scholar. Without his love and support, I probably never would have aspired to
complete this level of education. None of this would have been possible without

him and I am proud to call him my husband.

vi

TABLE OF CONTENTS

LIST OF TABLES ................................................................................................ IX
LIST OF FIGURES ............................................................................................. XI
CHAPTER 1: INTRODUCTION ............................................................................ 1
CHAPTER 2: LITERATURE REVIEW AND THEORETICAL FRAMEWORK ....... 7
THE INTERNET .............................................................................................. 7
Health Information and the Internet ................................................. 8

Cancer Information and the lntemet ................................................ 8

ONLINE COMMUNITIES ................................................................................. 11
Online Communities Defined ......................................................... 12
Health-Related Communities ......................................................... 15

Ethical Concerns with Studying Online Communities .................... 18

SOCIAL SUPPORT ....................................................................................... 20
Definition of Social Support ........................................................... 20

History of Social Support Research ............................................... 24

Measuring Social Support ............................................................. 26

Social Support Theories ................................................................ 29

Influences on Social Support ......................................................... 34

Impact of Social Support on Health ............................................... 38

Impact of Social Support on All-cause Mortality ............................ 38

Impact of Social Support on Condition Specific Health Outcomes 41

Sources of Social Support ............................................................. 43

ONLINE SOCIAL SUPPORT ........................................................................... 44
Characteristics of Online Support Groups ..................................... 47

Reasons for Utilizing Online Health Support Communities ............ 48

Importance of Studying Online Social Support .............................. 50

SUMMARY .................................................................................................. 53
Research Questions and Hypotheses ........................................... 57

CHAPTER 3: METHODOLOGY .......................................................................... 61
SAMPLE ..................................................................................................... 61

DATA COLLECTION ...................................................................................... 64
Independent Variables ................................................................... 65

Dependent Variable ....................................................................... 67

Reliability ....................................................................................... 68

DATA ANALYSES ......................................................................................... 69

vii

CHAPTER 4: RESULTS ..................................................................................... 71

POSTING ACTIVITY ...................................................................................... 71
RQ1: COMMUNITY MEMBER ROLE ............................................................... 76
RQZ: FIVE-YEAR RELATIVE SURVIVAL RATE ................................................. 79
R03: ADDITIONAL COMMUNITY CHARACTERISTICS ........................................ 83
Host ............................................................................................... 83
Openness ...................................................................................... 86

Post Volume .................................................................................. 89
Community Size ............................................................................ 92

H1: GENDER IN ONLINE SUPPORT COMMUNITIES .......................................... 95
H2 AND H3: SOCIAL SUPPORT PROVISION BY FIVE-YEAR SURVIVAL RATE ...... 98
SUMMARY ................................................................................................ 100
CHAPTER 5: DISCUSSION ............................................................................. 102
SOCIAL SUPPORT PROVISION WITHIN ONLINE SUPPORT COMMUNITIES ......... 103
GENDER AND ONLINE SUPPORT COMMUNITIES ........................................... 105
F IVE-YEAR SURVIVAL RATE AND SOCIAL SUPPORT PROVISION .................... 108
COMMUNITY MEMBER ROLE AND SOCIAL SUPPORT PROVISION TYPE ........... 111
COMMUNITY MEMBER ROLE AND FIVE-YEAR RELATIVE SURVIVAL RATE ....... 113
IMPACT OF OPENNESS ON ONLINE SUPPORT COMMUNITIES ......................... 114
COMMUNITY HOST AND SOCIAL SUPPORT PROVISION .................................. 116
POST VOLUME, GENDER PARTICIPATION, AND SOCIAL SUPPORT PROVISION .1 1 7
CONTRIBUTIONS FROM THIS STUDY ............................................................ 1 19
Limitations ................................................................................... 121
Future Research .......................................................................... 122

APPENDIX A: VAUX’S CATEGORIES OF SOCIAL SUPPORT MEASURES

AND THEIR RESPECTIVE INSTRUMENTS .................................................... 125
APPENDIX B: CODING PROTOCOL ............................................................... 126
WORKS CITED ................................................................................................. 137

Viii

LIST OF TABLES

Table 1: Community Descriptors ......................................................................... 73
Table 2: Frequency of Social Support Type by Community ................................ 74
Table 3: Emotional Support Types and their Frequency by Community ............. 75

Table 4: Informational Support Types and their Frequency by Community ........ 75

Table 5: Frequency of Community Member Roles across Communities ............ 77
Table 6: Frequency of Social Support by Role across Communities .................. 78
Table 7: Frequency of Social Support Type by Community Member Role .......... 79
Table 8: Frequency Of Gender Participation by Cancer Survival Rate ................ 80

Table 9: Frequency of Gendered Social Support by Cancer Survival Rate ........ 81

Table 10:
Table 11:
Table 12:
Table 13:
Table 14:
Table 15:
Table 16:
Table 17:
Table 18:
Table 19:
Table 20:

Table 21:

Frequency of Role Participation by Cancer Survival Rate ................... 82
Frequency of Social Support by Role by Cancer Survival Rate .......... 82
Frequency of Gender Participation by Community Host ..................... 84
Frequency of Gendered Social Support by Community Host .............. 84
Frequency of Role Participation by Community Host .......................... 85
Frequency of Social Support by Role by Community Host .................. 85
Frequency of Social Support Type by Community Host ...................... 86
Frequency of Gender Participation by Community Openness ............ 87
Frequency of Gendered Social Support by Community Openness ..... 87
Frequency of Role Participation by Community Openness ................. 88
Frequency of Social SuppOIt by Role by Community Openness ......... 88
Frequency of Social Support Type by Community Openness ............. 89

Table 22:
Table 23:
Table 24:
Table 25:
Table 26:
Table 27:
Table 28:
Table 29:
Table 30:
Table 31:
Table 32:
Table 33:
Table 34:

Table 35:

Frequency of Gender Participation by Community Post Volume ........ 90
Frequency of Gendered Social Support by Post Volume .................... 90
Frequency of Role Participation by Community Post Volume ............. 91

Frequency of Social Support by Role by Community Post Volume ..... 91

Frequency of Social Support Type by Community Post Volume ......... 92

Frequency of Gender Participation by Community Size ...................... 93
Frequency of Gendered Social Support by Community Size .............. 93
Frequency of Role Participation by Community Size .......................... 94
Frequency of Social Support by Role by Community Size .................. 94
Frequency of Social Support Type by Community Size ...................... 95
Frequency of Gender across Communities ......................................... 96
Frequency of Gendered Social Support by Community ...................... 97
Frequency of Social Support Type by Gender .................................... 98
Frequency of Social Support Type by Survival Rate ......................... 100

LIST OF FIGURES

Figure 1: Graphical Representation of Hypotheses Two and Three ................ 60

xi

Chapter 1: Introduction

In the year 2007, it is expected that approximately 1.4 million people will
be newly diagnosed with cancer (not including basal and squamous cell skin
cancers, as well as most non-invasive cancers) and that almost 560,000
Americans will die from cancer. In the US, cancer-related deaths account for
one out of every four deaths. In fact, cancer is the second leading cause of
death, only surpassed by heart disease. Although the average five-year relative
survival rate for all cancers is 66%, there are some forms of cancer that have
substantially higher or lower five-year relative survival rates (American Cancer
Society, 2007).

Previous research has indicated that cancer can be caused by both
internal and external factors. Although the risk of developing cancer increases
with age, cancer can occur in anyone regardless of age, race, or gender. Overall,
it is estimated that men have a slightly less than 50% lifetime risk for developing
cancer, while women have a slightly higher than 33% lifetime risk. Due to its
high prevalence and the fact that cancer knows no boundaries, most people will
experience cancer during their lifetime, either first hand as a patient or through a
family member, friend, or acquaintance (American Cancer Society, 2007).

Regardless of the average five-year relative survival rate of a specific form
of cancer, any cancer diagnosis can be a scary thing. These diagnoses can be
quite stressful or even traumatic. In order to help deal with the pressures
associated with illness, people often turn to others for social support. Although

definitions vary widely across the field, social support generally refers to “The

resources provided by other persons” (Cohen & Syme, 1985, p. 4). These
resources can be emotional, informational, or even instrumental. With the
widespread adoption of the lntemet, it is not surprising that many people turn to
online avenues for social support. Between email, chat rooms, instant
messaging, and listserv technology, as well as the abundant proliferation of the
use of social networking sites such as Myspace and Facebook, computer users
are becoming more savvy and comfortable connecting with others online.
Furthermore, Internet penetration in the US has been increasing over the years
and approximately 73% Of adults are now online, joining the almost 90% of
children ages 12 — 17. In addition, US broadband access is now available in the
homes of approximately 42% of Americans (Madden, 2006). As the number of
lntemet users increases, along with the availability of broadband access within
the home, it is reasonable to assume that more people will be turning to online
sources of social support.

There are countless reasons why people may seek social support in
online settings. Online support communities may be used to supplement
traditional offline methods of support, or it is possible that due to any number of
causes, online venues may be the only access of social support available to
certain individuals. In addition to the many people that actively participate in
online support communities, even more individuals lurk within these online
spaces and passively participate by reading archived messages. While those
lurking may not be directly participating in a social support exchange, support

may still be received in the form of information, or the understanding that others

are experiencing similar stressful situations. Regardless of the reason for turning
to the Internet for social support, because of the widespread availability of these
online communities, and due to number of people potentially either actively or
passively participating, it is imperative that we begin to understand more about
the types Of social support interactions occurring within this electronic
environment.

Despite the fact that there are numerous case studies that have
investigated various types of online health-related support communities, overall,
the literature in this field of study remains relatively weak due to a lack Of
generalizability of research findings. Much of the work in this field that has
attempted to understand what types of interactions are occurring within online
support communities has been focused primarily on case studies of individual
communities for various diseases and conditions (Dunham et al., 1998; Klemm,
Reppert, & Visich, 1998; Lasker, Sogolow, & Sharim, 2005; Mendelson, 2003;
Perron, 2002; White & Dorman, 2000). In addition, a recent investigation of
online breast cancer communities indicated that prior research in this area
centered primarily on individual communities and this was cited as one of the
problems associated with previous research in the field of online community
research (Barnett & Hwang, 2006). As little empirical work has been done in the
past to make comparisons across different health-related online social support
groups, it is imperative that these cross-community comparisons be conducted.
Not only are these comparisons essential in determining the different types of

support interactions that occur online, but until we can establish cross-community

comparisons, the generalizability of the online community research done in this
field is minimal, as we are unable to make generalizations about a larger
population from individual case studies.

This study seeks to explore the different forms of emotional and
informational support that people provide to one another during situations
involving cancers with varying five-year relative survival rates. Through
quantitative content analysis, this investigation identifies the frequency of
emotional and informational support provisions that appear within eight online
support communities focused on different forms of cancer. This investigation also
identifies several subtypes of emotional and informational support that are found
within online support communities. As we are unable to determine individual’s
perceptions of social support that is received, this investigation focuses solely on
the provision of social support. Focusing on social support provisions in online
communities is important because as previously mentioned, there are many
individuals who read through archived postings, but do not actively participate in
posting themselves. It is possible that these lurkers obtain social support through
passive involvement in the community through the reading of message postings
that were originally intended for other individuals. In addition, this investigation
also seeks to understand how factors such as the gender and role of community
members, the survival rate of cancer, and other community characteristics

impacts the types and amounts of social support that is provided within an online

support group.

In order to guide this investigation, this study utilizes Optimal Matching
Theory, a matching theory of social support, as a theoretical framework for
understanding social support interactions in the online realm. Despite the fact
that research in this field has yielded few strong theoretical approaches for social
support research, Optimal Matching Theory provides a suitable framework for
understanding the different types of online support interactions because it is not
dependent on an offline context. While the problems within the field of social
support are by no means close to being solved, it is essential that we begin to
understand social support in pragmatic ways, in this case, within an online
environment.

This research contributes to the existing body of knowledge in that it is
one of the first cross community comparisons of online health-related social
support communities, not only between different types of illnesses, but between
different community hosts. In addition, this serves as one of the first
investigations to look at the impact that community member role has on
participation and social support provision in online communities. Finally, this
investigation adds to the existing Optimal Matching Theory by demonstrating its
use in a new, digital realm, as well finding new factors that potentially influence
the discovery of “optimal matches” and more concretely determines the types of
support that are best matched to different circumstances.

Although looking strictly at online social support interactions found in
archived messages may only capture a slice of the total social support

interactions people engage in, due to the fact that online sources of social

support may potentially be the only source of support for some individuals, it is
important to understand these interactions in isolation. Through looking at online
instances of social support, we will be able to better understand the different
support needs of people, particularly in this digital medium. From this, it is
possible that we may be able to design better support services, both online and
offline, in the long run.

In the next section, chapter two, a discussion of the relevant literature
associated with this project is presented, followed by a series of research
questions and hypotheses that were developed to guide this investigation.
Chapter three provides an account of the methodology utilized in this study while
chapter four presents the results of this investigation. Finally, chapter five
concludes this work with a discussion of the results and their implications, as well
as an identification of the limitations of this study and a description of future work

needed in this field.

Chapter 2: Literature Review and Theoretical Framework

To begin this literature review, literature focused on the use of online
sources for general health and cancer information is presented, followed by a
discussion of online communities and the ethical considerations related to their
study. Next, this chapter turns to social support; definitions, theories, impacts on
health, and sources of social support. Finally, literature related to online social

support is presented.

The Internet

Over the course of the last 15 years, the lntemet and the World Wide Web
have become a pervasive part of our society’s infoscape. According to the Pew
lntemet & American Life project, Internet penetration has now reached
approximately 73% in the US and broadband access is available within an
estimated 42% of homes (Madden, 2006). In addition, statistics indicate that
more and more individuals from traditionally underrepresented demographics are
also utilizing the Internet as 53% of adults with yearly incomes less than $30,000
per year and 40% of adults with less than a high school education report using
the Internet (Madden, 2006). Due to the fact that an ovenrvhelming majority of our
population is online, including individuals from demographic groups thought not
to have pervasive access to the Internet, it is essential that we more thoroughly

investigate the types of online pursuits that people are engaging in.

Health InfomIa_tjon and the Internet

In the last decade, there have been numerous studies that have
demonstrated that the lntemet is a major source of health information. A recent
report from the Pew Internet & American Life project has indicated that 80% of
adult American lntemet users (113 million people) have used the lntemet to
search for at least one of 17 identified health topics (Fox, 2006). The lntemet is a
valuable tool for health-related information due to greater access to information,
increased privacy, and many other reasons (Ziebland et al., 2004). Estimates
state that as many as 10,000 health-related websites are being searched (Fox &
Rainie, 2000) and every day across the globe, over 12.5 million health-related
searches are conducted (Eysenbach, 2003). This makes up approximately 4.5%

of all conducted Internet searches (Eysenbach & Kohler, 2004).

Qa_ncer lnforma_t_ion and the Internet

The lntemet has become a major mode of dissemination for cancer-
related information. Results from the 2002 — 2003 Health Information National
Trends Survey (HINTS) survey indicated that when looking for cancer-related
information, the lntemet was the first information source accessed by
approximately 50% of American adults who use the Internet (Hesse et al., 2005).
Regarding the use of the Internet for cancer-related information, particularly for
breast cancer patients, accessing online medical information on the lntemet has
been linked to positive health outcomes (Fogel, Albert, Schnabel, Ditkoff, 8.
Neugut, 2002). In addition, there are so many sources of cancer-related

information on the Internet that there are several academic articles devoted to

highlighting some of the more popular and prominent resources (Junghans,
Sevin, lonin, & Seifried, 2004; Santoro, 2003).

In a recent investigation aimed at understanding cancer patients’ attitudes
about the lntemet, it was discovered that participants were keen on obtaining
cancer information from online sources, but were skeptical of unfamiliar sources
that provided information. When obtaining information from reputable online
sources, participants reported trust in the source and felt that they could
adequately evaluate the information. When looking for health-related content,
Medscape and WebMD were cited by cancer patients most frequently as their
favorite websites to access. In addition, it was discovered that favorite websites
listed by participants tended to be affiliated with large-scale organizations and
government agencies such as the National Cancer Institute and the American
Cancer Society (LaCoursiere, Knobf, & McCorkIe, 2005). Furthermore, research
on the use of the lntemet by patients with melanoma revealed that of those
participants who obtained information online related to melanoma, 94% indicated
that the information was useful, 67% reported that the information helped them to
understand their condition, and 45% stated that they would recommend the
lntemet to others who were looking for medical information (Sabel et al., 2005).
Finally, it has been suggested that cancer patients utilize the Internet for cancer
information through all stages of illness, specifically before visiting the doctor,
during investigations, after diagnosis, during the treatment selection stage,

before treatment, and for short and long term follow up (Ziebland et al., 2004).

The fact that the Internet iS one of the premier destinations for medical
information doesn’t necessarily ensure the quality of the information. One of the
inherent problems with online medical information is the possibility that there is a
plethora of inaccurate information available (Matthews, Camacho, Mills, &
Dimsdale, 2003). The Internet is an unregulated repository of information and at
times, the quality of information available may be suspect. As such,
investigations aimed at determining the quality of online information for a variety
of health topics have occurred over the last decade (lmpicciatore, Pandolfini,
Casella, & Bonati, 1997; Kunst, Groot, Latthe, Latthe, & Kahn, 2002; Meric et al.,
2002; Sutherland, Wildemuth, Campbell, & Haines, 2005).

A recent investigation of scoliosis information on the lntemet by Mathur et
al. (2005) found that like other similar investigations of different medical
conditions, the overall quality of scoliosis information was poor and the authors
suggested that “the problem that both patients and physicians encounter is not a
lack of information but rather an overload of information. The valid and valuable
information perhaps obscured by irrelevant and misleading information” (pg.
2,698). The notion that there is poor quality information online and that the
overall quality of online information is highly variable is echoed by several other
investigations (Bremner, Quinn, Quinn, & Veledar, 2006; Butler & Foster, 2003;
Greene, Appel, Reinert, & Palumbo, 2005; Croft & Peterson, 2002; Hanif et al.,
2006; Hargrave, Hargrave, & Bouffet, 2006; llic, Risbridger, & Green, 2004; Liu &
Liu, 2006; Maloney, llic, & Green, 2005; Selman, Prakash, & Khan, 2006; Tiller,

Rea, Silla, & Wood, 2006). In fact, a large review of studies focused on

10

determining the accuracy of medical information on the lntemet found that out of
the 79 studies that met inclusion criteria, 70% concluded that the accuracy of
medical information on the Internet is compromised, 21.5% were neutral, and
only 9% found the accuracy of information to be acceptable, although the authors
reported that in these instances, none of the studies used evidence-based
guidelines as a standard criteria (Eysenbach, Powell, Kuss, & Sa, 2002).
Recently, there have been many investigations in the literature indicating a need
to establish criteria, control, and evaluation measures to help identify accurate
health information and reputable sources through evaluation tools, labeling, or
other forms of credentialing (Bemstam, Shelton, Walji, & Meric-Bemstam, 2005;
Commission of the European Communities Brussels, 2002; Griffiths &
Christensen, 2005; Griffiths, Tang, Hawking, & Christensen, 2005; Mayer et al.,
2006). Despite this trend toward making it easier for health information
consumers to identify reputable and accurate websites, the current infoscape is
still unregulated. Although the general consensus among experts is that the
quality of online health information is not always exceptional, it is essential that
we focus some of our energies on moving past this to understand how the
lntemet is being used by people for enterprises other than seeking out
information. One use of the lntemet that goes beyond searching static websites

for information is participating in online communities.

Online Communities

Aside from the informational benefits of the lntemet, people use the

I ntemet for many different social enterprises. One such way that people engage

11

 

in social behaviors is through participation in online communities. In this next
section, the definition of an online community is presented along with a
discussion of how they are used for health-related purposes. Finally, the ethical

concerns of researching online communities are illustrated.

nline Communities Defined

L

Over the course of the last decade, the notion of community has been
studied from a variety of different academic perspectives. As communities have
been studied across multiple social science disciplines, it comes as no surprise
that the definition of community varies widely (Hamman, 2001 ). In the mid-
1950’s, Hillery (1955) attempted to understand what common themes were
central to different definitions of community. Through the comparison of 94
sociological definitions, Hillery found only one common thread; they all involve
people. According to the Encyclopedia of Community: From the Village to the
Virtual World, “the lesson here is that absolute definitions are not necessary; it
may be the fluidity of a core concept that makes it so useful. Community may be
thought of as a geographic place, shared hobbies or interests, a warm sense of
togetherness, interaction in a common space such as a chat room, and so forth.”
(Christensen & Levinson, 2003, p. xxxvii).

With the widespread penetration of the World Wide Web over the last
decade, the notion of what is considered to be a community has been constantly
challenged. As more and more people turn to the online realm for social matters,

scholars are forced to acknowledge that communities do not necessarily have to

be geographically bound and it is now commonly recognized that groups of

12

people can interact with one another in virtual or online communities that reside
within the lntemet. The term “virtual community” was coined by Howard
Rheingold (1993) in his book The Virtual Community: Homesteading on the
Electronic Frontier. When describing virtual communities, Rheingold stated that:

“People in virtual communities use words on screens to

exchange pleasantries and argue, engage in

intellectual discourse, conduct commerce, exchange

knowledge, share emotional support, make plans,

brainstorm, gossip, feud, fall in love, find friends and

lose them, play games, flirt, create a little high art and

a lot of idle talk. People in virtual communities do just

about everything people do in real life, but we leave

our bodies behind. You can’t kiss anybody and nobody

can punch you in the nose, but a lot can happen within

those boundaries” (Rheingold, 1993, pp. 3).

To define virtual communities, Rheingold stated that they are “social
aggregations that emerge from the Net when enough people carry on those
public discussions long enough, with sufficient human feeling, to form webs of
personal relationships in cyberspace” (Rheingold, 1993, pg. 5). Despite the fact
that these words were written over a decade ago, coupled with the fact that there
are many different ways that online communities have been defined, the spirit of
Rheingold’s definition of virtual community still rings true today.

According to the literature, there are many different ways in which online
communities have been defined and studied. In her book Online Communities:
Designing Usability, Supporting Sociability, Jenny Preece (2000) outlines several
different perspectives that can be used to frame an understanding of what an

online community is. The sociological perspective relies on the many decades of

research based on traditional communities within the field of sociology. Scholars

13

from this perspective tend to approach online communities as they would any
traditional community. The technology perspective approaches online
communities from a very techno-centric standpoint and tries to understand them
according to the type of technology that is used (i.e. chat rooms, bulletin boards,
listserv, etc.). This perspective tends to rely more on technologies used than on
actual social interactions and organization. The virtual worlds’ perspective views
online communities as an immersive experience. This perspective tends to focus
on the use of technologies such as MUDs and MOOs. Finally, the eCommerce
perspective holds a very wide view of online communities. Rather than believing
that online communities are dependent on a specific type of technology, the
eCommerce perspective is more concerned with what keeps people within a site
(aka stickiness).

Practically speaking, online communities in our current infoscape may
appear as any combination of listservs, newsgroups, bulletin boards, message
boards, chat groups, instant message systems, or web based portals where
people connect with one another virtually. With that being said, the presence of
these technologies does not necessarily indicate that an online community exists.
When surveying different definitions of community, several common elements
are found among them including the notion of membership, relationships
between members, commitment and reciprocity, shared values, collective goods,
and duration (Nonnecke & Preece, 2003). Without the presence of some of these
characteristics, an online space is not necessarily an online community. For

example, a bulletin board that was created for the purpose of exchanging

14

information related to a specific topic may not necessarily be an online
community if there are no reciprocal exchanges between members of the
community. Furthermore, an online community that is filled with predominantly
spam and very little applicable content may not be considered a community
either.

Over the last few decades, it has been suggested that membership and
participation in traditional communities, as we know them, has been decreasing.
Wellman & Gulia (1999) argue in rebuttal that these traditional communities are
not disappearing, rather, participation in them is taking new forms. Many
researchers have suggested that participation in online communities lures people
away from participating in traditional, offline communities, although evidence has
yet to prove this. In contrast, many others have suggested that online and offline
communities are interconnected (Nip, 2004). In a study that attempted to
distinguish the differences between a traditional community and its online
counterpart, it was found that the online and offline version of the same
community had different norms and goals (Nip, 2004). Despite this difference,
participation in online communities can increase the sense of belonging to the
offline community, which can lead to increases in offline community participation

(Nip, 2004).

Health-Related Communities

Of all of the different types of virtual communities, it seems that health-
focused communities are among the most interesting to study at this point in

time. As of March, 2007, over 37,000 electronic support groups, over 10,000

15

electronic fitness and nutrition groups, and over 2,700 electronic drug and
medication groups had been listed by YahoolGroups under the “health and
wellness” category. Recent estimates state that approximately 100 million
lntemet users are members of online communities (Rainie & Horrigan, 2005).
Due to the prevalence of online health-related information and support groups,
investigation of this topic is a necessity.

In the past, online communities have been studied through a variety of
methods including (but not limited to) ethnography, online surveys, focus groups,
and content analysis. Research has been conducted on a number of different
online support communities that focus on many different health-related topics
including Alzheimer’s disease (White & Dorman, 2000), depression (Houston,
Cooper, & Ford, 2002; Powell, McCarthy, & Eysenbach, 2003), cancer (Till,
2003; Ziebland et al., 2004), dementia (Glueckauf, Ketterson, Loomis, & Dages,
2004), and stroke (Pierce, Steiner, & Govoni, 2002), among several others. While
most of the research conducted has been largely exploratory in nature, there
have been a limited number of studies that have attempted to tie online support
community use to health outcomes.

Despite the positive rhetoric that online support communities have
received, current research suggests that we may not yet have reason to be so
optimistic. A meta-analysis conducted by Eysenbach, Powell, Englesakis, Rizo, &
Stern (2004) investigated the effects of online peer-to-peer interactions within
online health communities and support groups. This meta-analysis found that

while there was no evidence of negative health outcomes, there was no evidence

16

of positive health outcomes either. The Iack‘of measurable health benefits could
be due to the fact that the majority of research conducted in this area has been
exploratory. What has attempted to be descriptive in nature tends to suffer from
poor experimental design and too few participants. In addition, in cases where
statistically significant effects on health outcomes had been assessed, no
information was provided as to whether these effects were clinically significant.

While exploratory case studies investigating online community use of
disease-specific communities are plentiful in the literature, it is impossible to
make any broad generalizations across online communities. Coupled with the
fact that most of the research that has tied online community use to health
outcomes may have been flawed in the research design or statistical analysis, it
is essential that more deep and rigorous investigations of online support
communities be conducted.

Unfortunately, some research has highlighted instances where online
information and participation in online communities can have negative
consequences, such as in the case of websites and online support communities
for topics focused on harmful behaviors such as suicide, anorexia, and cutting
and self mutilation. Previous anecdotal and empirical research has indicated that
not only can individuals find information online instructing people how to behave
in such self-injurious behavior, but online communities exist for the promotion
and support of these behaviors (Abbate Daga, Gramaglia, Piero, & Fassino,
2006; Alao, Yolles, & Annenta, 1999; Berger, Lehmann, Larson, Alvemo, &

Tsao, 2005; Dobson, 1999; Fox, Ward, & O’Rourke, 2005; Mulveen & Hepworth,

17

2006; Norris, Boydell, Pinhas, & Katzman, 2006; Prior, 2004; Tierney, 2006;
Whitlock, Powers, & Eckenrode, 2006). In fact, a recent investigation of online
communities focused on self-injurious behaviors found 406 online communities
devoted to the topic by using the search terms “self-injury”, “self-harm”, “self-
mutilation”, and “cutting” (Whitlock et al., 2006). In addition to providing
information about topics such as suicide and other self-injurious behaviors, the
use of specific suicide-related content on the lntemet has been cited to have an

influence on individuals’ decisions to commit suicide, otherwise known as

cybersuicide (Rajagopal, 2004).

Ethical Concerns with Studying Online C_ommtmiti£

As with any research with human subjects, there is a tremendous
responsibility to ensure that research is done ethically. The practice of
conducting research in a computer-mediated environment greatly increases the
number of ethical considerations that a researcher must ponder. Areas in which
ethical dilemmas arise during the study of online communities include the
obtainment of informed consent and protecting community participants from
risks, malfeasance, and maintaining privacy (Flicker, Haans, & Skinner, 2004;
Rhodes, Bowie, & Hergenrather, 2003).

As mentioned above, the collection of informed consent is made
increasingly difficult when research is conducted in a computer-mediated
environment. It is virtually impossible in large communities to obtain informed

consent from all community members. To compensate for this, some schools of

18

thought believe that if conducted properly, research that only looks at archives of
posted messages does not need informed consent from posters.

To determine whether or not informed consent is required from community
members, the distinction between public and private spaces should be made.
Public spaces are open communities where any person can post messages and
search the archives without restriction. Private spaces are closed communities
where members must register with the group owner in order to join, post, and
search the archive (Perron, 2002). In addition, these private spaces may include
further restrictions, which place limitations on the types of activities that can
occur within the community.

Eysenbach & Till (2001) describe three criteria to use when determining
whether a community is to be considered public or private. First, if a subscription
or other form of permission is required to join a community, it is probably
considered private. Next, the number of members within a community can add to
the perceived level Of privateness. The larger the community, the less it is
private. Lastly, community norms ultimately dictate the privateness of a
community and codes of conduct and FAQs should be consulted.

In addition to ethical considerations when conducting research, there are
also considerations to be made in regards to reporting results. First, care should
be taken when identifying the specific community that is the focus of study. In
order to protect the privacy of community members, researchers must evaluate
whether or not to explicitly name the community. Second, identification of

individual users by name or pseudonym should be avoided unless informed

19

consent has been obtained. Third, the reporting of direct quotations from
community members should be avoided unless informed consent has been
obtained. With the use of Google and other search engines, not only is it possible
to search for an individual member online, it is possible to find the original
comment posted in an online community. These comments and online identities
could then possibly be traced back to the source (Eysenbach & Till, 2001).

Now that we have explored online communities in-depth by defining what
they are, how they are used for health-related matters, and identified the ethical
concerns for the investigation and reporting of results, a review of the social

support literature will now be presented.

Social Support
Social support, the idea that people give assistance to one another
through emotional and tangible means during times of need, is a concept that is
pervasive through the social science and medical literature. In this section,
definitions of social support are presented along with a discussion of social
support and its history of research, methods of measurement, theoretical

approaches, influences, impacts on health, and sources.

Definition of Social Support

When conducting a search for the term “social support” in Google
Scholar, a search engine designed by Google to specifically search scholarly
literature such as “peer-reviewed papers, theses, books, abstracts and articles,

from academic publishers, professional societies, preprint repositories,

20

universities and other scholarly organizations” (Google, 2007), as of March,
2007, over 520,000 hits were retrieved. In addition, as of March, 2007, the same
search in Pubmed, “a service of the US. National Library of Medicine that
includes over 16 million citations from MEDLINE and other life science journals
for biomedical articles back to the 19503” (Pubmed, 2007), retrieved over 32,750
hits. This serves as a great indicator of the popularity of social support research.
Social support has come to mean many different things to many different
people. For as many social support researchers as there are in the world, there
are without a doubt an almost equal number of definitions. Although there is no
concretely accepted definition, “social support generally refers to the perception
that assistance is or could be available from significant others (perceived
support), or to reports of actual transactions that typically do occur or have
occurred between people in times of trouble (received support)” (Thoits, 1992, p.
57). Social support definitions have run the gamut of specificity and the topic has
been as broadly defined as “the resources provided by other persons” (Cohen &
Syme, 1985, p. 4) and as narrowly defined as “an interpersonal transaction
involving one or more of the following: 1) emotional concern (liking, love,
empathy), 2) instrumental aid (goods or services), 3) information (about the
environment), or 4) appraisal (information relevant to self-evaluation)” (House,
1981, p.39). In his 2004 book Social Support and Physical Health: Understanding
the Health Consequences of Relationships, Bert Uchino summarizes the state of
social support research when he says “Social support is usually defined to

include both the structures of an individual’s social life (for example, group

21

memberships or existence of familial ties) and the more explicit functions they
may serve (for example, provision of useful advice or emotional support)”
(Uchino, 2004, p. 9-10). Another conception of social support is that it is an
interactional process in which two people interact with one another in both
support giving and receiving capacities (PeaI1in, 1985).

It has been suggested by researchers that due to the multifaceted
nature of social support, we should abandon attempts at defining social support
as a singular entity and rather focus our efforts on more precisely defining the
broader constructs that make up social support (Cohen, 1992; Vaux, 1988).
Because of the problems associated with the many different conceptions and
definitions of social support “common use of the term social support had
obscured the degree to which researchers examined fundamentally different
facets of close relationships and their benefits for participants” (Goldsmith, 2004,
p.12). Many researchers interested in more fully understanding social support do
tend to agree that social support is generally divided into structural (social
integration components) and functional (different types of support) components
(Helgeson, 2003; House & Kahn, 1985; Uchino, 2004). The functional aspect of
support is often further subdivided into different types of support. Adding to the
confusion of defining social support, the taxonomy of support types varies by
researcher (Cohen & Syme, 1985; Uchino, 2004; Vaux, 1988). Looking across
different support taxonomies, different candidates for types of support include
emotional, informational, companionship, belonging, validation, instrumental,

tangible, motivational, appraisal, etc. To make matters more confusing, not only

22

do different researchers use different taxonomies when describing functional
support, they also may define commonly agreed upon support types in different
ways.

To date, there has been little definitive agreement on how to
conceptualize, define, operationalize and measure social support (Hupcey, 1998;
Sarason, Sarason, & Pierce, 1990). What is largely agreed upon though is the
idea that social support is a multifaceted concept (Wilcox & Vemberg, 1985).
While some researchers claim that it is an umbrella term used for describing the
different ways social relationships impact health and well-being (Goldsmith,
2004; Sarason, Pierce, & Sarason, 1990), others have gone so far as to state
that “no single and simple definition of social support will prove adequate
because social support is a metaconstruct: a higher-order theoretical construct
comprised of several legitimate and distinguishable theoretical constructs” (Vaux,
1988).

Over the years, three primary approaches have been utilized when
investigating social support: social network, psychological, and communicative
approaches. The social network approach, perhaps the earliest method for
studying social support, looked at how different characteristics of ones social
network (density, size, and other factors) impacted health and well-being.
Psychological approaches were concerned with how individual psychological
factors such as the perception and satisfaction of support had an impact on
health and well-being. Finally communication approaches View social support as

exchanges that take place within a relationship. This more recent approach views

23

support as an interactional enterprise with a support provider and a support

receiver. (Burleson, Albrecht, Goldsmith, & Sarason, 1994)

History official Support Research

As previously demonstrated, information searches on the term “social
support” literally reveal thousands of academic research studies where social
support is measured as a variable. Further reading of many of these studies
reveal poorly conceptualized and defined notions of the construct of social
support. When looking deeper into the theoretical and conceptual work done in
the field of social support, we are shown a very muddled and murky history of
this well thought of, yet hardly understood concept. Through history, social
support has been conceptualized, defined, measured, and analyzed in more
ways than one could count. In fact, conflicting research findings within the social
support field have often been attributed to the fact that no one conceptualizes,
defines, measures, and analyses social support the same way (Barrera & Ainlay,
1983; Cohen & McKay, 1984; Goldsmith, 2004; Uchino, 2004; Vaux, 1988;
Wilcox & Vemberg, 1985). To further stir this muddled quagmire of research,
many empirical studies investigating social support interventions have been
atheoretical (Helgeson & Gottlieb, 2000; Turner & Shepherd, 1999). As time has
surged on the in field of social support research, the concept has continued to
grow more convoluted. Almost any form of social interaction has come to be
considered a form of social support (Hupcey, 1998). Despite the problems
surrounding the concept of social support, there is one thing that people can

agree on; while the mechanism by which social support works is largely

24

unknown, there is something about relationships with others that can have
positive effects on the health and well being of people.

Preliminary research on social support can be traced back over 100 years.
Durkheim, in his landmark book Suicide, originally published in 1897 and
translated into English in 1951, found that suicide was more prevalent in people
with fewer social ties (Durkheim, 1951 ). In another preliminary study in social
support, Thomas and Znaniecki (1920) documented the social disruption of
Polish immigrants as they moved from small rural villages to large urban US
Cities. Two seminal papers, published by Sidney Cobb and John Cassell in 1976,
fueled the flames of what was to become an explosion of social support
research.

Sydney Cobb defined social support as “information leading the subject to
believe that he is cared for and loved, esteemed, and a member of a network of
mutual obligations” (Cobb, 1976, p. 300). In his view, social support helps people
to cope with different crises, as well as aid people in adapting to change.
Through the course of this influential paper, Cobb cites various research studies
focused on different points of the lifecycle (from birth to death) to back up claims
that social support is a protective force in times of crisis. He cites research
suggesting that social support has positive impacts on the frequency of low birth
weight babies and the reduction in the amount of medication and time spent in
the hospital for patients, etc. John Cassell, on the other hand, looked at social

support from a biological perspective and considered it to be an important factor

25

in disease etiology. Cassell proposed that social support is an important factor in

a person’s biological resistance to disease (Cassell, 1976).

Measuring Social Support

As evidenced previously, there are an incredible number of investigations
focused on social support. While there is a whole host of ways social support has
been defined in the past, there are also a myriad number of ways that social
support has been measured. There currently exists a multitude of instruments
designed to measure different components of social support. In addition, there
are different sets of measures for structural and functional aspects of support.
For example, Vaux categorizes social support instruments according to the types
of things that they assess: measures of support network resources, measures of
supportive behavior, support appraisals, support incidents, help seeking and
support mobilization, and social participation (Vaux, 1988).

Instruments aimed at ascertaining measures of support network resources
are primarily interested in measuring social networks. The term social network is
used to describe the complete system of people that one is acquainted with. The
strength Of ties between network members can differ with some ties being quite
strong (such as familial ties) and others being quite weak (such as common
acquaintances). Social networks are quite important to the idea of social support
for it is within this network that people turn to for support. Therefore, instruments
aimed at measuring support network resources primarily attempt to investigate
the characteristics of size, density, intensity, boundaries, etc. Characteristics of

social networks such as size and density are thought to have an impact on the

26

support that one can receive (Brissette, Cohen, & Seeman, 2000; Vaux, 1988).
The underlying problem with these measures lies in the fact that each instrument
fails to measure all different aspects of social networks. Most rely on evaluating
one or two aspects such as size or density.

Those instruments that are aimed at measuring supportive behaviors do
just that. They attempt to determine what types of behaviors a person has
encountered that were seen as being helpful. This type of social support has
often been overlooked in favor of social network analysis or perceived support
approaches.

The largest category of social support measures, instruments aimed at
assessing support appraisals, seeks to determine how helpful supportive
behaviors are, or how satisfactory received support was seen by the recipient.
Measures of support incidents seek to identify characteristics from actual
occurrences Of supportive behaviors. Another category of instruments, help
seeking and support mobilization measures, attempt to determine differences
between individuals and their help seeking beliefs and behaviors.

Finally social participation measures seek to identify the extent to which
individuals participate in social endeavors within the community at large.
Essentially, these are measures of social integration. Social integration is a
multidimensional construct that often encompasses two different ideas. First of
all, social integration is thought to include a behavioral component that refers to
the amount of engagement one has in a wide range of activities and

relationships. Secondly, it is also thought to include a cognitive component that

27

refers to ones sense of community and identification with social roles (Brissette
et al., 2000). Over the years, social integration has been operationalized
differently along these two different conceptions. It has been measured through
variables such as marital status, contact with friends, roles within a community,
voluntary associations, religious participation, visits with friends and relatives,
social contacts, etc. While some researchers choose to define social integration
along behavioral lines, others define it along cognitive lines. Furthermore, many
researchers include both behavioral and cognitive components in their definition
of social integration. Due to the conflicts in defining social integration, it is difficult
to compare research that has attempted to isolate this measure. Based on
Vaux’s work, Appendix A illustrates the different social support instruments
based on they category of assessment (Vaux, 1988).

Despite the large number of social support instruments available to
researchers, “there are few or no data available on the psychometric qualities of
most of these measures or on their relationships to one another” (Cohen &
Syme, 1985, p.4). The majority of these social support measures attempt to
identify either psychological variables that ascertain perceptions of social support
or descriptive measures that focus on the properties of an individuals support
network. As such, of the overall set of social support literature, few investigations
analyze the actual support interactions that occur between people. In addition,
due to the fact that each instrument measures a different aspect of social
support, it is impossible to use any instrument in isolation and still claim it is a

valid measure of social support. Clearly, among all of the published studies

28

focused on social support, many of them are centering on and measuring
different things. This speaks to the fact social support probably is a multi-faceted
construct. Without better instruments and without a more holistic approach to
understanding social support, we will be unable to further our knowledge of social

support and the mechanisms through which it operates.

Social Support Theories

Over the years, there have been two general models that have
attempted to explain the mechanism through which social support impacts health
and well-being. The two prevailing mechanisms are the stress-buffer and direct
effects models of social support. In the stress-buffer model of social support,
social support acts as a buffer that protects individuals from stressful events and
conditions. In the direct effect model of social support, social support is thought
to have direct effects on a persons psychological and physical health and well-
being, independent of stress and coping processes (Uchino, 2004; Vaux, 1988).
Most of the social support theories that have been proposed are born from one of
these two research traditions. It should be noted that there is a well regarded
hypothesis regarding social support models known as the matching hypothesis.
This hypothesis (which generally falls under the stress-buffering model of social
support) posits that different types of support are best matched to different
stressors. Different support types can vary in the type of support offered, or the
source where support is coming from. This hypothesis has been the foundation

for some of the theories of social support. While the matching hypothesis has

29

received mixed results, it has received a lot of attention. The following is an

explanation of the more prominent theories regarding social support.

Transactional Stress and Coping Process

The Transactional Stress and Coping Process theory of social support
was developed by Lazarus and Folkman in 1984. This theory posits that the way
people interpret events and stressors has an impact on how stressful they are
perceived. Lazarus and Folkman give consideration to the fact that different
individuals placed in the same circumstance will perceive events differently, as
well as experience different reactions and coping strategies. Furthermore,
Lazarus and Folkman suggest that a transactional model “says that stress is
neither in the environment nor in the person but a product of their interplay”
(Lazarus & Folkman, 1984, p. 354). In order to interpret events, two types of
support appraisals are conducted. The first appraisal, known as the primary
appraisal, is used to determine whether or not a stressor is a threat. Each
encounter is identified as either irrelevant (when an encounter has no bearing on
the well-being Of the person), benign-positive (when an encounter has a positive
effect on the well-being of the person), or stressful (when an encounter is thought
to have a negative effect in terms of harm/loss, threat, or challenge, on the well-
being of the person). If an encounter is evaluated as being stressful during the
primary appraisal, the secondary appraisal process occurs. The secondary
appraisal is a complex process that determines what resources (either personal
or social) are available to help a person cope, which of those processes is likely

to successfully help a person cope, and the likelihood that a particular coping

3O

strategy can be employed effectively. Overall, the more negative the primary and
secondary appraisal is thought to be, the more stress a person will experience

(Lazarus & Folkman, 1984).

Stressor-Support Specificity Model
Developed by Cohen and McKay (1984) the stressor-support specificity

model is a matching model of support where it is believed that in order to reduce
stress, the appropriate method of support must be obtained. This model takes
into consideration the notion that it is impossible to adequately consider the
buffering process without taking into account the multifaceted nature of stressors
and support systems. In this model, there are three different types of support,
each of which is a different mechanism that allows personal relationships to
buffer the stress that a person experiences. These three types of support include
tangible (the offer of material resources), appraisal (assistance with defining a
situation as less stressful), and emotional support (increasing self esteem and a
sense of belonging). Cohen and McKay state that “Only the interpersonal
relationships that provide the appropriate forms of support will operate as
effective buffers” (p. 261 ). Cohen and McKay also take into consideration the fact
that different stressors may be best suited to more than one type of support.

(Cohen & McKay, 1984).

Optimal Matching Theory
The Optimal Matching Theory (OMT) developed by Cutrona in the early

1990’s is another matching model of social support. In this theory, five different

types of support are identified: emotional, network, esteem, tangible, and

31

informational support (Cutrona, 1990; Cutrona & Russell, 1990). Later work by
Cutrona grouped the initial five types of support into the categories of action-
facilitating (problem solving that includes information and tangible aid support)
and nurturant support (emotional and network support) (Cutrona & Suhr, 1992).
Within Optimal Matching Theory, Cutrona also identifies four dimensions
of stressors that may be best matched to different types of social support. These
dimensions include desirability (referring to the nature and strength of negative
emotions), controllability (extent to which a stressor can be controlled or
managed), duration of consequences (length of stressor effects), and life domain
(how a stressor affects a person) and were abstracted from theoretical and
empirical investigations of social support taxonomies (Cutrona & Russell, 1990).
These different characteristics determine what type of social support is required
in order to achieve positive health outcomes. In OMT, the effectiveness of a, type
of support is dependent on how well it meets the demands of a specific stressor.
Of the four dimensions, controllability, referring to the amount of control one
perceives to have over the stressor, was cited as being the most important
influence on the type of support that a person requires. For events with a
perceived high level of controllability, instrumental and informational support is
useful while events that are perceived to be uncontrollable require more
emotional support (Cutrona, 1990; Cutrona & Russell, 1990). Cutrona and
Russell (1990) characterize medical illnesses as negative, uncontrollable events
and because illness is most often times perceived to be uncontrollable, in health-

related circumstances, emotional support is often more beneficial. It is further

32

suggested that illness can have effects on multiple domains of a person’s life and
support can be best obtained by talking with others who have had the same
experiences in the same domains. While the exact matches between support
needs and support types have not been hammered out, it is possible that further
work in this field will rectify this problem.

Matching theories of social support have encountered problems over the
years. These problems are largely due to the fact that stressors are not easily
classified as controllable or uncontrollable and different types of support often
overlap (Uchino, 2004). Despite the fact that matching theories have had very
mixed results, they seem to make up some of the more promising theories of
social support. The recognition that one type of support is not necessarily best for
the same situation across people is an important conceptual improvement
(Goldsmith, 2004). Irrespective of the problems that matching theories have
faced, researchers still seem to favor them based on statistical modeling
evidence that different types of support make up the higher order concept of
social support and evidence that specific functional measures of the different
types of support demonstrate that social support matters in regards to health
(Uchino, 2004). Despite the fact that social support theory is still not concretely
delineated, it is essential that we being to understand social support in pragmatic
ways. Perhaps by investigating interactions, particularly from a functional support

perspective, we will be able to more concretely structure social support theories.

33

W

Like all other things in this world, it is difficult to understand social support
if one considers it to occur within a vacuum. Like all human behavior, social
support interactions are influenced by the context and environment in which they
occur. By understanding some of the potential influences on social support, we

may be able to better understand how it can be affected.

Perceived Support vs. Received Support

Within the social support field of research, the topic of perceived social
support (as opposed to support that has been received) has gained a lot of
momentum. A study by Barrera (2000) found that the perceived availability of
support is more related to successful coping than support that is actually
received. Previous research has shown consistently that perceived support
measures have very strong associations with positive health outcomes (Blazer,
1982; Wethington & Kessler, 1986). It has also been suggested that the
relationship between perceived and received support is complicated and the two
concepts are not highly related (DunkeI-Schetter & Bennett, 1990; Uchino, 2004).
Possible explanations for this low relationship include the fact that perceived
support is a cognitive representation of support that is actually received by a
person (Uchino, 2004). In terms of measuring perceived support, there are many
different measures that look at different aspects of perceived support. Some
measures focus on the availability of help, while others focus on satisfaction

(Sarason, Sarason, & Pierce, 1994).

34

Gender

Previous literature has indicated that sex/gender is related to social
support and has an effect on both the provision and receipt of support. It has
been previously suggested that more than any other social status variable,
gender is associated with many different factors (such as composition and size of
social networks, amount of support that is received, level of emotional exchange
within relationships) that may influence how social support is acquired and
experienced (Vaux, 1985, 1988).

Previous research has found many differences between the genders and
how gender relates to social support. For instance, studies have indicated that
men receive the majority of their support from their spouse, whereas women
receive the majority of their support from friends (Allen & Stoltenberg, 1995;
Antonucci & Akiyama, 1987; Depner & Ingersoll-Dayton, 1988; Fischer, 1982;
Kendler, Myers, & Prescott, 2005; Robinson, 1995; Vaux, 1985) although both
men and women typically seek out individuals of the same sex for support
(Cramer, Riley, & Kiger, 1991; Griffith; 1985; Soliman, 1993). Women also report
receiving more social support from family members (Allen & Stoltenberg, 1995;
Kendler et al., 2005).

It has also been established that women tend to have larger support
networks than men (Antonucci & Akiyama, 1987; Belle, 1987; Fischer, 1982;
Turner, 1994), are more involved in their social networks (Kendler et al., 2005;
Rosenthal, Gesten, & Shiffman, 1986), provide social support more often than
men (Belle, 1982; Fischer, 1982; Griffith, 1985; Kendler et al., 2005; Kessler,

McLeod, & Wethington, 1985; Trobst, Collins, Embree, 1994), and typically report

35

both seeking and receiving emotional support more often than men (Ashton &
Fuehrer, 1993; Burda, Vaux, & Schill, 1984; Butler, Giordano, & Neren, 1985).
Furthermore, previous research has also indicated that females are more likely
than males to join support groups (Krizek, Roberts, Ragan, Ferrara, & Lord,
1999) and that current members of support groups are more likely to be female
(Grande, Myers, & Sutton, 2006). In addition, a recent investigation of Alcoholics
Anonymous (AA), a support group for alcoholics, found that although there were
no significant differences between genders in their decision to receive treatment
for alcoholism, females were more likely to attend AA than males (Moos, Moos, &
Timko, 2006).

In an investigation by Turner (1994), it was discovered that women report
more instances of perceived support, which may possibly be due to the fact that
in this study (like previous findings), women reported more contact with their
support network with greater emotional involvement than men. Despite the
finding that women receive more support than men in this study, it was stated
that women also appear to have more negative support interactions than men,
possibly due to the increased contact and heightened emotional involvement.

In terms of gender differences within online support communities, Klemm,
Hurst, Dearholt, & Trone (1999) conducted an investigation of social support
within online breast and prostate cancer communities and discovered that in
regards to gender, men were more like to seek or provide information and
women were more likely to provide encouragement and support. This finding was

largely echoed by Seale, Ziebland, & Charteris-Black (2006) in their keyword

36

analysis of interviews and online support groups for cancer where it was found
that in online communities for breast and prostate cancer, men were more likely
to use keywords related to treatment, side effects, and diagnosis, as well as cite
information from other websites. Sullivan (2003) also found similar patterns of
social support provision in an investigation of online prostate and ovarian cancer
communities where men provided more informational support and women were
more emotion oriented. To summarize, Rosenthal, Gesten, & Shiffman (1986)
indicate that sex differences in social support “appear to be connected with
traditional sex role expectations that discourage men from engaging in intimate,
self-disclosing relationships while encouraging emotional closeness and intimacy

for women” (p. 482).

Other Influences on Social Support

Aside from the notion of perceived support and gender, the social support
literature points to several other potential influences on social support. For
example, it has been suggested that additional socio-demographic
characteristics of support recipients such as age and race are potential
influences on social support (Riley & Eckenrode, 1986). In addition, it has been
suggested that if support providers and receivers appraise a situation differently,
the support given may not be appraised by the recipient as being helpful (Dunkel-
Schetter & Skokan, 1990).

Another potential influence on social support is the determination Of
support type, as well as support timing. Before providing support, an individual

must recognize that support is needed by another and determine what type of

37

support might be most helpful (Dunkel-Schetter & Bennett, 1990). It has been
suggested that the timing of support is every bit as important as the type of
support provided (Jacobson, 1986). It is possible that if support is given at a time
when it is not expected, the support will not be appraised as being helpful and

may not be appreciated (Dunkel-Schetter & Skokan, 1990).

WWW

Despite the fact that the mechanism through which social support
operates is still largely unknown, in the past few decades, there have been an
incredible number of research studies linking measures of social support to either
mortality or health outcomes. In fact, some of the most influential studies
regarding social support and health outcomes were large scale longitudinal
epidemiological studies focused on all-cause mortality. Many research studies in
this area were conducted thirty years ago, but still remain some of the most
important work done in this field. To better understand the influences that social
support has on health outcomes, we shall look at both influences on all-cause

mortality, as well as disease specific influences.

Impact of Social Support on All-cause Mortaliy

Perhaps the most well known study investigating structural social support
measures and mortality was the Alameda County study conducted by Berkman
and Syme (1979). This was one of the first well controlled longitudinal studies
attempting to link social support and mortality (Uchino, 2004). All-cause mortality

is defined as death from a variety of causes, not one factor in particular. In 1965,

38

a large scale survey was conducted by the Human Population Laboratory, a part
of the California Department of Health. A total of 6,928 randomly selected
residents of Alameda County, California, participated in this study. Berkman and
Syme conducted a nine year follow-up to this initial study (1965 - 1974) looking
primarily at mortality rates for those who participated. Through various methods,
Berkman and Syme were able to locate 96% of the original 1965 sample.
Looking at marriage, contact with friends and relatives, membership in a church,
and group associations, both formal and informal, as measures of social and
community ties, it was found that those people who had fewer social and
community ties were more likely to have died during the nine year follow-up
period. In addition, it was determined that each of the four structural social
support measures predicted mortality independent of the other three. It was also
found after looking at age-adjusted relative risks for people who were most
isolated in comparison to those who were least isolated that men were 2.3 times
more likely to have perished during the follow-up period and women were 2.8
times more likely. Finally, it was found that the link between social ties and
mortality was independent of several factors taken into consideration in the initial
study. These factors included self-reported physical health status, year of death,
SES, and various health practices such as obesity, physical activity, smoking,
alcoholic beverage consumption, etc (Berkman & Syme, 1979).

Another influential longitudinal study linking social support and mortality
was the Tecumseh Community Health study by House, Robbins, and Metzner

(1982), which was a partial replication and extension of the work published by

39

Berkman and Syme in 1979. In this study, House, Robbins, and Metzner
compared data taken from portions of the Tecumseh Community Health study
(originally collected from 2,754 adults aged 35 — 69 in the years 1967 — 1969)
and compared it to current mortality information on the same participants. It was
found that men who had reported higher levels of social relationships in the
original study were significantly less likely to have died during the follow-up
period when controlled for age and a variety of risk factors. Although data for
women followed the same trend, the findings were not significant when controlled
for age and a variety of risk factors (House et al., 1982).

In 1982, another longitudinal study linking social support and mortality was
published by Dan Blazer. This study was a first of its kind due to the fact that it
only focused on an elderly population. Blazer composed a sample of 331 seniors
over the age of 65 in 1972 from Durham County North Carolina. These seniors
were assessed on three different measures of social support: perceived support,
roles and available attachments, and frequency of social interaction. Five years
after the initial assessment, mortality status of the original sample was
determined. Results indicated that all three measures of social support were
predictors of mortality in this elderly population (Blazer, 1982).

To further establish a relationship between social support and mortality,
researcher Bert Uchino, again in his book Social Support and Physical Health:
Understanding the Health Consequences of Relationships (Uchino, 2004),
conducted a review of over eighty published research articles linking social

support and all-cause mortality. In his investigation, approximately 80% of the

40

reviewed studies found that there were associations between either structural or
functional measures of social support and lower all-cause mortality rates. Uchino
found through his research that people who have low levels of social support
have between two and three times greater risk of death than those with high
levels of social support. Furthermore, he also found through reviews of social
support and mental health that people who have higher levels of social support
also live happier lives than those with low levels.

Although mortality research plays an important role in linking social
support and health outcomes, it fails to capture finer grained information
regarding the impacts of social support on health. For example, it has been
suggested that mortality studies fail to help us understand where in the course of

a disease social support has the most impact (Berkman, 1985).

W

In addition to investigating all-cause mortality, Uchino also investigated
whether or not social support has impacts on health outcomes for cardiovascular
disease, cancer, and HIV/AIDS. In his investigation, Uchino found ovenlvhelming
evidence that both structural and functional measures of social support have
positive impacts on cardiovascular mortality. Unfortunately, due to the small
number of research studies linking social support to cancer or HIV/AIDS
mortalities, Uchino was unable to draw any strong conclusions about the positive
impacts of social support on these diseases, although he indicated that the

preliminary findings were promising (Uchino, 2004).

41

Likewise, a similar, though smaller review of 17 articles focusing on the
impacts of social support on patients with heart failure found that social support
appears to be a good predictor of hospital readmissions and mortality in patients
with heart failure. It was suggested in this review that emotional support probably
plays an important role in this. Like other studies that have attempted to
aggregate results across other published research, conflicting results of the
impacts of social support on the health and well-being of patients with heart
failure was found. But, like other review studies, these conflicting findings could
potentially be explained by the differences in the conceptualization,
operationalization, and measurement of social support (Luttik, Jaarsma, Moser,
Sanden'nan, & Van Veldhuisen, 2005).

In addition to having impacts on heart failure, recent review investigations
have found links between social support and positive health outcomes. A recent
review investigation has suggested that social support may have positive impacts
on weight loss interventions (Verheijden, Bakx, van Weel, Koelen, & van
Staveren, 2005). Another recent review investigation has suggested that people
with low levels of social support in the workplace may have increased risk for
work-related musculoskeletal morbidity. In addition, there is limited evidence that
people with low levels of social support in the workplace may have increased risk
for musculoskeletal sickness absence, not returning to work after experiencing a
musculoskeletal problem, and restricted activity (Woods, 2005). Another recent
review of the effect that social support has on chronic disease management

found the social support has a positive impact on the management of chronic

42

diseases, particularly diabetes management, since investigations focusing on
social support and diabetes management comprised a majority of the studies

reviewed (Gallant, 2003).

Sources of Social Support

With the advent of the Internet, one thing has become clear. There are
multiple places to find support. While traditional sources of social support such as
immediate social networks and traditional support and self-help groups have
been recognized as offering social support for years, it has only been in the last
ten to fifteen years, with the rapid diffusion of the World Wide Web, that online
social support has been identified.

In terms of traditional sources Of offline social support, basic social ties
with friends and family members are generally what we think of when we think of
social support. In fact, for married people, the spouse is typically the first place
that people turn to for support during a crisis (Burke & Weir, 1977). A long line of
research has established that just being married has a positive impact on health.
People who are married have lower rates of many diseases and conditions
including heart attacks, strokes, tuberculosis, cirrhosis, and lung cancer. In
addition, marriage also has positive effects on mental health (Argyle, 1992).
Furthermore, it has also been suggested that friends can provide support in times
of great stress (Derlega, Barbee, & Winstead, 1994).

When people cannot find the support that they need from their existing
social network, people often times have a need to turn to other sources to find

the support they require. These other sources may often come in the form of

43

support or self-help groups. Support groups are groups of people that come
together for the express purpose of gaining social support from one another.
Support groups are a hybrid of group therapy and self-help groups and employ
professionals to disseminate information, provide group-skill training, and guide
group processes (Helgeson & Gottlieb, 2000). Support groups typically have six
to twelve members that share a similar stressor or other similarity and receive
expert information and training to Ieam to help one another to increase coping
and adjustment (Gottlieb, 1998). Support groups are often times sought out when
there is a deficiency in a persons natural support network (Helgeson & Gottlieb,
2000).

Of course traditional support groups and self-help groups are no longer
the only place for someone seeking social support to turn to when their social
network does not provide the support that is needed. We will now turn to an

overview of the literature concerning online health-related support groups.

medium

With the widespread diffusion of the World Wide Web, many people have
turned to online communities for support. In addition to the lntemet serving as a
widely accessible repository for cancer-related information, it has also become a
veritable meeting ground for individuals with cancer to seek support from one
another. Many research investigations in the last decade have documented the
existence of online support communities for use for many different types of
cancer including colorectal cancer (Klemm et al, 1998), breast cancer

(Lieberman et al., 2003; Owen, Klapow, Roth, Nabell, & Tucker, 2004; Owen,

44

Klapow, Roth, & Tucker, 2004; Till, 2003; Weinberg, Schmale, Uken, & Wessel,
1996; Winzlenberg et al., 2003), prostate cancer (Owen, Klapow, Roth, & Tucker,
2004), and leukemia (Ramos, Rai-Chaudhuri, & Neill, 2004).

Although not the first online support group, but perhaps the most well-
studied, the Comprehensive Health Enhancement Support System (CHESS)
designed by researchers at the University of Wisconsin - Madison, has come to
be thought of as perhaps the quintessential example of successful online social
support. CHESS is an online health-related support group focused on breast
cancer and is one of the few online communities operated and maintained by a
healthcare organization. CHESS was developed in 1987 by Dr. David Gustafson
whose wife had been diagnosed with breast cancer. After the diagnosis, Dr.
Gustafson and his wife were faced with numerous difficulties finding good
informational and emotional support. Drawing on previous research on social
support for patients and families experiencing serious illness, the aim of CHESS
was to provide comprehensive support and information services to patients and
their families. The overall goal for the researchers involved with CHESS was “to
see individuals and family units cope more effectively with their crisis, suffer less,
and feel like they’d made better decisions as a result of using CHESS” (CHESS,
2006).

CHESS offers a full range of condition-specific informational and social
support services to patients and families in the comfort of their own home.
Currently, CHESS has several different modules for different disease/conditions

including breast cancer, HIV/AIDS, and stress management, among others.

45

There are a limited number of modules that CHESS offers for a limited number of
conditions. For people who have a need to access CHESS, but do not have
access to a computer, CHESS will loan a computer for up to one year. Because
of the tightly controlled nature of CHESS, in order to gain access to the system,
healthcare providers must purchase a license to the system (CHESS, 2006).
Therefore, a limited number of people have access to CHESS when there is a
very large potential market of people who are currently restricted from using the
system.

Due to its strong affiliation with the University of Wisconsin - Madison,
CHESS may be a breeding ground for online health community research, but it is
not very prototypical of the majority of health-related online support groups.
Numerous grants have been awarded to CHESS over the years and as such,
they are able to offer services that most online support groups do not. In addition,
the highly restricted nature of CHESS means that it is very heavily regulated.
Because the service is administered through a health provider, access to the
CHESS system is tightly controlled. Due to the fact that CHESS is so heavily
regulated and purposefully administered, it does make a wonderful case study for
the potential of online social support groups, but again, it is not very
representative of the vast majority of online social support groups and the
research conducted on the community is not very generalizable. The lack of
generalizability for the research on CHESS ultimately requires more research to

be conducted on more prototypical online communities.

46

Characteristics of Online Support Groups

Most online social support groups utilize mailing lists or listservs centered
on particular health-related topics (White & Dorman, 2001). In comparison to
their face-to-face counterparts, online health support communities are heavily
utilized by sufferers of rare and debilitating conditions that may not be well
understood in the medical community, may be difficult or impossible to cure, or
have been commonly ignored by traditional healthcare such as chronic fatigue
syndrome, multiple sclerosis, etc. (Davison, Pennebaker, & Dickerson, 2000).
These online support communities can be used on their own or as a supplement
to traditional face-to-face support groups for those patients who are in need of a
greater amount of support (White & Dorman, 2001).

A study conducted by Cummings, Kiesler, & Sproull (2002) indicated that
people with low levels of offline support (from family members and friends) are
more likely than people who have high levels of offline support to participate and
remain active in an online support community. In addition, those people with low
levels of offline support particularly felt that they gained emotional and
informational benefits from their participation in the online community. An
investigation of online depression support groups has also found that users of
online support groups tend to have less social support (Houston et al., 2002).

An investigation aimed at understanding SibKids, an online support group
for children and adolescents who have a sibling with special needs, found that
there were three types of support offered through the online community;

emotional support, informational support, and social companionship (Tichon &

47

Yellowlees, 2003). These findings are similar to many other assessments of

online support communities found in the online support community literature.

Reasons for Utilizing Online Health Support Communities

Regardless of disease or condition, health-related online support
communities are used to find information, ask questions, share experiences, and
seek/provide emotional support (Eysenbach et al., 2004; Lamberg, 2003; Ramos
et al., 2004; White & Dorman, 2000). They allow community members to share
their knowledge and experiences in order to add to collective group knowledge.
Patients with diseases can offer to others valuable knowledge and insight
regarding their condition (Lester, Prady, Finegan, & Hoch, 2004). It has been
found that many users of online health communities report that the knowledge
they gained through community involvement has helped them to make better
decisions regarding treatment (Ramos et al., 2004). Winkelman and Choo (2003)
argue that patients with chronic diseases can be quite knowledgeable about their
condition and may be a great source of information for others. They further argue
that for a whole host of reasons, healthcare organizations should embrace this
and develop organizationally sponsored virtual communities for chronic patients.
By relying on their knowledge and experience of a condition, along with the
moderation of a healthcare professional, people suffering from chronic conditions
can help to create more positive health outcomes for people in similar situations.

While the majority of online health communities are member created, in
line with Winkelman and Choo’s advice, we are currently seeing growth in the

area of health provider sponsored communities. These communities are being

48

created not only by health professionals, but by the pharmaceutical industry as a
platform for advertising their own products, while providing useful services to the
public. Healthcare providers have the potential to offer virtual community
members a lot and should consider creating their own communities. Healthcare
providers have the medical expertise needed to provide quality health
lnforrnation, they can expand their community range and reach by offering quick
medical advice, and they can collect information gathered through data logs to
track the technical capabilities of patients in order to provide better online
services and product promotions (Butler, Friel, Lang, Park, & Santello, 2000).

Previous investigations have demonstrated that people are open to using
the lntemet for both health-related information and social support (Cudney &
Weinert, 2000; Monnier, Laken, & Carter, 2002; Owen, Klapow, Roth, Nabell, &
Tucker, 2004; Pierce et al., 2002). Due to the fact that the lntemet offers
confidentiality and ease of access (unlike talking with healthcare providers face-
to-face), the lntemet may also be an extremely attractive option for adolescents
who are seeking potentially sensitive or embarrassing health-related information
(Gray, Klein, Noyce, Sesselberg, & Cantrill, 2004).

Utilization of online health support communities has also been shown to
increase perceptions of the availability of social support (Barrera, Glasgow,
McKay, Boles, & Fell, 2002). Although a limited number of investigations have
tied online health support communities to health outcomes, there is some
evidence that participation in these communities can lead to positive health

outcomes (Lieberman et al., 2003). Through the use of computer-mediated

49

communication, online social support groups have many advantages over their

traditional face-to-face counterparts, though there are disadvantages as well.

WNW! Swim

Because online support groups are a relatively young phenomenon and
because they may offer considerable advantages over traditional face-to-face
support groups, it is important that we begin to study them from a social support
perspective. By understanding their advantages and disadvantages, we may be

better able to satisfy the needs of people seeking social support.

Advantages of Online Support

One of the most obvious advantages to the use of online support groups iS
the increased level of anonymity that the Internet affords (White & Dorman,
2001). In a traditional face-to-face setting, the chances of maintaining anonymity
are quite small, but through computer-mediated conversation and the use of non-
identifying pseudonyms, higher levels of anonymity can be maintained. Online
support groups can also increase anonymity by reducing sociodemographic
factors that would otherwise be obvious. These factors include race, gender, age,
social status, etc. (Madara, 1997; O’Brien, 1999; Rheingold, 1993; White &
Dorman, 2001). Because of this increased anonymity, people may be more
willing to engage in online discourse regarding potentially embarrassing,
sensitive, and/or taboo topics (Houston et al., 2002; Madara, 1997; Rhodes et al.,
2003; Weinrich, 1997; White & Dorman, 2001).

Another great benefit of online support communities is the increased

access provided by the medium. Through the use of asynchronous

50

communication, these online health communities have the ability to transcend
time and geographical constraints. They are available across the globe, all day,
every day (Braithwaite, Waldron, & Finn, 1999; Houston et al., 2002; Lamberg,
2003; Madara, 1997; Ramos et al., 2004; White & Dorman, 2001). This has great
implications for people who are homebound or possess other physical
disabilities. In addition, the benefits to people in rural areas who may not have
access to traditional support groups are great (Hill & Weinert, 2004; Madara,
1997). Lastly, the use of online support may be the only way for people with rare
conditions to connect with other people who suffer from the same (Davison et al.,
2000; White & Dorman, 2001).

Regarding cost, online support communities offer a low cost investment
both in terms of time and monetary resources. Members of online support
communities can contribute as much or as little time to participation as they wish.
In terms of financial resources, most online communities are free and the only
real barrier to access is the availability of a computer. While not every household
has a computer, the number of people with access to the lntemet is increasing.
Current estimates state that 73% of adults and almost 90% of teenagers have
access to the lntemet (Madden, 2006).

Another great advantage to online support groups is that participation in
online communities allows members to passively participate through lurking.
Lurking Occurs when a person reads messages within an online community, but
does not post themselves. It has been suggested that the majority of members in

online groups are Iurkers and that despite the fact that they do not actively

51

participate by posting, lurkers often feel a sense of community (Nonnecke &
Preece, 2003). Speaking as a testament to the number of lurkers in online
communities, an investigation by Perron (2002) indicated that many people who
posted for the first time on a mental illness support community prefaced their
contribution with the fact that they had been lurking for some time.

There are numerous reasons why people lurk. Among the top reasons for
lurking include the individual belief that posting isn’t necessary, needing to learn
more about the community, the belief that individuals were helping the
community by not posting, not being able to post due to software concerns,
and/or poor fit between the individual and the community (Preece, Nonnecke, &
Andrews, 2004). Although lurking can occur for many different reasons, it is
important to note that people who lurk can still derive support without actually
communicating with others within a community. Preece, Nonnecke, & Andrews
(2004) discovered in an investigation on lurking that both lurkers and active
participants primarily join online communities for the same reasons: to increase
their understanding of a particular topic. In addition, within online health
communities, it was found that both lurkers and active participants go to seek out

social support.

Disadvantages of Online Support

Before a person can access an online support community, two
requirements must be met; access to a computer and access to the lntemet.
Because there is a lack of universal access to the Internet, it is impossible for

many people in this world to gain access to online communities (Braithwaite et

52

al., 1999; Madara, 1997). While public access is becoming more and more
available in the US, there still exists the possibility of a “digital divide” among the
population (Madara, 1997).

In addition, in all computer-mediated settings, there exists a lack of
contextual and social cues which may make communication more difficult. Not
only can unintended meanings be interpreted through messages, but lack of
context can often make communication impossible to understand. Because of
this phenomenon, online community members are faced with the task of writing
messages that are both clear and concrete (Perron, 2002).

Finally, a threat to online support communities is the great potential for
researchers to exploit these communities through unethical research (Madara,
1997). Not only are researchers a potential threat, but so are advertisers,
marketers, and other people who may visit a community for personal gain. When
many community members make the decision to join an online community, there
is little thought that they may potentially be the subjects of a research project. In
fact, increasing evidence suggests that researchers that are lurking on online

communities may be seen as intruders (Eysenbach & Till, 2001).

Summag
As more people utilize the lntemet for communication and information
related purposes, we will continue to see an increase in the number of people
turning to online support groups. By researching online communities and
understanding what they are used for, how they are utilized, and what the

advantages and disadvantages are of social support in a computer-mediated

53

environment, we may better understand the topic of social support in general.
Due to the fact that messages in online communities are easily archived,
studying social support on the lntemet provides us with the ability to easily study
social support in general. In addition, drawing comparisons between support
interactions in electronic environments vs. real life environments will give us
further understanding of how online and Offline communities Operate.

Because the field of social support, after all this time, still lacks a solid
theoretical foundation, it is only reasonable that we continue to try to document
the types of support interactions that are occurring both online and offline in a
consistent manner. It is through continued research in both of these realms that
we will be able to sort out what exactly social support is, how it is manifested in
the real world, and how we can better assist people with real social support
needs. Establishing a more comprehensive understanding of social support, both
online and offline, has enormous implications for all aspects of how we aid
people with support needs. By having a more complete view of social support,
we will be able to design, structure, and implement support services more
efficiently and effectively both in the online and offline worlds.

This investigation sought to investigate online support communities for
cancers with varying five-year relative survival rates. In addition to understanding
who is utilizing these communities in terms of gender and community member
role, this study also sought to understand the different types of social support
interactions that are occurring within these online spaces. This study utilized

Optimal Matching Theory as a theoretical framework for understanding the

54

different types of social support that occur within eight cancer-related online
support groups. Optimal Matching Theory is an appropriate theoretical framework
for this study due to the fact that it is not context dependent and can help us to
understand the types of social support interactions that are occurring in an online
environment. In addition, by utilizing Optimal Matching Theory as a theoretical
framework, predictions related to different social support type frequencies within
each community could be made, serving to guide us in this investigation, as well
as testing the theory in this electronic context. Finally, Optimal Matching Theory
is a useful theoretical framework for helping to understand participation in online
support communities because it gives us insight into why some members of a
community persist over very long periods of time, whereas some members only
visit once or twice. For those members who participate in an online support
community over time, it is reasonable to assume that they do so because their
support needs are being met. This suggests that members who continue to
participate in an online community have found successful support matches. For
those members who only visit a community once or twice, it is also reasonable to
assume that they do so because their needs are not being met and a successful
support match has not been made.

This investigation of social support interactions within online health-related
communities differs from many other social support investigations in that social
support interactions are of primary interest, not specific measures of social
support as identified through the use of questionnaires or inventories. While this

is not a novel approach to studying social support, it is a far less utilized strategy.

55

Compared to the number of investigations of social support that exist in the
literature, the approach of describing and analyzing social support interactions
has been utilized in a relatively few number of studies focused on online health-
related support communities.

Although previous studies have utilized a similar research design to the
one in this investigation, these previous studies had several weaknesses that this
investigation attempts to overcome. First of all, prior studies lacked
generalizability as they were either case studies or investigations that included a
very small number of communities, making cross community comparisons either
impossible or difficult (Blank & Adams-Blodnieks, 2007; Dunham et al., 1998;
Klemm et al., 1998; Lasker, Sogolow, & Sharim, 2005; Mendelson, 2003; Owen,
Klapow, Roth, & Tucker, 2004; Perron, 2002; Tichon & Shapiro, 2003; Weinberg
et al., 1996; White & Dorman, 2000). To overcome this shortcoming in previous
literature, this investigation compared eight online support communities in an
effort to establish the beginnings of generalizability of this line of research.

Second, the methodologies of previous content analysis studies were
often compromised as no measures of inter-coder reliability were reported (Blank
& Adams-Blodnieks, 2007; Klemm et al., 1998; Mendelson, 2000; White &
Dorman, 2000;). This investigation attempted to overcome this failing in previous
literature by establishing sound methodological practices through the use of
multiple coders and the establishment of an acceptable level of intercoder

reliability.

56

Another shortcoming of previous research has been the use of very small
sample periods, making generalizability difficult. Similar investigations have
analyzed data from nine days (Klemm et al., 1998), two months (Klemm &
Wheeler, 2005; Lasker et al., 2005; Owen, Klapow, Roth, & Tucker, 2004), or
even three months (Mendelson, 2003; Pierce et al., 2002; Tichon & Shapiro,
2003). This investigation included one randomly selected week from each month
over a one year time period in an effort to expand the sample size and cover a
wider period of time. Finally, this study moves beyond similar research by trying
to understand how the role of a community member impacts social support

provision within online support communities.

Research Questions and Hypotheses

In order to better understand the ways in which social support is provided
in online support communities, as well as test Optimal Matching Theory, 3 series
of three research questions and three hypotheses were developed to guide this
investigation. These research questions and hypotheses specifically address
how different factors influence social support provisions within online support
communities, as well as whether or not controllability has the influence on
support provisions that OMT indicates.

. RQ1: What is the effect of community member role on online cancer
support communities?
a. Who are the users of online cancer support communities (patient,

family member, medical professional, cancer survivor, etc.)?

57

b. What types of social support do community members with different

roles provide and how frequently is each type provided?

0 RQ2: What is the effect of five-year relative cancer survival rate on online
cancer support communities?

a. Does the survival rate of a specific cancer impact online community
membership composition (both in terms of gender and community
member role)?

b. How does cancer survival rate impact the types and amounts of

emotional and informational support within communities?

. RQ3: Are there specific characteristics of communities that impact social
support provision in online cancer support communities?

a. Does the community host (cancer-specific vs. general use) impact
the types and frequency of social support provisions or the
membership of online cancer communities (both in terms of
community member role and gender)?

b. Does the private or public nature of a community impact the types
and frequency of social support provisions or the membership of
online cancer communities (both in terms of community member
role and gender)?

c. Does the volume of posts within a community impact the types and

frequency of social support provisions or the membership of online

58

cancer communities (both in terms of community member role and
gender)?

d. Does the size of a community impact the types and frequency of
social support provisions or the membership of online cancer
communities (both in terms of community member role and

gender)?

To further guide this investigation, three hypotheses were created. Based
on the research previously discussed regarding gender and social support,
Specifically research that indicates that women tend to provide social support
more frequently than men (Belle, 1982; Fischer, 1982; Griffith, 1985; Kendler et
al., 2005; Kessler et al., 1985; Trobst et al., 1994), are more likely to join support
groups (Krizek et al., 1999), that current members of support groups are more
likely to be female (Grande et al., 2006; Moos et al., 2006), and that women are
more likely than men to provide emotional support and men are more likely than
women to provide informational support (Klemm et al., 1999; Rosenthal et al.,
1986; Seale et al., 2006; Sullivan, 2003), it was hypothesized that: These
expectations were based on previous research indicating that.

0 H13: There will be a larger proportion of women in the online support
communities than men.

0 H1 b: Women will post more messages than men.

0 H 1c: Women will provide a greater proportion of emotional support than

men.

59

. H1d: Men will provide a greater proportion of informational support than
women.

Based on Optimal Matching Theory, as controllability was considered to
be the most important factor in influencing social support types, hypotheses two
and three were created utilizing five-year relative survival rate as a proxy for
controllability. Figure 1 provides a graphical depiction of these two hypotheses.

- H2: There will be a greater proportion of emotional support provision in
communities for cancers with low five-year relative survival rates as

Opposed to communities for cancers with high five-year relative survival

rates.

0 H3: There will be a greater proportion of informational support provision in
communities for cancers with high five-year relative survival rates as
opposed to communities for cancers with low five-year relative survival

rates.

Figure 1: Graphical Representation of Hypotheses Two and Three

 

 

 

 

Five-year relative + 4 lnfonnational
survival rate Support
— Frequency

 

 

 

 

 

Emotional
Suppon
Frequency

 

 

 

 

 

 

60

Chapter 3: Methodology

In order to address the research questions and hypotheses set forth in the
preceding chapter, a quantitative content analysis was conducted on 3,717
message postings located in eight different health-related online support
communities focused on four different types of cancer. As defined by Riffe, Lacy,
& Fico (2005, pg. 25) “quantitative content analysis is the systematic and
replicable examination of symbols of communication, which have been assigned
numeric values according to valid measurement roles and the analysis of
relationships involving those values using statistical methods, to describe the
communication, draw inferences about its meaning, or infer from the
communication to its context, both of production and consumption.” Because
research questions concerned determinations regarding the content of posts
made within online support communities, content analysis was deemed the
appropriate method to investigate these particular questions. It should be noted
that the methods used in this investigation were approved by the Michigan State

University institutional review board.

Sam Ie

When determining community hosts for this investigation, two contrasting
websites hosting online cancer support communities were chosen; Yahoo! and
ACOR (Association of Cancer Online Resources). Of the eight communities
located in this sample, four were hosted by YahoolGroups and four were hosted
by the ACOR. Yahool, one of the most recognizable names in the online realm,

“is a leading global Internet communications, commerce and media company that

61

offers a comprehensive branded network of services to more than 345 million
individuals each month worldwide” (Yahoo, 2005).Yahoo! is known for being one
of the world’s most prominent search engines and lntemet portals. In addition to
the many information services offered within the Yahoo! portal, as of February,
2007, over 37,000 electronic support groups, over 10,000 electronic fitness and
nutrition groups, and over 2,500 electronic drug and medication groups had been
listed by YahooIGroups under the “health and wellness” category. YahoolGroups
was selected for inclusion in this study because Yahoo! is a popular information
portal familiar to many Internet users. Due to the fact that they do not dedicate
their services to any particular topic, Yahoo! is a destination for a very diverse
audience and users do not necessarily Visit Yahoo! strictly for cancer-related
information.

Regarding ACOR, founded in 1996, it has become one of the largest and
most well known websites for online support communities focused on cancer, as
well as other cancer-related information, available on the lntemet. With
approximately 200 public and private listservs devoted to different types of
cancer and cancer-related issues, an average of 200,000 email messages are
sent through ACOR listservs per day (L-Soft, 2005). As such, ACOR has grown
to be an incredibly valuable asset in the online social support community. ACOR
was selected for inclusion in this study because the ACOR website is devoted
strictly to cancer-related information, resources, and support. Users of ACOR
utilize the site primarily for these purposes. As ACOR requires users to register

for the various communities before being able to access the community archives,

62

permission to utilize the archives for this investigation was obtained by Gilles
Frydman, the founder of ACOR.

The eight different communities included in this investigation were focused
on four different types of cancer with two communities for each cancer type. For
each cancer type, one community was hosted by ACOR and the other by
YahoolGroups. Cancer types were further divided into high and low five-year
relative survival rate conditions. To truly make the distinction between high and
low five-year relative survival rates, cancer types were chosen according to
cancers with the highest and lowest five-year relative survival rates as reported
by the American Cancer Society. The low five-year relative survival rate condition
included cancers with a less than 20% five—year relative survival rate, while the
high five-year relative survival rate condition included cancers with a more than
80% five-year relative survival rate. The high five-year relative survival rate
condition included communities focused on melanoma and thyroid cancer, two
types of cancer that have very high five-year relative survival rates (91.5% and
96.7% respectively across all stages) (American Cancer Society, 2007). The low
five-year relative survival rate condition included communities devoted to
pancreatic and lung cancer, two cancers with very low five-year relative survival
rates (5.0% and 15.0% respectively across all stages) (American Cancer Society,
2007)

This investigation utilized a stratified sample of twelve weeks of posts from
each of the eight online support communities. In order to control for seasonality,

one week (considered to be a seven day period starting on Sunday) was

63

randomly selected from each month between July 2005 and July 2006. Weeks
that were partially included in two different months were considered to fall in the
month in which the initial Sunday occurred. It is important to control for
seasonality as particular times of the year (particularly around holidays), may not
necessarily be representative of the types of interactions that occur throughout
the year.

Once the weeks were identified, all eight communities were evaluated
according to the same randomly selected time periods. In the event that it wasn’t
possible to stratify the sample due to small sample size, a census of message
postings from the corresponding one year time period was used. The final
sample included 3,717 posts across all eight communities. For a complete listing
of the number of postings located in each community, please refer to Table 1.
The content analyzed in this study consisted of individually posted messages
located in the eight different online communities and the unit of analysis was the
individual messages. All archived messages from our sampling period were

included in this investigation.

Data Collection

In order to content analyze the postings in this study, a coding protocol
based on the independent and dependent variables under investigation was
developed (Appendix B). This coding protocol clearly defined each variable, as

well as the coding procedures.

64

Independent Variables

For the purposes of this study, seven distinct independent variables were
identified: gender, community member role, five-year relative survival rate,
community host, community openness, community post volume, and community
Size. The first independent variable, community member gender, was coded as a
nominal level variable based on sex; male and female. In instances where
gender could not be identified, community members were coded as having an
unidentified gender. The gender of community members was identified using
contextual clues located within postings.

The second independent variable under investigation in this study was
community member role. Roles were coded as a nominal level variable and
community members were categorized as patient (person suffering from
disease), spouse/partner, child of patient, parent of patient, other family member
of patient, medical professional, volunteer/activist, friends, other, or unclassified.
Community member roles were identified using contextual clues located within
postings. Because there were so few community members who were considered
to be medical professionals, volunteer/activists, or friends of patients, these
categories were collapsed into the “other” category.

Our third independent variable, five-year relative survival rate, was
determined according to the statistics published yearly by the American Cancer
Society. For our purposes, five-year relative survival rate was coded with values
of low and high. The high five-year relative survival rate cancer communities of

interest in this investigation were focused on melanoma (91.5% five-year relative

65

survival rate across all stages) and thyroid cancer (96.7% five-year relative
survival rate across all stages), and the low five-year relative survival rate cancer
communities of interest are focused on pancreatic cancer (5.0% five-year relative
survival rate across all stages) and lung cancer (15.0% five-year relative survival
rate across all stages) (American Cancer Society, 2007). Because this variable
was determined prior to the start of this study, coders did not need to make a
determination regarding high vs. low five-year relative survival rates.

The fourth independent variable, community host type, was coded as a
nominal level variable with two values; general purpose and cancer-specific. AS
Yahoo! is a general purpose portal and has hundreds of thousands of online
communities spanning a wide variety of topics, all posts located in a
YahoolGroups community were coded as being hosted by a general purpose
host. Due to the fact that ACOR was developed as a cancer-specific domain and
all communities located within the site are cancer-specific, all posts located in an
ACOR community were coded as being hosted by a cancer-specific host.
Because this variable was determined prior to the start of this study, coders did
not need to make a determination regarding community host type.

Openness, the fifth independent variable in this investigation, describing
whether or not a community was public (meaning anyone had access to the
archives) or private (meaning lntemet users have to register with the community
before gaining access to the archives) was coded as a nominal level variable.

Because this variable was determined prior to the start of the study, coders did

66

not need to make a determination regarding whether a community was
considered to be public or private.

The sixth independent variable in this investigation, community size, was
based on the number of registered members that participated during the sample
period. Community size was coded on a nominal level after the coding of the
sample was finished. Those communities that had a large community size had
more than 100 active participants during the sample period, while those
communities that had a small community size had less than 65 active
participants during the sample period.

Finally, community post volume served as the seventh independent
variable in this investigation. Community post volume described the number of
posted messages during the sample period and was coded on a nominal level.
Communities were either coded as having a low post volume or a high post
volume. Those communities that had a low post volume had less than 300
postings included in the sample, while the communities that had a high post
volume had more than 500 postings included in the sample. Because this
variable was determined prior to the start of the study, coders did not need to
make a determination regarding whether a community was considered have a

high or low post volume.

Dependent Variable
In regards to our dependent variable, social support provisions, it is widely
accepted that there are three primary types of support that one can give/receive

including emotional, informational, and instrumental (direct assistance) support.

67

For the purposes of this investigation, support provision type was coded on a
nominal level according to the presence or absence of each type of support
provision and the different support subtypes. The different subtypes of
informational and emotional support were determined through thematic content
analysis. Each of the social support subtypes fit into the larger Optimal Matching
Theory as either being action-facilitation (known from here on as informational
support) or nurturant (known from here on as emotional support). It should be
noted that the emotional support subtypes developed through thematic content
analysis were guided by previous work by Klemm, Reppert, and Visich (1998)
and White & Dorman (2000). For a complete listing of the different subtypes of
emotional and informational support and their operationalizations, please see
Appendix B.

Regarding the notion of instrumental or tangible aid support, because of
the limits placed on human interaction within an electronic environment, the only
way to provide a person with instrumental support is to provide information.
Therefore, instrumental support in its purest form is not possible in an electronic
environment. To account for this, coders were trained to identify interactions
where instances of instrumental support were discussed between two community

members.

Reliability
Because this investigation utilized a content analysis methodology, the
establishment of intercoder reliability was essential. Intercoder reliability refers to

“the extent to which independent coders evaluate a characteristic of a message

68

or artifact and reach the same conclusion” (Lombard, Snyder-Duch, & Bracken,
2002, p. 589). In order to establish intercoder reliability, two investigators were
trained on the coding protocol. After extensive training and practice, each coder
independently analyzed a subsample consisting of 186 messages (5% of the
total sample). Previous work has indicated that 5% of the total sample is a
suitable sample Size for establishing intercoder reliability (Kaid & Wadsworth,
1989). The messages included in this subsample were randomly selected using
a random number generator. To be sure that an appropriate level of intercoder
reliability had been achieved, it was predetermined that all variables needed to
receive a Cohen’s Kappa agreement of at least .85 or higher in order for coding
to commence. Cohen’s Kappa is a conservative statistical index that accounts for
chance agreement between two coders (Lombard, Snyder-Duch, & Bracken,
2002). The .85 level was deemed to be acceptable because not only is Cohen’s
Kappa a conservative statistic, but a recent review of “rules of thumb” for
deciding appropriate levels of reliability determined that coefficients of “.90 or
greater would be acceptable to all, .80 or greater would be acceptable in most
situations, and below that, there exists great disagreement” (Neuendorf, 2002). In
addition, it has been suggested that most research reports reliability that is .80 or

higher (Riffe et al., 2005).

Data Analyses
Once the data had been collected, descriptive statistics were employed to
describe the frequencies of different types of social support activity according to

community member gender, role, five-year relative survival rate, community host,

69

community openness, community post volume, and community size. Chi-square
analyses were used to make comparisons within and between independent
variables in regards to social support type frequencies. Although social support
subtypes were reported, in an effort to make the data more meaningful, social
support subtypes were collapsed into the overarching categories of emotional
and informational support for data analysis purposes.

Now that the methodology of this study has been presented, in chapter
four, the results of this investigation are presented. To help structure the
presentation, results are presented according to the research question in which

they are related.

70

Chapter 4: Results

The primary objective of this investigation was to investigate online
support communities for cancers with varying five-year relative survival rates. In
addition to understanding who is utilizing these communities in terms of gender
and community member role, this study also sought to understand the different
types of social support provisions that are occurring within these online spaces.
The purpose of this chapter is to present the findings of this investigation. First,
an overall description of the posting activity within the communities is presented
followed by the primary results as structured by research questions and
hypotheses. A discussion of the results presented here will be presented in

chapter five.

Posting Activity

In total, there were 3,717 posts included in the sample period, written by
587 unique community members. Posting activity of community members varied
greatly among the 587 unique community members that posted during the
sample period. Number of posts by unique posters ranged from 1 to 245 posts
(M=6.3 posts per member, Median = 2.0 posts per member, SD = 14.52). For the
complete breakdown Of number of posts and unique posters per community, see
Table 1.

Overall, 2,255 messages (60%) in the total sample contained at least one
instance of social support provision. The remaining 40% of messages contained
instances of emotional and/or informational support seeking, messages that were

related to cancer, but did not contain social support, or off topic posts not related

71

to cancer. Of those messages that contained social support provisions, the
number of social support subtype themes ranged from one to six different themes
per message (M=1.5, SD=.85). Data analysis revealed significant differences in
the number of messages providing social support across the eight communities,
)(2 (7, n = 3,717) = 66.41, p<.05, and a one way analysis of variance revealed that
there were significant differences in the average number of themes for messages
that included social support across the eight communities, F(7, 2,247) = 7.46, p
<.05. This means that communities differed in the number of posts containing
social support provisions, as well as the average number of support themes
within messages.

At least one type of social support provision occurred in approximately half
or more of the postings within each community. The ACOR melanoma
community (high five-year survival rate condition) had the lowest percentage of
messages containing at least one social support provision (49%) and the
YahoolGroups pancreatic cancer community (low five-year survival rate
condition) had the highest percentage (72%). Table 1 shows a complete
breakdown of the percentage of messages within each community that contained

at least one instance of social support provision.

72

Table 1: Community Descriptors

 

 

 

 

 

 

 

Average
Number of
Social Support
Total % of Posts Provision
Number of Total Number of Containing Themes per
Posts Unique Posters Social Support Message per
(n=3,717) (n=587) Provisions Community
Yahoo: Melanoma 122 45 62% 1.53
Yahoo: Thyroid Cancer 944 53 51% 1.45
Yahoo: Pancreatic Cancer 272 48 72% 1.79
Yahoo: Lung Cancer 894 135 65% 1.56
ACOR: Melanoma 81 27 49% 1.58
ACOR: Thyroid Cancer 286 64 60% 1.33
ACOR: Pancreatic Cancer 523 101 61 % 1 .64
ACOR: Lung Cancer 595 114 65% 1.25

 

 

Regarding social support provision type, 1,311 messages (58%) included
only instances of emotional support provision, 687 messages (30%) contained
only instances of informational support provision, and 257 messages (11%)
contained instances of both emotional and informational support provision. Table
2 provides a complete breakdown of the type of support included in messages

containing social support within each community.

73

Table 2: Frequency of Social Support Type by Community
Type of Social Support Provision Present

 

 

I

 

., B ., s
«I 8 B 8 I“ 8 g 8
E (u 0 r: E (D O c:
O o o to O o o to
5A we =6 ”a 6A vs =3 03
6 f2 '6 g 8 a: 900 75 3 '5 Is 3 ‘- 2’00
2 II S T " a ‘9 2 ll 51" T °° : ‘°
”C III grr __,rr ..: LII grr __,rr
8 v l- 5 m 5 O' 5 m V l- 5 m 5 .. 5
.c b" 0'. o 0 Di 0'. g
to O o .c 0 O m
> 6 2 9 < o O 32’
>- (D < O
>- <
Informational Support Only 33% 12% 47% 17% 30% 55% 32% 52%

Emotional Support Only 50% 80% 41% 73% 45% 33% 50% 39%

Both Informational and
Emotional Support 1 7% 8% 12% 10% 25% 12% 19% 9%

 

 

Total 100% 100% 100% 100% 100% 100% 101%" 100%

* Does not add up to 100% due to rounding error.

 

 

 

 

 

 

 

 

 

As for the subtypes of support that were provided during the sample
period, Tables 3 and 4 provide a complete listing of the emotional and
informational support subtypes and their frequencies in each community. Data
indicate that in all eight communities, encouragement was the most frequently
represented emotional support subtype while welcoming and offers for contact
were the least frequently represented. Regarding informational support,
consistently across all eight communities, mediated information and medical
information were the most frequently represented informational support subtypes

while live assistance information was consistently the least frequently

represented.

74

Table 3: Emotional Support Types and their Frequency by Community

 

Emotional Support Subtypes

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

g .9 g .2
'O “" 'O ""

2 'B A 8 8’ A 2 '3 8 a
23 Bag 2 6§3 538 as Bar: 2 633 3 58‘
£11 'ié‘u‘u’é‘ié‘r 6%? 8"? I753"? 63%? 668i?
”5 Docuocooc “5 “05"05 00C
8 E V 8 v 5 V 25 8 8 V 0 "
fi 9 5 >' o < O <
>- >- < <

Prayer/Spiritual 1 7% 10% 25% 20% 9% 5% 1 7% 8%

Understanding 9% 19% 20% 14% 15% 15% 1 1% 1 1%

Sympathy 6% 16% 13% 3% 1% 14% 15%

We'COming 2% 3% 1% 3% 1% 4%

E"°°“'ageme”t 64% 61% 35% 47% 71% 77% 51% 59%

Offers for Contact

9% 1% 3% 3% 3% 2% 7% 2%

Total 101%" 100% 100% 100% 101%" 100% 101%" 99%”

* Does not add up to 100% due to rounding error.

Table 4: Informational Support Types and their Frequency by Community

Informational Support Subtypes
co .9 35 m .2 5
E 2 *5 8 8 E 2 i: ‘5 as 8
O O a) V m 0 O V a) O m
g A 3. L a 8 ‘IT 0 ’5 S A IL>"~ "T D (R. O 6‘
'5 B If 8 ‘- 8 S 2’00 6 B I5 E 8 E 2’0
Ell ”SW 0.; 3"? Ell -'h an. 3%.
'o' 5 8 o 5 'o' 8 7‘. 5 (g 5 5 8 (g 8 T! 5
O J: o c o I: c m
.5 >9 .1: «I o 8 (<32) m 8 «I O
9 >“3 o 5 < o < o 0
>- <
Mediated 40% 24% 54% 25% 21 % 38% 29% 45%
Live Assistance 8% 7% 1% 5% 6% 17% 8%
Disease/Sym ptom 1 2% 1 6% 14% 9% 31 % 1 8% 6% 8%
Medical Information 22% 42% 23% 44% 31% 28% 28% 26%
General Information 18% 12% 8% 17% 17% 10% 20% 14%
Total 100% 101 %* 100% 100% 100% 100% 100% 101%

 

 

 

 

 

 

 

 

 

 

* Does not add up to 100% due to rounding error.

75

 

 

RQ1: Community Member Role

The first research question concerned the effect of community member
role on online cancer support communities. Results indicate that the largest
group of posters during the sample period was patients (34%). In addition, there
were also large numbers of individuals who were classified as a spouse/partner
(18%) or child of a patient (18%). Overall, 17% of unique posters were
unclassifiable and the remaining community members were either parents of
patients (2%), other family members of patients (8%) or were classified as “other’
(4%). A significant relationship between community member role and community
participation was found, x2 (42, n = 587) = 181.27, p<.05, meaning that the
demographic composite of community members in terms of role varied by
community. This finding should be taken with a grain of salt as more than 10% of
the cells within this table had-fewer than five cases. Table 5 provides a complete

breakdown of the role of unique posters for each community.

76

Table 5: Frequency of Community Member Roles across Communities+

 

Communij/ Member Role

 

 

 

 

 

 

 

 

 

 

38 938 $53 9:52 2,93 338.925? .8231

fig ‘5" g" on QEEHEEE 5'};

av as :5 '05 eevé’edv v

C O = o- «a 2

D o. C O

a) 0

Yahoo: Melanoma 13% 9% 6% 3% 14% 4% 5%
Yahoo: Thyroid Cancer 3% 21% 2% 29% 5%
Yahoo: Pancreatic Cancer 10% 2% 9% 15% 16% 14%
Yahoo: Lung Cancer 26% 14% 25% 35% 7% 31% 19%
ACOR: Melanoma 6% 4% 5% 2% 14% 7% 10%
ACOR: Thyroid Cancer 13% 20% 2% 1% 7% 4% 19%
ACOR: Pancreatic Cancer 8% 11% 26% 26% 21% 20% 24%
ACOR: Lung Cancer 20% 19% 25% 18% 7% 18% 5%
Total 99%* 100% 100% 100% 99%" 100% 101%*

 

* Does not add up to 100% due to rounding error.

+ More than 10% of cells in this table have fewer than five cases per cell.

x2 (42, n = 587) = 131.272, p<.05.

In addition to making up the majority of unique posters, patients also

posted the majority of the posts containing social support provisions during the

sample period. Of the 2,255 messages containing social support, 1,067

messages (47%) were posted by patients, 409 messages (18%) were posted by
spouse/partners, 266 messages (12%) were posted by children of patients, 271

messages (12%) were posted by other family member of patients, 27 messages

(1%) were posted by parents of patients, 66 messages (3%) were posted by

individuals classified as “other”, and 149 of messages (7%) were posted by

community members whose role was unclassifiable. The relationship between

community member role and the posting of messages with social support

provision was found to be significant x2 (42, n = 2,255) = 1,255.11, p<.05,

meaning that the frequency of posting messages with social support provisions

77

 

by different groups of community member roles varied by community. Please see

Tables 5 and 6.

Table 6: Frequency of Social Support by Role across Communities

 

 

 

 

 

 

 

 

 

 

 

Community Member Role
._ +- -°-' C
a A .. p E A i A :5: 3% A
’53 .58 £§ 2% ES 113525 E9
5% EV 8:: o: 3.1: as: 6.1%,
C v C 3 v 2 v C .c Q v
3 V 8 1E § 5 E
(D 0 a g
Yahoo: Melanoma 8% 5% 1% 2% 7% .4% 3%
Yahoo: Thyroid Cancer 1% 43% 3% 59%
Yahoo: Pancreatic Cancer 5% ,3% 11% 12% 32% 32%
Yahoo: Lung Cancer 38% 16% 40% 46% 24% 11%
ACOR: Me'anoma 3% 2% 2% 1% 15% 1% 2°/o
ACOR: Thyroid Cancer 15% 11% 1% 7% 2% 33%
ACOR: Pancreatic Cancer 2% 8% 26% 26% 11% 16% 17%
ACOR: Lung Cancer 28% 16% 16% 13% 25% 3%
Total 100% 101%* 100% 100% 99%" 100% 101%*

 

* Does not add up to 100% due to rounding error.
x2 (42, n = 2,255) = 1,255.11, p<.05.

Regarding social support provision, data indicate that patients,

 

spouse/partners, children of patients, and parents of patients provided a greater
percentage of purely emotional support than purely informational support or
informational and emotional support combined. Not including those classified as
“other", family members of patients who were not spouse/partners, parents, or
children provided the greatest percentage of purely informational support (55%)
and the lowest percentage of purely emotional support (34%), while children of
patients provided the largest percentage of purely emotional support (72%) and

parents of patients provided the smallest percentage of purely informational

78

support (11%) across all communities. Data indicate that community member
role was significantly related to social support provision type, x2 (12, n = 2,255) =
233.32, p<.05. This indicates that the different types of social support provision
varied by community member role. Please see Table 7 for a complete for a

complete breakdown of social support provisions by role.

Table 7: Frequency of Social Support Type by Community Member Role

 

 

 

 

 

 

 

 

 

 

 

 

Community Member Role
8 E ‘5' >,
£8 8: ga éa EA EEE? _A
8 5! .g 8 Q 8 9: 8 ~5- R Li? B .2 R 2 8
g" m}? 811 on HP; 58m" 52
3 '- ..
8 8 33 O
Informational Support Only 46% 23% 28% 19% 11% 55% 80%
Emotional Support Only 46% 65% 57% 72% 70% 34% 15%
Both Informational and 8% 1 1% 15% 9% 19% 1 1% 5%
Emotional Support
Total 100% 99%* 100% 100% 100% 100% 100%

 

* Does not add up to 100% due to rounding error.
x2 (12, n = 2,255) = 233.32, p<.05.

R92: Five- Year Relative Survival Rate

The second research question addressed the effect of five-year relative
survival rate on online cancer support communities. In terms of gender and
participation in high and low survival rate communities, results indicate that these
variables were not significantly related, x2 (2, n = 587) = .07, p>.05, meaning that

men and women were just as likely to belong to high survival rate communities

79

as they were to low survival rate communities. Table 8 provides a complete
breakdown of unique community member gender and participation in high and

low five-year relative survival rate communities.

 

 

 

Table 8: F requencLof Gender Participation by Cancer Survival Rate
Five-Year Relative Survival
Rate
Low High
(n=398) (n=189)
Male 20% 1 9%
Female 66% 66%
Unclassified 14% 1 5%
TOtal 1 00% 1 00%

 

 

 

 

 

x (2, n = 587) = .07, p>.05.

Despite the fact that gender and participation in communities for varying
survival rate cancers was not found to be significantly related, data analysis
indicated that there was a significant relationship between gender and social
support provision according to survival rate, x2 (2, n = 2,255) = 47.19, p<.05.
While women posted a larger proportion of messages containing social support
provision than men regardless of survival rate, males posted a larger proportion
of messages with social support provisions in the high survival rate condition
(29%) than they did in the low survival rate condition (16%), whereas women
posted a larger proportion of messages with social support provisions in the low
survival rate condition (77%) than they did in the high survival rate condition
(67%). Table 9 provides a complete breakdown of gendered postings that

included social support provisions by survival rate.

80

Table 9: Frequency of Gendered Social Support by Cancer Survival Rate

 

 

 

 

 

 

Survival Rate
Low High
(n=1,483) (n=772)
Male 16% 29%
Female 77% 67%
Unclassified 7% 5%
Total 100% 101%

 

 

* Does not add up to 100% due to rounding error.
x2(2. n = 2,255) = 47.19, p<.05.

Regarding community member role and participation in high vs. low
survival rate communities, results indicate that patients made up the largest
proportion of community members in low survival rate communities (58%). In the
high survival rate condition, patients (23%), spouses/partners (22%) and children
of patients (25%) made up almost equal proportions. The relationship between
community member role and participation in high vs. low survival rate
communities was found to be significant, x2 (6, n = 587) = 104.97, p<.05,
indicating that the demographic composite of community members in terms of
role varied by the survival rate of the community. Furthermore, not only was there
a significant relationship between community member role and membership in
high and low survival rate communities, but there was also a significant
relationship between community member role and contributions of messages that
contain social support provisions within high and low survival rate communities,
x2 (6, n = 2,255) = 349.48, p<.05, meaning that the frequency in which different
community member role groups posted messages with social support varied by
the survival rate of the community. Patients by far contributed the largest
proportion of messages that contain social support provisions to the high survival

rate condition (82%), while patients and spouse/partners contributed the largest

81

proportions of postings to the low survival rate condition (29% and 26%
respectively). Please see Tables 10 and 11 for a complete breakdown of
community member roles and their participation and frequency of posting

messages with social support in high and low survival rate communities.

Table 10: Frequency of Role Participation by Cancer Survival Rate

 

 

 

 

 

 

Five-year Relative Survival
Rate
Low High
(n=398) (n=189)

Unclassified 19% 16%
Patient 58% 23%
Spouse/Partner 8% 22%
Child of Patient 3% 25%
Parent of Patient 5% 1 %
Other Family Member of Patient 4% 10%
Other 4% 3%
Total 101 %* 100%

 

 

* Does not add up to 100% due to rounding error.
x2(6, n = 587) = 104.97, p<.05.

Table 11: Frequency of Social Support by Role by Cancer Survival Rate

 

 

 

Five-year Relative
Survival Rate
Low High
(n=1,483) (n=772)

Unclassified 7% 5%
Patient 29% 82%
Spouse/Partner 26% 4%
Child of Patient 17% 1%
Parent of Patient .2% 3%
Other Family Member of Patient 18% 1%
Other 3% 3%

Total 100% 99%*

 

 

 

 

 

* Does not add up to 100% due to rounding error.
x2 (6, n = 2,255) = 707.79, p<.05.

82

RQ3: Additional Community Characteristics

The final research question concerned whether or not there were
additional community characteristics that impact online cancer support
communities. To determine whether or not there were features of communities
that were significantly related to social support provision and the demographic
makeup of the communities in terms of gender or community member role, four
features were identified and analyzed. These features included the community
host, openness of the community (whether it is public or private), volume of
postings within the community, and the size of the community in terms of the
number of registered community members. Each of these four variables are now

discussed.

Host

The relationship between gender and participation in communities with
different hosts was found to be significant, x2 (2, n = 587) = 16.34, p<.05, as was
the relationship between gender and frequency of posting messages with social
support in communities with different hosts, x2 (2, n = 2,255) = 8.04, p<.05. This
means that the gender breakdown of community members varied by community
host, as did the frequency with which differently gendered community members

posted messages with social support messages. Please see Tables 12 and 13.

83

Table 12: Frequency of Gender Participation by Community Host

 

 

 

 

 

 

Community Host
Yahoo! ACOR
(n=281) (n=306)
Male 17% 22%
Female 63% 69%
Unclassified 20% 9%
Total 100% 100%

 

 

xz(2, n = 587) = 16.34, p<.05.

Table 13: Frequency of Gendered Social Support by Community Host

 

 

 

 

 

 

Community Host
Yahoo! ACOR
(n=1340) (n=915)
Male 19% 23%
Female 74% 73%
Unclassified 7% 5%
Total 100% 101%"

 

 

* Does not add up to 100% due to rounding error.
x2 (2. n = 2,255) = 8.04, p<.05.

Regarding community member role and participation in communities with
different hosts, no significant relationship was found, x2 (6, n = 587) = 3.98,
p>.05, indicating that community members with different roles were no more
likely to belong to Yahoo! than they were to ACOR. Despite this fact, a significant
relationship was found between community member role and frequency of
posting messages with social support provisions in communities with different
hosts, x2 (6, n = 2,255) = 21.69, p<.05, meaning that the frequency that different
community member role groups posted messages with social support varied by

community host. Please see Tables 14 and 15.

84

Table 14: Frequency of Role Participation b Community Host

 

 

 

 

 

 

 

Community Host
Yahoo! ACOR
(n=281) (n=306)
Unclassified 18% l 15%
Patient 33% 36%
Spouse/Partner 1 5% 20%
Child of Patient 20% . 16%
Parent of Patient 3% 2%
Other Family l
Member of 8% l 7%
Patient
Other 3% 4%
Total 1 00% 1 00%

 

x (6, n = 587) = 3.98, p>.05.

Table 15: Frequency of Social Support by Role by Community Host

 

 

 

 

 

Community Host
Yahoo! ACOR

(n=1340) E91 5)
Unclassified 6% 8%
Patient 51 % 43%
Spouse/Partner 1 7% 20%
Child of Patient 12% 12%
Parent of Patient 1% 1 %
Other Family
Member of 1 1% 13%
Patient
Other 2% 4%
Total 100% 101%*

 

 

 

* Does not add up to 100% due to rounding error.
x2 (6, n = 2,255) = 21.69, p<.05.

Finally, results indicate that a greater proportion of messages containing
only emotional support were posted in Yahoo!, while a greater proportion of
messages containing only informational support were posted in ACOR. The
relationship between community host and the frequency of postings with strictly
emotional support provision, strictly informational support provision, and postings

with both types of support provision was found to be significant, x2 (2, n = 2,255)

85

= 178.25, p<.05, meaning that the types of social support messages in online

cancer communities varied by community host. Please see Table 16.

Table 16: Frequency of Social Support Type by Community Host

 

 

 

 

 

 

 

Community Host
Yahoo! ACOR
(n=1,340) (n=915)
Informational Support Only 21% 45%
Emotional Support Only 69% 42%
Both Informational and 10% 13%
Emotional Support
Total 100% 100%

 

x2(2. n = 2,255) = 178.25, p<.05.

Openness

Data analysis indicated that although males made up an equal proportion
of community members in public and private communities, a larger proportion of
females belonged to private communities. The relationship between gender and
participation in communities that were public vs. private was found to be
significant, x2 (2, n = 587) = 17.46, p<.05, meaning that the participation of males
and females in online support communities for cancer varied by whether or not
the community was public or private. Results also indicate that females posted a
greater proportion of messages with support provisions in private communities
(76%) than they did in public communities (53%), while males posted a greater
proportion of messages with support provisions in public communities (31 %) than
they did in private communities (19%). The relationship between gender and the

posting of messages containing instances of social support provision in public or

86

private communities was also found to be significant, x2 (2, n = 2,255) = 83.18,

p<.05. Please see Tables 17 and 18.

Table 17: Frequency of Gender Participation by Community Openness

 

 

 

 

 

 

 

 

Openness _ _
Public Private
(n=93) (n=494)
Male 19% 19%
Female 53% 69%
Unclassified 28% 1 2%
Total 1000/0 100%

 

x2(2, n = 587) = 17.46, p<.05.

Table 18: Frequency of Gendered Social Support by Community Openness

 

 

 

 

 

Openness
Public Private
(n=272) (n=1983)
Male 31 °/o 19%
Female 53% 76%
Unclassified 16% 5%
Total 100% 100%

 

 

 

 

 

x2 (2, n = 2,255) = 83.18, p<.05.

Regarding community member role and participation in communities that
were public vs. private, no significant relationship was found, x2 (6, n = 587) =
10.13, p>.05. This indicates that members with different community roles were
just as like to belong to public communities as they were to private communities.
Despite the fact that the relationship between community member role and
participation in private vs. public communities was not found to be significant, a
significant relationship between community member role and the posting of
messages that contain social support provisions between public and private

communities was found, x2 (6, n = 2,255) = 189.23, p<.05, indicating the posting

87

of messages with social support by different genders varied by the openness of a

community. Please see Tables 19 and 20.

Table 19: Frequency of Role Participation b Community Openness

 

 

 

Openness

Public Private

(n=93) (n=494)
Unclassified 25% 1 5%
Patient 23% 37%
Spouse/Partner 16% 18%
Child of Patient 20% 17%
Parent of Patient 2% 2%
Oth r F mil
Meribe? of Patient 10% 7%
Other 4V0 3%)
Total 100% 99%*

 

 

 

 

 

* Does not add up to 100% due to rounding error.
x2 (6, n = 587) = 10.13, p>.05.

Table 20: Frequency of Social Support by Role by Community Openness

 

 

 

 

 

 

Openness
Public Private
(n=272) (n=1983)

Unclassified 7% 7%
Patient 19% 51 %
Spouse/Partner 18% 18%
Child of Patient 14% 1 1%
Parent of Patient 1 % 1 %
Other Famil

Member of Patient 32% 9%
Other 9% 2%
Total 100% 99%*

 

 

* Does not add up to 100% due to rounding error.
x2 (6. n = 2,255) = 189.23, p<.05.

Finally, data indicate that a larger proportion of messages containing
purely emotional support provisions were posted in private communities (60%)
than in public communities (43%), while a larger proportion of messages

containing purely information support were posted in public communities (43%)

88

than in private communities (29%). The relationship between openness and
social support provision type was found to be significant, x2 (2, n = 2,255) =
29.44, p<.05, meaning that the types of social support messages in online cancer

communities varied by community openness. Please see Table 21.

Table 21: Frequency of Social Support Type by Community Openness

 

 

_ _ Openness
Public Private
(n=272) (n=1983)

 

Informational Support Only

Emotional Support Only

Both Informational and
Emotional Support

 

Total

43%

43%

1 3%

 

99%*

 

29%

60%

11%

1 00%

 

 

* Does not add up to 100% due to rounding error.
x2 (2. n = 2,255) = 29.44, p<.05.

Post Volume

In terms of post volume, a significant relationship was found between
gender and participation in communities with varying degrees of post volume, x2
(2, n = 587) = 7.35, p<.05, meaning that the demographic composite of
community members in terms of gender varied by the post volume of a
community. In addition, a greater proportion of females were found to have
posted social support provisions in communities that had a high post volume,
whereas males posted a greater proportion of messages with support provisions
in communities with low post volumes. The differences in gender and the posting
of support provisions by community post volume was found to be significant, )(2

(2, n = 2,255) = 43.53, p<.05. Please see Tables 22 and 23.

89

Table 22: Frequency of Gender Participation by Community Post Volume

 

 

 

 

 

 

 

Post Volume____
Low High
(n=184) (n=403)
Males 18% 20%
Females 62% 68%
Unclassified 20% 12%
Total 100% 1000/0

 

x"’(2, n = 587) = 7.35, p<.05.

Table 23: Frequency of Gendered Social Support by Post Volume

 

 

 

 

 

 

 

Post Volume,—
Low High
(n=484) (n=1,771)
Males 27% 19%
Females 62% 76%
Unclassified 1 1% 5%
Total 100% 100%

 

x2 (2, n = 2,255) = 43.53, p<.05.

In addition to the significant relationships between gender and community
post volume, significant relationships were also found between community
member role and participation in communities with varying degrees of post
volume, )(2 (6, n = 587) = 19.28, p<.05, and community member role and the
frequency of posting messages that contain social support provisions in
communities with varying post volume, x2 (6, n = 2,255) = 159.95, p<.05. This
means that the demographic composite of community members in terms of role
varied by community post volume and that the frequency of different community
member role groups posting messages with social support also varied

community post volume. Please see Tables 24 and 25.

90

Table 24: Frequency of Role Participation by Community Post Volume

 

 

 

 

 

 

Post Volume
Low High
(n=184L (n=403)

Unclassified 23% 14%
Patient 38% 33%
Spouse/Partner 1 2% 20%
Child of Patient 12% 20%
Parent of Patient 3% 2%
Other Family
Member of 8% 8%
Patient
Other 5% 3%
Total 101 %* 100%

 

 

* Does not add up to 100% due to rounding error.
x2(6, n = 587) = 19.28, p<.05.

Table 25: Frequency of Social Support by Role by Community Post Volume

 

 

 

Post Volume
High
Low (n=1 ,771
(n=484) )

Unclassified 10% 6%
Patient 38% 50%
Spouse/Partner 1 3% 20%
Child of Patient 8% 13%
Parent of Patient 2% 1%
Other Family
Member of 20% 10%
Patient
Other 10% 1 %
Total 101%* 101%*

 

 

 

 

 

* Does not add up to 100% due to rounding error.
x2(6, n = 2,255) = 159.95, p<.05.

Finally, data indicate there was a greater proportion of messages
containing purely emotional support provisions in the high post volume condition
(63%) than there was in the low post volume condition (40%). Conversely, there
was a greater proportion of messages containing purely informational support

provisions in the low post volume condition (46%) than there was in the high post

91

volume condition (26%). Results indicate that the differences between post
volume and social support provision type are significant, x2 (2, n = 2,255) =
89.88, p<.05. This means that the types of social support messages in online

cancer communities varied by community post volume. Please see Table 26.

Table 26: Frequency of Social Support Type by Community Post Volume

 

 

 

Post Volume
Low High
(n=484) (n=1,771)
Informational Support Only 46% 26%
Emotional Support Only 40% 63%

Both Informational and

0 o
Emotional Support 14 /° 11 /o

 

 

 

 

Total 100% 100%
x2(2. n = 2,255) = 89.88, p<.05.

 

Commnity §Lz_e_

Data indicate that no there was no significant relationship between gender
and participation in communities of different sizes, x2 (2, n = 587) = 4.00, p>.05.
This suggests that males and females were just as likely to belong to large
communities as they were to small communities. Despite the lack of relationship
between gender and membership in communities of different sizes, the
relationship between gender and the frequency of posting messages that contain
social support provisions in communities of different sizes was found to be
significant, x2 (2, n = 2,255) = 58.95, p<.05. Please see Tables 27 and 28.
Overall, males posted a larger proportion of messages with support provisions in

small communities (28%) than they did in large communities (15%), whereas

92

females posted a larger proportion of messages with support provisions in large

communities (79%) than they did in small communities (65%).

Table 27: Frequency of Gender Participation by Community Size

 

 

 

 

ngmunity Size
Small Large
(n=237) (n=350L
Males 18% 21%
Females 650/0 670/0
Unclassified 18% 12%
Total 100% 100%

 

 

 

 

x2 (2, n = 587) = 4.00, p>.05.

Table 28: Frequency of Gendered Social Support by Community Size

 

 

 

Community Size
Small Large
(n=463) (n=1650)
Males 28% 1 5%
Females 650/0 790/0
Unclassified 7% 6%
Total 100% 100%

 

 

 

 

 

x2(2, n = 2,255) = 58.95, p<.05.

In addition to the significant relationships between gender and
participation and posting of support provisions in communities of different sizes,
significant relationships were also found between community member role and
participation in communities of different sizes, x2 (6, n = 587) = 54.93, p<.05, as
well as between community member role and the frequency of posting messages
that contain social support provisions in communities of different sizes, x2 (6, n =
2,255) = 371.10, p<.05. This indicates that the demographic composite of
community members in terms of role varied by community size as did the
frequency of posting social support messages by different community member

roles. Please see Tables 29 and 30.

93

Table 29: Frequency of Role Participation b Community Size

 

 

 

 

Community Size
Small Large
(n=237) (n=350)
Unclassified 19% 1 5%
Patient 47% 26%
Spouse/Partner 10% ‘ 23%
Child of Patient 9% 23%
Parent of Patient 4% 1%
h r F mil

Enter: be? of Platient 6% 9%
Other 5% 3%
Total 1 00% 100%

 

 

 

 

 

xz(6, n = 587) = 54.93, p<.05.

Table 30: Frequency of Social Supportgy Role by Community Slze

 

 

 

 

Community Size
Small Large
(n=149) (n=1,067)
Unclassified 5% 8%
Patient 66% 33%
Spouse/Partner 7% 26%
Child of Patient 4% 17%
Parent of Patient 2% .2%
h r F mil

filter: be:3 of Patient 10% 14%
Other 5% 2%
Total 1 00 % 1 00%

 

 

 

 

fie, n = 2,255) = 371.10, p<.05.

Finally, data indicate that there was no significant relationship between
community size and the frequency of postings with strictly emotional support
provision, strictly informational support provision, and postings with both types of
support provision, x2 (2, n = 2,255) = 1.98, p>.05. This indicates that each type of
social support provision was just as likely to appear in small vs. large

communities. Please see Table 31.

94

Table 31: Frequency of Social Support Type by Community Size

 

 

 

 

 

 

CommunitySize
Small Large
(n=968) Q1=1,287)
Informational Support Only 29% 31 %
Emotional Support Only 60% 57%
Both Informational and o o
Emotional Support 11 /° 12 /°
Total 100% 100%

 

 

x2(2, n = 2,255) = 1.98, p>.05.

H1 : Gender in Online Support Communities

The first hypothesis in this investigation was a multipart hypothesis dealing
with the gender of participants in online support communities. The first part of this
hypothesis stated:

. H1a: There will be a larger proportion of women in the online

support communities than men.

During the sample period, as mentioned previously, there were 587
unique community members who posted messages in the eight different
communities. The majority of each community was made up of females. Of the
587 unique community members, 389 were female (66%), 114 were male (19%),
and 84 were unclassified (14%), thus supporting hypothesis 1a. Table 32
provides a complete breakdown of the gender of unique posters by community.
Significant differences were found between gender and community membership
indicating that these two variables are significantly related, x2 (14, n = 587) = 40.

65, p<.05.

95

Table 32: Frequency of Gender across Communities

 

 

 

 

 

 

 

 

 

 

 

 

Gender
Male Female Unclassified

(n=1 14) (n=389) (n=84)
Yahoo: Melanoma 11% 5% 14%
Yahoo: Thyroid Cancer 8% 10% 6%
Yahoo: Pancreatic 0 o 0
Cancer 5 /o 7 /o 17 /0
Yahoo: Lung Cancer 18% 23% 31 %
ACOR: Melanoma 6% 4% 5%
ACOR: Thyroid Cancer 7% 13% 8%
ACOR: Pancreatic o 0
Cancer 26% 17 /o 6 /o
ACOR: Lung Cancer 19% 21% 13%
Total 100% 100% 100%

 

 

 

 

 

x2 (14, n = 587) = 40.55, p<.05.

The second part of this multifaceted hypothesis stated that:

- H1b: Women will post more messages than men.

Results indicate that in addition to making up the majority of unique
posters, females also posted the majority of the posts during the sample period,
as well as the majority of messages containing social support. Out of the total
sample, 2,689 messages (72%) were posted by females, 799 messages (21%)
were posted by males, and 229 (6%) were posted by community members
whose gender was unclassified, thus supporting hypothesis 1 b. Significant
differences were found in the number of messages containing social support
provisions posted by males, females, and individuals whose gender was
unclassified across the different communities, x2 (14, n = 2,255) = 218.08, p<.05.
Table 33 provides a complete breakdown of the total number of messages that

included social support provisions posted in each community by gender.

96

Table 33: Frequency of Gendered Social Support by Community

 

Gender

 

Male
(n=463)

Female

(n=1 ,650)

Unclassified
(n=142)

 

Yahoo: Melanoma
Yahoo: Thyroid Cancer

Yahoo: Pancreatic
Cancer
Yahoo: Lung Cancer

ACOR: Melanoma
ACOR: Thyroid Cancer

ACOR: Pancreatic
Cancer
ACOR: Lung Cancer

Total

8%
30%
1 1%

7%

3%

8%
1 5%

19%
101%*

2%
20%
7%
31 %
2%
8%
15%
16%
101%*

10%
11%

21 %
27%
1 %
4%
3%
23%
100%

 

 

 

 

 

 

* Does not add up to 100% due to rounding error.
x2 (14, n = 2,255) = 218.08, p<.05.

The third and fourth parts of this hypothesis stated:

0 H1 0: Women will provide a greater proportion of emotional

support than men.

. H1 d: Men will provide a greater proportion of informational

support than women.

Regarding social support provision, data indicate that both males and
females provided a greater percentage of purely emotional support (51% and
61% respectively) than purely informational support (35% and 28% respectively)
or informational and emotional support combined (14% and 11% respectively).
Those whose gender was unidentified provided a greater percentage of purely
informational support (48%) than purely emotional support (43%) or combined
informational and emotional support (9%). Please see Table 34 for a complete
for a complete breakdown of social support provisions by gender.

Despite the fact that emotional support provisions were given more

frequently by males and females, females provided a greater percentage of

97

emotional support (61 %) than did males (51%) or the unclassified group (43%).
The group providing the largest percentage of purely informational support were
those people whose gender was unclassified (48%) as compared to males (35%)
and females (28%). When looking only at males and females, males tended to
provide a greater proportion of informational support than females and females
tended to provide a greater proportion of emotional support than males. Results
indicate that these differences were significant, x2 (4, n = 2,255) = 37.32, p<.05,

thus supporting hypotheses 1c and 1d.

Table 34: Frequency of Social Support Type by Gender

 

 

 

Gender
Male Female Unclassified
(n=463) (n=1,650) (n=142)
Informational Support Only 35% 28% 48%
Emotional Support Only 51 % 61 % 43%
Both Informational and 14% 1 1% 9%
Emotional Support
Total 1 00% 1 00% 1 00%

 

 

 

 

 

 

x2 (4. n = 2,255) = 37.32, p<.05.

H2 and H3: Social Support Provision by Five- Year Survival Rate

Based on Optimal Matching Theory, the final two hypotheses in this
investigation dealt with social support provision types in communities for varying
five-year survival rates. These hypotheses stated:

. H2: There will be a greater proportion of emotional support provision in

communities for cancers with low five-year relative survival rates as

98

opposed to communities for cancers with high five-year relative

survival rates.

. H3: There will be a greater proportion of informational support
provision in communities for cancers with high five-year relative
survival rates as opposed to communities for cancers with low

five-year relative survival rates.

Results indicate that the majority of messages that included social support
provisions in this sample were from communities for low-five-year relative
survival rate cancers (66%) as opposed to the high five-year relative survival rate
cancers (34%). Results show that both high and low survival rate communities
had a greater percentage of emotional support as opposed to informational
support or combinations of emotional and informational support. Despite this fact,
counter to hypotheses two and three, the low five-year relative survival rate
communities were found to have a larger percentage of purely informational
support (33%) than the high five-year relative survival rate communities (25%)
and the high five-year relative survival rate communities had a larger percentage
of purely emotional support (65%) than the low five-year relative survival rate
communities (54%). Please see Table 35.

Data indicate the differences in the five-year relative survival rate and the
frequency of postings with strictly emotional support provision, strictly

informational support provision, and postings with both types of support provision

99

were significant, x2 (2, n = 2,255) = 21.78, p<.05, therefore H2 and H3 were not

supponed.

Table 35: Frequency of Social Support Type bi Survival Rate

Five-Year Relative
Survival Rate

 

 

 

 

 

 

 

Low High

(n=1483) (n=772)
Informational Support
Only 33% 25%
Emotional Support 55% 65%
Only
Both Informational and 12% 11%
Emotional Support
Total 99%* 101%*

 

* Does not add up to 100% due to rounding error.
x2 (2. n = 2,255) = 21.78, p<.05.

Sir—mm

In summary, results from this investigation indicate that social support
provisions are common as they were present in 60% of the posted messages
during this sample period. Regarding our research questions, community
member role, five-year survival rate, and community characteristics such as
community host, openness of the community, community post volume, and
community post size were, for the most part, all found to have interesting
relationships with participation in communities, the posting of messages with
social support provisions, and the types of social support provided within online
support communities. In addition, although support was found for the first
hypothesis as females were found to make up the largest demographic of online

community members, post the most messages, and provide a larger proportion

100

of emotional support than males, while males post a larger proportion of
informational support than females, no support was found for hypotheses two
and three that stated there would be a larger proportion of emotional support
provisions in low survival rate communities and a larger proportion of
informational support provisions in high survival rate communities. Next, chapter

five presents a more in-depth discussion of the pertinent research findings.

101

Chapter 5: Discussion

At the outset, this investigation sought to investigate online support
communities for cancers with high and low five-year relative survival rates. In
addition to understanding who is utilizing these communities in terms of gender
and community member role, this study also sought to understand the different
types of social support provisions that are occurring within these online spaces.
The purpose of this final chapter is to more fully explain the major findings from
this investigation as outlined in the previous chapter.

Overall, findings from this investigation indicate that social support
provisions are abundant within online support communities for cancers with high
and low five-year survival rates. While females make up the largest demographic
in these online communities, males actively participate as well. Results indicate
that the first hypothesis regarding gender in online communities was supported
as females make up the largest demographic of online community members,
post the most messages, and provide a larger proportion of emotional support
than males, while males post a larger proportion of informational support than
females. Unfortunately, no support was found for hypotheses two and three that
stated there would be a larger proportion of emotional support provisions in low
survival rate communities and a larger proportion of informational support
provisions in high survival rate communities.

This investigation also found that community member role was
significantly related to social support provision as patients and immediate family

members tended to provide more emotional support, while other family members

102

and individuals classified as “other” provide more informational support.
Furthermore, significant relationships between role and the participation in high
and low five-year relative survival rates were found as well between role and the
posting of messages with social support provisions in high and low five-year
survival rate communities.

In terms of characteristics of communities that impact online support
groups, this investigation found that the public or private nature can have
significant effects on the types of social support provision within online support
communities as emotional support seems to be provided in larger proportion in
private communities and informational support is presented in larger proportions
In public communities. It was also found that there were very real differences
between social support provision types for the two community hosts. Finally the
post volume of a community was found to impact the gender and social support
provisions of online community members and the size of the community was
found to impact gendered participation.

In this chapter, these pertinent findings are highlighted along with a
discussion regarding possible explanations for the results, as well as an
assessment of the importance of the findings and their implications. Following the
discussion of the research findings, the limitations of this study are presented

along with a discussion of future directions for research.

Social Support Provision wiLhin inine Support Communities

Results from this investigation indicate that social support provisions are

quite common within online support communities for cancers with high and low

103

 

five-year relative survival rates. Data indicate that overall, social support
provisions were present in 60% of the sample. When looking at all eight
communities individually, it was found that the percentage of messages that
contained social support provision within a particular community during the
sample period ranged from 49% - 72%. Regarding the type of social support
provision, emotional support was by far the most prevalent as 58% of the
messages that contained social support provision only contained instances of
emotional support and 30% of messages with social support provisions
contained only informational support provisions. Interestingly enough, only 11%
of messages with social support provisions contained instances of both
informational and emotional support.

The fact that social support provisions are so prevalent within online
support community messages is unsurprising as this is presumably why many
people turn to online support groups. People who are not interested in receiving
and/or providing social support would not necessarily belong to an online support
group in the first place. This finding indicates that online social support groups
are efficaciously performing one of their fundamental objectives: to provide social
support.

The finding that social support provisions are prevalent within online
support communities is important since it indicates that online support
communities are fulfilling their purpose. These online communities are places
where people go to provide and receive social support. This has great

implications for the marketing of online social support communities for cancer

104

from both the online community and the healthcare provider perspectives.
Finding that online support communities really do provide social support is of
great benefit to the online communities themselves as it may be a springboard
for marketing online services to the general public and the healthcare community.
In addition, if healthcare providers are confident that online support communities
actually do provide necessary emotional and informational support to community
members, they may be more prone to recommending that individuals utilize
these groups for social support. Particularly for patients who may have small
support networks or may live in locations where offline support groups are not
available, there are many reasons why a healthcare provider may recommend
that patients join an online support group, even if it is just to lurk and read

archived postings.

Gender and Online Support Communities

At the outset of this investigation, a multipart hypothesis was created
stating that based on the literature concerning social support and gender, there
would be a larger proportion of women in the online support communities than
men, women would post more messages than men, women would provide a
greater proportion of emotional support than men, and that men would provide a
greater proportion of informational support than women. Data provided support
for all four parts of this multifaceted hypothesis.

Based on previous research suggesting women are more likely to join
traditional offline support groups (Krizek et al., 1999) and that current members

of traditional support groups are more likely to be female (Grande et al., 2006;

105

Moos et al., 2006), it was expected that women would participate in online
support groups more frequently than men. Of the 587 unique community
members who posted messages in the eight online support communities during
the sample period, 66% were women, 19% were men, and 14% were
unclassified. As those community members identified as women made up two-
thirds of the unique community members included in this investigation, our
expectation that the majority of community members would be female was met.
Clearly, like in traditional offline support groups, men are less likely to participate
in online support groups. While we are unable to draw conclusions as to why
men participated in online support communities less frequently than women, it is
possible that as previous research suggests, they do rely more on their spouses
for support and less on sources outside the home.

In addition to the fact that women currently make up the largest
percentage of online support community members, the multipart hypothesis
concerning gender also stated the expectation that women would also post the
majority of messages within online support groups. This expectation was met as
out of the 3,717 messages included in this sample, 72% were written by females,
21% were written by males, and only 6% were written by individuals whose
gender was unclassified. This finding supports the conclusion that women
provide social support more often than men as indicated by previous research
(Belle, 1982; Fischer, 1982; Griffith, 1985; Kendler et al, 2005; Kessler et al.,
1985; Trobst et al., 1994). If the members of an online support community are

viewed as members of a social network, this finding also supports the conclusion

106

 

that women are more involved in their social networks (Kendler et al., 2005;
Rosenthal et al., 1986).

The final two parts in our multipart hypothesis relating to gender indicated
that it was expected that women would provide a greater proportion of emotional
support than men, whereas men would provide a greater proportion of
informational support than women. These expectations were based on previous
research indicating that women are more likely than men to provide emotional
support and men are more likely than women to provide informational support
(Klemm et al., 1999; Rosenthal et al., 1986; Seale et al., 2006; Sullivan, 2003).
Results indicate that both of these expectations were met as a larger proportion
of men (35%) provided instances of purely informational support than did women
(28%) and a larger proportion of women (61%) provided purely emotional support
than did men (51%). It did come as a surprise though that both genders provided
a greater proportion of purely emotional support than purely informational
support. These findings suggest that within online support communities,
regardless of gender, emotional support provisions are far more common.

The finding that women are more active within online support groups has
very practical implications for the design of online support communities. Although
we are unable to ascertain exactly why women participate more in both online
and offline support groups than men, it is possible that either men do not see as
much value in support groups and choose not to participate in them, or it is
possible that the overwhelming presence of women may discourage more

frequent participation from men. Additionally, it could also be that men are

107

socialized over their lives to not seek out support in social ways. Because we are
unable to see what the demographics of the Iurkers are, it is also possible that
men may actually frequent online support communities as often as women, but
they may choose not to actively participate. Overall, these findings related to
gender possibly speak to the need to solicit more males to actively participate
within online support communities. This could be done through the creation of
online support groups specifically for men. Or, it is possible that if the men who
are currently involved in online support communities become more active, more
men may be encouraged to participate. Through the encouragement of active
male participation in online support communities, other men seeking social
support may benefit due to the fact that there would be more people that they

could relate to providing social support.

Five- Year Survival Rate and Social Support Provision

Overall, emotional support was provided more frequently than
informational support in both high and low five-year survival rate communities.
The fact that emotional support was more prevalent in both high and low survival
rate communities suggests that there are more people providing emotional
support in online communities than informational support. Reasons for why there
is more emotional support rather than informational support is unclear, although it
is possible that either online support community members go elsewhere to obtain
information, online community use lends itself to be more effective for emotional
support, or perhaps there are more instances of emotional support because

when a person requests emotional support, more people respond to comply,

108

whereas when a person requests informational support, when the information is
given, there is no need for community members to continue to respond.

In this investigation, as indicated by the second and third hypotheses, it
was expected that five-year survival rate would significantly affect social support
provision. Based on Optimal Matching Theory, using five-year relative survival
rate as a proxy for controllability, it was hypothesized that there would be a
greater proportion of emotional support provisions in low five-year survival rate
communities than in high five-year survival rate communities. It was further
hypothesized that there would be a greater proportion of informational support
provisions in high five-year survival rate communities than in low five-year
survival rate communities.

Contrary to the stated hypotheses, low five-year survival rate communities
contained a greater proportion of purely informational support (33%) than the
high five-year survival rate communities (25%) and high five-year survival rate
communities contained a greater proportion of purely emotional support (65%)
than did the low five-year survival rate communities (55%). This indicates that
there was no support for hypotheses two and three, although the relationship
between social support type and five-year relative survival rate was found to be
significant, meaning that the assumption of a relationship between social support
provision and five-year relative survival rates was correct, but the direction of the
relationship was incorrect.

Possible reasons to explain why hypotheses two and three were not

supported include the idea that while Cutrona was correct in asserting a

109

relationship between controllability and social support provision type in Optimal
Matching Theory, she misunderstood the nature of the relationship. Data from
this investigation suggest that while controllability (in this case, survival rate) is
significantly related to social support provision, informational support is provided
in greater proportion in low survival rate communities than in high survival rate
communities, while emotional support is provided in greater proportion in high
survival rate communities than in low survival rate communities. It is possible that
in low survival rate situations, people experiencing low five-year survival rate
cancers are desperate for information as opposed to those people experiencing
high five-year survival rate cancers. Furthermore, in high five-year survival rate
situations, it is possible that emotional support is provided in greater proportion
than in low five-year survival rate conditions because information isn’t as
necessary or perhaps because these individuals are more hopeful for a positive
outcome.

Another possible explanation for the fact that no support was found for
hypotheses two and three relates to the idea of individual self-efficacy. It is
possible that individuals experiencing high survival rate cancers have higher
levels of self-efficacy in regards to dealing with cancer and feel less of a need for
informational support, whereas individuals experiencing low survival rate cancers
have lower levels of self-efficacy when dealing with cancer and need more
informational support.

These findings speak to the need for medical professionals to provide

copious amounts of information to patients who have been diagnosed with a low

110

survival rate cancer and their families. While emotional support is vital for any
type of cancer diagnosis, there seems to be a greater need to provide
informational support during these low survival rate situations, than in high
survival rate situations. In addition, these findings also suggest that online
support communities for low five-year survival rate cancers would benefit from
having informational resources that community members could utilize. Whether
these informational resources take the shape of FAQ pages located on the
community homepage, links to external websites, links to specific message
postings that were particularly informative, or even a collaborative informational
document such as a wiki, easily accessible informational sources could be quite
beneficial to all community members, particularly new members just joining a

community.

Community Member Role and Social Supflrt Provision Type

Results from this investigation indicate that there is a significant
relationship between community member role and social support provision type.
Data indicate that patients and their immediate family members
(spouses/partners, children of patients, and parents of patients) all provided
larger proportions of purely emotional support than they did purely informational
support. Conversely, those community members who were not immediate family
members (classified as other family members of patients or “other”) provided
larger proportions of purely informational support.

One possible explanation for the finding that patients and immediate

family members provide a greater proportion of purely emotional support as

111

opposed to purely informational support stems from the notion that these
individuals may be better able to provide emotional support due to their trials and
tribulations of dealing with cancer. As these individuals are experiencing cancer
either themselves or through an immediate family member, they may better
understand the emotional experiences involved with cancer and be better suited
to provide emotional support. Family members who are not immediately related
to patients and those people classified as “other” may not be as well equipped to
provide emotional support because they are farther removed from the cancer
experience. Therefore, individuals who are not patients or immediate family
members may be best suited to providing informational support to others.

The finding that patients and immediate family members of cancer
patients provide a greater proportion of emotional support than informational
support is important in that it indicates that there may be a need to encourage
family members not in a patients’ immediate family, as well as individuals
classified as “other’, to provide emotional support to others, particularly those in
similar situations. One of the benefits of online support communities is that
individuals can connect with others in similar situations. While patients are
providing emotional support in abundance, it could be beneficial for non-
immediate family members and those classified as “other’ to receive social
support provisions from other individuals with the same role. If non-immediate
family members and those individuals classified as “other” are not providing a

great frequency of emotional support, individuals in similar situations looking for

112

emotional support from people with similar experiences may be lacking the social

support that their counterparts could provide.

ggmmunity Member Role and Five- Year Relative Survival Rate

Another interesting finding from this study was that despite the fact that
patients make up the largest proportion of unique community members from low
five-year relative survival rate communities (58%) and do not necessarily make
up a significantly large proportion of the high survival rate communities, they by
far provide the largest proportion of messages containing social support
provisions in high five-year survival rate communities (82%).

This finding could possibly be explained by the idea that in high survival
rate communities, patients are considered to be the authority on the specific type
of cancer, thus are more qualified to provide social support as they have coped
with the illness and most likely have a good prognosis. In low survival rate
communities, it is possible that patients are not considered the authority and that
spouses, family members, and individuals with other roles may also be viewed as
being equally qualified to provide social support. It is also possible that perhaps
patients in low survival rate communities are just sicker than patients in high
survival rate communities and are thus too ill to post more often.

This finding has great implications in that it suggests that it may be
essential to develop online support communities for high survival rate cancers
that are specifically oriented to non-patients. Again, one benefit of online support
communities is that they provide people with an opportunity to connect with

others in similar situations. When patients provide the bulk of support provisions

113

in communities for cancers with high five-year survival rates, those members who
are not patients may miss out on valuable insights that other non-patients could
possibly provide. By creating online communities specifically for individuals who

are not patients, we may be able to foster more activity from this group.

Impact of Openness on Online Support Communities

Another interesting finding from this investigation is that there is a
significant relationship between the openness of a community and the posting of
messages containing social support provisions from males and females. Results
indicate that males post a larger proportion of messages with social support
provisions in public communities (31%) than in private communities (19%).
Conversely, women post a larger proportion of messages with social support
provisions in private communities(76%) than in public communities (53%).

Data also indicate that there is a significant relationship between
openness and the types of social support provided in messages. Results show
that a larger proportion of messages with purely informational support are posted
in public communities (43%) as opposed to private communities (29%), whereas
a larger proportion of messages with purely emotional support are posted in
private communities (60%) as opposed to public communities (43%).

The most intuitive explanation for the differences in social support
provision type as it relates to community openness is that people are more
comfortable posting personal, emotional support provisions in private spaces and
that informational support lends itself to public spaces where the information will

reach a broader audience. Unfortunately, due to the limitations of the utilized

114

methodology, it is impossible to tell whether or not this is the case. An alternative
explanation for the patterns found in gendered posting by public and private
communities could be that women are potentially more comfortable posting
messages in an environment where they know that there is a greater level of
privacy while men are more comfortable in a public environment.

To explain social support provision types as they occur in public vs.
private communities, it is possible that there is a larger proportion of purely
emotional support provision in private communities because women (who tend to
post emotional support provisions) post most frequently in private communities.
Likewise, it is possible that there is a larger proportion of purely informational
support provision in public communities because men (who tend to post
informational support provisions) post most frequently in public communities.
While this doesn’t explain why men and women tend to post where they do, it is a
reasonable explanation as to why the types of support are found in communities
based on openness.

The finding that there is a significant relationship between community
openness and the posting of social support provisions has great implications for
the design of online support communities. Clearly, the openness of a community
has a very real effect on the types of messages that people post. The decision to
leave a community public or make it private should not be taken lightly. Findings
from this investigation suggest that when trying to develop a support group for

emotional support, the community should be made private, whereas when trying

115

to develop a support community for the purposes of informational support, it is

possible that the community could benefit from being left public.

Community Host and Social Support Provision

Data reveal that there is a significant relationship between community host
and social support provision type. Specifically, there is a greater proportion of
purely informational support provision in the ACOR communities (45%) than in
the Yahoo! communities (21 %) and there is a greater proportion of purely
emotional support provision in the Yahoo! communities (69%) than in the ACOR
communities (42%).

This relationship could be possibly explained by the very different nature
of the two community hosts. As mentioned in chapter three, ACOR is a
community host specifically intended for cancer-related purposes. When people
visit ACOR, they do so in order to find cancer-related information and social
support. On the other hand, Yahoo! is a community host that is not cancer-
specific. People who join online cancer support communities through Yahoo! may
already utilize other Yahoo! based communities or web-based services for a
variety of different reasons. It is quite possible that a greater proportion of
informational support provisions were found in ACOR communities as opposed
to Yahoo! communities because there is a greater expectation that informational
support will be available through ACOR. This could be due to perceptions that
people at ACOR have more expertise with cancer and are more qualified to
provide informational support versus the people at Yahoo! If this in fact were the

case, as social support provisions are ideally matched to social support seeking,

116

it makes sense that there would be more informational support seeking at ACOR
and therefore more informational support provisions. As there may potentially be
a lack of perceived expertise at Yahoo! communities, it also makes sense that
there would be less informational support seeking and therefore a greater
frequency of emotional support provisions in these communities.

The finding of the significant relationship between community host and
social support provision has many implications for the use of online support
communities. If in fact there is a greater proportion of informational support at
ACOR due to its cancer-specific nature, this implies that those individual who are
seeking informational support would be best served joining a cancer-specific
community. If this finding were to apply to all communities, cancer-specific
community hosts would be best served by marketing the frequency and quality of
informational support provision available to their members. In addition,
healthcare providers would be best served to steer new patients to cancer-

specific communities when looking for informational support.

Post Volumel Gender Participationl and Social Support Provision
Data from this investigation indicate that females post a larger proportion
of messages containing social support provision to communities that have a high
post volume (76%) as opposed to a low post volume (62%), whereas males post
a larger proportion of messages to communities with a low post volume (27%) as
opposed to a high post volume (19%).
This finding could possibly be explained by suggesting that males are

more comfortable posting in small volume communities, whereas females are

117

more comfortable posting in large volume communities. This possible
explanation potentially has merit as previous research has indicated that women
tend to have larger support networks than men (Antonucci & Akiyama, 1987;
Belle, 1987; Fischer, 1982; Turner, 1994) and therefore may be more
comfortable contributing to communities that have larger volumes of posts.

Data also indicate that there is a larger proportion of purely emotional
support in high post volume communities (63%) as opposed to low post volume
communities (40%), while there is a larger proportion of purely informational
support in low post volume communities (46%) as opposed to high post volume
communities (26%). A possible explanation for this finding is that the number of
emotional support provisions may create high post volume communities due to
the fact that one specific request for emotional support elicits multiple emotional
support provisions from different community members. On the other hand, the
number of informational support requests may create low volume communities
because one specific request for informational support from a community
member may only elicit a small number of responses due to the fact that once
one person has provided the necessary information, there is no need for others
to keep responding.

It is also possible that the differences in post volume could be due to the
posting patterns of the different genders. As males post more often in low post
volume communities and they tend to provide informational support more often
than emotional support, this could possibly explain why there is more

informational support provision in low post volume communities. Conversely, as

118

females post more often in high post volume communities and they tend to
provide emotional support in a greater proportion than informational support, this
could explain why there is more emotional support provision in high post volume
communities.

If in fact there are fewer informational support provisions in low post
volume communities due to the fact that informational support only really requires
a small number of support provisions to satisfy someone seeking informational
support, this speaks to the notion that online support communities, regardless of
post volume, should have an informational resource that community members
can utilize to answer basic questions regarding the specific disease and illness.
As mentioned previously, whether this informational resource takes the shape of
a FAQ page located on the community homepage, links to external websites,
links to specific message postings that were particularly informative, or even a
collaborative informational document such as a wiki, this type of informational

source could be quite beneficial to all community members.

Contributions from this Study
Results from this investigation are important as they give us very real
prescriptive suggestions as to how online support communities should be
structured for various circumstances. As research in the field of online support
communities has been largely descriptive in nature, this investigation begins to
allow for the understanding of how issues such as the demographics of
community members and the characteristics of a community can have real life

impacts on the types of social support provisions that are found online. This can

119

have real life consequences as the community members who utilize these
services have very real support needs since they are dealing with cancer either
themselves or through a family member or friend. When dealing with online
support communities, researchers need to understand that we are dealing with
people’s lives and potentially their health. It is through further investigations such
as this that we will be better able to understand how support communities should
be designed to more effectively help the community members who are utilizing
them.

This investigation also contributes to the existing body of literature
pertaining to online support communities as it marked the creation of a reusable
coding protocol. Aside from those communities focused on cancer, this coding
protocol could be used to evaluate social support provisions in a variety of
health-related online support communities.

Regarding theoretical contributions, this investigation adds to Optimal
Matching Theory by upholding the idea that controllability has a significant
relationship with the type of social support provided. In the original papers on
Optimal Matching Theory, as previously indicated, Cutrona specified that
emotional support was more important in low controllability circumstances,
serving as the basis for the predictions that emotional support messages would
be more prevalent in low survival rate communities than in high survival rate
communities and informational support messages would be more prevalent in
high survival rate communities than in low survival rate communities. This

investigation adds to OMT through the finding that theses hypotheses were not

120

supported, and that at least in online support groups, emotional support is
provided in greater frequency in both low and high controllability conditions. This
suggests that either there are more messages within online support communities
that elicit emotional support responses or that there is something about
emotional support seeking that elicits more responses than informational support
seeking within online communities. In addition to adding to OMT though
increasing our understanding of how controllability is related to social support
provision, this investigation adds to OMT by demonstrating its use in an

electronic environment.

Limitations

As with all research investigations, this study is not without limitations.
Perhaps the largest limitation to this investigation is the fact that this study was
strictly a content analysis of posted messages within the eight online support
communities for cancers with high and low five-year survival rates. Due to the
limitations of content analysis, this study does not provide a deeper
understanding of peoples’ motivations to participate and provide social support
within online support communities.

A second limitation of this investigation is that while social support
provisions were captured, it is unclear whether the support that was provided
was viewed as useful and beneficial by the support recipient. Again, due to the
limitations of content analysis, this was not possible. Also, because this

investigation excluded any sources of social support outside of the online

121

communities (i.e., in emails, over the phone, in person, etc.), this investigation
only captured a narrow slice of the total social support available to an individual.

Another limitation of this study relates to the fact that as this investigation
did not focus at all on social support seeking, no conclusions can be made as to
how frequently social support is actually sought. In addition, the fact that this
study did not tie online support community use to actual health outcomes is
another limitation of this investigation.

Finally, a limitation of this investigation is that no direct quotations were
reported. Direct quotations would have enriched the results section of this study
as they could have been used to fully demonstrate how different types of
emotional support and informational support actually appear within online support
communities. This limitation was a direct result of the restrictions placed by the
institutional review board at Michigan State University.

Despite these limitations, this study still is meaningful and important in that
steps have been taken to move beyond descriptive case study analysis of
individual support communities. This investigation serves as the first step in a
long line of research that aims to not only improve the quality of research in this
field, but seeks to understand what can be done to more effectively support the

social support needs for different types of people as they experience cancer.

Future Research

As with any research investigation, results from this study raise a whole
host of new and interesting questions regarding social support within online

health-related communities. Since this investigation served as the foundation for

122

what will no doubt become a long line of research involving online health-related
communities, these questions will help to focus the future direction of this
research stream.

To begin with, this investigation revealed differences in the types of social
support provided within public and private communities. Future work should
focus on better understanding why people choose to participate in public and/or
private communities, how the level of openness impacts community
demographics, and why there are differences between social support interactions
within these public and private spaces. Because the decision to make an online
support community public or private has potentially significant consequences for
the interactions that occur within online communities, this decision should not be
made lightly and the more researchers can understand the implications of
community openness, the better off online communities will be.

Next, this investigation indicated that people with different demographics
(regarding both gender and community member role) had fundamental
differences in the types and frequency of social support messages that were
provided. Future work should seek to understand why different sources provide
different types of social support. By better understanding why different sources
provide the types of social support that they do, health-related services may be
better able to design effective resources for different types of people.

Although this investigation provided much data concerning the frequency
of social support messages, no data concerning the quality of emotional or

informational support was of obtained or analyzed. As the quantity vs. quality

123

issue is of vital importance when dealing with peoples’ health and quality of life,
future work should seek to include measures of social support quality. In addition,
future work should focus on the entire support process as it occurs online by
studying social support seeking in conjunction with the provision of social
support. For example, future work should seek to address how frequently social
support is sought and how often responses with social support messages are
provided. By more fully understanding the social support process within online
support communities, we may be better able to provide effective and efficient
support services both online and offline.

Finally, future work in this field should attempt to tie online support
community use to actual health outcomes. If we truly wish to understand the
impact of online social support communities within the health context,
understanding how the social support that is received within an online
environment impacts the actual health outcomes of community members is a

must.

124

Appendix A: Vaux’s Categories of Social Support

Measures and their Respective Instruments

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

4: {’5’ 2
“6,3. 8 '5 9 - 2 8 3 8 .5
2 ° 2 3 E 3 E 3 '5 ,5 E = 7, a
a z a 5 o 5 n5 5 4:, efi '5 .e
3 "-' 3 '0 3 fi 3 a t o g g 3 ,2
0 8. ° 3 3 m (D Q 0 CD 0) o t
5 a.“ 2 '0 < E .9 s .1“
3 3 o
(D a) I
Social Network Inventory Arizona Social Social Efficacy of Community
Questionnaire of Socially Support Interview Support Help Support
(SNQ) Supportive Schedule (ASSIS) Incidents Seeking
Behavior (SSI) (EHS)
(ISSB)
Social Social Social Support Network
Relationship Support Resources (SSR) Orientation
Scale (SRS) Behaviors Scale
(SSB) (NOS)
Social Support Social Relationship
Questionnaire Scale (SRS)
(SSQ)
Social Support Daily Interaction
Network Rating Form (DIRF)
Interview
(SSNI)
Arizona Social Social Support
Support Questionnaire
Interview (SSQ)
Schedule
(ASSIS)
Social Support Perceived Social
Resources Support (PSS)
(SSR)
Family Relationship
Index (FRj
Work Relationship
Index (WRI)
Interpersonal
Support Evaluation
List (ISEL)
Social Provisions
Scale (SPS)
Provision of Social
Relationships (PSR)
Revised Kaplan
Scale (RKSL
Social Support
Appraisals (SSA)

 

 

 

 

 

 

125

 

Appendix B: Coding Protocol

Introduction

This investigation seeks to identify the types and amounts of social
support provisions contained within and across different cancer-related online
support communities. For the purposes of this investigation, a content analysis
will be conducted on posts from eight different cancer-related online

communities; four hosted by Yahoo!Groups and four hosted by the Association of .1.
Online Cancer Resources (ACOR). These eight communities are focused on four T
different types of cancer, two that have high five-year relative survival rates 1

(melanoma and thyroid cancers) and two that have low five-year relative survival 1

rates (pancreatic and lung cancers). In our sample, we will be investigating one
community focused on each type of cancer located in both Yahoo!Groups and
ACOR. The URLs to the communities are as follows:

Yahoo!Groups:

. Melanoma httpz/lhealth.groups.yahoo.com/group/melanoma/

o Thyroid Cancer http://heaIth.groups.yahoo.com/grouplthyroid-cancer/

. Pancreatic Cancer httpzllgroups.yahoo.com/group/pancreatic_cancer/

0 Lung Cancer
http://health.groups.yahoo.com/group/Lung_Cancer_Support/

ACOR:

o Melanoma http://Iistserv.acor.org/archives/melanoma.html

o Thyroid Cancer http://listserv.acor.org/archives/thyroid-onc.html

o Pancreatic Cancer http://|istserv.acor.org/archives/pancreas-onc.html

. Lung Cancer http://Iistserv.acor.org/archives/Iung-nsclc.html

This protocol outlines a set of systematic procedures to be used in the
coding of posted messages. Definitions and examples of all variables to be
coded are included. The units of observation and the units of analysis in this
investigation are individual community postings. Each post will be evaluated
against the coding protocol. The following procedures may be applied to evaluate
postings from any online health-related support community.

126

Conceptual Definitions

Online Community

An online community is conceptualized as collections of more than one
person existing over a sustained period of time that have interdependence
among one another in terms of common goals, interests, and/or identities. These
collections of people must possess at least one channel of communication (in
this case, the lntemet) that can be used for the exchange of communication
regarding their mutual goals, interests, identities, or interdependencies.

Functionally, we will consider online communities to be any combination of
listservs, newsgroups, bulletin boards, message boards, chat groups, instant
message systems, or web based portals on the lntemet where people go to
give/receive social support and/or information from other community members on
a particular topic. These are the commonly accepted forms of online
communities. Online communities are characterized by archives of posts which
may or may not be publicly accessible.

Messages postings

Message postings are defined as individual posted messages, posted by
one person, located within an online community. Although messages can only be
posted by an individual person, they may be written on behalf of several people.
A message posting can range in length from one word to several pages of
communications. Postings may include text and in some cases even images. It is
not uncommon for previous content to be copied in to messages, particularly
when a person replies to a specific message. For the purposes of coding, only
new content (not content copied from previous messages) will be analyzed for
each post. Although not always, old content is typically designated with a “<”
symbol at the beginning of a line. In circumstances where new responses are
embedded within old content, only the new responses will be coded. Also,
content that is a part of a subject or signature line will not be coded. Signature
lines are considered to be closing words or phrases stated immediately before a
persons name or alias. Examples include “sincerely”, “yours”, “best regards”, “in
God’s name”, “peace be with you”, etc.

Controllability

In this investigation, five-year relative survival rates serve as a proxy for
controllability. Controllability is conceptualized as the degree to which a particular
disease or condition can be managed. For the purposes of this investigation,
controllability is predetermined through the selection of online communities to be

127

investigated. Communities were selected along varying levels of controllability.
Therefore, coders do not need to code messages for controllability.

Gender

Gender is a social construct used to describe the self identified masculinity
or femininity of a person. In this investigation, gender will be operationalized into
two nominal categories based on sex; male and female. Because posters will
most likely not self-identify their gender, gender will only be coded when it is
contextually identified. Contextual cues to a person’s gender may include
discussion of being a mother or father, discussion of a spouse, etc. After coding
is complete, coders will go back and “fill in the blanks” by matching gender with
individual posters. If a poster never self-identifies their gender, they will be
designated as unclassifiable.

Role of communig member

The role of a community member is defined as the relationship an
individual has to the community topic. For the purpose of this investigation,
community role will be operationalized as a nominal level variable with values
such as patient, caregiver, medical professional, or a friend or family member of
patient. Because posters will most likely not self-identify their role in each post
they write, roles will only be coded when they are identified. After coding is
complete, coders will go back and “fill in the blanks” by matching self-identified
roles with individual posters. If a poster never self-identifies their community role,
they will be designated as unclassifiable. In the event that a person is coded as
two different roles, one of which is a patient, the patient categorization will
“trump” all other community member roles.

Social support

Social support encompasses the notion that that there are stressors in life
and people may receive from others either the perception that assistance is or
could potentially available or actual exchanges of emotional or instrumental
assistance in times of need. These coding procedures for the presence of social
support could be applied to any online support community, not just the eight
being evaluated in this investigation. Each posting will be read and evaluated for
the presence or absence of each of the different subtypes of social support.
Because we are only interested in the total number of messages that each type
of social support is present, each type of social support will only be counted once
per message. For our purposes, we are only interested in the presence or
absence of each subtype of social support. While a particular message may
contain three different instances of support type A and two instances of support
type B, we are only interested in the fact that the particular message contained
support types A and B, not the overall instances of A and B.

128

Variable Definitions and Coding Information

Demo ra hic and Administrative Variables Variables 1 - 9

V1.

V2.

V3.

V4.

V5.

V6.

V7.

V8.

Posting identification number - This is the unique number assigned to
each individual posting by the host community.

Cancer Type
0 = Melanoma
1 = Thyroid

2 = Pancreatic
3 = Lung

Community Host
0 = Yahoo!Groups

1 = ACOR
Openness
O = Public

1 = Private

Post Volume

0 = Low

1 = High
Community Size
O=Smml

1 = Large

Usemame of the message poster - As found in the header information

Gender of the message poster — As identified by contextual clues within
the posted messages and then matched back.

O=Mme

1 = Female

2 = Unknown

129

V9. Role
0 = Unclassified
1 = Patient
2 = Spouse/Partner
3 = Child of patient
4 = Parent of patient
5 = Other family member of patient
6 = Medical professional
7 = Volunteer/Activist
8 = Friends
9 = Other

Social Support Variables (Variables 10 - 21)

Emotional Support Variables (V10 - V15)

Emotional support is defined as support interactions where emotional
appeals are given from one person to another. These emotional appeals may
include expressions of understanding, reassurance, sending “positive energy”,
compassion, and sympathy. In addition to overt expressions of emotional
support, miscellaneous support materials such as poems, prayers, inspirational
stories, and relations of personal experience for support purposes are considered
to be Instances of emotional support.

V10. Prayer] Spiritual Support
The prayer/spiritual support subtype is defined as emotional support
interactions where offers to pray for someone, spiritual advice, quotations
of scripture, or other spiritual matters are provided by one person.

Examples of Prayer/ Spiritual Support may look something like:
“You are in my prayers.”

“I’m sending you positive energy.”

“Matthew 3:10 says...”

“Sometimes God calls us home.”

To Code:
0=No
1 = Yes

130

V11.

V12.

Understanding Support

The understanding support subtype is defined as emotional support
interactions where one person offers words of understanding or relates
personal experiences to another in order to demonstrate that they identify
with the support recipient. In order to qualify as understanding support,
there needs to be some personal appeal from one person to another
indicating that the support provider has an understanding of the recipient’s
situation.

Examples of Understanding Support may look something like:

0 “I understand where you are coming from.”

o “I experienced the same thing with my mother.”

0 “When my husband was diagnosed, I also felt the same thing"

. “I know exactly what you are going through... when l was
diagnosed...”

. “When I first went on chemo, I felt even sicker as well.”
0 “I would have done the same thing”
To Code:
O=No
1 = Yes
Sympathy Support

The sympathy support subtype is defined as emotional support
interactions where one person offers words sympathy to another.

Examples of Sympathy Support may look something like:
. “I am so sorry for your loss.”

0 “I am sorry that this has happened to you.”

- “Words cannot describe the sadness I feel for you.”

Note: Interactions where someone says “I’m sorry” for something non-
support related will not be coded (i.e., “I’m sorry your pen ran out of ink”).

To Code:
O=No
1 = Yes

131

V13.

V14.

Welcoming/Belonging Support

The welcoming/belonging support subtype is defined as emotional support
interactions where one person offers words of welcome or belonging to
new or established community members. These interactions typically
appear when a new community member introduces him/herself to the
community.

Examples of Welcoming Support may look something like:
“Welcome to the community.”

“We are so glad to have you here.”

“You have found a home in us.”

“You are a member of this community”

“You belong here with us”

“You are one of us"

To Code:
O=No
1 = Yes

Encouragement Support

The encouragement support subtype is defined as emotional support
interactions where one person offers words of encouragement another.
This category also includes instances of jokes, anecdotes, inspirational
stories or images, etc.

Examples of Encouragement Support may look something like:
“Don’t give up.”

“I support you in your decision.”

“You can get through this.”

“You are not alone.”

“Just take it one day at a time.”

“I’m glad to hear you are doing well.”

“It is wonderful that XXXX is doing well.”

Images or stories meant to make someone smile.

Positive responses to support provisions, images, stories, etc.

To Code:
O=No
1 = Yes

132

V15. Offers for Direct Contact
The Offers for Direct Contact support subtype is defined as an interaction
where one person directly offers their services to another through various
channels such as telephone, email, instant message, blog, etc.

Examples of Offers for Direct Contact may look something like:

0 “Feel free to contact me at any time.”

0 “My phone number is XXX-XXXX, call me whenever you need me.”

0 “My email address is asdf@asdf.com, email me whenever you need
someone to talk to.”

To Code:
O=No
1 = Yes

Informational Support Variables (V16 - V20)

lnforrnational support is defined as any support interaction where
informational aid is directly given from one person to another.
Informational aid will be coded according to the different types of
assistance.

DECISION RULE: As V16 — V19 ARE SUBTYPES OF V20 (GENERAL INFORMATIONAL
SUPPORT), ANY ONE SUPPORT INTERACTION THAT COULD BE CLASSIFIED AS BOTH
INFORMATIONAL SUPPORT AND ONE OF THE SUBTYPES WILL BE CODED ACCORDING To
THE SUBTYPE.

V16. Mediated Resource Information (Informational Support Subtype)
Mediated resource information is defined as any information one person
offers to another that is related to mediated resources such as books,
websites, movies, TV shows, other online communities, chat rooms, etc.
Any type Of printed resource that is related to the topic at hand is
considered to be a part of this category. Interactions that describe the
content of a printed resource listed are still to be coded as printed
resource information.

Examples of printed resource information may look like:

0 “There is a great website on Condition X located at www...’

- “Check out community X, they have great pain management advice”

0 “The book )OOOOOO( written by John Smith has a lot of good
information about...”

0 “The Today Show is having a spot on cancer tomorrow morning”

To Code:
O=No
1 = Yes

133

V17.

V18.

Live Assistance Resource Information (lnfonnational Support

Subtype)
Live assistance resource information is defined as any information one

person offers to another that is related to live resources available offline
regarding a particular disease / condition.

Examples of live assistance resource information may look like:

. “The National Diabetes Network offers free disease management
seminars”
“There is a support hotline offered through X that you can call for help”
“Contact this number to find your local hospice provider XXX-XXX-
XXXX”

0 “There is a Specialist in your area named Dr. Smith”

To Code:
O=No
1 = Yes

Disease I Symptom Information (Informational Support Subtype)
Disease and Symptom information is defined as any information one
person offers to another that is related to characteristics of a disease or
symptoms associated with a particular disease / condition.

Examples of disease / symptom information may look like:
“Symptoms of Condition X include”

“Dry mouth is a symptom association with Condition X”
. “HIV/AIDS kills XXXX number of people a year"

c “There are X number of different forms of MS”

NOTE: Personal accounts and opinions regarding experienced symptoms
are not to be considered a part of this category. Only information that is
seen as unbiased Should be considered. Finally, interactions that include
unreliable sources of information are not considered to be informational

support.

To Code:
O=No
1 = Yes

134

V19.

V20.

Medication l Treatment Information (Informational Support Subtype)
Medication / Treatment information is defined as any information one
person offers to another that is related to different medications,
treatments, or therapies related to the disease/condition. This includes
information regarding disease management.

Examples of medication / treatment information may look like:

“There is a new type of treatment called X”

“Medication X has been known to cause the symptoms you describe”
“Have you heard about the new drug for your condition called X?”
“Lipitor is a drug that is used to lower cholesterol”

NOTE: Personal accounts of what medications a person is currently taking
are not to be considered a part of this category. Also, personal opinions
about medications are not to be considered a part of this category. Finally,
interactions that include unreliable sources of information are not
considered to be informational support. Examples of what NOT to code
may include:

. “I use Lipitor to lower cholesterol”

0 “My doctor put me on insulin to help control my blood sugar”

To Code:
O=No
1 = Yes

General Cancer Informational Support
lnforrnational support is defined as support interactions where unbiased
information is given regarding anything that is on topic.

Examples of informational support may look something like:

“There are several different types of wheelchairs, including X and Y”
“Band-Aid produces a bandage that is waterproof"

“There are gluten-free communion wafers commercially available”
“X is a grocery store that caters to people with Special dietary needs”
“November is national breast cancer awareness month”

“There will a conference on XXXXXX next October”

NOTE: Interactions that include unreliable sources of information are not
considered to be informational support. For example:

. “I’ve heard that zinc is good for fighting colds”

0 “I’ve been told that Nabisco makes a gluten free product”

. “I’ve always been told that carrots are good for your eyes”

To Code:
O=NO
1 = Yes

135

 

Instrumental Support (V21)

V21.

Instrumental Support

This category is designed to code any support interaction that cannot be
defined as either emotional support or informational support. This leaves
room for the third type of support, instrumental support. Examples include
offers of non-information related goods and services.

Examples of Instrumental support may look something like:
0 “I’m sending you money for treatment”
a “I’m sending you X”

To Code:
O=No
1 = Yes

136

Works Cited

Abbate Daga, G., Gramaglia, C., Piero, A., & Fassino, S. (2006) Eating disorders
and the lntemet: cure and curse. Eating and Weight Disorders, 11(2), e68
— 71.

Alao, A. O., Yolles, J. C., & Amienta, W. (1999). Cybersuicide: The Internet and
suicide. The American Journal of Psychiatry, 156(11), 1836 — 1837.

Allen, S. F., & Stoltenberg, C. D. (1995). Psychological separation of Older
adolescents and young adults from their parents: An investigation of
gender differences. Journal of Consulting and Development, 73(5), 542 -
546.

American Cancer Society. (2007). Cancer facts and figures 2007. Atlanta:
American Cancer Society.

Antoucci, T. C., & Akiyama, H. (1987). An examination of sex differences in
social support among Older men and women. Sex Roles, 17(11/12), 737 —
749.

Argyle, M. (1992). Benefits produced by supportive social relationships. In H. O.
F. Veiel & U. Baumann (Eds), The meaning and measurement of social
support (pp. 13 — 32). New York: Hemisphere Publishing Company.

Ashton, W. A., & Fuehrer, A. (1993). Effects of gender and gender role
identification of participant and type of social support resource on support
seeking. Sex Roles, 28(7/8), 461 — 476.

Barnett, G. A., & Hwang, J. M. (2006). The use of the lntemet for health
information and social support: A content analysis of online breast cancer
discussion groups. In M. Murero & R. E. Rice (Eds), The Internet and
health care: Theory, research, and practice (pp. 233 - 253). Mahwah, NJ:
Lawrence Erlbaum Associates.

Barrera, M. (2000). Social support research in community psychology. In J.
Rappaport & E. Seidman (Eds.), Handbook of community psychology (pp.
215 — 245). New York: Kluwer Academic/Plenum.

Barrera, M., & Ainlay, S. L. (1983). The structure of social support: A conceptual

and empirical analysis. Journal Of Community Psychology, 11(2), 133 —
143.

137

Barrera, M., Jr., Glasgow, R. E., McKay, H. G., Boles, S. M., & Feil, E. G. (2002).
Do Internet—based support interventions change perceptions of social
support? An experimental trial of approaches for supporting diabetes self-
management. American Journal of Community Psychology, 30(5), 637 —
654.

Belle, D. (1982). The stress of caring: Women as providers of social support. In
L. Goldberger & S. Breznitz (Eds.), Handbook of stress: Theoretical and
clinical aspects (pp. 496 — 505). New York: Free Press.

Belle, D. (1987). Gender differences in the social moderators of stress. In R. C.
Barnett, L. Biener, & G. K. Baruch (Eds), Gender and stress (pp. 257 -
277). New York: Free Press.

Berger, B. D., Lehrrnann, J. A., Larson, G., Alvemo, L., & Tsao, C. I. (2005).
Nonpsychotic, nonparaphilic self—amputation and the lntemet.
Comprehensive Psychiatry, 46(5), 380 — 383.

Berkman, L. F. (1985). The relationship of social networks and social support to
morbidity and mortality. In S. Cohen & S. L. Syme (Eds.), Social support
and health (pp. 241 — 262). Orlando, FL: Academic Press.

Berkman, L. F., & Syme, S. L. (1979). Social networks, host resistance, and
mortality: A nine-year follow-up study of Alameda County residents.
American Journal of Epidemiology, 109(2), 186 - 204.

Bemstam, E. V., Shelton, D. M., Walji, M., & Meric-Bemstam, F. (2005).
Instruments to assess the quality of health information on the World Wide
Web: What can our patients actually use? International Journal of Medical
Informatics, 74(1), 13 — 19.

Blank, T. O., & Adams-Blodnieks, M. (2007). The who and the what of usage of
two cancer online communities. Computers in Human Behavior, 23(3),
1249 — 1257.

Blazer, D. (1982). Social support and mortality in an elderly community
population. American Journal of Epidemiology, 115(5), 684 — 694.

Braithwaite, D. O., Waldron, V. R., & Finn, J. (1999). Communication of social
support in computer-mediated groups for people with disabilities. Health
Communication, 11(2), 123 - 151.

Bremner, J. D., Quinn, J., Quinn, W., & Veledar, E. (2006). Surfing the net for
medical information about psychological trauma: An empirical study of the
quality and accuracy of trauma-related websites. Medical Informatics and

the Internet in Medicine, 31(3), 227 — 236.

138

Brissette, l., Cohen, S., & Seeman, T. (2000). Measuring social integration and
social networks. In S. Cohen, L. Undenrvood, & B. H. Gottlieb (Eds.),
Social support measurement and intervention: A guide for health and
social scientists (pp. 53 — 85). New York: Oxford University Press.

Burda, P. C., Jr., Vaux, A., & Schill, T. (1984). Social support resources:
Variations in sex and sex-role. Personality and Social Psychology Bulletin,
10(1),119 —126.

Burke, R. J., & Weir, T. (1977). Marital helping relationships: The moderators
between stress and well being. Journal of Psychology, 95, 121 - 130.

Burleson, B. R., Albrecht, T. L., Goldsmith, D. J., & Sarason, I. G. (1994).
Introduction: The communication of social support. In B. R. Burleson, T. L.
Albrecht, & I. G. Sarason (Eds.), Communication of social support:
Messages, interactions, relationships, and community (pp. xi - xxx).
Thousand Oaks, CA: Sage Publications.

Butler, L., & Foster, N. E. (2003). Back pain online: A cross-sectional survey of
the quality of web-based information on low back pain. Spine, 28(4), 395 -
401.

Butler, R. J., Ill, Friel, D. F., Lang, R. D., Park, G., & Santello, S. A. (2000).
Healthcare provider-sponsored virtual communities: The benefits of the
electronic village. Journal of Healthcare Information Management, 14(1),
65 - 71.

Butler, T., Giordano, S., & Neren, S. (1985). Gender and sex-role attributions as
predictors of utilization and natural support systems during personal stress
events. Sex Roles, 13(9/10), 515 — 524.

Cassell, J. (1976). The contribution of the social environment to host resistance.
American Journal of Epidemiology, 104(2), 107 — 123.

CHESS. (2006). CHESS: Comprehensive Health Enhancement Support System.
Retrieved on March 28, 2007, from
http://Chess.chsra.wisc.edU/Chess/default.htm.

Christensen, K., & Levinson, D. (Eds.). (2003). Encyclopedia of community: From
the village to the virtual world (Vols. 1 — 4). Thousand Oaks, CA: Sage
Publications, Inc.

Cobb, S. (1976). Social support as a moderator of life stress. Psychosomatic
Medicine, 38(5), 300 — 314.

139

Cohen, S. (1992). Stress, social support, and disorder. In H. O. F. veiel & U.
Baumann (Eds), The meaning and measurement of social support (pp.
109 — 124). New York: Hemisphere Publishing Company.

Cohen, S., & McKay, G. (1984). Social support, stress and the buffering
hypothesis: A theoretical analysis. In A. Baum, S. E. Taylor, & J. E. Singer
(Eds), Handbook of psychology and health, Vol. IV: Social psychological
aspects of health (pp. 253 - 267). Hillsdale, NJ: Erlbaum.

Cohen, S., & Syme, S. L. (1985). Issues in the study and application of social
support. In S. Cohen & S. L. Syme (Eds), Social support and health (pp. 3
- 22). Orlando, FL: Academic Press.

Commission of the European Communities, Brussels. (2002). eEurope 2002:
Quality criteria for health-related Websites. Journal of Medical lntemet
Research, 4(3): e15.

Cramer, L. A., Riley, P. J., & Kiger, G. (1991). Support and antagonism in social
networks: Effects of community and gender. Journal of Social Behavior
and Personality, 6(4), 991 - 1005.

Croft, D., & Peterson, M. (2002). An evaluation of the quality and contents of
asthma education on the World Wide Web. Chest, 121(4), 1301 - 1307.

Cudney, S. A., & Weinert, C. (2000). Computer-based support groups. Nursing in
cyberspace. Computers in Nursing, 18(1), 35 — 43; quiz 44—36.

Cummings, J., Kiesler, S. B., & Sproull, L. (2002). Beyond hearing: Where real
world and online support meet. Group Dynamics: Theory Research, and
Practice, 6(1), 78 - 88.

Cutrona, C. E. (1990). Stress and social support: In search of Optimal matching.
Journal of Social and Clinical Psychology, 9(1), 3 — 14.

Cutrona, C. E., & Russell, D. W. (1990). Type of social support and Specific
stress: Toward a theory of optimal matching. In B. R. Sarason, I. G.
Sarason, & G. R. Pierce (Eds), Social support: An interactional view (pp.
319 — 366). New York: John Wiley.

Cutrona, C. E., & Suhr, J. A. (1992). Controllability of stressful events and

satisfaction with spouse support behaviors. Communication Research,
19(2), 154 — 174.

140

Davison, K. P., Pennebaker, J. W., & Dickerson, S. S. (2000). Who talks? The
social psychology of illness support groups. American Psychologist, 55(2),
205 - 217.

Depner, C. E., & lngersoII—Dayton, B. (1988). Supportive relationships in later
life. Psychology and Aging, 3(4), 348 - 357.

Derlega, V. J., Barbee, A. P., & Winstead, B. A. (1994). Friendship, gender, and
social support: Laboratory studies Of supportive interactions. In B. R.
Burleson, T. L. Albrecht, & l. G. Sarason (Eds), Communication of social
support: Messages, interactions, relationships, and community (pp. 136 —
151 ). Thousand Oaks, CA: Sage Publications.

Dobson, R. (1999). lntemet sites may encourage suicide. British Medical Journal,
319(7206), 337.

Dunham, P. J., Hurshman, A., Litwin, E., Gusella, J., Ellsworth, C., & Dodd, P. W.
(1998). Computer-mediated social support: Single young mothers as a
model system. American Journal of Community Psychology, 26(2), 281 -
306.

DunkeI—Schetter, C., & Bennett, T. L. (1990). Differentiating the cognitive and
behavioral aspects of social support. In B. R. Sarason, l. G. Sarason, & G.
R. Pierce (Eds), Social support: An interactional view (pp. 267 — 296).
New York: Wiley.

DunkeI—Schetter, C., & Skokan, L. A. (1990). Determinants of social support
provision in personal relationships. Journal of Social and Personal
Relationships, 7(4), 437 — 450.

Durkheim, E. (1951). Suicide: A study in sociology. London: The Free Press.

Eysenbach, G. (2003). The impact of the lntemet on cancer outcomes. CA: A
Cancer Journal for Clinicians, 53(6), 356 — 371.

Eysenbach, G., & Kohler, C. (2004). Health—related searches on the lntemet.
Journal of the American Medical Association, 291 (24), 2946.

Eysenbach, 6., 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(7449), 1166.

141

Eysenbach, 6., Powell, J., Kuss, O., & Sa, E. (2002). Empirical studies
assessing the quality of health information for consumers on the World
Wide Web: A systematic review. Journal of the American Medical
Association, 287(20), 2691 — 2700.

Eysenbach, G., & Till, J. E. (2001). Ethical issues in qualitative research on
lntemet communities. British Medical Journal, 323(7321), 1103 - 1105.

Fischer, C. S. (1982). To dwell among friends: Personal networks in town and
City. Chicago: University of Chicago Press.

Flicker, S., Haans, D., & Skinner, H. (2004). Ethical dilemmas in research on
lntemet communities. Qualitative Health Research, 14(1), 124 — 134.

Fogel, J., Albert, S. M., Schnabel, F., Ditkoff, B. A., & Neugut, A. I. (2002).
Internet use and social support in women with breast cancer. Health
Psychology, 21(4), 398 — 404.

Fox, N., Ward, K., & O’Rourke, A. (2005). Pro-anorexia, weight—loss drugs and
the lntemet: An ‘anti-recovery’ explanatory model of anorexia. Sociology
of Health & Illness, 27(7), 944 - 971.

Fox, S. (2006). Online health search 2006. Retrieved on March 8, 2007, from
http://207.21.232.103/Ddfs/PIP Online Health 2006.Ddf.

Fox, 8., & Rainie, L. (2000). The online healthcare revolution: How the web helps
Americans take care of themselves. Retrieved March 30, 2007, from
httpzllwww.pewintemet.org/pdfs/PlP_Health_Report.pdf.

Gallant, M. P. (2003). The influence of social support on chronic illness self—
management: A review and directions for research. Health Education &
Behavior, 30(2), 170 — 195.

Glueckauf, R. L., Ketterson, T. U., Loomis, J. S., & Dages, P. (2004). Online
support and education for dementia caregivers: Overview, utilization, and
initial program evaluation. Telemedicine Journal and E-Health, 10(2), 223
- 232.

Goldsmith, D. J. (2004). Communicating social support. New York: Cambridge
University Press.

Google. (2007). About Google scholar. Retrieved on March 17, 2007, from
http://scholar.qoogle.com/intl/en/scholar/about.html.

Gottlieb, B. H. (1998). Support groups. In H. S. Friedman (Ed), Encyclopedia of
mental health. San Diego, CA: Academic Press.

142

 

Grande, G. E., Myers, L. B., & Sutton, 8. R. (2006). How do patients who
participate in cancer support groups differ from those who do not?
Psycho-oncology, 15(4), 321 - 334.

Gray, N. J., Klein, J. D., Noyce, P. R., Sesselberg, T. S., & Cantrill, J. A. (2004).
Health information seeking behaviour in adolescence: The place of the
Internet. Social Science and Medicine, 60(7), 1467 - 1478.

Greene, D. L., Appel, A. J., Reinert, S. E., & Palumbo, M. A. (2005). Lumbar disc
herniation: Evaluation of information on the lntemet. Spine, 30(7), 826 —
829.

Griffith, J. (1985). Social support providers: Who are they? Where are they met?
And the relationship of network characteristics to psychological distress.
Basic and Applied Social Psychology, 6(1), 41 — 60.

Griffiths, K. M., & Christensen, H. (2005). Website quality indicators for
consumers. Journal of Medical Internet Research, 7(5):e55.

Griffiths, K. M., Tang, T. T., Hawking, D., & Christensen, H. (2005). Automated
assessment of the quality of depression websites. Journal of Medical
Internet Research, 7(5):e59.

Hamman, R. B. (2001). Computer networks linking network communities. In C.
Werry & M. Mowbray (Eds), Online communities: Commerce, community
action, and the virtual university (pp. 71 — 95). Upper Saddle River, NJ:
Prentice—Hall.

Hanif, F., Sivaprakasam, R., Butler, A., Huguet, E., Pettigrew, G. J., Michael, E.
D., et al. (2006). lnforrnation about liver transplantation on the World Wide
Web. Medical Informatics and the Internet in Medicine, 31 (3), 153 — 160.

Hargrave, D. R., Hargrave, U. A., & Bouffet, E. (2006). Quality of health
information on the lntemet in pediatric neuro—oncology. Neuro—oncololgy,
8(2), 175 - 182.

Helgeson, V. S. (2003). Social support and quality of life. Quality of Life
Research, 12(Supplement 1), 25 — 31.

Helgeson, V. S., & Gottlieb, B. H. (2000). Support groups. In S. Cohen, L.
Underwood, & B. H. Gottlieb (Eds), Social support measurement and
intervention: A guide for health and social scientists (pp. 221 — 245). New
York: Oxford University Press.

143

Hesse, B. W., Nelson, D. E., Kreps, G. L., Croyle, R. T., Arora, N. K., Rimer, B.
K., et al. (2005). Trust and sources of health information: The impact of
the Internet and its implications for health care providers: Findings from
the first Health Information National Trends Survey. Archives of Internal
Medicine, 165(22), 2618 — 2624.

Hill, W. G., & Weinert, C. (2004). An evaluation of an online intervention to
provide social support and health education. Computers, Informatics,
Nursing: CIN, 22(5), 282 — 288.

Hillery, G. A., Jr., (1955). Definitions of community: Areas of agreement. Rural
Sociology, 20(2), 111 - 123.

House, J. S. (1981). Work stress and social support. Reading, MA: Addison—
Wesley.

House, J. S., & Kahn, R. L. (1985). Measures and concepts of social support. In
S. Cohen & S. L. Syme (Eds), Social support and health (pp. 83 — 108).
Orlando, FL: Academic Press.

House, J. 8., Robbins, C., & Metzner, H. (1982). The association of social
relationships and activities with mortality: Prospective evidence from the

Tecumseh community health study. American Journal of Epidemiology,
116(1),123 - 140.

Houston, T. K., Cooper, L. A., & Ford, D. E. (2002). lntemet support groups for
depression: A 1-year prospective cohort study. The American Journal of
Psychiatry, 159(12), 2062 — 2068.

Hupcey, J. E. (1998). Clarifying the social support theory-research linkage.
Journal of Advanced Nursing, 27(6), 1231 - 1241.

llic, D., Risbridger, G., & Green, S. (2004). Searching the lntemet for information
on prostate cancer screening: An assessment of quality. Urology, 64(1),
112 - 116.

Impicciatore, P., Pandolfini, C., Casella, N., Bonati, M. (1997). Reliability of
health information for the public on the World Wide Web: Systematic
survey of advice on managing fever in children at home. British Medical
Journal, 314(7098), 1875 — 1879.

Jacobson, D. E. (1986). Types and timing of social support. Journal of Health
and Social Behavior, 27(3), 250 — 264.

Junghans, T. B., Sevin, I. F., lonin, B., & Seifried, H. (2004). Cancer information
resources: Digital and online sources. Toxicology, 198(1-3), 177 — 193.

144

Kaid, L. L., & Wadsworth, A. J. (1989). Content analysis. In P. Emmert & L. L.
Barker (Eds), Measurement of communication behavior (pp. 197 — 217)
New York: Longman.

Kendler, K. S., Myers, J., & Prescott, C. A. (2005). Sex differences in the
relationship between social support and risk for major depression: A
longitudinal study of opposite-sex twin pairs. American Journal of
Psychiatry, 162(2), 250 — 256.

Kessler, R. C., McLeod, J. D., & Wethington, E. (1985). The costs of caring: A
perspective on the relationship between sex and psychological distress. In
I. G. Sarason & B. R. Sarason (Eds), Social support: Theory, research,
and applications. The Hague, Nethertands: Martinus Nijhoff.

Klemm, P., Hurst, M., Dearholt, S. L., & Trone, S. R. (1999). Gender differences
on lntemet cancer support groups. Computers in Nursing, 17(2), 65 — 72.

Klemm, P., Reppert, K., & Visich, L. (1998). A nontraditional cancer support
group: The lntemet. Computers, Informatics, Nursing: CIN, 16(1), 31 — 36.

Klemm, P., & Wheeler, E. (2005). Cancer caregivers online: Hope, emotional
roller coaster, and physical/emotionaI/psychological responses.
Computers, Informatics, Nursing: CIN, 23(1), 38 - 45.

Krizek, C., Roberts, C., Ragan, R., Ferrara, J. J., & Lord, B. (1999). Gender and
cancer support group participation. Cancer Practice, 7(2), 86 - 92.

Kunst, H., Groot, D., Latthe, P. M., Latthe, M., & Khan, K. S. (2002). Accuracy of
information on apparently credible websites: Survey of five common health
topics. British Medical Journal, 324(7337), 581 — 582.

LaCoursiere, S. P., Knobf, M. T., & McCorkle, R. (2005). Cancer patients’ self-
reported attitudes about the lntemet. Journal of Medical Internet
Research, 7(3):922.

Lamberg, L. (2003). Online empathy for mood disorders: Patients turn to lntemet
support groups. Journal of the American Medical Association, 289(23),
3073 — 3077.

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 mailinglist. Journal of Medical Internet
Research, 7(1):e10.

Lazarus, R. 8., & Folkman, S. (1984). Stress, appraisal, and coping. New York:
Springer Publishing Company.

145

Lester, J., Prady, S., Finegan, Y., & Hoch, D. (2004). Learning from e—patients at
Massachusetts general hospital. British Medical Journal, 328(7449), 1188
- 1190.

Lieberman, M. A., Golant, M., Giese—Davis, J., Winzlenberg, A., Benjamin, H.,
Humphreys, K., et al. (2003). Electronic support groups for breast
carcinoma: A clinical trial of effectiveness. Cancer, 97(4), 920 - 925.

Liu, Y., & Liu, M. (2006). Osteosarcoma: Evaluation of information on the
lntemet. Telemedicine Journal and e—Health, 12(5), 542 — 545.

L-Soft. (2005). L-Soft Interviews Gilles Frydman, President of ACOR. Retrieved
on March 30, 2007, from httpzllwww.Isoft.com/news/qa-issue1-2005-
us.asp.

Lombard, M., Snyder-Duch, J., & Bracken, C. C. (2002). Content analysis in
mass communication: Assessment and reporting of intercoder reliability.
Human Communication Research, 28(4), 587 — 604.

Luttik, M. L., Jaarsma, T., Moser, D., Sandennan, R., & Van Veldhuisen, D. J.
(2005). The importance and impact of social support on outcomes in
patients with heart failure: An overview of the literature. Journal of
Cardiovascular Nursing, 20(3), 162 - 169.

Madara, E. J. (1997). The mutual—aid self—help online revolution. Social Policy,
27(3), 20 - 26.

Madden, M. (2006). lntemet penetration and impact. Retrieved on March 17,
2007, from httD://www.Dewintemetcrolodfs/PIP lntemet Impactbdf.

Maloney, S., IIic, D., & Green, S. (2005). Accessibility, nature and quality of
health information on the Internet: A survey on osteoarthritis.
Rheumatology, 44(3), 382 — 385.

Mathur, S., Shanti, N., Brkaric, M., Sood, V., Kubeck, J., Paulino, C., et al.
(2005). Surfing for scoliosis: The quality of information available on the
Internet. Spine, 30(23), 2695 — 2700.

Matthews, S. C., Camacho, A., Mills, P. J., & Dimsdale, J. E. (2003). The lntemet
for medical information about cancer: Help or hindrance?
Psychosomatics, 44(2), 100 — 103.

Mayer, M. A., Karkaletsis, V., Archer, P., Ruiz, P., Stamatakis, K., 8. Leis, A.

(2006). Quality labelling of medical web content. Health Informatics
Journal, 12(1), 81 — 87.

146

Mendelson, C. (2003). Gentle hugs: lntemet listservs as sources of support for
women with lupus. Advances in Nursing Science, 26(4), 299 — 306.

Meric, F., Bemstam, E. V., Mirza, N. Q., Hunt, K. K., Ames, F. C., Ross, M. I., et
al. (2002). Breast cancer on the World Wide Web: Cross sectional survey
of quality of information and popularity of Websites. British Medical
Journal, 324(7337), 577 — 581.

Monnier, J., Laken, M., & Carter, C. L. (2002). Patient and caregiver interest in
lntemet—based cancer services. Cancer Practice, 10(6), 305 - 310.

Moos, R. H., Moos, B. S., & Timko, C. (2006). Gender, treatment and self-help in
remission from alcohol use disorders. Clinical Medicine & Research, 4(3),
163 — 174.

Mulveen, R., & Hepworth, J. (2006). An interpretative phenomenological analysis
of participation in a pro-anorexia lntemet Site and its relationship with
disordered eating. Journal Of Health Psychology, 11(2), 283 - 296.

Neuendorf, K. A. (2002). The content analysis guidebook. Thousand Oaks, CA:
Sage Publications.

Nip, J. (2004). The relationship between online and offline communities: The
case of the queer sisters. Media, Culture, & Society, 26(3), 409 - 428.

Nonnecke, B., & Preece, J. (2003). Silent participants: Getting to know Iurkers
better. In C. Lueg & D. Fisher (Eds), From usenet to cowebs: Interacting
with social information spaces (pp. 110 — 132). London: Springer.

Norris, M. L., Boydell, K. M., Pinhas, L. & Katzman, D. K. (2006). Ana and the
lntemet: A review of pro-anorexia websites. International Journal of Eating
Disorders, 39(6), 443 — 447.

O’Brien, J. (1999). Writing in the body: Gender (re)production in online
interaction. In M. Smith & P. Kollok (Eds), Communities in cyberspace
(pp. 76 — 104). New York: Routledge.

Owen, J. E., Klapow, J. C., Roth, D. L., Nabell, L., & Tucker, D. C. (2004).
Improving the effectiveness of adjuvant psychological treatment for
women with breast cancer: The feasibility of providing online support.
Psycho-oncology, 13(4), 281-292.

Owen, J. E., Klapow, J. C., Roth, D. L., & Tucker, D. C. (2004). Use of the
lntemet for information and support: Disclosure among persons with
breast and prostate cancer. Journal of Behavioral Medicine, 27(5), 491 —
505.

147

Pearlin, L. I. (1985). Social structure and processes of social support. In S.
Cohen & S. L. Syme (Eds), Social support and health (pp. 43 — 60).
Orlando, FL: Academic Press.

Perron, B. (2002). Online support for caregivers of people with a mental illness.
Psychiatric Rehabilitation Journal, 26(1), 70 — 77.

Pierce, L. L., Steiner, V., & Govoni, A. L. (2002). ln-home online support for
caregivers of survivors of stroke: A feasibility study. Computers,
Informatics, Nursing: CIN, 20(4), 157 — 164.

Powell, J., McCarthy, N., & Eysenbach, G. (2003). Cross—sectional survey of
users of lntemet depression communities. BMC Psychiatry, 3(1), 19.

Prior, T. I. (2004). Suicide methods from the lntemet. American Journal of
Psychiatry, 161(8), 1500 — 1501.

Preece, J. (2000). Online communities, designing usability and supporting
sociability. Chichester, England: John Wiley & Sons, LTD.

Preece, J., Nonnecke, B., & Andrews, D. (2004). The top five reasons for lurking:
Improving community experiences for everyone. Computers in Human
Behavior, 20(2), 201 - 223.

Pubmed. (2007). Pubmed. Retrieved on March 17, 2007, from
http://www.ncbi.nlm.nih.qovlentrez/querv.fcqi?DB=pubmed.

Rainie, L., & Horrigan, J. (2005). Trends 2005. Retrieved March 30, 2007, from
httpzllpewresearch.Org/assetS/files/trend$2005.pdf.

Rajagopal, S. (2004). Suicide pacts and the Internet. British Medical Journal,
329(7478), 1298 - 1299.

Ramos, J. D., Rai—Chaudhuri, A., & Neill, R. W. (2004). International online
discussion lists on chronic myelogenous leukaemia. British Medical
Journal, 328(7449), 1177 — 1 178.

Rheingold, H. (1993). The virtual community: Homesteading on the electronic
frontier. Reading, MA: Addison-Wesley.

Rhodes, S. D., Bowie, D. A., & Hergenrather, K. C. (2003). Collecting
behavioural data using the World—Wide Web: Considerations for
researchers. Journal Of Epidemiology and Community Health, 5 7(1 ), 68 —
73.

148

Riffe, D., Lacy, S., & Fico, F. G. (2005). Analyzing media messages: Using
quantitative content analysis in research. Mahwah, NJ: Lawrence Erlbaum
Associates.

Riley, 0., & Eckenrode, J. (1986). Social ties: Subgroup differences in costs and
benefits. Journal of Personality and Social Psychology, 51(4), 770 — 778.

Robinson, N. S. (1995). Evaluating the nature of perceived support and its
relation to perceived self—worth in adolescents. Journal of Research on
Adolescence, 5(2), 253 — 280.

Rosenthal, K. R., Gesten, E. L., & Shiffman, S. (1986). Gender and sex role
differences in the perception of social support. Sex Roles, 14(9/10), 481 —
499.

Sabel, M. S., Strecher, V. J., Schwartz, J. L., Wang, T. S., Karimipour, D. J.,
Orringer, J. S., et al. (2005). Patterns of lntemet use and impact on
patients with melanoma. Journal of the American Academy of
Dermatology, 52(5), 779 — 785.

Santoro, E. (2003). lntemet and information on breast cancer: An overview.
Breast, 12(6), 424 — 431.

Sarason, B. R., Pierce, G. R., & Sarason, I. G. (1990). Social support: The sense
of acceptance and the role of relationships. In B. R. Sarason, l. G.
Sarason, & G. R. Pierce (Eds), Social support: An interactional view (pp.
97 — 128). New York: John Wiley & Sons.

Sarason, B. R., Sarason, I. G., 8. Pierce, G. R. (1990). Traditional views of social
support and their impact on assessment. In B. R. Sarason, I. G. Sarason,
8. G. R. Pierce (Eds), Social support: An interactional view (pp. 9 - 25).
New York: John Wiley & Sons.

Sarason, l. G., Sarason, B. R., & Pierce, G. R. (1994). Relationship—specific
social support: Toward a model for the analysis of supportive interactions.
In B. R. Burleson, T. L. Albrecht, & l. G. Sarason (Eds), Communication
of social support: Messages, interactions, relationships, and community
(pp. 91 —112). Thousand Oaks, CA: Sage Publications.

Seale, C., Ziebland, S., & Charteris—Black, J. (2006). Gender, cancer experience
and lntemet use: A comparative keyword analysis of interview and online
cancer support groups. Social Science & Medicine, 62(10), 2577 — 2590.

Selman, T. J., Prakash, T., & Khan, K. S. (2006). Quality of health information for
cervical cancer treatment on the lntemet. BMC Womens Health, 6.9.

149

Soliman, A. M. (1993). Choice of helpers, types of problem and sex of helpers of
college students. International Journal for the Advancement of
Counseling, 16(2), 67 - 79.

Sullivan, C. F. (2003). Gendered cybersupport. A thematic analysis of two online
cancer support groups. Journal of Health Psychology, 8(1), 83 - 103.

Sutherland, L. A., Wildemuth, 8., Campbell, M. K., & Haines, P. S. (2005).
Unraveling the web: An evaluation of the content quality, usability, and
readability of nutrition web sites. Journal of Nutrition Education and
Behavior, 37(6), 300 — 305.

ThoitS, P. A. (1992). Social support functions and network structures: A
supplemental view. In H. O. F. veiel & U. Baumann (Eds), The meaning
and measurement of social support (pp. 57 - 62). New York: Hemisphere
Publishing Corporation.

Thomas, W., & Znaniecki, F. (1920). The polish peasant in Europe and America.
New York: Knopf.

Tichon, J. G., & Shapiro, M. (2003). The process of sharing social support in
cyberspace. Cyberpsychology & Behavior, 6(2), 161 - 170.

Tichon, J., & Yellowlees, P. (2003). lntemet social support for Children and
adolescents. Journal Of Telemedicine and Telecare, 9(4), 238 - 240.

Tierney, S. (2006). The dangers and draw of online communication: Pro-anorexia
websites and their implications for users, practitioners, and researchers.
Eating Disorders, 14(3), 181 — 190.

Till, J. E. (2003). Evaluation of support groups for women with breast cancer:
Importance of the navigator role. Health and Quality of Life Outcomes,
1(1), 16.

Tiller, G., Rea, S., Silla, R., & Wood, F. (2006). Burns first aid information on the
lntemet. Burns, 32(7), 897 — 901.

Trobst, K. K., Collins, R. L., & Embree, J. M. (1994). The role of emotion in social
support provision: Gender, empathy and expressions of distress. Journal
of Social and Personal Relationships, 11 (1 ), 45 - 62.

Turner, G., & Shepherd, J. (1999). A method in search of a theory: Peer

education and health promotion. Health Education Research, 14(2), 235 -
247.

150

Turner, H. A. (1994). Gender and social support: Taking the bad with the good?
Sex Roles, 30(7/8), 521 — 541.

Uchino, B. N. (2004). Social support and physical health: Understanding the
health consequences of relationships. New Haven, CT: Yale University
Press.

Vaux, A. (1985). Variations in social support associated with gender, ethnicity
and age. Journal of Social Issues, 41(1), 89 — 110.

Vaux, A. (1988). Social support: Theory, research, and intervention. New York:
Praeger.

Verheijden, M. W., Bakx, J. C., van Weel, C., Koelen, M. A., & van Staveren, W.
A. (2005). Role of social support in lifestyle—focused weight management
interventions. European Journal of Clinical Nutrition, 59(Supplement 1),
S179 — S186.

Weinberg, N., Schmale, J., Uken, J., & Wessel, K. (1996). Online help: Cancer
patients participate in a computer—mediated support group. Health &
Social Work, 21 (1 ), 24 - 29.

Weinrich, J. D. (1997). Strange bedfellows: Homosexuality, gay liberation, and
the lntemet. Journal of Sex Education and Therapy, 22(1), 58 - 66.

Wellman, B., & Gulia, M. (1999). Virtual communities as communities: Net
surfers don’t ride alone. In M. Smith & P. Kollok (Eds), Communities in
cyberspace (pp. 167 — 194). New York: Routledge.

Wethington, E., & Kessler, R. D. (1986). Perceived support, received support,
and adjustment to stressful life events. Journal of Health and Social
Behavior, 27(1), 78 — 89.

White, M., & Dorman, S. M. (2000). Online support for caregivers: Analysis of an
lntemet Alzheimer mailgroup. Computers in Nursing, 18(4), 168 - 179.

White, M., & Dorman, S. M. (2001). Receiving social support online: Implications
for health education. Health Education Research, 16(6), 693 — 707.

Whitlock, J. L., Powers, J. L., & Eckenrode, J. (2006). The virtual cutting edge:

The Internet and adolescent self-Injury. Developmental Psychology, 42(3),
407 — 417.

151

Wilcox, B., & Vemberg, B. (1985). Conceptual and theoretical dilemmas facing
social support research. In I. G. Sarason & B. R. Sarason (Eds), Social
support: Theory, research and applications (pp. 3 — 20). The Hague,
Netherlands: Martinus Nijhoff.

Winkelman, W. J., & Choo, C. W. (2003). Provider-sponsored virtual
communities for chronic patients: Improving health outcomes through
organizational patient-centered knowledge management. Health
Expectations, 6(4), 352 — 358.

Winzelberg, A. J., Classen, C., Alpers, G. W., Roberts, H., Koopman, C., Adams,
R. E., et al. (2003). Evaluation of an lntemet support group for women with
primary breast cancer. Cancer, 97(5), 1164 — 1173.

Woods, V. (2005). Work-related musculoskeletal health and social support.
Occupational Medicine, 55(3), 177 — 189.

Yahoo. (2005). The history of Yahoo! — How it all started. Retrieved on March 30,
2007, from http://docs.yahoo.com/infO/misc/historv.html.

Ziebland, S., Chapple, A., Dumelow, C., Evans, J., Prinjha, S., & Rozmovits, L.

(2004). How the lntemet affects patients’ experience of cancer: A
qualitative study. British Medical Journal, 328(7439), 564.

152

 

   

lIllilllllllllll