University of Southern California Dissertations and Theses

Social support in Cambodia: the role of peer educators in behavior change

(USC Thesis Other)

Social support in Cambodia: the role of peer educators in behavior change

PDF

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content Running head: SOCIAL SUPPORT IN CAMBODIA 1

Social Support in Cambodia:
The Role of Peer Educators in Behavior Change
Katherine M. Pieper

A Dissertation Presented to the
Faculty of the USC Graduate School
University of Southern California
In Partial Fulfillment of the
Requirements for the Degree
DOCTOR OF PHILOSOPHY
(COMMUNICATION)

SOCIAL SUPPORT IN CAMBODIA 2

Table of Contents
Abstract ............................................................................................................................................3
Acknowledgements ..........................................................................................................................4
Chapter 1: Review of Literature ......................................................................................................8
Chapter 2: Methods ........................................................................................................................37
Chapter 3: Results ..........................................................................................................................48
Chapter 4: Discussion ....................................................................................................................63
References ......................................................................................................................................76
Endnotes .........................................................................................................................................89
Tables and Figures .........................................................................................................................91

SOCIAL SUPPORT IN CAMBODIA 3

Abstract
In developing countries, maternal and child health is often a primary area where health
systems fail. To address this need, organizations offer interventions to train community
members on disease prevention and effective health practices. This study assesses the results of
one such intervention among men and women in Cambodia. This study examines the role of
village volunteers in providing health training to residents of a single province in Cambodia.
The role of social support in these training sessions is investigated as a means of explaining the
adoption of new behaviors and acquisition of information. Study participants were men and
women (of childbearing age) in rural villages in Cambodia. Participants were interviewed
individually regarding their knowledge and behaviors related to a number of health outcomes.
Results demonstrated that peer educators are an important source of health information for their
community. The impact of social support is less clear. Findings are discussed in terms of cross-
cultural adaptation of measurement with attention to future intervention contexts. Self-efficacy
theory (Bandura, 1977) and the Health Belief Model (Janz & Becker, 1984; Stretcher &
Rosenstock, 1997) are used as a framework for understanding how efficacy support may
contribute to health outcomes.

SOCIAL SUPPORT IN CAMBODIA 4

Acknowledgements
It is fitting that a dissertation about social support has involved a number of people who
provided guidance and encouragement. Completing this dissertation is the culmination of a
nearly ten-year process, a journey that took me places I never expected and taught me more than
I could have hoped.
To my dissertation committee, Dr. Stacy Smith and Dr. Michael Cody of the Annenberg
School for Communication and Journalism, and Dr. JoAnn Farver of the Department of
Psychology at the University of Southern California, I am humbled and so grateful that despite
numerous delays, you were willing and gracious enough to see this project to its conclusion. Dr.
Cody and Dr. Farver were particularly generous and flexible with their time, which contributed
mightily to my success.
Dr. Stacy Smith deserves mention, not only for advising me through the dissertation
process, but for thirteen years of guidance—beyond just methods, statistics and research. Stacy
has offered so many opportunities to learn first-hand and always accepted failures with grace and
successes with joy. For her professional advisement, I am profoundly grateful. Even more so,
through her friendship, support, and encouragement, I have been truly blessed. Thank you, Stacy.
The administration of the Annenberg School for Communication and the Ph.D. program,
in particular Dr. Larry Gross and Dr. Tom Goodnight, have been supportive in both tangible and
intangible ways, which made it possible for me to complete this degree program despite
considerable interruption. Similarly, Anne Marie Campian has been an extremely important
resource and advocate. I am truly grateful for their help.
My dissertation would not have been possible to complete without the support of
numerous colleagues and friends at World Relief. Thanks are due to several individuals at
SOCIAL SUPPORT IN CAMBODIA 5

World Relief Baltimore: Stephan Baumann, who encouraged the use of this data set, and Melanie
Morrow and Rachel Hower who designed the evaluation itself and allowed me to include several
measures on social support. Even more thanks are owed to those at World Relief Cambodia:
Tim Amstutz for his early support of this project; Geof Bowman for allowing my ongoing
involvement; Sivan Oun for her leadership; Bunthouen Phan for translation and a place to stay in
Kampong Cham; Vichet Lor for help designing measures and for providing raw data; Engchy
Kin for translation assistance; SreyLeack Heng for assistance with data entry; and to so many
others from both the Sokhapheap Phum Youeng project and World Relief more broadly who
gave their time, expertise and energy to making this project possible.
Others working in Cambodia or affiliated with World Relief deserve mention as well. To
Katie Williams, who provided necessary work release and also moral support and friendship,
Kerstin Pless, who gave my belongings a place to live, and Larissa Peters for making sure I
could take time away without worrying, thank you all. To the men and women of the Cambodia
Partnership Team: Ron Sadlow, Matt and Mindy Gandy, Marganne Pearce, Mike Newkum,
Lynn Jue, Steve Beirn, Jim and Elva Weaver, Nate Irwin, Joshua Harber, Jay Clark, Holly
McCallum, and others I may not have mentioned, thank you for your generosity and your
support of my personal goals. I am so grateful for your prayers, your friendship and your
kindness to me over the past five years. To Scott White at Lake Avenue Church, a special thank
you for allowing this germ of an idea to sprout into a full grown project so many years down the
line. Your wisdom and support throughout this process have been essential.
To my Cambodian family, Joke van Opstal, Jonathan, Elisa, Luka, Zoe, Anna, Sara and
Hedia, thank you for welcoming me into your home and lives and for your support and love
throughout this process. My Cambodian brothers and sisters, including Bunnath Nop, Nareth
SOCIAL SUPPORT IN CAMBODIA 6

Peang, Sokha Min, Romroth Choun, Sopheap Kim, Uddom Kim, and so many more I cannot list,
are owed thanks for their kindness and care for me, and for teaching me more about the
Cambodian culture and way of life than I could describe in this document. I am blessed to know
you.
More personally, my family has been willing to give and sacrifice in so many ways to
help me reach my goals. My gratitude to and for them is immeasurable. My parents, Richard and
Laura Pieper, never stopped believing that I could and would complete this project, and their
faith—even when I was unsure, unwilling, or on a different continent—has been a constant
source of amazement and encouragement to me. I could not have finished this project and this
degree without their love and, occasionally, their stubborn insistence that I get to work. My
grandmother, Susan Kruse, has been unwavering in her love and assistance, and unfailingly
generous to me. My siblings, Phil Pieper and Liz and Dan Waldron have also provided
encouragement and needed humor in this process, and Victoria and Natalie Waldron have added
support in their own unique way.
The friends who have supported me on this journey are numerous. My fellow USC
students and friends, Marcia Dawkins, Elaine Chan, Becky Herr-Stephenson, and Jeff Hall, who
began this process with me and eventually proved that it is possible to graduate. Marc Choueiti ,
Amy Granados, Rebecca Shapiro and Sarah Erickson helped me learn to work in a team and sent
me to Cambodia with excitement and enthusiasm. Katie Dunn, Grace Lee Ollerenshaw, Amber
Sutton, Barbra Bowman, Cyndi Older, Darla Lansu, and others, who prayed and believed for me,
even when I couldn’t. To Dr. Susannah Fisher, who got there first and cheered me from the
finish line, I look forward to our next milestones. Ravi Prakash and Erisa Hines fed me, cheered
SOCIAL SUPPORT IN CAMBODIA 7

for me, laughed at me and made sure I took breaks. All of these people, and so many more who
walked with me through this process have my gratitude, if not a special mention here.
To the One who set my feet on this path, who journeyed with me, and of whom I can say:
I am “confident of this, that He who began a good work in you will carry it to completion until
the day of Christ Jesus”: to Him be all glory and all honor and all praise. Amen and amen.
SOCIAL SUPPORT IN CAMBODIA 8

Social Support in Cambodia: The Role of Peer Educators in Behavior Change
Chapter 1: Review of Literature
Kofi Annan, in his farewell speech to the United Nations, argued: “We have to give our
fellow citizens, not only within each nation but in the global community, at least a chance to
share in our prosperity” (BBC.com, December 11, 2006, ¶29). By doing so, Annan reaffirmed
the UN Millennium Development Goals (MDGs), originally set in 2000. The Millenium
Development Goals included, among others, reducing child and maternal mortality by the year
2015. In 2006, nearly 9.7 million children died before their fifth birthdays, and over half a
million women die each year due to pregnancy-related causes (UNICEF, 2007). Some of the
poorest nations in the world are also most at risk for women’s and children’s premature deaths
(UN “Investing in Development Report,” 2005). Health promotion activities have served to help
reduce mortality rates, but there is more work to be done. The purpose of this study is to examine
to what extent the use of peer educators in a maternal and child health intervention affected
knowledge and behaviors that contribute to healthier, longer-lived women and children.
This study concerns a health intervention that took place over three years in Cambodia.
Ranked 136th on the UN Human Development Index (HDI) for 2008, this small Southeast Asian
nation receives low rankings across nearly all development indicators. Political turmoil
throughout the latter half of the 20
th
Century is primarily to blame for such grim numbers, which
reveal that Cambodians still face crippling poverty. Twenty-six percent of the population earns
less than $1.25 per day, placing them below the generally accepted global poverty line
(UNICEF, 2011). Rural areas are often the most impoverished, and as urban dwellers represent
only around 22% of the population of over 14 million (CIA Factbook, 2011), many citizens face
hardship. Cambodia still has a high under-5 child mortality rate (88 per 1,000 live births;
SOCIAL SUPPORT IN CAMBODIA 9

UNICEF, 2011), and ranks among the worst countries for child health in the world (UNICEF,
2007). Maternal health is similarly dismal; 426 out of 100,000 mothers perish after the birth of
their children (WHO, 2011). In addition to these statistics, Cambodia has an adult HIV
prevalence rate of 0.8%; approximately 75,000 people are living with HIV in this small nation
(CIA Factbook, 2009).
A number of international organizations are working in Cambodia to reverse these trends,
but must overcome significant barriers. For one, the spread of disease may be exacerbated by the
influence of traditional Khmer culture which survives to this day. Contemporary values and
beliefs (Ledgerwood, 1994), including a variety of uninformed childrearing and risky sexual
practices, are still informed by religion and culture. Though ancient Buddhist rules and
conventions may not guide every aspect of present-day behavior, they are part of a moral code
that is instilled in varying degrees upon young Khmer women by their elders (Ledgerwood,
1994). At the core of these teachings is the idea that a woman must respect and serve her
husband, and Khmer culture emphasizes a woman’s primary obligation as caretaker and
homemaker (Frieson, 2001; Derks, 2008). Despite changing circumstances, migratory labor
opportunities, and the prospect of higher education, women are still encouraged to retain such
traditional beliefs (Derks, 2008). Such a view often affects educational attainment, as girls are
more often removed from school to care for younger siblings than boys (Gorman, Dorina, &
Kheng, 1999; Fordham, 2003). Practically speaking, following traditional beliefs often places the
burden of caring for their family’s health on females, while at the same time, women must also
be attentive to their own health, though they may be ill-equipped to do so.
A second barrier to the success of interventions is that promising techniques used to reach
the population, such as community participation, may require considerable investment of time
SOCIAL SUPPORT IN CAMBODIA 10

and resources. NGOs recognize that a fundamental aspect of the success of their work is
achieving a sense of community ownership over the programs they implement (Jacobs & Price,
2003). Community participation is a process which addresses health needs, and involves the
work of individuals and community members to collectively assess their problems and organize
to create and implement solutions, including monitoring and evaluation of the strategies used
(Zakus & Lysack, 1998, p. 2). One method of spreading health information in local communities
is through the use of peer educators, whose methods may consist of “informal tutoring in
unstructured settings, one-to-one discussions and counseling” (Turner & Shepherd, 1999, p.
236). These personal contacts are thought to be uniquely suited to training community members
because they are often more similar across a range of attributes (i.e., gender, age, ethnicity,
education) than an expert might be (Albarracín, Albarracín, & Durantini, 2008; Durantini,
Albarracín, Mitchell, Earl, & Gillette, 2006). Peer educators, or as they are sometimes called,
home visitors (Olds & Kitzman, 1993) or lay health advisors (Earp & Flax, 1999), are a common
means of information dissemination across multiple health topics, and may even provide more
than simply health information to intended recipients—they engage in supportive relationships
which promote behavior change.
These relationships are a prime example of the well-established and complex link
between social ties and health (Cohen, Gottlieb, & Underwood, 2000; Mitchell & Hurley, 1981).
A broader concept, termed social support, is commonly used to describe the process through
which some intangible, protective, or helpful resource is provided via social contact (Cohen,
Gottlieb, & Underwood, 2000). In fact, social support has a well-documented relationship to
individual health. Social support and social integration in developed nations are positively
related to overall mortality and, less consistently, negatively related to morbidity of illnesses
SOCIAL SUPPORT IN CAMBODIA 11

such as coronary artery disease and myocardial infarction (Berkman, 1984; Cohen, 1988;
Wallston, Alagna, DeVellis, & DeVellis, 1983). Social support has been demonstrated to have
important effects on individual health behaviors, including decrease in disease outcomes (i.e.,
diabetes, hypertension, kidney disease) among elderly individuals (Tomaka, Thompson, &
Palacios, 2006), and on infant birth weight (Feldman, Dunkel-Schetter, Sandman, & Wadhwa,
2000). Enhanced social circumstances have even been associated with better mental health
scores (Williams, Ware, & Donald, 1981), and particularly with the experience of stressful life
events (Lin, Ensel, Simeone, & Kuo, 1979). However, these relationships are drawn from large
population-based studies which are correlational in nature, and may be vulnerable to spurious
variables or time order confusion. According to Tomaka et al. (2006), what is clear from
previous research is that social ties have a general influence on health outcomes. Thus, while it is
difficult to know the precise mechanism whereby social support impacts health, it is clear that
relationships play a role in maintaining well-being.
The present investigation addresses the role of social support in the success of a
preventative health intervention. Utilizing a unique sample of rural Cambodian villagers, the
study measures self-perceptions of received social support alongside indicators used to evaluate a
three year maternal and child health project. In 2007, World Relief, a faith-based non-profit
headquartered in the U.S., received large scale funding from the U.S. Agency for International
Development (USAID) to expand an existing maternal and child health outreach program from
one provincial district to a total of six districts in the same Cambodian province. World Relief’s
program, called Sokhapheap Phum Yoeung (Our Healthy Village), or SPY, recruited men and
women from local villages to serve as peer educators. These individuals were part of “care
groups” in which they learned health information and were charged with visiting and teaching
SOCIAL SUPPORT IN CAMBODIA 12

the lessons they learned to households with members within target demographic range:
households with children under age five or men and women aged 15 to 49.
The goal of this study is to better understand how the elements of social support
incorporated into the SPY program are related to gains in health information and behavior
change evidenced by those living in the areas where the intervention took place. Social support
was not measured until the final project review and cannot be compared with initial levels of
support. Therefore, this study is not designed to assess whether the project had an impact; rather,
the aim is to determine the effect of social support on knowledge gain and behavior change in
regard to different health outcomes at the end of the intervention. Of particular interest is how
different types of social support may influence individual’s knowledge gain and behavioral
outcomes for particular intervention topics. Two further goals of this study are first, to provide
guidance for future interventions on how best to incorporate principles of social support into
program design for maximum community impact and, second, to test cross-cultural measures of
social support in a field setting.
Social support measures are not new, and have been used since at least the 1970s to
describe and evaluate the role of relationships in mental and physical well-being (Cohen &
Syme, 1985). Yet, there is still not a great deal of standardization in the measurement of the
concept (Dean, et al., 1994). Social support is typically regarded as a multidimensional construct
(Newcomb, 1990; Wallston, et al., 1983), and at a global level, can be viewed as “one aspect of a
general beneficial human propensity of personal attachment and reciprocal social exchange”
(Newcomb, 1990, p. 485). In relation to health, it can also be broadly viewed as “referring to any
process through which social relationships might promote health and well-being” (Cohen,
Gottlieb, & Underwood, 2000, p. 4). Social support occurs across the life span in different
SOCIAL SUPPORT IN CAMBODIA 13

capacities, in infancy and childhood with parents, in the teen years with peers, and as adults with
a variety of different sources (Newcomb, 1990).
One of the major conceptual issues surrounding the study of social support is the nature
of the construct. Researchers often delineate the construct as perceived, received, or even
intended (Aaronson, 1989; Cutrona, Cohen, & Igram, 1990; Dunkel-Schetter & Skokan, 1990;
Schaefer et al., 1981). For instance, Schaefer et al. (1981) define perceived social support as “an
evaluation or appraisal of whether and to what extent an interaction or pattern of interactions, or
relationship is helpful” (p. 384). Such conceptual discrepancies may have distinct effects on
health or health behaviors, as they capture different aspects of the interactions in which social
support is offered, provided, or accepted. It is clear, though, that social support can be viewed as
the product of a transaction (Berkman, 1984; Dunkel-Schetter & Skokan, 1990; Sarason, Pierce,
& Sarason, 1990), and as a resource that is provided or received by the participants in that
interaction (Hobfoll, Freedy, Lane, & Geller, 1990). Ultimately, social support is rarely
something easily quantified within a social interaction (Sarason, Pierce, & Sarason, 1990), and
instead reflects the degree of correspondence between what one party needs and what another
party provides.
Some researchers have considered social support as both a characteristic and function of
social networks, which are distinct from the concept of social support (Berkman, 1984; Cohen,
1988; Schaefer, Coyne, & Lazarus, 1981). Social networks can be viewed as the vehicle by
which social support is disseminated, and also as a structure for the flow of information.
Networks provide opportunities and social or other resources for interventions. The type of
health information presented within a network can either enhance or deter good health (Ertel,
Glymour, & Berkman, 2009). Similarly, Cohen (1988) describes social integration, which is “a
SOCIAL SUPPORT IN CAMBODIA 14

structural index of social ties” (p. 271), differentiated from social networks by the lack of formal
measures of network composition (i.e., density, size, multiplexity, etc.) and focused on the type
of ties that exist in the network (i.e., marital status, family and friends, group participation, and
church/religious attendance; Cohen, 1988, p. 271). For practical reasons, the present study does
not measure the structure or function of social networks among respondents, and instead focuses
on a single source of social support and its influence on individual health behaviors.
In general, social support occurs in three contexts, often in response to specific needs:
situational, intrapersonal, and interpersonal, with a variety of factors influencing the degree of
support occurring in each case (Cohen,1988; Sarason, Pierce, & Sarason, 1990). Most models
imply that social support is helpful to individuals because it enhances or facilitates the
environmental and psychological processes which help solve problems or relieve emotional
disturbance (Cutrona, 1990, p.4). However, researchers generally agree that, while the receipt of
support is important, and that matching support to need may increase favorable outcomes
(Cutrona, 1990), simply identifying that support is available within the network, even if it is not
received, is advantageous for individuals (Cohen, Underwood, & Gottlieb, 2000).
Although general social support is beneficial for overall well-being, individuals have
different support needs, and varying situations may require alternate types of support. Several
personality, situational, and individual difference variables impact both the kind of support
received and the outcome of such support. For instance, Sarason, Sarason, and Shearin (1986)
found that general social support can function as a trait-like variable. In particular, the number
of supportive relationships a person perceives generally remains consistent, even over the course
of several years. The authors note that the stability of these relationships is likely due to co-
SOCIAL SUPPORT IN CAMBODIA 15

occurring personality traits which make those who are high in this type of general social support
fundamentally different from people low in social support.
Other studies confirm that personality plays a role. Lu (1995) found that among adults
living in Taiwan, gender, locus of control and social resources had direct impacts on social
support receipt, such that younger, better educated women reported receiving more social
support. Similar, and derivative of, locus of control, the concept of mastery is understood as the
degree to which a person’s life chances are perceived to fall within one’s locus of control
(Hobfoll, Shoham, & Ritter, 1991, p. 333). In a study of low-income African-Americans in
Pennsylvania, mastery at initial testing was related to perceptions that more tangible and advice
support was available a year later. Additionally, those women who perceived higher levels of
available tangible social support at the study’s outset had greater levels of mastery at follow-up,
suggesting that learning to take advantage of offered support impacts later confidence in seeking
help (Green & Rodgers, 2001). Hobfoll et al. (1991) found that among women in Israel, mastery
interacted with stressful situations in the receipt of social support. The pregnant women who
were high in mastery received less social support following a low stress event (vaginal delivery)
than low mastery women. When experiencing a high stress event (Caesarean delivery), high
mastery women received more social support than low mastery women in the same condition,
suggesting that women with high mastery receive support when they need it most. As a personal
resource, feelings of control over one’s life can impact how individuals respond to the helpful
actions of others, particularly in high stress situations. Although personality measures were not
measured in the current investigation, the studies above are instructive for thinking about how
different types of social support may be related to health outcomes.
SOCIAL SUPPORT IN CAMBODIA 16

Aside from personality and situational variables, demographics also matter for support
receipt. Gender, in particular, has a consistent effect on the nature and type of social support
individuals receive, and it has differential impacts on health. Antonucci and Akiyama (1987)
found that elderly women in the U.S. reported discussing health among their social networks
more often than men. In a large-scale panel study conducted in the U.S., Umberson, Chen,
House, Hopkins, and Slaten (1996) found that women report receiving more social support from
their adult children, and score higher than men on social support and informal social integration.
In their review of large scale studies regarding social relationships and overall mortality, House,
Landis, and Umberson (1988) noted that across multiple large surveys in both the U.S. and
Sweden, the impact of social relationships is often more significant for men’s health, particularly
mortality. In fact, in small communities, social isolation (or a very low level of social
relationships) has serious consequences for health and general mortality for men. This
corresponds with Thoits’(1995) review; women have a greater propensity to seek out social
support than men do, report more perceived social support than men, and, like men, appear to
benefit from relationships with other women (House, Landis, & Umberson, 1988). While not all
examinations of general morbidity and mortality find a relationship between social support,
gender, and health, many early studies did not analyze gender differences, or even include
women (Shumaker & Hill, 1991). Aside from all-cause mortality, Kaplan and Hartwell (1987)
suggest that norms and support related to health behaviors may differ by gender. Women’s
reliance on multiple sources of social support may be one explanation for the differential impact
of disease outcomes for men and women.
A related factor is marital status. For men, simply having a spouse is the greatest single
influence on the quality of support reported (Antonucci & Akiyama, 1987; House, Landis, &
SOCIAL SUPPORT IN CAMBODIA 17

Umberson, 1988), and men report being more satisfied with levels of support provided through
marriage. This is not to say that women’s satisfaction with social support is unrelated to being
married (Vaux & Harrison, 1985), but women rely more than men on same-sex friends and
relatives for support. Shumaker and Hill (1991) remark on the highly consistent negative
relationship evidenced between mortality and morbidity and marital status for white men across
multiple large studies, which is not replicated among women. One explanation for this may be
that gender roles typically encourage women more than men to care for their spouse’s health.
Across a large, diverse sample in the U.S., married men were more likely than unmarried men to
report that other people attempt to regulate their health (Umberson et al., 1996). Among elderly
Nepalese men, social support from a spouse improved subjective well-being (though this is not
an explicit measure of health), but it is likely that women underreported social support for
cultural reasons (Chalise et al., 2007, p. 310). However, in some cases, such as post-partum diet
and exercise, support from husbands is an important and consistent influence on enacting healthy
behaviors for females (Thornton et al., 2006). For example, marital status indirectly affects infant
birth weight via social support (Feldman et al., 2000). In her review article, Thoits (1995) states
that “the simplest and most powerful measure of social support appears to be whether a person
has an intimate, confiding relationship or not” (p. 64). This relationship is most typically a
spouse or significant other, though friends and relations may serve as a less powerful source of
this type of support.
An individuals’ socio-economic status clearly impacts their health directly as well as via
social support. In certain domains, interventions have had the largest impact among some of the
poorest individuals (Olds & Kitzman, 1993). Israel et al. (2002) found that among low-income
women in Detroit, higher levels of instrumental support (i.e., providing tangible assistance of
SOCIAL SUPPORT IN CAMBODIA 18

money, transportation and child care) predicted better health outcomes over and above the
impact of chronic stress. While emotional support also had beneficial effects on health, Israel et
al. (2002) found that only at medium levels and only in the absence of instrumental support was
this effect noticeable. From this, the researchers concluded that for mothers facing economic
strain, instrumental support offers protection against the negative outcomes of stress (p. 356).
Culture also plays a role in how individuals experience and provide social support to
others. Asians, in particular, may hold different views of the self than other groups (Kim,
Sherman, Ko & Taylor, 2006; Morling, Kitayama, & Miyamoto, 2003; Taylor, et al., 2004).
Asian cultural norms which value interdependency and relationships over self-interest may mean
that individuals put their own needs after those of the collective. In fact, one reason Asians or
Asian-Americans give for not seeking social support is the fear of upsetting group harmony or
losing face (Kim, Sherman, Ko, & Taylor, 2006; Kim, Sherman, & Taylor, 2008; Taylor et al.,
2004; Taylor et al., 2007). As such, the cultural context in which a supportive interaction occurs
will impact how socially supportive behaviors or communications are received, or when they are
offered (Chalise et al., 2007; Kim, Sherman, & Taylor, 2008).
Taylor et al. (2007) found that Asians (i.e., Korean, Japanese, and Chinese) use social
support for coping with problems in different ways than Western cultures. Namely, Asians
evidenced a preference to seek out social support which did not require them to self-disclose
their problems. In their study, Taylor et al. (2007) explored whether seeking implicit support or
explicit support produced more anxiety in participants. Asians and Asian-Americans experienced
more distress, (including increased cortisol responses) when they were asked to seek explicit
social support from peers with whom they had a close relationship rather than when they sought
implicit support, while European-Americans in the sample evidenced no differences. Moreover,
SOCIAL SUPPORT IN CAMBODIA 19

Asians are less likely to seek out social support in general and report relying less on emotional
support than European Americans (Taylor et al., 2004). However, these studies do not consider
that unique Asian cultures (i.e., Cambodian) may express social support differently. Taken
together, these findings suggest that the kind of social support that peer educators in the SPY
program provided fit the characteristics Asian individuals typically find most helpful: it was
implicitly offered, and did not require them to directly disclose the problem for which they were
seeking support. Yet, it is difficult to know precisely what the Cambodian experience of social
support may be, and whether the findings from studies that broadly characterize Asian
respondents may generalize to this specific sample.
The aforementioned studies measured or manipulated individual factors which might
cause a person to seek social support from their social network. Aside from the use of perceived,
existing support from relationships, researchers have also sought to increase the types and
amount of social support provided to individuals through their networks, via a number of social
support interventions (Cohen, Underwood, & Gottlieb, 2000). Particularly for groups with weak
networks or other environmental stressors, augmenting social support may have beneficial
effects on health. Gottlieb (1981) argues that new sources of support will be necessary when an
individual’s current social contacts reinforce unhealthy behaviors, such as in the case of
alcoholics (p. 223). One important consideration in the development of social support
interventions is sociocultural context (Gottlieb, 1981). Others have stated that in order to be
successful, researchers must stipulate what they hope to achieve through their work (Mitchell &
Hurley, 1981, p. 284). Previous interventions have attempted to bolster the helping influence of
already-existing sources of aid (i.e., hairdressers, community workers, etc.). Of interest in the
SOCIAL SUPPORT IN CAMBODIA 20

present study is how new sources of social support introduced into networks as knowledgeable
advisors or peer educators may help to influence health outcomes.
Combining the established benefits of social support for health behavior change with the
strengths of peer educators should therefore be particularly effective. Berkman (1984) describes
potential pathways through which social networks and their functions (such as social support)
can impact overall health, particularly morbidity and mortality. Social contacts may provide
advice, services, or access to medical care, or communicate positive values related to medical
care. They may provide care directly to individuals in a network, or other tangible assistance to
promote improved or good health. Finally, social networks and social contacts may transmit or
reinforce norms related to health. In some cases, these may include negative or risky health
behaviors (i.e., smoking, alcohol use), but could also promote positive behaviors (i.e., condom
use, good nutrition) that members would enact to maintain their group identity (Berkman, 1984,
p. 428-429). To the extent that peer educators utilize social networks or their social position in
these ways, it is realistic to expect that these individuals will have positive effects on individual
health knowledge or behavior.
Peer education in the home also has benefits. Home visits provide health workers or
volunteers with insight into the context of a woman’s health-related behaviors (Olds & Kitzman,
1993). They also promote a longstanding and dependable relationship which can act as a conduit
for social support, and may include sharing information, changing existing behaviors, or linking
individuals with available health services (Eckenrode & Hamilton, 2000). Israel (1985) describes
characteristics of individuals in the social network which are useful for health interventions—
such as those in the present investigation. Labeled ‘natural helpers’, these are people “who are
respected and trusted, and who listen well and are empathic, sufficiently in control of their own
SOCIAL SUPPORT IN CAMBODIA 21

life circumstances, and responsive to the needs of others” (Israel, 1985, p. 68). Israel (1985)
argues that interventions should not only increase the number of “meaningful relationships” (p.
70) between the natural helper and others in the network, but that relationships between network
members should also increase for interventions to work. Additionally, rather than emphasizing
what services individuals can access, or only interpersonal or self-help solutions, interventions
using natural helpers should focus on the root causes of problems in the community, and if
necessary, promote changes in service provision (p.75). In this way, social support can have
broad impact in the wider community as well as help create individual behavior change.
General health can be bolstered through home-based support. Earp, Ory, and Strogatz
(1982) found that after a year, hypertensive individuals who received in-home visits from a
medical practitioner maintained better diastolic blood pressure than those who received regular
care through clinical visits. Additional social support in the form of family members (or other
supportive individuals) was not significantly more effective in reducing blood pressure than
visits by health professionals alone. Previous interventions in the area of home visiting for
pregnant women and parents of low birth weight or pre-term infants demonstrate the
effectiveness of some of these ideas. Although a variety of outcome measures were used (i.e.,
children’s cognitive development, physical health, maternal care giving, pre-term labor, low
birth weight, health care utilization), the effects of home visits appear to be greatest for teenage
parents, and for those who are unmarried or poor (Olds & Kitzman, 1993; Norr et al., 2003).
Interventions which provided only social support were the least effective, leading the authors to
suggest that the positive health benefits of social support are most evident when support is paired
with an emphasis on health-related behaviors (Olds & Kitzman, 1993, p. 64). In fact, the most
successful programs are the most comprehensive, addressing multiple needs of low-income, at-
SOCIAL SUPPORT IN CAMBODIA 22

risk families, and targeting specific behaviors for change (Olds & Kitzman, 1993; Ramey &
Ramey, 1993). Eckenrode and Hamilton (2000) observe that beyond their life circumstances,
families must have indicated a particular need, or understand the reason they receive home visits
in order for the successful development and maintenance of the home visiting relationship.
Social support interventions have been demonstrably effective at bolstering healthy
outcomes. Despite the accumulated research, very few of these studies have considered evidence
from large scale interventions conducted outside the U.S. Moreover, they have particularly
failed to examine how Asian or collectivist cultures respond to social support interventions (i.e.,
Israel et al., 2002; Olds & Kitzman, 1993). Though Kaniasty and Norris (2000) found that
individuals from a more collectivistic cultural orientation may be more likely to offer or take
advantage of social support in their community for individual problems, their results are more
applicable to post-emergency situations (i.e., following a hurricane) and stressful life events (i.e.,
undesirable or traumatic incidents in the past twelve months) than to health behavior change or
knowledge acquisition through peer educators. House, Landis, and Umberson (1988) note that
across multiple prospective studies in the U.S. and Sweden, smaller rural communities exhibit
stronger and more pervasive social support, but this finding has not been extended to developing
countries. Neither study specifically examined how Asian cultural norms might impact support
exchanges.
The health promotion literature demonstrates that peer education and home visiting have
effectively impacted individual health behaviors in interventions outside the U.S. In his study of
family planning behavior adoption among village women in Bangladesh, Kincaid (2000)
compared the impact of home visits versus group meetings on family planning attitudes and use
of modern contraceptive methods. Trained health workers either met with women in one-on-one
SOCIAL SUPPORT IN CAMBODIA 23

interaction, in groups, or not at all. Across all dependent measures, women who received any
kind of intervention (home visit or group meeting) were more likely to use modern contraceptive
methods than women who received no information. After two years of meetings, women who
took part in group meetings had a significantly greater rate of adoption of modern contraceptive
methods, significantly more knowledge, and significantly more discussion of family planning
with their husbands than women who received individual home visits. In fact, the rate of change
to modern contraceptives was five times greater for women who met with a trainer as part of a
group rather than individually. The rate of change for women who received no training was
negative (Kincaid, 2000). It is clear from this one study that peer education and social group
meetings can be a valuable means of increasing family planning behaviors among rural women.
Aside from contraception, the use of village health volunteers for behavior change has
been effective across multiple intervention topics. Interpersonal communication through village
volunteers in rural Bangladeshi villages was responsible for a rise in consumption of vegetables
high in vitamin A (Hussain, Aaro, & Kvale, 1997). In Zimbabwe, Waterkeyn and Cairncross
(2005) evaluated the impact of health clubs on sanitation and hygiene improvement in rural
areas. The health clubs had a significant impact on hygiene, with club members reporting
increases in the number of nutrition gardens in villages, more covered fecal disposal, and
improvement in hygienic drinking water practices. Kelly (2004) and his colleagues (Kelly et al.,
1991, 1992, 1997) have also demonstrated that peer educators who are highly similar and
popular have an important influence on altering risk behaviors, particularly in communities at
highest risk for HIV. Though the majority of these studies are from the U.S., meta-analyses
confirm that similarity on some dimension (i.e., age, gender, ethnicity) is persuasive in the area
of global HIV and AIDS behavior change; similar peers have a greater impact than more diverse
SOCIAL SUPPORT IN CAMBODIA 24

sources of information (Albarracin et al., 2003). Theories of social cognition also demonstrate
that peers and similar others serve as models for behavior, with Bandura (1990) lending strong
support for this method of intervention.
The health status of rural villages is also affected by the knowledge and norms that
residents hold and their ability to receive information from external sources. Goldman, Pebley,
and Beckett (2001) demonstrated that women in Guatemalan villages were more likely to believe
that poor hygiene was the source of diarrhea in children when they had more interpersonal
contacts outside their village or were in communication with more outside contacts. Even
informal contact in certain contexts can serve to promote or counteract health norms in rural
areas. An additional factor that contributes to the health of women and children is social capital.
In a study of women in Ethiopia, low social capital (i.e., ability to borrow money, participation in
village leadership committees, and membership in community organizations) was associated
with higher under-five mortality (Fantahun et al., 2007). Perhaps serving as community health
educators or feeling that a health intervention or group is a strong relational connection will help
women to feel that they are better able to improve their family’s health.
Evidence suggests that cultural differences impact the degree to which social support is
effective for individuals. Cambodian religious beliefs and cultural orientation are in line with
collectivist values of maintaining harmony in personal relationships and are cognizant of the
needs of others (Derks, 2008; Ledgerwood, Ebihara, & Mortland, 1994), all of which have been
noted as having an impact on social support seeking (Kim, et al., 2006; Kim, Sherman, & Taylor,
2008; Taylor et al., 2004, 2007). It is reasonable to expect that Cambodian villagers may already
be reliant upon social support in their community and particularly open to support from peers in
regard to health behaviors. However, Cambodians may be unlikely to request support for health
SOCIAL SUPPORT IN CAMBODIA 25

needs for fear of losing face or disrupting group harmony. For instance, in Nepal, Chalise et al.
(2007) indicated that elderly women in their sample may have underreported the amount of
social support they provided to their husbands due to religious beliefs or cultural taboos (p. 310).
In previous studies (Kim, et al., 2006; Kim, Sherman, & Taylor, 2008; Taylor et al., 2004, 2007),
researchers have focused primarily on support seeking, and not how receipt of social support
might impact health behaviors. Additionally, these studies were conducted primarily in Western
settings, or with college students in more developed Asian countries. Thus, although it is
possible to narrowly generalize these findings to certain aspects of this intervention, it is also
likely that a rural, cross-cultural, and economically disadvantaged population might view social
support in different and meaningful ways.
The present investigation aims to bridge the gap between specific health interventions in
least developed countries and the broad social support literature that demonstrates the beneficial
impact of social relationships on health in population-based studies in most developed nations.
The inclusion of measures of social support in an otherwise health-focused survey provides a
means to examine which elements may be of particular use in designing and implementing health
interventions. Utilizing peer educators (or lay health advisors) replicates previous work on the
effectiveness of similar and knowledgeable health professionals, but also attempts to specify
what is important about the social aspects of these community volunteers. This study also
extends the application of social support measures to the development literature by incorporating
simple measures of social support in the context of a program evaluation. Moreover, the cross-
cultural setting illuminates how non-Western participants in a close-knit social network may rely
on social support for health information.
SOCIAL SUPPORT IN CAMBODIA 26

Three distinct elements of social support are of interest in this investigation (Schwarzer,
Dunkel-Schetter, & Kemeny, 1994). Others have argued that, just as social support is
multidimensional in nature, so is the concept of health, and different types of support may
influence specific health outcomes in different ways (Mitchell & Hurley, 1981). Additionally, for
one-to-one health interventions, it is important to understand what types of social support are
exchanged, and how clients perceive their interactions with project staff, and whether the
provision of support has a broader impact on the social network (Eckenrode & Hamilton, 2000,
p. 264). Additionally, network members who provide support may be more effective if they can
offer a wide range of support in a way that is sensitive to the receiver’s needs and feelings
(Cutrona, 1990).
As a general measure of support, perceived social support may be received more when an
individual’s own health is threatened, rather than that of a family member (Dunkel-Schetter,
Folkman, & Lazarus, 1987). For instance, expressions of love and support were rated as the best
form of support among chronically ill patients (Lehman & Hemphill, 1990). In other cases it
does not appear to impact health behavior (Aaronson, 1989). Yet, according to Schaefer et al.,
(1981), perceived social support can directly measure the amount of actual help given to a
person, and can link the level of support to outcomes, particularly in regard to health. When
general social support is considered as frequency of contact with individuals in the social
network, preventive health behavior for indirect risks (such as seeking medical care,
immunizations, and nutrition) is significantly predicted by the rate of interaction with non-kin
network members (Langlie, 1977). Men who reported having more individuals they could rely
on for social support were less likely to report practicing risky sexual behaviors (i.e., unprotected
intercourse; Ekstrand & Coates, 1990). Talking with a close friend every day significantly
SOCIAL SUPPORT IN CAMBODIA 27

impacted the likelihood that an adolescent would seek care for a sexually transmitted disease
(Lowery, Chung, & Ellen, 2005). Among Pakistani women, family support for post-natal child
care had a protective effect on mothers’ depression (Rahman, Iqbal & Harrington, 2003).
Wallston et al. (1983) suggest that it is not the strength or type of ties in a network that influence
utilizing health services, instead it is “the norms and values relevant to seeking care endorsed by
one’s network” (p. 376). Perceived social support may have a stronger link to health outcomes
than other conceptions of support. Hupcey (1998, p. 1234) notes that perceptions, though often
inaccurate, influence how satisfied individuals are with the support they receive, and the
outcome of this support.
A second dimension, instrumental support, has been operationalized elsewhere as
“tangible assistance or aid” (Dunkel-Schetter, Folkman, & Lazarus, 1987, p. 73), or as “direct
aid” (Schaefer et al., 1981, p. 385). Lack of tangible support has been related to depression and
negative morale (Schaefer et al., 1981). Elsewhere, students evaluating a situation in which
tangible support was offered viewed it as less helpful than emotional support (Dunkel-Schetter et
al., 1987). The type of situation might have influenced this effect; participants evaluated the
support offered to someone who had just received information about the death of a close relative.
However, in other situations, such as birth outcomes and child health, solely providing emotional
support may be insufficient to produce changes in behavior (Eckenrode & Hamilton, 2000). The
role of instrumental support in changing health behaviors may be of particular interest to
development workers or donors who must design interventions, and often decide how to
negotiate the delicate balance of whether to provide instrumental support and if so, what type.
Finally, a third dimension is efficacy support. When faced with implementing a new
behavior, self-efficacy is an individual’s “self-belief in one’s capabilities to use those skills well”
SOCIAL SUPPORT IN CAMBODIA 28

(p. 288). Anecdotal evidence from Kincaid’s (2000) investigation suggests that volunteers in a
family planning intervention experienced increased self-efficacy as a result of their participation.
Among college students, self-efficacy and perceived risk were significantly related to HIV
prevention behaviors (Goldman & Harlow,1993). From a theoretical perspective, Bandura
(2001) argues that an individuals’ perceived self-efficacy impacts the degree to which they will
alter their behavior (p. 288). Bandura (1977) suggests that the process of learning via feedback
and analysis from different events occurs over long periods of time. As part of this process,
individuals form two types of event-related expectancies: outcome and efficacy. In the former,
people understand how certain behaviors relate to the production of results. In the latter, they
assess their ability to execute the necessary actions to achieve the anticipated outcome.
Efficacy expectations are not static with regard to behavior, but can be enhanced via four
main categories of learning processes, three of which are relevant to this study: performance
accomplishments, vicarious experience, and verbal persuasion (Bandura, 1977, 1982). Although
people’s efficacy expectations can be altered through any of these pathways, each has unique
strengths and weaknesses. It is important to note, though, that each process can increase self-
efficacy, and that behavior corresponds to the level of self-efficacy change, regardless of the
method by which it is improved (Bandura, 1982). Performance accomplishments, with roots in
personal mastery, can help to establish strong feelings of self-efficacy after repeated success in
achieving desired outcomes. In the case of vicarious experience, or modeling, similarity of
models to observers can increase the personal relevance of information, and thus enhance the
impact of the symbolic experience (Kazdin, 1974). Verbal persuasion may impart the weakest
results in terms of efficacy expectations, as individuals may or may not believe what others tell
them in regard to their own abilities. In conjunction with receiving provisional aid, verbal
SOCIAL SUPPORT IN CAMBODIA 29

persuasion may encourage individuals to expend greater effort in achieving desired outcomes
(Bandura, 1977). In this investigation, this effect may be evidenced in relationships between
efficacy and tangible support.
Self-efficacy is of particular value when preventative health behaviors are considered.
Bandura (1990) writes that in order to protect themselves from HIV infection, people need to
understand how the virus is transmitted, need guidance on how to change behaviors, and to
believe that they have the ability to move from concern to action. This may include emphasizing
the recurrent nature of making self-protective choices, for example, when sexual or drug use
practices are in question. Social forces are a key factor in personal risk-taking, and interventions
for behavior change should utilize peers as models for change. According to Bandura (1990, p.
14), “…people judge their own capabilities, in part, from how well those whom they regard as
similar to themselves exercise control over situations.” Similar and indigenous peers can be
instrumental in helping others to develop stronger beliefs in their own abilities, particularly as
peers from one’s own network often have more sustainable effects than external influences (p.
15). For the present study, peer educators from an individual’s own village should be especially
important to the creation or strengthening of efficacy beliefs.
Concepts of self-efficacy and efficacy support are related to other conceptions of health
behavior and health promotion (Stretcher et al., 1986; Stretcher & Rosenstock, 1997). These
relationships are most evident in the Health Belief Model (HBM; Janz & Becker, 1984; Stretcher
& Rosenstock, 1997). Like Bandura’s (1977) theory, the HBM is based on expectation,
specifically value-expectancy concepts which guide behavior. Where value is considered as a
desire to avoid ill-health, expectation is the belief that certain actions can prevent that illness
(Stretcher & Rosenstock, 1997). In the HBM, three groups of determinants are hypothesized to
SOCIAL SUPPORT IN CAMBODIA 30

guide health behaviors. The first is the salience or relevance of health issues to individuals
(health concern). Second, the belief that one is susceptible or vulnerable to a serious health
problem (perceived threat) is important, because the perceived severity of that threat also entails
any number of social consequences (Stretcher & Rosenstock,1997). Finally, the belief that
perceived threats can be reduced by following a particular recommendation, with acceptable cost
to the individual, which may be financial or be represented by other barriers which must be
overcome (Rosenstock, Stretcher, & Becker, 1988). While early iterations of the model did not
include self-efficacy, Stretcher and Rosenstock (1997) now state that it is particularly relevant
for health issues which must be addressed through lifestyle changes. Rosenstock et al. (1988)
believe that the concept can be useful for exploring pathways for chronic illness or unhealthy
behaviors to change. Others (i.e., Janz & Becker, 1984) have conceived of self-efficacy as one
type of perceived barrier to engaging in health-related behaviors, as lack of self-efficacy may
impede behavior change. In the case of contraception usage, studies have found that perceptions
of confidence in ability to use contraception resulted in higher contraceptive use, and
interventions to improve efficacy in relation to family planning measures resulted in greater
intent to use contraception, though later use was not measured (Stretcher et al., 1986). There
appears to be great value in conceiving of self-efficacy, and by extension, efficacy support,
within the broader framework of the HBM, particularly in relation to long-term health behavior
changes.
Given the pervasive fatalism found in Cambodian culture, feelings of lack of control or
ability may influence how individuals evaluate and respond to support provided by health
educators. The Health Belief Model provides a framework for understanding the health decisions
of Cambodians, particularly in regard to perceived threats. One key aspect of the current
SOCIAL SUPPORT IN CAMBODIA 31

intervention was to make health threats more salient to those who might otherwise be ignorant or
unaware of their severity (i.e., diarrheal illnesses, exclusive breastfeeding). In addition, social
support from knowledgeable peers should be important for bolstering feelings of competence in
addressing family health problems. Ultimately, although self-efficacy was not explicitly
measured in this study, efficacy support measures should provide an indication of whether and
how much support from similar peers can help to alter health behaviors or knowledge in line
with the HBM and self-efficacy theory.
The social support measures in this study were incorporated into regular monitoring and
evaluation of intervention activities, specifically the final program evaluation. As a condition of
receiving funding, the SPY program was required to check its progress against health standards;
continued improvement guaranteed funding for the life of the intervention. The project was
evaluated on a yearly basis to comply with these requirements and performed sufficiently well to
retain grant monies for the three-year period. The present study, then, does not focus on the
overall impact of the project on health behaviors and knowledge gain, since its impact was
demonstrated by program managers and duly reported (World Relief, 2010). Instead, the
contribution of social support to the success of health education is considered. Hypotheses and
research questions were formulated with this relationship in mind.
In this study, the aspect of interest regarding social support provision is individuals’
perceptions of the amount of support provided to them by community health educators. Though
the project was designed to offer similar amounts of support to all community members,
individual differences among care group members (i.e., kin or pre-existing relationships; travel
distance; rejection of assistance, etc.) might have resulted in differences in perceived support
among recipients. During the final evaluation, a general measure of how supported or helpful
SOCIAL SUPPORT IN CAMBODIA 32

volunteers were to individuals was used to assess feelings regarding interactions between
educators and receivers. This study was considered ancillary to the process of determining
project impact for funding reports. As such, data collection was done using commonly accepted
procedures for large scale projects and not according to more stringent sampling methods.
Additionally, questionnaires were limited in length to ensure respondents did not need to provide
too much time to interviewers, who were tasked with completing the evaluation as quickly as
possible. While the study is able to advance hypotheses and research questions in line with
previous research, it should still be considered an exploratory investigation of the relationship
between social support in a large scale, cross-cultural health intervention.
Cambodian villages and social life vary with regard to migration and external
connections to diverse sources of information (i.e., other villages, district cities, Phnom Penh,
etc.; Derks, 2008) and this may mean that residents have multiple sources of support regarding
health education topics. To determine the impact of community health workers, individuals were
queried only on the support provided by a single peer. Outcome measures for health also varied
by the age of the respondent or their children, mainly due to the survey technique utilized, and
consisted of questions on topics such as newborn care, treatment for diarrhea, HIV prevention
and family planning. Social support has a demonstrated positive effect on a variety of health
outcomes (Cohen, Underwood, & Gottlieb, 2000; Gottlieb, 1981), and when provided through
home-based interventions has improved aspects of maternal and child health (Olds & Kitzman,
1993; Norr et al., 2003). Given the evidence, it is likely that a relationship between the social
support provided through peer education should be evidenced in this study as well. The first
hypothesis states:
SOCIAL SUPPORT IN CAMBODIA 33

H
1
: Individuals who report higher levels of self-perceived social support will report
greater knowledge of health behaviors across multiple intervention topics (i.e., newborn
care, complementary feeding, diarrheal treatment, HIV prevention, family planning) than
those individuals who report lower levels of self-perceived social support.
Although improvement across all health indicators was desired, it may be the case that
some health issues are benefited more from social support. Among Latin American women
living in the U.S., social support from their networks, especially husbands, has been shown to
result in healthier post-partum behaviors, such as weight loss and infant care, though lack of
female friends and relatives to provide guidance was one factor which limited knowledge
(Thornton et al., 2006). More broadly, maternal care giving improved after home visitors offered
support and training to parents, and this effect was particularly noticeable among teenage parents
(Olds & Kitzman, 1993). The social support questions asked specifically about visits received in
the previous three months, and mothers who had given birth during that time period were
included in the study. These participants may demonstrate a particularly important effect
between social support and health behaviors. The second set of hypotheses relate to this
relationship, especially the impact of the intervention and social support on first-time mothers:
H
2a
: Women with higher levels of self-perceived social support will report more healthy
behaviors related to newborn care than those women who received no visits or low support.
H
2b
: First-time mothers with high levels of self-perceived social support will report more
healthy behaviors related to newborn care than those first-time mothers who received no
visits or low support.
Demographics may also play a role in the effectiveness of social support for good health.
As mentioned earlier, multiple demographic factors are related to the receipt of social support,
SOCIAL SUPPORT IN CAMBODIA 34

one of which is gender. In their review, Shumaker and Hill (1991) found that while there is a
relationship between social support and general health (overall morbidity and mortality) for
women, it appears to be weaker than what exists for men, and more complex in nature. Dunkel-
Schetter et al. (1987) found that women in their study had more sources of social support than
men, but did not report significantly higher levels of informational, tangible, or emotional
support. Women seek out social support more often than men (Thoits, 1995) and talk about
health more often with other social network members than men (Antonucci & Akiyama, 1987).
Women also prefer to receive health information from peers (Hurdle, 2001). Again, it seems that
for men, support primarily comes from a spouse whereas women rely on a more diverse network
of support. The third hypothesis deals with gender differences in perceived social support
among this sample, and the impact of gender on HIV-related knowledge.
H
3a
: Women will report higher levels of self perceived social support than men will.
H
3b
: Women with greater self perceived social support will report more knowledge of
HIV prevention than men with high self-perceived social support.
Beyond gender, age may also play a role in how social support offered through a health
intervention may impact behavior change. Interventions with younger mothers in disadvantaged
circumstances have been shown to be one of the most fruitful areas for the impact of social
support (Olds & Kitzman, 1993). Education may be another factor which plays a role in
individuals’ understanding of health threats, especially for Cambodian women who may not
receive the full benefits of schooling (Gorman, Dorina, & Kheng, 1999; Fordham, 2003). Within
the HBM, educational attainment is thought to influence perceptions of susceptibility, severity
and costs of health behaviors (Stretcher & Rosenstock, 1997, p. 114). The first research question
pertains to demographic differences:
SOCIAL SUPPORT IN CAMBODIA 35

RQ
1
: What is the relationship between demographic variables (i.e., age, education),
social support, and knowledge acquisition and behavior change across multiple
intervention topics (i.e., newborn care, complementary feeding, diarrheal treatment,
HIV and AIDS prevention, family planning)?
Consistent with self-efficacy theory (Bandura, 1977) and the Health Belief Model (Janz
& Becker, 1984), support which helps improve health-related feelings of self-efficacy should
result in behaviors designed to improve health outcomes. Self-efficacy should be particularly
important in the case of infant care and feeding. This is a domain in which new mothers may
have limited experience, or may be an infrequent practice. The subject of breastfeeding is often
subject to cultural beliefs and misnomers, or impacted by poverty or work (King & Ashworth,
1987). For these reasons, increasing self-efficacy in this domain should have an effect on
breastfeeding behaviors. Bandura (1977, 1982) also suggests that increases in self-efficacy
impact changes in behavioral performance. While this study does not explicitly measure self-
efficacy, it does measure the support given which should enhance self-efficacy. The final
hypothesis and research question are related to the role of efficacy support on health outcomes,
first among the group of participants who should evidence the strongest relationship between
social support and health, and also more generally:
H
4
: Efficacy support will have a stronger relationship with newborn care knowledge
and behaviors than perceived or tangible support for women.
While efficacy support is of interest in one specific type of health behavior performance,
the SPY program sought to increase knowledge and change behavior across a number of topics
(i.e., HIV/AIDS, family planning, breastfeeding, diarrheal illness, etc.). Some evidence exists
that social support may not impact all health behaviors or health outcomes equally (Langlie,
SOCIAL SUPPORT IN CAMBODIA 36

1977; Wallston et al., 1983; House, Landis, & Umberson, 1988; Tomaka, Thompson, &
Palacios, 2006). It is likely that the amount or type of social support required to alter behavior
for a particular topic is not equivalent across these intervention points. Certain topics, such as
HIV/AIDS or family planning, may require negotiation with husbands or partners (Gorman,
Dorina, & Kheng, 1999; Prybylski & Alto, 1999; Derks, 2008) and represent ongoing behavioral
choices. Using a condom to prevent disease or pregnancy is not a one-time behavior change, but
must be sustained over a long period of time. In contrast, antenatal care choices occur at a single
point in time, or are spaced across multiple births (over a series of years), and may be significant
only in those instances. Though in each case healthy choices can have long-lasting effects,
social support may be more important for long-term, continuous behavior change rather than
episodic decision-making. It is also important to understand how the nature of the social support
provided may be related to change. Perceived social support may be measured as the emotional
aspects of relationships (Umberson et al., 1996), expressions of love and helpfulness (Lehman &
Hemphill, 1990), or even as the social resources that are available to an individual (Cohen,
Gottlieb & Underwood, 2001). Many of these studies find that perceived social support, broadly
defined, is related to health and well-being. Israel et al. (2002) found that for low-income
women, instrumental support in the form of tangible aid was a stronger negative predictor of
depressive symptoms than emotional support. The final research question aims to illuminate how
differences in the type of social support provided may be related to different health outcomes.
RQ
2
: How is each type of social support related to maternal and infant health outcome
measures for women?

SOCIAL SUPPORT IN CAMBODIA 37

Chapter 2: Method
Participants
Men and women age 15 to 49 (n = 684) were randomly sampled and surveyed in Khmer
(native Cambodian language). Participants were identified using random sampling of
households. Individuals were eligible for the survey if they fit one of six pre-determined
population groups. When mothers were surveyed, they were asked to respond to questions based
on the health of their youngest child. Participants were recruited from villages located in six
operational health districts of Kampong Cham province, Cambodia. See Figure 1. The
population of Kampong Cham province was estimated at 1.68 million in 2008 (National Institute
of Statistics, 2008), the largest provincial population in the country.
Mean age of participants was 30.34 years (SD = 7.69) and participants had an average of
1.96 children (SD = 0.84). On average, participants had completed some primary education. Of
those surveyed, 51.2% (n = 349) had completed grades one through three, 26.1% (n = 178) had
finished grades four through six and only 4% (n = 26) had finished grade seven through nine.
Just under 20% (n = 128) had no formal education. Sample trends approximate population
statistics regarding levels of Cambodian education (NIS, 2008). Although some respondents
might have spent time at university, given the survey population, the percentage would have
been negligible. This option was not given.
A total of 570 women (83.3%) and 114 (16.7%) men completed the surveys. In regard to
occupation, 62% of the sample (n = 421) identified their work as farming, 15% (n = 103) said
they owned a business or sold something, nearly 15% (n = 101) were involved in housekeeping
or child care, and 7.5% (n = 51) performed some other kind of job, including teaching or
working for the government.
SOCIAL SUPPORT IN CAMBODIA 38

Stratification groups included: women with children under five months (n = 114 , M age
= 25.96, SD = 5.39); women with children between six and nine months (n = 114, M age = 27.52,
SD = 6.17); women with children under 24 months (n = 114, M age = 29.79, SD = 6.49); women
with children under 5 (n = 114, M age = 30.84, SD = 6.88); women age 15 to 49 (n = 114, M
age= 36.05, SD = 7.98); men age 15-49 (n = 114, M age = 31.86, SD = 8.64). Sixty-four percent
of the population is between ages 15 and 64, with a life expectancy of 63 years (CIA Factbook,
2012).
Sampling Procedure
Data collection occurred between July 13 and 20, 2009. Lot Quantity Assessment
Sampling (LQAS) was utilized to select and interview subjects. LQAS methodology has been
widely used to measure the coverage and impact of development projects around the world
(Cakir, Uner, Temel, & Akin, 2008; Robertson & Valadez, 2006), and has grown in popularity
among health system researchers in the past few years. LQAS minimizes sample size, but
provides a simplistic hypothesis test to determine if goals have or have not been reached
(Robertson & Valadez, 2006).
1

Staff from the Sokhapheap Phum Youeng (SPY) project conducted sampling and
interviewing procedures. Sampling primarily took place at the household level, though surveys
were completed by individual interview. This was done to maximize efficiency during data
collection; multiple respondents could be found at a single house (i.e., one mother of a child
under two, one man age 15 to 49). For the purpose of this project, a household consisted of “a
group of people who share the same kitchen or hearth or a group of people who eat from the
same cooking pot” (WR Cambodia and Adventist Development and Relief Association (ADRA)
Cambodia, 2007, p. 3). The definition is consistent with the population census of Cambodia
SOCIAL SUPPORT IN CAMBODIA 39

(National Institute of Statistics, 2008, p. 11). According to the most recent population census at
the time of data collection, average household size was 4.7 people (National Institute of
Statistics, 2008).
To identify possible participants, community leaders assisted with the development of a
village map, which was subsequently divided into subsections (30-50 households each). One
subsection was randomly selected and the center of the section was determined and used for
household selection.
2
Survey teams selected additional households at a predetermined random
sampling interval to reduce homogeneity in the sample.
3
Households were surveyed only if they
met the criteria for belonging to one target group. Only when there was an individual living in
the household whose profile corresponded to one of the parallel sample groups (i.e., mother of a
child under six months, etc.), was the household included in the sample. If a household did not
meet the requirements, the survey team moved to the next target household.
LQAS procedure (Robertson & Valadez, 2006) involves stratifying the sample by
population groups (n = 6) and, for this investigation, surveying 19 people per population group in
each coverage area (n = 6). A sample of 19 provides health evaluators with a sample that can
adequately estimate intervention coverage and impact (Robertson & Valadez, 2006). To attain
this sample, in each randomly selected village, one respondent from each of the six population
groups was surveyed. Nineteen villages were selected per catchment area, providing a total of 19
respondents across all six population groups and supervision areas. These population groups
were: women with children under five months; women with children between six and nine
months; women with children under 24 months; women with children under 5; women age 15 to
49; men age 15 to 49. See Table 1.
Measures
SOCIAL SUPPORT IN CAMBODIA 40

The focus of this investigation was on the role of social support in knowledge gain and
behavior change. Measures will be presented by first discussing social support variables, then
those items related to health outcomes. Although the intervention included measures other than
those presented here, what is discussed here are the variables used for evaluation purposes.
Several measures of social support were used in the current study, and all respondents
answered these items. First, individuals were asked they had been visited by a peer educator in
the past three months. The intervention specified that individuals should be visited by a
volunteer peer educator of the same gender once every three months. Overall, 58.8% (n = 401)
had received a visit, and 41.2% (n = 281) had not met with a peer educator. This was considered
the first, most basic measure of social support. Some respondents were volunteers with the
program. When this occurred, they responded to questions based on their interactions with the
volunteer leader who met with them monthly. If one of the leaders was chosen randomly, she
responded based on assistance given by the person who coordinated volunteer leaders, and those
individuals were instructed to answer questions based on support received from staff. This was
considered volunteer status, or whether a respondent served as a project volunteer, and was
another crude measure of social support. Respondents were asked, “Do you volunteer with SPY
as one of following?” Responses included: no (95.6%, n = 654), Women’s Health Educator
(4.2%, n = 29), and Care Group Leader (0.1%, n = 1). At the analysis level, the latter two
categories were collapsed.
Social support measures were developed independently, and loosely based on items used
by Dunkel-Schetter, Folkman, and Lazarus (1987) and Hall (2010) to measure similar
dimensions of social support. Other measures had not been used cross-culturally prior to this
survey and were judged to be only slightly applicable to the context of the survey and home visit
SOCIAL SUPPORT IN CAMBODIA 41

aspect of the project. Response items to all social support questions were scaled. This was the
first time the SPY project had employed these type of measures.
4
Means and standard deviations
for social support variables are provided in Tables 5-7.
Individuals were asked about three dimensions of received social support and asked to
think about a single person who provided this support (i.e., the peer educator). First, participants
were asked about how much instrumental support they were provided. Instrumental support
was defined as assistance in performing tasks, accessing services, or providing supplies
necessary for keeping a family healthy. Individuals rated how much instrumental support they
received. For all items, responses ranged from none to a lot (i.e., 0 = none, 1 = a little, 2 = some,
3 = a lot). First, participants were asked, “how much did this person help you perform tasks to
keep your family healthy?” (M = 2.40, SD = 0.61, range = 0-3). Second, the individual was
asked, “how much did this person help you by giving you some things you needed to keep your
family healthy” (M = 2.29, SD = 0.61, range = 0-3). Finally, respondents were asked, “how
much did this person help you to access different services, like the health center or voluntary,
confidential counseling and testing (VCCT), you needed to keep your family healthy” (M = 2.30,
SD = 0.69, range = 0-3). During training for data collection, these questions received a great
deal of focus, as it was necessary to differentiate instrumental support from informational
support (or teaching) provided as a regular aspect of project visits.
Another kind of support, efficacy support, was defined as support that made the
respondent feel they could solve problems on their own. Similar to the instrumental support
items, responses ranged from none to a lot (i.e., 0= none, 1 = a little, 2= some, 3 = a lot).
Individuals first answered the question, “After talking with this person, how much did you feel
that you had the knowledge to improve your family’s health” (M = 2.25, SD = 0.65, range = 1-3).
SOCIAL SUPPORT IN CAMBODIA 42

Next, they were asked, “After talking with this person, how much did you feel that you could do
things on your own to keep your family healthy” (M = 2.20, SD = 0.66, range = 0-3). Finally,
individuals were asked, “After talking with this person, how much did you feel that when your
family was sick you could find help or provide help to them” (M = 2.34, SD = 0.69, range = 0-3).
At the analysis level, instrumental social support and efficacy support measures were combined.
Respondents were also asked about other perceived social support. Several items
measuring this construct were adapted from the UCLA Social Support Index (Schwarzer,
Dunkel-Schetter, & Kemeny, 1994) to be used in a cross-cultural setting and with this particular
project. Original items used a seven point scale (0 = not good/not at all to 7 = very good/very
much), which was retained for these measures. Individuals first identified “How much did you
feel that this person cared about the health of you and your family” (M = 5.18, SD = 1.40, range
= 1-7). Next, participants responded to the question “How easy was it to ask questions of this
person” (M = 5.68, SD = 1.38, range = 1-7). Individuals also rated the general helpfulness of the
other person, using a seven-point scale of not at all helpful (0) to very helpful (7; M = 5.83, SD =
1.33, range = 2-7). Participants also answered “Overall, how important do you think this
person’s input is in helping you make health decisions” (M = 6.05, SD = 1.24, range = 1-7).
These answers were collapsed into a general perceived social support scale (alpha = 0.80, M =
22.84, SD = 4.19, range = 8-28).
Dependent variables in this study consisted of items measuring health knowledge and
behavioral outcomes. A series of eight intervention topics were covered by project staff over the
three year project cycle. Topics were chosen to address indicators devised by USAID and
provided as guidelines for all funded projects. Of these, six intervention topics were used to
address hypotheses regarding health information knowledge and behavior change. Those topics
SOCIAL SUPPORT IN CAMBODIA 43

were: breastfeeding, antenatal care, diarrhea treatment, acute respiratory infection (ARI), birth
spacing/family planning, and HIV prevention. Topics were identified and questions formulated
using baseline surveys and indicators provided by USAID for maternal and child health projects.
Proportions, means and standard deviations for each response are found in Tables 2 through 4.
The first outcome emphasized breastfeeding in the first hour after delivery. This was
measured by asking mothers (women with children 0 to 6 months old) when they first breastfed
their newborn. Responses were: within one hour post-delivery (n = 87, 77%); more than one
hour after delivery (n = 12, 10.6%); one to three days post-delivery (n = 11, 9.7%); more than
three days post-delivery (n = 3, 2.7%); did not breastfeed; don’t know. During analysis, these
items were collapsed due to small cell sizes to only two options: on the same day as delivery (n =
99, 87.6%) and within three days post-delivery (n = 14, 12.4%). While other measures were
included on the final survey (i.e., food given in previous day or week; current breastfeeding),
there was little variability in the responses and they could not be used for analysis. See Table 2.
Antenatal care dealt with delivery and post-delivery hygienic activities for live births.
The SPY project did not track mother or child mortality data over the course of the project.
Rather, this variable was measured in regard to practice. Practice questions were directed to
women with children under six months of age (n = 114), and each mother was asked to identify
what was done with her newborn immediately after birth. Response items were yes or no
questions measuring: immediate breastfeeding, immediate drying, immediate wrapping, clean
cord care, and hand washing prior to cord care. At analysis, these measures were summed to
create an index of post-delivery behaviors (M = 3.54, SD = 1.39, range = 5). See Table 2.
Diarrhea treatment was measured by asking mothers with children under age 5 (n = 114)
about correct preparation of oral rehydration therapy (ORT) packets for use with a child
SOCIAL SUPPORT IN CAMBODIA 44

experiencing diarrhea. Preparing the solution involves using one liter of purified water and
dissolving the entire ORT packet fully into the water. Villagers often boil water or leave plastic
bottles in the sun to purify water. Interviewers indicated whether mothers described the entire
process correctly or incorrectly. Correct responses included describing each step of the process
in the right order (i.e., boil water, use 1 packet of ORS, dissolve packet fully). When women
provided correct responses for the entire process, these were tabulated; 61.8% gave a correct
answer (n = 68). See Table 3.
The ability of mothers to identify Acute Respiratory Infection (ARI) in their children was
ascertained using questions directed to mothers with children under 5 years of age (n = 114).
Respondents were asked to name the symptoms of ARI (i.e., difficulty breathing, chest rising
and falling, cough, high fever, nostrils flaring, don’t know). For analysis, women’s responses
were combined to form a scale of ARI symptom knowledge (M = 3.14, SD = .98). Because the
program did not provide medical treatment, correctly naming the symptoms of ARI indicates that
women know when to seek treatment for an infected child at a local health care center. See
Table 3.
Birth spacing/family planning for future children and family size was measured by asking
mothers of children under age two (n = 114) and women age 15 to 49 (n = 114) which
contraceptive method they were currently using. The response options were: pill, injection,
condom, tubal ligation, IUD, other, none. These forms of birth control (responses other than
other) were combined to form a single measure that represented use of any kind of family
planning method. See Table 4.
For subsample groups men age 15 to 49, women age 15 to 49, and mothers of children
under age two and under age five, several questions asked about HIV prevention. Respondents
SOCIAL SUPPORT IN CAMBODIA 45

were first asked, “Have you ever heard of HIV/AIDS?” Those who answered positively were
then asked about HIV prevention. Eighteen possible responses were queried with an individual
yes or no question: nothing; abstain from sex; use condoms; limit sex to one partner/stay faithful
to one partner; limit number of sexual partners; avoid sex with prostitutes; avoid sex with
persons who have many partners; avoid intercourse with persons of the same sex; avoid sex with
persons who inject drugs intravenously; avoid blood transfusions; avoid injections; avoid
kissing; avoid mosquito bites; seek protection from traditional healer; avoid sharing razors,
blades; other; don’t know. While SPY staff marked questions as correct or incorrect on the final
surveys (based on curriculum and WR values), to limit cultural or other bias about sexual
behavior in the responses, during analysis the responses were restricted to six clearly correct or
incorrect items. Incorrect answers were scored as negative values and were: nothing (n = 0),
avoid mosquito bites (n = 1), and avoid kissing (n = 10). Correct answers were scored as
positive values and were: abstinence (n = 138), limit sex to a single partner (n = 317), and use
condoms (n = 433). These responses were then summed to form an index (M = 1.99, SD = .96).
See Table 4.
Translation and Instrument Training
To obtain survey data, project staff members were trained in sampling and interviewing
procedures prior to data collection. These staff had been responsible for training the community
groups of volunteer educators during the intervention. Training occurred over four days in July
2009. Training consisted of reviewing village sampling procedures, interviewing techniques,
and questionnaires. For health questions, interviewers were told to read certain response items,
but several questions were open-ended so that responses could be later coded as ‘correct’ or
‘incorrect’ based on project values. Interviewers were asked not to prompt an answer, but were
SOCIAL SUPPORT IN CAMBODIA 46

allowed to ask additional questions if a participant had difficulty remembering. If the individual
was still unable to respond, the interviewer was instructed to use don’t know. When questions
were skipped, interviewers attempted to ask the question again at the end of the survey. If that
was not possible or the respondent could not answer, it was noted that a question had not been
answered.
Social support questions required additional training. During training, the principal
investigator reviewed the purpose of adding the items to the final survey, discussed the three
different social support constructs, and discussed each individual question to ensure that it made
sense to interviewers. Interviewers were instructed to read the scaled response items when
asking questions.
During data collection, oversight and technical assistance came from district supervisors,
with daily debriefings and routine spot checks of data collection teams to review techniques and
problems. Data entry was done by staff using hand tabulation procedures before being input into
digital format. Staff members were trained prior to data entry, including the desired answers to
survey questions, enabling data entry staff to determine which answers constituted a positive
value for each question or indicator on the survey. Answers were deemed “correct” when they
corresponded to previously determined health criteria taught by volunteers. Curriculum was
developed by WRC staff to be consistent with USAID standards. Thus, answers labeled
“correct” were determined by project staff and not by the researcher, who received electronic
copies of data after coding and not the response sheets. While it was possible to compute correct
answers from raw scores, missing data or incorrectly entered scores could not be verified with
original answer sheets. Additionally, questionnaires were prepared and delivered in the Khmer
SOCIAL SUPPORT IN CAMBODIA 47

language, and any notes or open ended responses required translation and were not provided to
the researcher.
Translation
For all health measures, questions were formulated in English and translated to Khmer by
staff of WRC and Adventist Relief and Development Association (ADRA) Cambodia. When
survey questions were revised for clarity, the Khmer versions of questions were altered first, then
English versions. Items in English approximate the literal Khmer translation of questions. All
survey respondents were interviewed in Khmer, the official language of Cambodia.
Social support measures were written in English. They were translated by WRC staff and
checked by an additional translator for clarity. As these measures were unique to the final
survey, they were submitted to the project management team for review. At that time, questions
were once again revised for clarity and translation issues, with questions finalized in Khmer with
English translation. The revised measures were then used for the final evaluation.
All measures are reported using the approximate English translation of Khmer text unless
otherwise noted.

SOCIAL SUPPORT IN CAMBODIA 48

Chapter 3: Results
Analysis Plan
Two types of data were utilized in this investigation. Social support scales were
measured continuously, but the majority of the health outcomes were measured using categorical
responses. Several of these variables (i.e., HIV prevention knowledge, ARI symptoms, and
newborn care) were later summed and treated as continuous variables during analysis. Means
and standard deviations for these scales and for each item are reported in Tables 2-4. All
outcome items were checked for homogeneity of variance and against assumptions of the normal
distribution—most were not normally distributed, but did have homogenous variance. It is
possible that individuals who answered one of the items correctly might have correspondingly
better knowledge on other items tapping the same intervention outcome. When these categorical
items were later summed to form a scale (i.e., HIV prevention, ARI symptoms, newborn care,
etc.) the data were not normally distributed.
First, social support items were factor analyzed. Following this procedure, analyses were
conducted using t-test and chi-square analyses to determine whether being visited by a peer
educator (no vs. yes) produced either mean differences or significant differences in cell counts
between outcome variables. To explore differences via the perceived social support or tangible
social support scale, ANOVA tests were utilized to examine main effects and interaction
between the independent variables. For categorical variables, binary logistic regression was
employed. Finally, to test the influence of demographic variables (i.e., age, education, volunteer
status), multiple regression was used for continuous outcome variables, while binary logistic
regressions were again utilized for categorical outcomes.
SOCIAL SUPPORT IN CAMBODIA 49

Social support variables also violated assumptions of the normal distribution, and were
skewed positively, though each scale did achieve homogenous variance. One explanation for this
may be that individuals who did not receive any visits were not asked about the social support
they received from health educators. Individuals who responded may have reported that any
visit provided some degree of social support, creating a ceiling effect. For t-tests, Levene’s test
for homogeneity of variance was used to ensure there was not a significant difference between
group variances. For tests revealing significant between group differences, results will be
reported below. Median splits were used for scaled social support variables when high and low
comparisons were made, in an effort to equalize sample size as much as possible.
For regression procedures testing hypotheses regarding social support, forced entry
methods were employed. When multiple regression procedures were used with demographic
variables, backward stepwise models were employed. This technique may be limiting from a
theoretical perspective, as stepwise techniques privilege mathematical associations rather than
theoretical significance and remove control over variable selection from researchers (Field,
2005). However, because this study is in many respects exploratory, stepwise methods may
provide a good indication of what predictors could be important to consider in future
interventions. Where appropriate, once stepwise techniques had been used to eliminate non-
significant predictors, regression models were re-run using hierarchical or forced entry methods.
Although overall sample size was quite large, the actual number of respondents for each
outcome variable was often small. Participants varied across analyses because of the different
subsamples used for each outcome variable (range = 34-682). While in ideal cases alpha levels
should be set more conservatively (i.e., p < .05), because several samples were small, alpha
levels which were less than .10 are reported as exploratory findings. However, these results
SOCIAL SUPPORT IN CAMBODIA 50

should be interpreted with caution, noting that there is a risk of rejecting the null hypothesis
when the data may support it. When necessary, Tukey post-hoc tests were used to determine
between-group mean differences for interactions or multiple groups. The Tukey test was chosen
for its conservative estimates and to restrict Type I error rates (Field, 2005).
Factor Analysis
Because social support measures had not been previously employed, exploratory factor
analysis was used to determine whether social support items loaded onto the hypothesized three
factors. As the measures were designed to tap different dimensions of the same underlying
construct and some degree of correlation was expected, principal components analysis and
oblique rotation were utilized (Costello & Osborne, 2005). Examination of scree plots and
limiting factors to those with eigenvalues over 1.0 revealed that only two factors should be
retained. The first factor, which loaded positively, seemed to correspond to a combined
instrumental and efficacy support. The second, with negative factor loadings, indicated a
perceived social support scale. Together, the scales accounted for a combined 55% of the
variance, which is a range consistent with other studies on this topic (i.e., Dunkel-Schetter,
Folkman & Lazarus, 1987). Factor loadings, communalities, and eigenvalues are reported in
Table 8. Additional factor analysis was conducted with the goal of determining whether a three-
factor structure was possible. Eigenvalues for the third factor were lower than one, and factor
loadings did not conform to the hypothesized dimensions.
Analyses were conducted using only two scaled indices: perceived social support and
what was labeled tangible social support. These scales were created by calculating the sum of
items that (a) loaded onto a single factor with a loading greater than .40, (b) did not load onto
another factor with a loading greater than .40 (Costello & Osborne, 2005). For one item, factor
SOCIAL SUPPORT IN CAMBODIA 51

loading was almost equivalent for each factor, though negative for factor one. The question was
initially hypothesized to tap the combined efficacy and instrumental support scale. While this
item had the lowest loading for each factor, the difference between this item and those on the
first factor was less than the same difference for the second factor. This item was therefore used
in creating the tangible social support scale.
Self and Child Health Awareness
The first hypothesis stated: Individuals who report higher levels of self-perceived social
support will report greater knowledge of health behaviors across multiple intervention topics
(i.e., newborn care, complementary feeding, diarrheal treatment, HIV prevention, family
planning) than those individuals who report lower levels of self-perceived social support. Three
particular outcomes were considered as representative of self and child health awareness. First,
knowledge of ARI symptoms measured whether women could identify when their child might be
ill and need treatment. Knowledge of ORS indicated whether mothers could correctly prepare a
treatment for their child’s diarrhea. Finally, women’s use of contraception was queried.
To examine differences in the continuous variable index of ARI knowledge, first an
independent samples t-test was conducted, using visited by a peer educator as the grouping
variable. The difference between the groups was not significant, t (109) = -1.66, p = .10, r = .02.
Because those individuals who had not received a visit did not respond to scaled social support
questions, an interaction could not be tested. Instead, a 2 (Low vs. High Perceived Social
Support) X 2 (Low vs. High Tangible Social Support) ANOVA with knowledge of ARI
symptoms was performed. The resulting interaction was not significant, F (1, 59) = .68, p = .42,
η
2
= .01. A significant main effect for tangible social support on knowledge of ARI symptoms
emerged, F (1, 59) = 6.77, p < .05, η
2
= .10. Women who received higher levels of tangible
SOCIAL SUPPORT IN CAMBODIA 52

social support reported more knowledge of ARI symptoms (M = 3.51, SD = .64) than women
who reported receiving lower levels (M = 3.00, SD = 1.23). A main effect for perceived social
support was non-significant, F (1, 59) = 2.49, p = .12, η
2
= .04.
Categorical variables were examined using chi-square and binary logistic regression.
First, a chi-square analysis with knowledge of ORS and visited by a peer educator was
performed, with a significant result, X
2
(1, 110) = 4.54, p < .05, φ = .20. Sixty-eight percent (n =
52) of women who spoke with a peer educator correctly named all three steps, whereas only
23.5% (n = 16) of those who had not been visited could correctly describe the process.
Binary logistic regression was used to predict whether perceived social support and
tangible social support would impact knowledge of ORS. Both social support variables were
dummy coded with low levels of social support as zero and high levels as one, which was
continued for all logistic regression analyses using these two variables. An interaction term was
built with the median split groupings for both tangible and perceived social support. The model
including the interaction term was not significant, Χ
2
(3, n = 61) = 3.35, p = .34, r
2
= 05. The
interaction term was not a significant predictor, β = -.39, SE = 1.51, exp(B) = .68, p = .79.
Neither tangible social support (β = -1.49, SE = 1.19, exp (B) = .23, p = .21, nor perceived social
support (β = .76, SE = .73, exp (B) = 2.15, p = .29) were significant predictors of knowledge of
ORS.
Similar to the above procedures, a chi-square analysis was performed for the impact of
peer educator visit on use of contraceptives. The resulting equation was not significant, X
2
(1, n
= 159) = 2.55, p = .11, φ = -.13. Forty-six percent of women who had spoken with a peer
educator (n = 43) were using some form of contraception, compared with 59.1% of women who
had not spoken to a peer educator (n = 39). Binary logistic regression was used to determine
SOCIAL SUPPORT IN CAMBODIA 53

how scaled measures of social support were related to use of contraceptives. The resulting
model was non-significant, X
2
(3, n = 53) = 1.15, p = .76, r
2
= .02. Neither the interaction or
tangible social support nor perceived social support have an impact on the use of contraceptives
among women in this sample.
Overall, these results do not support the first hypothesis, that individuals who report
higher levels of social support will engage in more knowledge of health behaviors. Peer
educator visits appear to be the most effective, though not significant, means of improving
knowledge, followed by tangible social support. Perceived social support does not significantly
contribute to knowledge gain.
Newborn Care
Hypothesis 2a stated: Women with higher levels of self-perceived social support will report
more healthy behaviors related to newborn care than those women who received no visits or low
support. Similarly, the fourth hypothesis stated: efficacy support will have a stronger
relationship with newborn care knowledge and behaviors than perceived or tangible support for
women. Because efficacy support and instrumental support did not factor into separate variables,
hypothesis four was tested alongside the second hypothesis, using tangible social support as a
comparison to perceived social support.
Two primary dependent variables were used to measure newborn care: an index of
newborn care behaviors and a single item measuring immediate breastfeeding. Measurement of
antenatal care consisted of an index composed of five post-delivery behaviors (immediate
breastfeeding, immediate drying, immediate wrapping, clean cord care, and hand washing prior
to cord care). Differences in the outcome measures for those who had received a peer educator
SOCIAL SUPPORT IN CAMBODIA 54

visit were tested with an independent samples t-test for index of newborn care and a chi-square
analysis for the categorical outcome, immediate breastfeeding.
Mothers who had been visited by a peer educator reported significantly more antenatal
care behaviors, t (112) = -2.74, p < .01, r = 25. Mothers who had received a visit (n = 81)
reported an average of 3.77 (SD = 1.36) behaviors, whereas those who had not received a visit (n
= 33) reported only an average of three behaviors (SD = 1.32).
An ANOVA with perceived social support and tangible social support as independent
variables and antenatal care behaviors as the dependent variable revealed a significant
interaction, F(3, 68) = 3.51, p < .10, η
2
= .05, though this should be treated as an exploratory
finding only. When mothers had low perceived social support, differences in tangible support
contributed little to the number of antenatal care behaviors reported. However, women with
high levels of both perceived and tangible social support (M = 4.15, SD = .23) reported
significantly more antenatal care behaviors than women with high levels of perceived social
support and low levels of tangible social support (M = 2.75, SD = .48). See Figure 2. In
addition, a main effect for tangible social support emerged, F (1, 68) = 4.10, p < .05, η
2
= .06.
Similar to the interaction effect, women with high levels of tangible social support demonstrated
more antenatal care behaviors (M = 3.89, SD = .22) than women who had low levels of tangible
social support (M = 3.17, SD = .29). No significant main effect emerged for perceived social
support on antenatal care.
As mentioned earlier, at the analysis level, the variable for immediate breastfeeding was
collapsed to reflect either breastfeeding on the same day of delivery, or between one and three
days post-delivery. The resulting chi-square analysis approached significance, X
2
(1, n = 113) =
3.70, p = .05, φ = -.18. Ninety-one percent (n = 74) of women who had received a visit from a
SOCIAL SUPPORT IN CAMBODIA 55

peer educator breastfed their newborn on the same day as delivery, whereas 78.1% (n = 25) of
women who did not receive a visit practiced immediate breastfeeding.
A logistic regression equation with immediate breastfeeding as the dependent variable
and perceived social support, tangible social support, and the interaction term as independent
predictors was conducted. The overall model achieved significance, X
2
(3, n = 72) = 8.41, p <
.05, r
2
= .12. However, perceived social support was not a significant predictor (β = -.07, SE =
1.31, exp(B) = .93, p = .96) of immediate breastfeeding, nor was tangible social support (β = -
.72, SE = .99, exp(B) = .49, p = .47), or the interaction term (β = -19.97, SE = 6996.69, exp(B) =
.00, p = .99). In other words, while the tested model was a better fit to the data than the constant
alone, the predictors themselves did not significantly predict change in the outcome variable.
Based on this and the above evidence on antenatal care behaviors, the fourth hypothesis is only
partially supported.
Hypothesis 2b stated, first-time mothers with high levels of self-perceived social support will
report more healthy behaviors related to newborn care than those first-time mothers who
received no visits or low support. To address the second hypotheses, only women who had one
child were included in the analysis. A significant difference was found for peer educator visit on
the index of newborn care behavior, t (49) = -2.09, p < .05, r = .29. Women who received a visit
in the previous three months (n = 38) reported more healthy behaviors following delivery (M =
2.77, SD = 1.01) than women who did not receive a visit (n = 13, M = 2.68, SD = 1.45).
A chi-square analysis with peer educator visit and immediate breastfeeding was
significant, X
2
(1, n = 51) = 12.97, p < .01, φ = -.50. This result may be due to low expected cell
sizes; only three women who had received a peer educator visit (7.9%) reported waiting longer
that one day to begin breastfeeding. In contrast, seven women (53.8%) who had not been visited
SOCIAL SUPPORT IN CAMBODIA 56

waited at least one day post-delivery to begin breastfeeding. To adjust for low expected
frequencies, Fisher’s exact test was employed, but the result remained significant at the p < .01
level.
Turning to other forms of social support, an ANOVA with perceived social support and
tangible social support as independent variables was used with the index of newborn care as the
dependent variable. Although small sample size was an issue, the resulting interaction was
significant, F (3, 30) = 9.63, p < .01, η
2
= .24. Interestingly, those first-time mothers who had
low levels of both perceived and tangible social support (M = 4.11, SD = .44, n = 9) or had high
levels of perceived and tangible social support (M = 4.00, SD = .31, n = 18) reported the highest
number of post-delivery behaviors. Those with low levels of tangible social support but high
levels of perceived social support evidenced the lowest number of post-delivery behaviors (M =
1.50, SD = .93) and those with high tangible support but low perceived social support fell
between the two (M = 2.80, SD = .59). There were no significant main effects for either
perceived social support, F (3, 30) = 1.32, p = .26, η
2
= .04, or tangible social support, F (3, 30)
= .94, p = .34, η
2
= .03.
To examine the categorical outcome, immediate breastfeeding, logistic regression was
used, with perceived social support, tangible social support and an interaction term as
independent predictor variables. With small sample sizes, the model was notable as an
exploratory finding, X
2
(3, n = 34) = 7.29, p < .10, r
2
= .56. However, though model fit was
superior to the null model, none of the predictors were significant.
Similar to the first hypothesis, the second set of hypotheses is only partially supported.
Once again, peer educator visits appear to be having the greatest impact on behavior and
knowledge changes. Tangible and perceived social support, when provided in concert, do seem
SOCIAL SUPPORT IN CAMBODIA 57

to influence outcomes, but the evidence is negligible that either type of support offers much
benefit when provided alone. In the case of new mothers, the findings are particularly difficult to
interpret, as provision of tangible support seems to outweigh the effect of perceived social
support on post-delivery behaviors.
HIV Prevention
Hypothesis 3b related to knowledge of HIV prevention behaviors. As mentioned earlier,
to measure HIV prevention knowledge, six primary indicators were used. Three of these
(condom use, abstinence, and faithfulness to a single partner) were correct answers, while three
additional responses (avoid kissing, avoid mosquito bites and nothing) were incorrect. Answers
were summed, with correct answers scored as +1 and incorrect as -1 (range = 3). To determine
the influence of social support on HIV Prevention Knowledge, an independent samples t-test was
used with peer educator visit as the independent variable. Then, a 2 (gender) X 2 (low vs. high
perceived social support) X 2 (low vs. high tangible support) ANOVA was conducted with HIV
prevention knowledge as the dependent measure.
The t-test with peer educator visit as the independent variable and HIV prevention as the
dependent variable was significant, t (402) = -4.55, p < .01, r = 22. Men and women who were
visited by a peer educator in the previous three months knew more ways to prevent HIV (M =
2.14, SD = .75, n = 232) than those who had not received a visit (M = 1.80, SD = .74, n = 172).
However, an additional t-test to compare men and women who had received a visit from a peer
educator revealed no significant difference in knowledge of HIV prevention, t (230) = 1.20, p =
.23.
The three-way interaction between gender, perceived social support and tangible social
support with HIV prevention knowledge was not significant, F (7, 150) = .543, p = .46, η
2
= .004.
SOCIAL SUPPORT IN CAMBODIA 58

No two-way interactions emerged between the social support variables or with gender. Finally,
there were no significant main effects for gender, perceived social support, or tangible social
support.
The third hypothesis, that women with higher levels of self-reported social support would
indicate more knowledge of HIV prevention, was not supported. Although being visited by a
peer educator was related to greater knowledge of HIV prevention, no significant differences by
gender for higher levels of support were found, either for peer educator visits, tangible or
perceived social support. Thus, while knowledge of HIV prevention was improved among those
who were sampled, it is not likely that this improvement was due to the provision of high social
support.
Demographic Variables
To answer the remaining hypothesis, that women will report higher levels of self
perceived social support than men, a series of chi-square analyses were used. First, peer
educator visits were analyzed. Women reported significantly more visits than men, X
2
(1, n =
682) = 34.21, p < .01, φ = .22. Only one-third of men (n = 38) said that a peer educator had been
visited in the past three months, whereas 63.7% of women (n = 363) had been visited. Next,
differences in perceived social support and tangible social support by gender were examined.
For tangible social support, a significant chi-square value emerged, X
2
(1, n = 334) = 4.18, p <
.05, φ = .11. More women reported high levels of tangible social support (59.1%, n = 182) than
men (38.5%, n = 10). Finally, the resulting chi-square for perceived social support was not
significant, X
2
(1, n = 364) = .62, p = .43, φ = .43. Men and women both reported almost equal
proportions of support: 51.4% of men (n = 19) and 58.1% of women (n = 190) said they received
high levels of perceived social support. These findings suggest that women often report
SOCIAL SUPPORT IN CAMBODIA 59

receiving more social support than men, but not across all types of social support. Hypothesis
three is supported.
The first research question concerned the role of demographic variables in the
relationship between social support and health outcomes. For this analysis, only those outcomes
which were related to social support were used. Thus, contraceptive use was excluded from this
analysis, as there were no significant relationships between scaled social support variables and
this outcome. Additionally, as there were no significant relationships between HIV prevention
knowledge and scaled social support variables, only peer educator visits were analyzed with
demographic variables. First, a correlation matrix was created for demographic variables (i.e.,
age, education, number of children), health outcomes, and social support variables. As many of
the demographic variables were dichotomous, point biserial or biserial correlations were used.
Then, an analysis between significantly correlated demographic variables and social support
items was performed for dependent health outcomes. For continuous variables, ANOVA was
used to test interaction effects, and for categorical variables, binary logistic regression was
employed.
Significant correlations between several demographic variables and health outcomes
emerged. Volunteer status (n = 111) was significantly related to ARI symptom knowledge, r
pb
=
.25, p < .01, knowledge of ORS, r
pb
= .19, p < .05, and marginally related to HIV prevention
knowledge, r
pb
= .09, p < .10. Volunteer status was also significantly related to peer educator
visits, r
pb
= .14, p < .01. However, interaction effects between volunteer status, social support,
and ARI symptom knowledge could not be tested. All volunteers who responded to ARI
knowledge questions (n = 6) indicated that they received high levels of tangible social support
and had received a peer educator visit. Put differently, there were no volunteers who reported
SOCIAL SUPPORT IN CAMBODIA 60

that they had not received a visit or had low levels of tangible social support. Interactions
between these variables could not be computed. A similar problem emerged for ORS knowledge,
and further relationships between volunteer status, social support and this outcome variable
could not be analyzed.
Finally, to test the impact of volunteer status on the relationship between peer educator
visit and HIV prevention knowledge, a 2 (visited or not visited) X 2 (volunteer vs. non-volunteer)
ANOVA with HIV prevention knowledge as the dependent variable was conducted. No
significant interactions between volunteer status and peer educator visits emerged, F(1, 400) =
.92, p = .34, η
2
= .002. A main effect for peer educator visits was significant, F(1, 400) = 4.41, p
< .05, η
2
= .01. As evidenced before, receiving a visit from a peer in the previous three months
resulted in greater knowledge of HIV prevention for men and women (M = 2.14, SD = .75) than
for those who did not receive a visit (M = 1.80, SD = .74). There was no significant effect for
volunteer status.
Age was marginally related to index of ARI symptom knowledge, r = -.17, p < .10 and for
social support items, age was significantly related to peer educator visits, rho = .08, p < .05.
Despite these associations, a test of the interaction using a 4 (age quartiles) x 2 (visited vs. not
visited) ANOVA with ARI symptom knowledge as the dependent variable was not significant,
F(3, 103) = .23, p = .87, η
2
= .01. No significant main effects for age or peer educator visits
emerged.
Number of children was marginally correlated with immediate breastfeeding, r
pb
= -.17, p
< .10, but this relationship was previously analyzed when testing hypothesis two, and was not
further examined. It appears that certain demographic categories are related to social support
SOCIAL SUPPORT IN CAMBODIA 61

and to health outcomes, but that they do not offer an explanation for the full range of these
relationships.
The second research question was tested with a series of correlations. Social support
items were correlated with health outcome measures. Across all outcomes, peer educator visit
was significantly correlated with the most health outcomes. Significant relationships between
peer educator visits and HIV prevention, r
pb
= .22, p < .01, antenatal care behaviors, r
pb
= .25, p
< .01, and ORS knowledge, r
pb
= .20, p < .05. Peer educator visits was also marginally related to
immediate breastfeeding, r
pb
= -.18, p < .10.
In contrast, tangible social support was only significantly correlated with ARI symptom
knowledge, r
pb
= .29, p < .05, and marginally significantly related to ORS knowledge, r
pb
= .22, p
< .10 and antenatal care, r
pb
= .22, p < .10. Perceived social support was only significantly
related to immediate breastfeeding, r
pb
= -.24, p < .05.
Ultimately, few relationships involving demographic, social support, and dependent
health outcome variables emerged in response to the first research question. Volunteer status is
one variable that seems to be related to many health outcomes, but its interaction with social
support remains unclear. Additionally, age appears to be the most important demographic
variable, though once again, no significant interaction with social support emerged.
In terms of the second research question, peer educator visits seem to be the most
important type of social support required to improve dependent health outcomes. Tangible social
support also appears to be influential for knowledge variables, including recognizing symptoms
of ARI and preparing treatment (ORS knowledge). However, perceived social support is only
significantly related to immediate breastfeeding, a finding that did not emerge when perceived
social support was tested alongside tangible social support. Thus, there is still more work to do
SOCIAL SUPPORT IN CAMBODIA 62

in order to determine with any clarity the full range of relationships between different types of
social support and health outcomes.

SOCIAL SUPPORT IN CAMBODIA 63

Chapter 4: Discussion

The main purpose of this study was to investigate how social support in natural
community interventions is related to various maternal and child health outcomes for people in
developing nations. In particular, this study addresses how social support provided through peer
educators facilitated health behavior change and knowledge gain among men and women in rural
Cambodia. Data from the evaluation of a three-year intervention in Kampong Cham province,
Cambodia were utilized to explore these questions. Social support was measured in three
distinct ways at the final evaluation: through a visit from a peer educator, as perceived social
support, and as tangible social support. Six health outcomes were measured: antenatal care,
immediate breastfeeding, knowledge of acute respiratory infection (ARI), knowledge of oral
rehydration solution (ORS), contraceptive use, and knowledge of HIV prevention.
The results demonstrate that, despite the complex nature of the construct of social
support, individuals do change health knowledge and behaviors in part because of the
interpersonal aspects of information sharing. Three general conclusions can be drawn from the
findings, and will be discussed separately below with regard to relevant hypotheses. First, peer
educators are an important source of health information. Second, social support does not have a
global impact on health outcomes. Finally, social support in cross-cultural contexts must be
evaluated differently than in U.S.-based studies.
Peer Educator Visits
General social support, measured as receiving or not receiving a visit from a peer
educator in the previous three months, has a discernible influence on health knowledge and
behaviors. The first hypothesis predicted that higher levels of social support would result in
better outcomes for several variables related to self and child care relative to individuals with
SOCIAL SUPPORT IN CAMBODIA 64

lower support. For knowledge about preparing oral rehydration solution (ORS), receiving a visit
from a peer educator resulted in women reporting significantly greater knowledge than women
who were not visited. Peer educator visits were not related to knowledge of ARI symptoms or
use of contraception. It appears that peer educator visits provide necessary information for
outcomes with very specific steps. However, when the focus is on recognizing symptoms of ill-
health, the results are not as strong, though differences were in the hypothesized direction. In the
case of contraception-related knowledge, visits give women an opportunity to make an informed
choice rather than following proscribed behavior.
The second hypothesis predicted the role of social support in antenatal care behaviors for
all mothers in the sample and for first-time mothers only. Women who received peer educator
visits reported significantly more healthy post-delivery behaviors and significantly more of these
women breastfed their infant on the same day as delivery. Among first-time mothers, similar
results occurred; women who received visits reported more healthy antenatal care behaviors. The
number of first-time mothers who breastfed their infants on the same day as delivery was
marginally higher among women who received a peer educator visit, though it is difficult to be
completely confident in this finding due to small sample size.
For both hypothesis one and two, the findings echo Kincaid’s (2000) work in Nepal as
well as other U.S.-based studies (Earp, Ory & Strogatz, 1982) on the significance of peer
educators for health behavior change. Despite Kincaid’s (2000) results that peer educators are
less effective than social support provided through group training meetings, it seems that some
beneficial form of support is being provided through home visits. Perhaps, as in other studies,
peer education was successful when it served to link individuals to health services (Eckenrode &
Hamilton, 2000), or targeted specific behaviors for change (Ramey & Ramey, 1993).
SOCIAL SUPPORT IN CAMBODIA 65

Turning to the third hypothesis, which predicted differences in received social support
and HIV prevention knowledge between women and men, the results are mixed. Women did
report a greater number of peer educator visits than men. Receiving a peer educator visit also
resulted in significantly more knowledge of HIV prevention—but this finding was true among
both men and women. Existing research on social support suggests that women, more than men,
generally receive health information from a source in their social network, not their spouse
(Antonucci & Akiyama, 1987; Hurdle, 2001). However, in this study, although women reported
receiving more visits than men, among all men and women who received a visit, there was no
significant difference in HIV prevention knowledge. Some part of the didactic process that
occurs during peer educator visits is either sufficient to produce change among both men and
women or provides some degree of support that was not measured. It may be important in future
studies to explicitly test what occurs in these interactions and why it matters for health behavior.
The finding that HIV prevention knowledge increased among the full sample is less
surprising. A wealth of research on the utility of peer educators for HIV risk reduction exists,
examining peer education in particular (Kelly et al., 1991, 1992, 1997; Kelly, 2004). That the
effect is not significantly different for men and women somewhat echoes the findings of
Albarracin et al.’s (2008) meta-analysis. That meta-analysis reported that peer educators who
were similar in age had a more pronounced effect on knowledge gain than those who were
ethnically similar or of the same gender (p. 531). Volunteers for the SPY program were chosen
for their similarity to the members of their communities, which seems, in the case of HIV
prevention knowledge, to have had a clear effect.
Finally, the second research question asked how social support was related to health
outcomes. Of all types of support, peer educator visits were related to the largest number of
SOCIAL SUPPORT IN CAMBODIA 66

outcomes, particularly knowledge outcomes (i.e., ORS preparation). One explanation for the
findings regarding peer educator influence is that visits themselves represent a kind of
informational support (Starker, 1986; Kaniasty & Norris, 2000). By sharing information with
villagers, peer educators provide a form of support distinct from both perceived and tangible
social support. In this regard, it is possible that information is the most desired type of support
from these peer educators; family members may fill other kinds of support needs (Cutrona, 1990)
once the individual is aware of what healthy behaviors they should enact.
The Health Belief Model (HBM; Janz & Becker, 1984; Stretcher & Rosenstock, 1997)
may provide another framework for understanding these results, as well as other social support
findings. Peer educator visits may illustrate every stage of the model, from illuminating salient
health threats to villagers, highlighting an individual’s vulnerability to those risks, and lastly,
enlightening community members on how to reduce susceptibility to these health issues. The
lesson plans used by peer educators were designed to educate individuals about the health beliefs
related to remedies (World Relief, 2004). As the HBM is an expectancy-based model, focusing
on broad health risks may have socialized villagers to the likelihood that their own children faced
infection. In doing so, peer educators may have also motivated and educated individuals to
address those risks. The pattern of results suggests that, as the HBM predicts, women are
learning to recognize threats to their children’s health (i.e., knowledge of ARI symptoms) and
how to address them (e.g., knowledge of ORS preparation).
This study both confirms and extends findings from other investigations which used peer
educators to promote health behavior change. Both in the U.S. (Earp, Ory, & Strogatz, 1982;
Kelly, 2004; Kelly et al., 1991; 1992; 1997; Olds & Kitzman, 1993) and abroad (Hussein, Aaro
& Kvale, 1997; Kincaid, 2000), peer educators have been used to create change across various
SOCIAL SUPPORT IN CAMBODIA 67

health indicators. Across most of the measured knowledge and behaviors in this investigation,
peer educator visits were either marginally or significantly related to improved outcomes. Such
findings demonstrate that local community members can contribute to key changes in health
behaviors. While public health campaigns or media-driven information may also be effective, in
regions where broad coverage via media is not possible, utilizing local volunteers may be a
crucial solution to addressing population-wide health deficits.
Several factors may have limited the findings in regard to peer educators. First, volunteer
educators were assigned to meet with households once every three months across the lifespan of
the intervention. Health topics varied according to a pre-determined curriculum, so some of the
health knowledge and behaviors which were examined in the final evaluation may not have been
recently reviewed by educators. Additionally, this evaluation did not measure the frequency of
contact with a volunteer; individuals who live in close proximity to or who encounter village
health workers might have benefitted from more frequent communication with them. Future
research should consider what aspects of the peer relationship are most important for knowledge
gain, as well as how to refine the peer education process for maximum success of interventions.
Lack of Global Impact
Social support, as a multi-dimensional construct, was initially operationalized in the
present study as consisting of three dimensions. After factor analysis of the social support items,
however, only two clear constructs emerged: perceived social support and what was called
tangible social support. As a construct, tangible support has been used in previous studies,
generally to describe the presence of instrumental support (Kaniasty & Norris, 2000). For this
investigation, tangible support was used to describe the items tapping both instrumental and
efficacy support, which emerged as a single factor rather than two constructs, as anticipated.
SOCIAL SUPPORT IN CAMBODIA 68

Social support, measured as tangible support, does not have an equal or consistent impact across
intervention topics. As above, the findings regarding the scaled social support items will be
reviewed in terms of relevant hypotheses.
In terms of the first hypothesis, tangible support was significantly related only to
knowledge of ARI, but not to knowledge of ORS or contraception use. Similarly, for the third
hypothesis, no significant relationship between tangible support and HIV prevention knowledge
occurred. Perhaps because two of the indicators (i.e., ORS, HIV) with non-significant
relationships were knowledge-based measures, peer visits alone are sufficient information
acquisition to address health issues in the future. Tangible support was not related to
contraception use, and adoption rates for contraceptives among women in this sample seemed to
be around 50%. It may be the case that neither support nor peer education drives adoption, but
rather that women are independently deciding whether to prevent pregnancy.
For the second hypothesis, regarding antenatal care and breastfeeding, a significant
interaction between tangible and perceived social support occurred. High levels of tangible
support in conjunction with high levels of perceived social support resulted in significantly more
healthy antenatal care behaviors. Consistent with this, tangible support alone was also
significantly related to performing more healthy post-delivery behaviors.
Despite the findings, outlined earlier, that peer educator visits do create impact in
multiple areas, other types of social support do not have such robust effects. The pattern of
results for the impact of tangible support, in particular, across different health outcomes is still
unclear. That tangible support does matter for health is not wholly unexpected. Israel et al.’s
(2002) finding that higher levels of instrumental support among low-income women in Detroit
predicted overall health demonstrates that tangible support can have more general than specific
SOCIAL SUPPORT IN CAMBODIA 69

impact. Perhaps for health issues, such as ARI, clear instructions and directions on how to
recognize symptoms were perceived as more helpful than emotional support. The significant
interaction between tangible and perceived social support for ARI symptom knowledge reiterates
Israel et al.’s (2002) findings that the impact of emotional support on health becomes negligible
when instrumental support is provided at high levels.
Theoretically, Bandura’s (1977) conception of self-efficacy, in concert with the HBM,
does provide some interpretation for these results. As noted above, villagers who perceived
tangible support as part of their interactions with volunteers may have developed a greater
understanding of the health risks they and their children faced. In regard to self-efficacy,
Bandura (1977) stipulates that the generality of efficacy expectations may vary for each person.
That is, some individuals may experience self-efficacy in regard to multiple behaviors while
others feel efficacious only in specific situations. Thus, it is possible that tangible support
provided through this intervention increased efficacy expectations for some individuals, but only
with regard to certain health outcomes. Tangible support is most significantly related to
recognizing symptoms of ARI, but only marginally related to preparing treatments for diarrheal
illness and not to knowledge of prevention of HIV. Examining these findings suggests that
individuals may fall on a spectrum of efficacy beliefs regarding their ability to enact healthy
behaviors, and that tangible support may be useful only at certain points along that continuum.
These findings may be particularly useful to health practitioners and intervention
directors in developing nations. Increasingly, sustainable development models are privileged by
funders, and often stress capacity building of local individuals for ongoing change (Edwards,
Hulme, & Wallace, 1999). In such cases, a better understanding of the psychological
mechanisms, such as self-efficacy, related to empowerment, and the corresponding impact of
SOCIAL SUPPORT IN CAMBODIA 70

such constructs on behavior, may be of use to practitioners. As future health interventions might
consider employing tangible support, it is important to consider not just why, but also when,
tangible or perceived social support will create change. Future research should continue to
provide clarity on the relationship between tangible support and behavior change, especially the
underlying explanations for why this concept is important.
Cross-Cultural Evaluation
Across all the findings regarding perceived and tangible social support, one final
conclusion was notable. The construct of social support—particularly perceived social
support—must be measured with a specific eye to the culture in which it is being evaluated.
Although Taylor et al. (2004; 2007) demonstrated that Asian or Asian-American respondents did
not respond in a similar manner to White individuals on measures of social support, he and his
colleagues did not propose alternative items. For this evaluation, existing measures had to be
adapted and developed. Those measures did not appear to tap either the full range of support,
nor the expected dimensions of the construct that were hypothesized to have the greatest impact.
Findings regarding tangible social support were reviewed in the last section, and turning
to the relationship between perceived social support and health knowledge and behavior, several
hypotheses were tested but not confirmed by the data. In answer to the second research question,
perceived social support was correlated with immediate breastfeeding. However, addressing the
second hypothesis, when explicitly tested as a significant predictor of immediate breastfeeding,
perceived social support did not significantly interact with tangible social support or impact
immediate breastfeeding on its own. In fact, across all hypotheses and research questions,
perceived social support was not significantly related to any other outcome variable.
SOCIAL SUPPORT IN CAMBODIA 71

The reasons for the lack of relationship between perceived social support and other
variables may be, in part, because the construct itself was not particularly relevant to the study
sample. Taylor and his colleagues have shown that emotional support is regarded as the least
helpful among Asian respondents (Taylor et al., 2004). Moreover, explicitly seeking support
produces more distress among Asian and Asian-American students (Taylor et al., 2007).
Informational support provided through visits, or tangible support focused on skill development,
may have been perceived as general helpfulness rather than addressing an individual’s emotional
needs. In contrast, when perceived support, measured in terms of emotional help, was provided,
individuals might have viewed this as being related to a specific, demonstrated weakness.
Because face-saving is a cultural value among this population, such an attribution may have been
embarrassing or stigmatizing for individuals. Chalise and his colleagues (2007) noted that
cultural forces in their study may have suppressed the amount of social support women reported
providing to their husbands. A similar demand characteristic, though in the reverse direction,
may be operating here.
Theoretical mechanisms, like the HBM or self-efficacy, which emphasize the role of
information and feelings of mastery in overcoming obstacles, provide little in the way of
explanation for these findings. However, turning to the social support literature, perceived social
support may serve to buffer stress in relation to health outcomes (Cutrona, 1990). In fact, this
kind of emotional support may serve a protective role in helping mothers remain healthy,
creating an alternative pathway to healthy children, in which mothers are able to provide better
care because their own health is secure. However, more research on this is necessary, as the
findings provided in this study are not sufficient to demonstrate such a relationship.
SOCIAL SUPPORT IN CAMBODIA 72

Future studies should consider how cultural differences might bias responses and pretest
measures accordingly. For instance, emotional support may be less relevant as a measure of
social support, and, as in other studies (House, Landis, & Umberson, 1988), social integration in
rural community networks would be more informative. Personality differences may play a role
in how individuals express or feel supported (see, for example, Lu, 1995; Hobfoll, Shoham, &
Ritter, 1991), which was not examined in this investigation. There is also a need for measures of
social support which are flexible enough or specifically designed to capture the same construct as
it is expressed in multiple cultures. If social support is intended for use in other interventions,
accurately capturing how local people view the concept and value its expression will be crucial
for ensuring success of health programs. In many ways, these findings serve less as an indication
of how to measure perceived social support, and more of a cautionary tale for those interested in
developing new items tapping the construct.
Limitations
The present investigation is limited by several factors. First, and perhaps most important
to the findings themselves, the measures in this study were developed for the present intervention
and were not pretested before use. While several social support questions were derived from
previous work (Dunkel-Schetter et al., 1987; Hall, 2010), others were crafted specifically for this
intervention. The demands of the field setting did not allow for pretesting; health outcome
measures had been included in previous evaluations and were not pretested by field staff. Thus,
it is probable that the measures were insufficient to tap into the constructs of interest, particularly
in the area of perceived social support.
There are a number of cultural factors which may explain why these measures of social
support were less effective than hoped. First, as Taylor et al. (2004) demonstrate, Asian
SOCIAL SUPPORT IN CAMBODIA 73

respondents favor implicit over explicit support. It is possible that the implicit help offered by
peer educators was in fact beneficial for respondents, but that these measures did not capture the
full range of support provided. Also, the concept of face-saving (Hinton, 1998) is an important
element of Cambodian social interaction. Cambodians try to maintain harmonious exchanges
with others, and negative consequences occur when an individual loses face or causes someone
else to lose face. As the project staff were responsible for completing survey interviews,
measures may have invoked some feeling in respondents that they were ‘reporting’ on the
volunteer to the organization and thus created bias or reluctance. Evaluating volunteers poorly
might have been considered an act that risked the volunteers losing face with the organization,
and thus negative consequences to the self. Individuals may have rated provision of social
support higher to avoid such a situation. The ceiling effect among social support measures
suggests that this might be the case.
Secondly, traditional values of femininity stress the role of the woman in maintaining the
home and caring for the family. The moral code that is impressed upon young Khmer girls
stresses their obligation to their family and their home (Ledgerwood, 1994). In fact, a woman’s
ability to keep her family healthy reflects on her standing in the community, and the ‘merit’ she
gains within the hierarchical structure of family and social life (Derks, 2008). With regard to
questions about their feelings of efficacy to complete health tasks, women may have felt
uncomfortable reporting to the data collection teams that they felt any kind of doubt in their
abilities.
Next, the sampling method utilized was primarily chosen for its efficiency, and not for
statistical power. Several groups had very small sample sizes, particularly when tests were
restricted to only those who had reported high or low levels of social support. Sampling limited
SOCIAL SUPPORT IN CAMBODIA 74

responses in other ways, such as in the case of those hypotheses related to mothers of only one
child. The health intervention did not track child or infant mortality and mothers who had given
birth but whose children were not living were not included in the subsample of mothers with
children younger than six months. By restricting the interviews in such a way, if women did not
practice healthy behaviors and the result was the death of an infant, they were intentionally
excluded from the survey. This may have eliminated a powerful indicator of intervention
success, as well as an opportunity to gauge situations in which the influence of social support
might have been important.
The present investigation also restricted the locus of social support only to the peer
educator. While useful to estimate project success, doing so provides only limited information
on the range of social support provided to individuals and how broader sources of support might
matter for health behavior change. For instance, it is possible that social support from peer
educators is useful for behavior that must be performed alone or which takes place in different
contexts. Perhaps the social context of childbirth means that more instant support or instruction
is important for behaviors that take place following delivery. In those cases, social support
provided by a peer educator may be less important than what is provided by the individuals
present during and immediately after childbirth. Additionally, young mothers in Cambodia
typically live with their parents. The type of care given to very young children may be
influenced by women of a different generation, with a different set of beliefs or knowledge
regarding infants. While the primary caregiving role is allocated to the mother, she is embedded
in a relational network that may be much more influential than a peer educator alone.
Finally, this study would have been strengthened by two methodological alterations.
First, the findings might have benefitted from the use of more qualitative measures of social
SOCIAL SUPPORT IN CAMBODIA 75

support. Although the sponsoring organization collected stories of “Most Significant Change” to
comply with funding guidelines, individuals who provided stories were not selected randomly
and not included in the study. The true nature of social support provided by peer educators
might have been more clearly delineated by asking respondents to describe the nature of their
relationship and interactions with volunteers. Qualitative measures might also have allowed for
different comparisons between social support and health outcome measures, some of which
might have illuminated why social support was important for creating change. Second, measures
should have been developed for use specifically with this nontraditional survey technique. The
nature of the LQAS procedure limited sample size for health outcomes, but also for social
support measures. Individuals who did not receive a visit from a peer educator did not answer
questions about other types of social support; for some outcomes, this reduced the sample size to
a level where statistical power had to be considered.
Ultimately, this study provides an exploratory examination of the role of social support in
a community-based health intervention outside the U.S. It contributes to existing findings about
the importance of peer educators for health promotion, and illuminates two key areas for future
study. Those are, first, clarifying which health outcomes are most responsive to social support
interventions. Second, this research stresses the importance of developing measures that tap
cross-cultural social support values. Although the findings are limited in key areas, the overall
results point to the ongoing potential for local volunteers to create meaningful change in their
own villages.

SOCIAL SUPPORT IN CAMBODIA 76

References
Aaronson, L.S. (1989). Perceived and received social support: Effects on health behavior during
pregnancy. Nursing Research, 38(1), 4-9.
Albarracín, D., McNatt, P.S., Klein, C.T.F., Ho, R.M., Mitchell, A.L. & Kumkale, G.T. (2003).
Persuasive communications to change actions: An analysis of behavioral and cognitive
impact in HIV prevention. Health Psychology, 22(2), 166-177.
Albarracín, J., Albarracín, D. & Durantini, M. (2008). Effects of HIV-prevention interventions
for samples with higher and lower percents of Latinos and Latin Americans: A meta-
analysis of change in condom use and knowledge. AIDS Behavior, 12, 521-543.
Antonucci, 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.
BBC.com. (December 11, 2006). Full text: Kofi Annan’s final speech. Retrieved October 6,
2011 from http://news.bbc.co.uk/2/hi/6170089.stm.
Bandura, A. (1977). Self-efficacy: Toward a unifying theory of behavioral change.
Psychological Review, 84(2), 191-215.
Bandura, A. (1982). Self-efficacy mechanisms in human agency. American Psychologist, 37(2),
122-147.
Bandura, A. (1990). Perceived self-efficacy in the exercise of control over AIDS infection.
Evaluation and Program Planning, 13, 9-17.
Bandura, A. (2001). Social cognitive theory of mass communication. Media Psychology, 3,
265-299.
Berkman, L.F. (1984). Assessing the health effects of social networks and social support.
Annual Review of Public Health, 5, 413-432.
SOCIAL SUPPORT IN CAMBODIA 77

Cakir, B., Uner, S., Temel, F., & Akir, L. (2008). Lot quality survey: an appealing method for
rapid evaluation of vaccine coverage in developing countries – experience in Turkey.
BMC Public Health, 8, 240.
Chalise, H.N., Saito, T., Takahashi, M., & Kai, I. (2007). Relationship specialization amongst
sources and receivers of social support and its correlations with loneliness and subjective
well-being: A cross sectional study of Nepalese older adults. Archives of Gerontology
and Geriatrics, 44, 299-314.
The World Factbook. (2012). Central Intelligence Agency. Retrieved October 5, 2012 from
https://www.cia.gov/library/publications/the-world-factbook/geos/cb.html.
Cohen, S. (1988). Psychosocial models of the role of social support in the etiology of physical
disease. Health Psychology, 7(3), 269-297.
Cohen, S. & Syme, S.L. (1985). Issues in the study and application of social support. In S.Cohen
& L. Syme (Eds.), Social support and health (pp. 3-22). San Francisco: Academic Press.
Cohen, S., Underwood, L.G., & Gottlieb, B.H. (2000). Social relationships and health. In S.
Cohen, L.G. Underwood, & B.H. Gottlieb (Eds.) Social support measurement and
intervention: A guide for health and social scientists (pp. 3-28). New York: Oxford
University Press.
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., Cohen, B., & Igram, S. (1990). Contextual determinants of the perceived
supportiveness of helping behaviors. Journal of Social and Personal Relationships, 7,
553-562.
SOCIAL SUPPORT IN CAMBODIA 78

Dean, K., Holst, E., Kreiner, S., Schoenborn, C., & Wilson, R. (1994). Measurement issues in
research on social support and health. Journal of Epidemiology and Community Health,
48, 201-206.
Derks, A. (2008). Khmer women on the move: Exploring work and life in urban Cambodia.
Honolulu, HI: University of Hawai'i Press.
Dunkel-Schetter, C. & Skokan, L.A. (1990). Determinants of social support provision in
personal relationships. Journal of Social and Personal Relationships, 7, 437-450.
Dunkel-Schetter, C., Folkman, S., & Lazarus, R.S. (1987). Correlates of social support receipt.
Journal of Personality and Social Psychology, 53(1), 71-80.
Durantini, M.R., Albarracín, D., Mitchell, A.L., Earl, A.N., & Gillette, J.C. (2006).
Conceptualizing the influence of social agents of behavior change: A meta-analysis of the
effectiveness of HIV-prevention. Psychological Bulletin, 132(2), 212-248.
Earp, J. L., Ory, M.G., & Strogatz, D.S. (1982). The effects of family involvement and
practitioner home visits on the control of hypertension. American Journal of Public
Health, 72(10), 1146-1154.
Earp, J.L. & Flax, V.L. (1999). What lay health advisors do: An evaluation of advisors’
activities. Cancer Practice, 7(1), 16-21.
Eckenrode, J. & Hamilton, S. (2000). One-to-one support interventions: Home visitation and
mentoring. In S. Cohen, L.G. Underwood, & B.H. Gottlieb (Eds.), Social support
measurement and intervention: A guide for health and social scientists (pp. 246-277).
New York: Oxford University Press.
SOCIAL SUPPORT IN CAMBODIA 79

Edwards, M., Hulme, D. and Wallace, T. (1999) NGOs in a global future: marrying local
delivery to worldwide leverage. Public Administration and Development 19, 117-
136. Retrieved September 23, 2012 from http://www.futurepositive.org/edwards.php.
Ekstrand, M.L., & Coates, T.J. (1990). Maintenance of safer sexual behaviors and predictors of
risky sex: The San Francisco men's health study. American Journal of Public Health,
80(8), 973-977.
Ertel, K.A., Glymour, M.M., & Berkman, L.F. (2009). Social networks and health: A life
course perspective integrating observational and experimental evidence. Journal of
Social and Personal Relationships, 26(1), 73-92.
Fantahun, M., Berhane, Y., Wall, S., Byass, P., & Hogberg, U. (2007). Women’s involvement in
household decision-making and strengthening social capital—crucial factors for child
survival in Ethiopia. Acta Paediatricia, 96, 582-589.
Feldman, P.J., Dunkel-Schetter, C., Sandman, C.A., & Wadhwa, P.D. (2000). Maternal social
support predicts birth weight and fetal growth in human pregnancy. Psychosomatic
Medicine, 62, 715-725.
Frieson, K. (2001). In the shadows: Women, power and politics in Cambodia (Occasional Paper
26). Victoria, BC: Centre for Asia-Pacific Initiatives.
Goldman, J.A. & Harlow, L.L. (1993). Self-perception variables that mediate AIDS-preventive
behavior in college students. Health Psychology, 6, 489-498.
Goldman, N., Pebley, A.R., & Beckett, A. (2001). Diffusion of ideas about personal hygiene and
contamination in poor countries: evidence from Guatemala. Social Science & Medicine,
52, 53-69.
SOCIAL SUPPORT IN CAMBODIA 80

Gorman, S., Dorina, P., & Kheng, S. (1999). Gender and development in Cambodia: An
overview (Working Paper 10). Phnom Penh, Cambodia: Cambodia Development
Resource Institute.
Gottlieb, B.H. (1981). Preventive interventions involving social networks and social support. In
B.H. Gottlieb (Ed.), Social Networks and Social Support (pp. 201-232). Beverly Hills,
CA: Sage Publications.
Green, B.L. & Rodgers, A. (2001). Determinants of social support among low-income mothers:
A longitudinal analysis. American Journal of Community Psychology, 29(3), 419-441.
Hall, J.A. (2010). An exploratory study of communication, gender-role conflict, and social
support of parents of children treated at Children’s Hospital. Journal of Psychosocial
Oncology, 28(5), 511-525. DOI: 10.1080/07347332.2010.49846.
Hinton, A. L. (1998). Why did you kill? The Cambodian genocide and the dark side of face and
honor. The Journal of Asian Studies, 57(1), 93-122.
Hobfoll, S.E., Shoham, S.B., & Ritter, C. (1991). Women's satisfaction with support and their
receipt of aid. Journal of Personality and Social Psychology, 61(2), 332-341.
Hobfoll, S.E., Freedy, J., Lane, C., & Geller, P. (1990). Conservation of social resources: Social
support resource theory. Journal of Social and Personal Relationships, 7, 465-478.
House, J.S., Landis, K.R., & Umberson, D. (1988). Social relationships and health. Science,
241(4865), 540-545.
Hupcey, J.E. (1998). Clarifying the social support-theory linkage. Journal of Advanced Nursing,
27, 1231-1241.
SOCIAL SUPPORT IN CAMBODIA 81

Hussain, A., Aaro, L.E., & Kvale, G. (1997). Impact of a health education program to promote
consumption of vitamin A rich foods in Bangladesh. Health Promotion International,
12(2), 103-109.
Israel, B.A. (1985). Social networks and social support: Implications for natural helper and
community level interventions. Health Education Quarterly, 12(1), 65-80.
Israel, B.A., Farquhar, S.A., Schulz, A.J., James, S.A., Parker, E.A. (2002). The relationship
between social support, stress, and health among women on Detroit’s eastside. Health
Education & Behavior, 29(3), 342-360.
Jacobs, B. & Price, N. (2003). Community participation in externally funded health projects:
Lesson from Cambodia. Health Policy and Planning, 18(4), 399-410.
Janz, N.K. & Becker, M.H. (1984). The Health Belief Model: A decade later. Health Education
Quarterly, 11(1), 1-47.
Kaniasty, K. & Norris, F. H. (2000). Help-seeking comfort and receiving social support: The
role of ethnicity and context of need. American Journal of Community Psychology,
28(4), 545-581.
Kaplan, R.M. & Hartwell, S.L. (1987). Differential effects of social support and social network
on physiological and social outcomes in men and women with Type II diabetes mellitus.
Health Psychology, 6(5), 387-398.
Kazdin, A.E. (1974). Covert modeling, model similarity, and a reduction of avoidance behavior.
Behavior Therapy, 5, 325-340.

SOCIAL SUPPORT IN CAMBODIA 82

Kelly, J. A., St. Lawrence, J.S., Diaz, Y. E., Stevenson, L.Y., Hauth, A.C., Brasfield, T.L.,
Kalichman, S.C., Smith, J.E., & Andrew, M.E. (1991). HIV risk behavior reduction
following intervention with key opinion leaders of population: An experimental analysis.
American Journal of Public Health, 81(2), 168-171.
Kelly, J.A., St. Lawrence, J.S., Stevenson, L.Y., Hauth, A.C., Kalichman, S.C., Diaz, Y.E.,
Brasfield, T.L., Koob, J.J., Morgan, M.G. (1992). Community AIDS/HIV risk reduction:
The effects of endorsements by popular people in three cities. American Journal of
Public Health, 82(11), 1483-1489.
Kelly, J.A., Murphy, D.A., Sikkema, K.J., McAuliffe, t.L., Roffman, R.A., Solomon, L.J., et al.
(1997). Randomized, controlled, community-level HIV prevention intervention for sexual
risk behavior among homosexual men in U.S. cities. Lancet, 35, 1500-1504.
Kelly, J. A. (2004). Popular opinion leaders and HIV prevention peer education: resolving
discrepant findings, and implications for the development of effective community
programs. AIDS Care, 16(2), 139-150.
Kim, H.S., Sherman, D.K., Ko, D., & Taylor, S.E. (2006). Pursuit of comfort and pursuit of
harmony: Culture, relationships, and social support seeking. Personality and Social
Psychology Bulletin, 32(12), 1595-1607.
Kim, H.S., Sherman, D.K., & Taylor, S.E. (2008). Culture and social support. American
Psychologist, 63(6), 518-526.
Kincaid, D.L. (2000). Social networks, ideation, and contraceptive behavior in Bangladesh: a
longitudinal analysis. Social Science & Medicine, 50, 215-231.
SOCIAL SUPPORT IN CAMBODIA 83

King, J. & Ashworth, A. (1987). Historical review of the changing pattern of infant feeding in
developing countries: The case of Malaysia, The Caribbean, Nigeria and Zaire. Social
Science and Medicine, 25(12), 1307-1320.
Langlie, J.K. (1977). Social networks, health beliefs, and preventive health behavior. Journal of
Health and Social Behavior, 18(3), 244-260.
Ledgerwood, J. (1994). Gender symbolism and culture change: Viewing the virtuous woman in
the Khmer story "Mea Youeng". In M. Ebihara, C.A. Mortland, & J. Ledgerwood (Eds.),
Cambodian culture since 1975: Homeland and Exile. Ithaca, NY: Cornell University
Press.
Ledgerwood, J., Ebihara, M., & Mortland, C., (1994). Introduction. In M. M. Ebihara, C.A.
Mortland, and J. Ledgerwood (Eds.), Cambodian culture since 1975: Homeland and
exile. Ithaca: Cornell University Press.
Lehman, D.R. & Hemphill, K.J. (1990). Recipients’ perceptions of support attempts and
attributions for support attempts that fail. Journal of Personal and Social Relationships,
7, 563-574.
Lin, N., Ensel, W.M., Simeone, R.S., & Kuo, W. (1979). Social support, stressful life events,
and illness: A model and an empirical test. Journal of Health and Social Behavior, 20(2),
108-119.
Lowery, L.M., Chung, S., & Ellen, J.M. (2005). Social support and sexually transmitted disease
related healthcare utilisation in sexually experienced African-American adolescents.
Sexually Transmitted Infections, 81, 63-66.
Lu, L. (1995). Correlates of social support: Personal characteristics and social resources.
Counselling Psychology Quarterly, 8(2), 173-181.
SOCIAL SUPPORT IN CAMBODIA 84

Mitchell, R.E. & Hurley, D.J. Jr. (1981). Collaboration with natural helping networks: Lessons
from studying paraprofessionals. In B.H. Gottlieb (Ed.) Social networks and social support
(pp. 277-298). Beverly Hills, CA: Sage Publications.
Morling, B., Kitayama, S., & Miyamoto, Y. (2003). American and Japanese women use different
coping strategies during normal pregnancy. Personality and Social Psychology Bulletin,
29(X), 1-14. DOI: 10.1177/0146167203256878.
National Institute of Statistics (2008, August). General population census of Cambodia 2008
Summary. National Institute of Statistics, Ministry of Planning, Phnom Penh, Cambodia.
Nazzar, A., Adongo, P.B., Binka, F.N., Phillips, J.F. & Debpuur, C. (1995). Developing a
culturally appropriate family planning program for the Navrongo experiment. Studies in
Family Planning, 26(6), 307-324.
Newcomb, M.D. (1990). Social support by many other names: Toward a unified
conceptualization. Journal of Social and Personal Relationships, 7(4), 479-494.
Norr, K.F, Crittenden, K.S., Lehrer, E.L., Reyes, O., Boyd, C.B., Nacion, K.W., & Wantanabe,
K. (2003). Maternal and infant outcomes at one year for a nurse-health advocate home
visiting program serving African Americans and Mexican Americans. Public Health
Nursing, 20(3), 190-203.
Olds, D.L. & Kitzman, H. (1993). Review of research on home visiting for pregnant women and
parents of young children. The Future of Children, 3(3), 53-92.
Rahman, A., Iqbal, Z., & Harrington, R. (2003). Life events, social support and depression in
childbirth: Perspectives from a rural community in the developing world. Psychological
Medicine, 33, 1161-1167.
SOCIAL SUPPORT IN CAMBODIA 85

Ramey, C.T. & Ramey, S.L. (1993). Home visiting programs and the health and development of
young children. The Future of Children, 3(3), 129-139.
Robertson, S.E. & Valadez, J.J. (2006). Global review of health care surveys using lot quality
assurance sampling (LQAS), 1984-2004. Social Science and Medicine, 63, 1648-1660.
Rosenstock, I.M., Stretcher, V.J., & Becker, M.H. (1988). Social learning theory and the health
belief model. Health Education Quarterly, 15(2), 175-183.
Sarason, I.R., Pierce, G.R., & Sarason, B.R. (1990). Social support and interactional processes:
A triadic hypothesis. Journal of Social and Personal Relationships, 7(4), 495-506.
Sarason, I.G., Sarason, B.R., & Shearin, E.G. (1986). Social support as an individual difference
variable: Its stability, origins, and relational aspects. Journal of Personality and Social
Psychology, 50(4), 845-855.
Schaefer, C., Coyne, J.C., & Lazarus, R.S. (1981). The health-related functions of social
support. Journal of Behavioral Medicine, 4(4), 381-406.
Schwarzer, R., Dunkel-Schetter, C. & Kemeny, M. (1994). The multidimensional nature of
received social support in gay men at risk of HIV Infection and AIDS. American Journal
of Community Psychology, 22(3), 319-339.
Shumaker, S.A. & Hill, D.R. (1991). Gender differences in social support and physical health.
Health Psychology, 10(2), 102-111.
Stretcher, V.J., DeVellis, B.M., Becker, M.H., & Rosenstock, I.M. (1986). The role of self-
efficacy in achieving health behavior change. Health Education Behavior, 13, 73-92.
Stretcher, V.J. & Rosenstock, I.M. (1997). The Health Belief Model. In A. Baum, S. Newman, J.
Weinman, R. West, & C. McManus (Eds.), Cambridge handbook of psychology, health
and medicine (pp. 113-117). Cambridge: Cambridge University Press.
SOCIAL SUPPORT IN CAMBODIA 86

Taylor, S.E., Sherman, D.K., Kim, H.S., Takagi, J.J., & Dunagan, M.S. (2004). Culture and
social support: Who seeks it and why? Journal of Personality and Social Psychology, 87
(3), 354-362.
Taylor, S.E., Welch, W.T., Kim, H.S., & Sherman, D.K. (2007). Cultural differences in the
impact of social support on psychological and biological stress responses. Psychological
Science, 18(9), 831-837.
Thoits, P.A. (1995). Stress, coping, and social support processes: Where are we? What next?
Journal of Health and Social Behavior (Extra Issue), 53-79.
Thornton, P.L., Kieffer, E.C., Salabarria-Pena, Y., Odoms-Young, A., Willis, S.K., Kinn, H., &
Salinas, M.A. (2006). Weight, diet, and physical activity-related beliefs and practices
among pregnant and postpartum Latino women: The role of social support. Maternal and
Child Health Journal, 10(1), 95-104.
Tomaka, J., Thompson, S., & Palacios, R. (2006). The relation of social isolation, loneliness,
and social support to disease outcomes among the elderly. Journal of Aging and Health,
18(3), 359-384.
Turner, G. & Shepherd, J. (1999). A method in search of a theory: peer education and health
promotion. Health Education Research, 14(2), 235-247.
Umberson, D., Chen, M.D., House, J.S., Hopkins, K., & Slaten, E. (1996). The effect of social
relationships on psychological well-being: Are men and women really so different?
Sociological Review, 61(5), 837-857.
United Nations Programme on HIV/AIDS. (2007). AIDS epidemic update. Geneva,
Switzerland: Author.
SOCIAL SUPPORT IN CAMBODIA 87

UN Millennium Project 2005. Investing in development: A practical plan to
achieve the millennium development goals. Overview. New York: Author.
United Nations Children’s Fund. (2007, December). The state of the world’s children 2008. New
York: Author
UNICEF. (2011). Cambodia country statistics. Retrieved October 11, 2011 from
http://www.unicef.org/infobycountry/cambodia_statistics.html.
Vaux, A. & Harrison, D. (1985). Support network characteristics associated with support
satisfaction and perceived support. American Journal of Community Psychology, 13(3),
245-267.
Wallston, B.S., Alagna, S.W., DeVellis, B.M., & DeVellis, R.F. (1983). Social support and
physical health. Health Psychology, 2(4), 367-391.
Waterkeyn, J. & Cairncross, S. (2005). Creating demand for sanitation and hygiene through
Community Health Clubs: A cost-effective intervention in two districts in Zimbabwe.
Social Science & Medicine, 61(9), 1958-1970.
Williams, A.W., Ware, J.E., & Donald, C.A. (1981). A model of mental health, life events, and
social supports applicable to general populations. Journal of Health and Social Behavior,
22, 324-336.
World Health Organization. (2011). Maternal mortality in 1990-2008. Retrieved October 19,
2011from http://www.who.int/gho/maternal_health/countries/en/.
World Relief. (2004). The Care Group difference: A guide to mobilizing community-based
volunteer health educators. Baltimore, MD: Author.
SOCIAL SUPPORT IN CAMBODIA 88

World Relief Cambodia and Adventist Development and Relief Association (ADRA) Cambodia.
(2007, May). Baseline survey report: Sokhapehap Phoum Yeung (Our Healthy Village).
Phnom Penh, Cambodia: Author.
World Relief Cambodia and Adventist Development and Relief Association (ADRA) Cambodia.
(2010, January). Third year report: Sokhapehap Phoum Yeung (Our Healthy Village).
Phnom Penh, Cambodia: Author.
Zakus, J.D.L. & Lysack, C.L. (1998). Revisiting community participation. Health Policy &
Planning, 13(1), 1-12.

SOCIAL SUPPORT IN CAMBODIA 89

Endnotes
1
The goal of LQAS methodology is to determine whether a program has reached its target
coverage of knowledge or behavior change by determining the percentage of people who have
been reached by an intervention. The procedure requires dividing the survey zone into
catchment areas so that different areas can be targeted depending on LQAS results. For the
current project, staff used six geographic areas as the catchment zones. These corresponded with
six operational health districts (OD) in Kampong Cham province, Cambodia. Kampong Cham
province is located along the eastern border of Cambodia and Vietnam. Operational health
district boundaries are determined by the national Ministry of Health, using population and
geography as guidelines. Kampong Cham province has a total of nine ODs, but only six were
reached by the Sokhapheap Phum Youeng (SPY) project. These districts are Memut, Ponhea
Kreak-Dambae, Prey Chhor-Kong Meas, Srey Sentour-Kong Meas, Chamkaleu-Steung Trong,
and Chung Prey-Bathey.
2
Because village homes are not numbered and streets are not well organized, an
alternative sampling method had to be used. One member of the survey team threw a pen on the
ground, and the team began with the house indicated by the direction faced by the tip of the pen.
In cases where no one was home, teams used the ‘next front door method’ which involves
selecting homes based on which has the nearest front door to the house initially selected.
3
Neighboring households are often comprised of family members; for example, one
family may build a home on only half of a plot of land, giving the other half to a son or daughter
upon their marriage. Thus, selecting a larger sampling interval eliminates the possibility that
interviewers would unintentionally choose family members or non-independent data points for
the sample.
SOCIAL SUPPORT IN CAMBODIA 90

4
There was some confusion regarding the subjectivity of the measures. The principal
investigator explained that the purpose of the social support measure was to determine how the
respondent felt regarding the volunteer visits, rather than searching for a “correct” answer.
When social support measures were posed to respondents, if the individual did not understand
how to answer the question, the interviewer explained that the person was to choose from the
scaled response items and clarified the numerical values, if necessary. It is also possible that this
was the first time respondents had answered using these types of scaled items. If respondents
were confused by the response items, interviewers were instructed to clarify the meaning of none
as “receiving no support or visits,” a little as “receiving 1-2 visits or instances of help,” some as
“receiving 3-5 visits or instances of help,” and a lot as “receiving more than 5 visits or many
instances of help.” For other scaled items, interviewers were instructed to read each of the
numbers (one through seven) to the respondent, and not to subjectively determine what a
respondent meant when a question was answered, for example, with “good” or “important.”
Interviewers were also given a standard example to use when explaining the meaning of certain
questions to an uncertain respondent. They were told to describe a visit in which a volunteer
provided instruction on oral rehydration therapy, in which the volunteer gave information,
demonstrated the use of oral rehydration solution, or told the individual where they could
purchase a solution packet. Because individuals were assumed to have familiarity with oral
rehydration therapy by the final survey, the term was used during interviews. If participants did
not understand the example, the interviewer explained it again, including a definition of oral
rehydration therapy.
SOCIAL SUPPORT IN CAMBODIA 91

Table 1. List of Intervention Topics and Sample Size by Stratification Group.

Intervention Topic

Breastfeeding
Newborn
Care
ORS
Knowledge
Knowledge
of ARI
Symptoms
Use of
Contraception

HIV
Prevention
Knowledge

Women with children 0-5 months 119 119

Women with children 6-9 months

Women with children under 2
years

119 119

Women with children under 5
years

119 119

119

Women 15-49 years

119 119

Men 15-49 years
119

For all stratification groups, n = 119.
SOCIAL SUPPORT IN CAMBODIA 92

Table 2. Breastfeeding and Antenatal Care Variables among Mothers with Children Under
6 Months.

n
Ever Breastfed

114
Yes

1 (0.01%)
No

1 (0.01%)
Not applicable (currently breastfeeding)

112
(98.2%)

Currently Breastfeeding

228
Yes

223
(97.8%)
No

5 (2.2%)

Length of time between delivery and breastfeeding

113
1 hour

87 (77%)
> 1 hour

12 (10.6%)
1-3 days

11 (9.7%)
More than 3 days

3 (2.7%)

Index of Antenatal Care Behaviors
c

114

Immediate breastfeeding

86
a
(77.5%)
Immediate drying & wrapping

44
b

(39.3%)
Immediate cord care

32
a

(28.8%)
Other cord care

19 (16.8%)

Don't know

10 (8.8%)

a
For this variable, n = 111.
b
For this variable, n = 112.

c
Responses to each were summed to form index of antenatal care behaviors (M =
3.54, SD = 1.39, range = 0-5).

SOCIAL SUPPORT IN CAMBODIA 93

Table 3. Diarrheal Illness and ARI Symptom Knowledge Among Mothers of Children Under 5
Years.

n
ORS Procedure Steps

110

Boil 1 liter or use pure water

87
a
(78.4%)

Use 1 package of ORS

84 (76.4%)

Dissolve packet fully

75 (68.2%)

Describe the 3 steps of the ORS procedure

110

Incorrect

42 (38.2%)

Correct

68 (61.8%)

Never heard of ORS

--

Heard of ARI?

114

Yes

111 (97.4%)

No

3 (2.6%)

Index of ARI Knowledge
b

111

Difficulty breathing

102 (91.9%)

Chest rising and falling

86 (77.5%)

Cough

66 (59.5%)

High fever

83 (74.8%)

Nostrils flaring

12 (10.8%)

Other

18 (16.2%)

Don't know/no answer

3 (2.7%)

a
For this variable, n = 111.

b
Responses were summed to form index of ARI symptom knowledge (M = 3.14, SD = .98,
range = 0-5)

SOCIAL SUPPORT IN CAMBODIA 94

Table 4. Contraceptive Use and HIV Prevention Knowledge Among Mothers of Children
Under 2 Years, 5 Years, and Men and Women age 15 to 49.

n
Using Contraception 113

No 65 (57.5%)

Yes 48 (42.5%)

Want a child in the next two years 114

No 99 (86.8%)

Yes 13 (11.4%)

Other 2 (1.8%)

Contraceptive currently used 159

Pill 29 (18.2%)

Injection 37 (23.3%)

Condoms 10 (6.3%)

Tubal Ligation 3 (1.9%)

Intrauterine Device (IUD) 4 (2.5%)

Vasectomy 0 (0.0%)

Other 32 (20.1%)

None 44 (38.9%)

Heard of HIV/AIDS 456

Yes 451 (98.9%)

No 5 (1.1%)

Index of HIV prevention knowledge
a

449

Nothing
0
b
(0.0%)

Abstinence 139 (31%)

Condom Use 435 (96.9%)

Faithfulness 319 (71%)

Avoid kissing 10 (2.2%)

Avoid mosquito bites 1 (0.2%)

a
Other response options were offered, but these six received the most responses and were the only
responses used to create the index of HIV prevention knowledge(M = 1.99, SD = .76, range = 0-3).
b
For this variable, n= 405.

SOCIAL SUPPORT IN CAMBODIA 95

Table 5. Volunteer Status and Peer Educator Visits Among Full Sample.
n
Do you volunteer with the SPY program? 682
Yes 30 (4.32%)
No
652 (95.6%)

Talked with a Care Group member in last 6 months? 678
Yes 399 (58.8%)
No 279 (41.2%)

SOCIAL SUPPORT IN CAMBODIA 96

Table 6. Item Means and Standard Deviations for Tangible Social Support Variables.

n
Mean
(SD)
Range
How much did this person help by giving you tasks to keep your family healthy? 397

2.40
(.61)
3
How much did this person help you by giving you things you needed? 400

2.29
(.69)
3
How much did this person help you to access different services? 397

2.30
(.74)
3
After talking with this person, how much did you feel that you had the knowledge to
improve your family's health?
384

2.25
(.65)
3
After talking with this person, how much did you feel that you could do things on your
own to keep your family healthy?
392

2.20
(.66)
3
After talking with this person, how much did you feel that when your family was sick
you could find help or provide help to them?
388

2.34
(.69)
3

SOCIAL SUPPORT IN CAMBODIA 97

Table 7. Item Means and Standard Deviations for Perceived Social Support Variables.

n
Mean
(SD)
Range
How much did you feel that this person cared about the health of your family? 399

5.18
(1.40)
7
How easy was it to ask questions of this person? 402

5.68
(1.38)
7
Rate the overall helpfulness of this person 385

5.83
(1.33)
7
Overall, how important do you think this person's input is in helping you make health
decisions?

401
6.05
(1.24)
7

SOCIAL SUPPORT IN CAMBODIA 98

Table 8. Factor Analysis of Social Support Scales

Component
1 2 Extraction
How much did this person help you by giving you some
things you needed to keep your family healthy?
.780 -.343 .557
How much did this person help you by giving you tasks to
keep your family healthy?
.746 -.331 .611
How much did this person help you to access different
services, like the health center or VCCT, you needed to
keep your family healthy?
.736 -.338 .543
After talking with this person, how much did you feel that
you had the knowledge to improve your family’s health?
.730 -.380 .533
After talking with this person, how much did you feel that
you could do things on your own to keep your family
healthy?
.622 -.332 .388
After talking with this person, how much did you feel that
when your family was sick you could find help or provide
help to them?
.599 -.543 .441
Overall helpfulness of this person .353 -.804 .594
Overall, how important do you think this person’s input is
in helping you make health decisions?
.406 -.785 .573
How much did you feel that this person cared about the
health of you and your family?
.460 -.764 .648
How easy was it to ask questions of this person? .321 -.755 .618

Extraction Method: Principal Component Analysis.
Rotation Method: Oblimin with Kaiser Normalization.

SOCIAL SUPPORT IN CAMBODIA 99

Table 9. Pearson and Point-Biserial Correlations Between Social Support, Demographics, and Health Outcome
Variables.

Measure One Two Three Four Five Six Seven Eight Nine Ten Eleven Twelve
1. Peer Educator Visit
--

2. Perceived Social Support Scale
.02 --

3. Tangible Social Support Scale
-.01 .32
b
--

4. Immediate Breastfeeding
-.18
-
.24
a

-.06 --

5. Index of Post-Delivery Behaviors
.25
b
.07 .22
a
-.25
a
--

6. HIV Prevention Knowledge
.22
b
.03 .09 -- -- --

7. Contraceptive Use
-.13 -.03 .06 -- -- .00 --

8. Index of ARI Symptom Knowledge
.16 -.01 .29
a
-- -- .34
b
-- --

9. ORS Knowledge
.20
*
.09 .22 -- -- .27
b
-- .26
b
--

10. Age
-.00 .03 .06 -.03 .02 -.05 .01 -.17 -.11 --

11. Education
-.02 .03 .07 -.00 -.12 .07 -.02 .02 .06
-
.18
b

--

12. Volunteer Status
.14
b
.01 .06 -.08 .07 .09 -.04 .25
b
.19
a
.08
b
-.01 --
13. Number of Children
.06 .07 -.03 -.17 .03 -.05 .06 -.15 .07 .63
b
-.24
b
.04

a = p < .05, b = p < .01

SOCIAL SUPPORT IN CAMBODIA 100

Table 10. Contrasts in peer educator visits for dependent health outcome variables.

Peer Ed Visit
No Yes

Variable M (SD) M (SD) t(df)

ARI Knowledge
2.91
(.88)
3.24
(1.01)
-1.66
(109)

HIV Knowledge
1.80
(.74)
2.14
(.75)
-4.55
(402)
b

Index of Post-Delivery Behaviors (all)
3.00
(1.32)
3.77
(1.36)
-2.73
(112)
b

Index of Post-Delivery Behaviors (first time
mothers)
2.77
(1.01)
3.68
(1.45)
-2.10
(49)
a

a = p < .05, b = p < .01

SOCIAL SUPPORT IN CAMBODIA 101

Table 11. Differences in Dichotomous Health Outcome Variables by Type of Social Support.

Low High Total Low High Total No Yes Total
ORS Knowledge 11
b
30
a
41 13 35 48 16
b
52
a
68
% within SS variable 26.8% 73.2% 59.1% 79.5% 72.7% 47.1% 68.4% 61.8%
Total % 15.3% 41.7% 56.9% 19.7% 53.0% 14.5% 47.3%
Contraceptive use 13 22 35 13 21 34 39 43 82
% within SS variable 48.1% 44.9% 46.1% 46.4% 52.5% 50.0% 59.1% 46.2% 51.6%
Total % 17.1% 28.9% 19.1% 30.9% 24.5% 27.0%
Immediate Breastfeeding 28 42 70 23 45 68 25 74 99
% within SS variable 84.8% 97.7% 92.1% 88.5% 91.8% 90.7% 76.1% 91.4% 87.6%
Total % 36.8% 55.3% 30.7% 60.0% 22.1% 65.5%
Note: cells with different subscripts differ from each other at the p < .05 level.
Variable Name
Perceived Social Support Tangible Social Support Peer Educator Visit
SOCIAL SUPPORT IN CAMBODIA 102

Table 12. Logistic regression of social support variables on dependent health outcomes.
95% CI for exp(B)
Dependent Included B(SE) SE Wald OR Lower Upper

Knowledge of ORS
a
Constant 1.49 1.45 1.06 4.44 -- --
PSS*TSS -.39 1.51 .07 .68 .04 12.93
PSS .59 .99 .35 1.80 .26 12.50
TSS -1.49 1.19 1.55 .23 .02 2.36

Use of Contraception
b
Constant
-.81 1.16 .49

.44
-- --
PSS*TSS 1.05 1.22 .74

2.86 .26 31.47
PSS .41 .96 .18 1.50 .23 9.79
TSS .41 .96 .18
1.50
.23 9.79
Immediate Breastfeeding
c
Constant -1.23 1.46 .71 .29 -- --
PSS*TSS -19.97 6996.69 .00 .00 .00 .00
PSS -.07 1.31 .003 .93 .07 12.11
TSS -.72 .99 .52 .49 .07 3.44
Immediate Breastfeeding
d
Constant -19.53 28420.77 .00 .00 -- --
First-time mothers only PSS*TSS -1.67 29958.12 .00 .19 .00 .00
PSS 19.12 28420.77 .00 201934486.80 .00 .00
TSS -1.67 1.39 1.43 .19 .01 2.91
a: R
2
= .02 (Cox & Snell), .03 (Nagelkerke). Model X
2
(3) = 1.15, p = .77, -2LL = 72.31.
b: R
2
= .05 (Cox & Snell), .08 (Nagelkerke). Model X
2
(3) = 3.35, p = .34, -2LL = 68.83.
c: R
2
= .11 (Cox & Snell), .25 (Nagelkerke). Model X
2
(3) = 8.41, p = .04, -2LL= 32.89.
d: R
2
= .19 (Cox & Snell), .43 (Nagelkerke). Model X
2
(3) = 7.29, p < .10, -2LL = 13.01.
SOCIAL SUPPORT IN CAMBODIA 103

Table 13. Mean differences in health outcome variables by type of social
support.

Perceived Social Support

Low High

Variable Name

M (SD)

M (SD)

HIV Prevention Knowledge

Low TSS

2.21 (.78)

2.04 (.75)

Male

2.42 (.79)

2.00 (.71)

Female

2.16 (.78)

2.05 (.78)

High TSS

2.19 (.79)

2.28 (.64)

Male

3.00 (.00)

2.25 (.50)

Female

2.12 (.78)

2.28 (.65)

ARI Symptom Knowledge

Low TSS

3.20 (1.21)

2.57 (1.27)

High TSS

3.67 (.50)

3.47 (.67)

Antenatal Care

Low TSS

3.59 (1.54)

2.75 (1.83)

High TSS

3.64 (1.39)

4.15 (1.06)

Antenatal Care--First-time Mothers Only

Low TSS

4.11 (1.17)

1.50 (2.12)

High TSS

2.80 (1.92)

4.00 (1.14)

SOCIAL SUPPORT IN CAMBODIA 104

Table 14. Gender Differences in Social Support.

Men Women Total

Visited by Peer Educator 38 363 401
% within gender 33.9% 63.7% 58.8%
Total % 5.6% 53.2%
Not Visited by Peer Educator 74 207 281
% within SS variable 66.1% 36.3% 41.2%
Total % 10.9% 30.4%

High PSS 19 190 209

% within SS variable 51.4% 58.1% 57.4%

Total % 5.2% 52.2%

Low PSS 18 137 155

% within SS variable 48.6% 41.9% 42.6%

Total % 4.9% 37.6%

High TSS 10 182 192

% within gender 38.5% 59.1% 57.5%

Total % 3.0% 54.5%

Low TSS 16 126 142

% within gender 61.5% 88.7% 42.5%

Total % 4.8% 37.7%

SOCIAL SUPPORT IN CAMBODIA

Figure 1. Map of Sokhapheap Phum Youeng

Chamkar Leu
Steung Trong
Srei Santhor-
Kong Meas
Prey Chhor
Kong Meas
Cheung Prey—
Bantheay
SOCIAL SUPPORT IN CAMBODIA
Sokhapheap Phum Youeng (SPY) project area.
Chamkar Leu –
Trong
Ponhea
Kreak- Dambae
Prey Chhor-
Kong Meas
105

Memut
SOCIAL SUPPORT IN CAMBODIA 106

Figure 2. Interaction between Perceived Social Support and Tangible Social Support on
Antenatal Care Behaviors.

2
2.5
3
3.5
4
4.5
Low PSS High PSS
High
TSS
Low
TSS

Abstract (if available)

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

The effect of Twitter social support on health outcomes and its mediators: a randomized controlled trial of a social support intervention on Twitter for patients affected by cancer

PDF

The formation and influence of online health social networks on social support, self-tracking behavior and weight loss outcomes

PDF

Benefits of social networking in online social support groups

PDF

The role of social control in promoting healthy eating behavior among Chinese immigrants: an ecological approach

PDF

Social media and health: social support and social capital on pregnancy-related social networking sites

PDF

School-based Health Centers 2.0: Using social media to increase utilization of UMMA Wellness Center Services

PDF

Communication and social support of parents of children treated at Childrens Hospital Los Angeles

PDF

Building personal wellness communities: meaningful play in the everyday life of a network society

PDF

Developing and testing a heuristic-systematic model of health decision making: the role of affect, trust, confidence and media influence

PDF

Relationship formation and information sharing to promote risky health behavior on social media

PDF

Culture, causal attribution, and social support-seeking in Asian college students

PDF

Diffusion of a peer‐led suicide preventive intervention in secondary schools: strategies to increase effectiveness of peer‐led interventions

PDF

Voice, participation & technology in India: communication technology for development in the mobile era

PDF

“You move, therefore I am”: the combinatorial impact of kinesthetic motion cues on social perception

PDF

Contagious: social norms about health in work group networks

PDF

Peers as institutional agents: acquiring social capital through peer interactions

PDF

The social functions of celebrity gossip among high school girls

PDF

A network analysis of online and offline social influence processes in relation to adolescent smoking and alcohol use

PDF

Network influences in health initiatives: multimedia games for youth in Peru

PDF

Immigrant health in context: a communication ecology approach to understanding health behavior

Asset Metadata

Creator Pieper, Katherine M. (author)

Core Title Social support in Cambodia: the role of peer educators in behavior change

School Annenberg School for Communication

Degree Doctor of Philosophy

Degree Program Communication

Publication Date 11/23/2012

Defense Date 10/15/2012

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

Tag Cambodia,health,OAI-PMH Harvest,peer education,social support

Language English

Contributor Electronically uploaded by the author (provenance)

Advisor Smith, Stacy L. (committee chair), Cody, Michael J. (committee member), Farver, Jo Ann M. (committee member)

Creator Email kpieper@usc.edu

Permanent Link (DOI) https://doi.org/10.25549/usctheses-c3-119053

Unique identifier UC11292668

Identifier usctheses-c3-119053 (legacy record id)

Legacy Identifier etd-PieperKath-1340.pdf

Dmrecord 119053

Document Type Dissertation

Rights Pieper, Katherine M.

Type texts

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

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

Repository Name University of Southern California Digital Library

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