Trinity University

Digital Commons @ Trinity

Psychology Faculty Research Psychology Department

2010

Peer-Facilitated Cognitive Dissonance versus Healthy Weight

Eating Disorders Prevention: A Randomized Comparison
Carolyn Becker
Trinity University, [email protected]

Chantale Wilson
Trinity University

Allison Williams
Trinity University

Mackenzie Kelly
Trinity University

Leda McDaniel
Trinity University

See next page for additional authors

Follow this and additional works at: https://digitalcommons.trinity.edu/psych_faculty

Part of the Psychology Commons

Publication Details
Body Image

Repository Citation
Becker, C.B., Wilson, C., Williams, A., Kelly, M., McDaniel, L., Elmquist, J. (2010). Peer-facilitated cognitive
dissonance versus healthy weight eating disorders prevention: A randomized comparison. Body Image,
7(4), 280-288. doi: 10.1016/j.bodyim.2010.06.004

This Post-Print is brought to you for free and open access by the Psychology Department at Digital Commons @
Trinity. It has been accepted for inclusion in Psychology Faculty Research by an authorized administrator of Digital
Commons @ Trinity. For more information, please contact [email protected].
Authors
Carolyn Becker, Chantale Wilson, Allison Williams, Mackenzie Kelly, Leda McDaniel, and Joanna Elmquist

This post-print is available at Digital Commons @ Trinity: https://digitalcommons.trinity.edu/psych_faculty/8

Running Head: PEER-FACILITATED COGNITIVE DISSONANCE 1

Peer-Facilitated Cognitive Dissonance versus Healthy Weight Eating Disorders Prevention: A

Randomized Comparison

Carolyn Black Becker, Chantale Wilson, Allison Williams, Mackenzie Kelly, Leda McDaniel,

Joanna Elmquist

Trinity University

In press: Body Image

Please do not distribute without permission of authors.

Author Note

The authors would like to thank the sororities of Trinity University for their ongoing

collaboration and support.

Correspondence concerning this article should be addressed to Carolyn Becker, PhD, FAED

Department of Psychology, Trinity University, San Antonio, TX 78212.

Email: [email protected]
Running Head: PEER-FACILITATED COGNITIVE DISSONANCE 2

Abstract

Research supports the efficacy of both cognitive dissonance (CD) and healthy weight

(HW) eating disorders prevention, and indicates that CD can be delivered by peer-facilitators,

which facilitates dissemination. This study investigated if peer-facilitators can deliver HW when

it is modified for their use and extended follow-up of peer-facilitated CD as compared to

previous trials. Based on pilot data, we modified HW (MHW) to facilitate peer delivery,

elaborate benefits of the healthy-ideal, and place greater emphasis on consuming nutrient dense

foods. Female sorority members (N=106) were randomized to either two 2-hour sessions of CD

or MHW. Participants completed assessment pre- and post-intervention, and at 8-week, 8-month,

and 14-month follow-up. Consistent with hypotheses, CD decreased negative affect, thin-ideal

internalization, and bulimic pathology to a greater degree post-intervention. Both CD and MHW

reduced negative affect, internalization, body dissatisfaction, dietary restraint, and bulimic

pathology at 14 months.

Keywords: eating disorders; prevention; cognitive dissonance; healthy weight; dissemination

Running Head: PEER-FACILITATED COGNITIVE DISSONANCE 3

Peer-facilitated cognitive dissonance versus healthy weight eating disorders prevention:

A randomized comparison

Eating disorders (EDs) are associated with a wide range of medical complications

involving all major systems of the body (Kaplan & Woodside, 1987; Mitchell & Crow, 2006),

and in severe cases EDs may lead to death (Herzog et al., 2000). Further, even the most

empirically supported treatment for eating disorders (i.e., cognitive behavioral therapy for

bulimia nervosa) is only effective for about 60% of those completing treatment (Fairburn et al.,

1995). The seriousness of these disorders coupled with the dearth of empirically supported

treatments make effective prevention efforts crucial.

Early ED prevention programs typically adopted a universal psychoeducation approach

and, unfortunately, proved largely ineffective at reducing ED pathology (Stice & Shaw, 2004;

Stice, Shaw, & Marti, 2007). Indeed the early history of ED prevention was marked by failure to

improve attitudes and behaviors (Pearson, Goldklang, & Striegel-Moore, 2002). Researchers,

however, recently have made marked progress in developing more effective ED prevention

programs. One source of this progress has been the development of cognitive dissonance-based

interventions (CD: see Stice, Shaw, Becker, & Rhode, 2008 for review on CD and empirical

support). Cognitive dissonance theory states that people will work to resolve inconsistencies

between their beliefs and actions (Festinger, 1962). CD induces dissonance by having

participants engage in activities opposing the thin-ideal standard of female beauty. Theoretically,

these actions are inconsistent with participants’ assumed investment in this culturally reinforced

ideal, and thus produce dissonance. In order to resolve dissonance, participants decrease their

investment in the thin-ideal, which should reduce other risk factors implicated in the

development of ED pathology (see Stice, 1994; Stice, 2001).

Running Head: PEER-FACILITATED COGNITIVE DISSONANCE 4

CD is one of very few ED prevention programs to meet the American Psychological

Association’s (APA) criteria for an efficacious intervention (APA, 1995), meaning that CD has

outperformed no-intervention control groups, outperformed at least one alternative intervention,

and results have been replicated by independent researchers. CD has been studied by five

independent labs (see Stice, Shaw, et al., 2008 for review of all studies; see also Green, Scott,

Diyankova, Gasser, & Pederson, 2005; Matusek, Wendt, Wiseman, 2004; Mitchell, Mazzeo,

Rausch, & Cooke; 2007), and has been found to reduce the following ED risk factors: negative

affect, internalization of the thin-ideal standard of female beauty, body dissatisfaction,

maladaptive dietary restraint, and bulimic pathology (Stice, Shaw et al., 2008). Two meta-

analyses also found that ED prevention programs producing the largest effects were those that

were interactive (not didactic), multi-session (not single session), and dissonance-based (not

psychoeducational) (Stice & Shaw, 2004; Stice et al., 2007). Moreover, in one trial CD reduced

onset of EDs to 6% at 3 years compared to 15% for attention control in young women with body

dissatisfaction (Stice, Marti, Spoor, Presnell, & Shaw, 2008). CD also has proven effective in

mixed risk populations. Becker and colleagues also have demonstrated via a series of studies that

CD reduces ED pathology (typically measured via the diagnostic items of the Eating Disorder

Examination-Questionnaire) when delivered to collegiate women who had not been selected

because of body dissatisfaction (Becker, Smith, & Ciao, 2005; Becker, Smith, & Ciao, 2006;

Becker, Bull, Schaumberg, Cauble, & Franco, 2008; Perez, Becker, & Ramirez, in press). This is

important because many social systems prefer to deliver programs more universally as opposed

to selecting (and possibly stigmatizing) high risk individuals. Becker and colleagues also

demonstrated that undergraduate students could be trained to effectively lead CD to their mixed

risk peers (Becker et al., 2006; Becker, Bull, Schaumberg et al., 2008).
Running Head: PEER-FACILITATED COGNITIVE DISSONANCE 5

Another ED prevention program that has garnered some significant empirical support is

the Healthy Weight (HW) intervention. Developed by Stice and colleagues initially as a placebo

control group against which to compare CD, HW also has been found to reduce ED risk factors

such as body dissatisfaction, negative affect, dieting, and bulimic pathology (Stice, Chase,

Stormer, & Appel, 2001; Stice, Shaw, Burton, & Wade, 2006). Like CD, HW also has been

found to reduce onset of eating disorders at 3 year follow-up, and long term changes in risk

factors are comparable for both interventions at 3 years (Stice, Marti et al., 2008).

Whereas the goal of CD is for participants to reject the thin-ideal, HW promotes the

balancing of caloric intake and output in order to attain a healthy body weight. In HW,

participants evaluate eating and exercise habits through self-monitoring and then commit to

small, manageable changes to their diet and exercise for homework assignments. Once

participants experience progress in changing unhealthy habits and successfully balance intake

and output, they theoretically feel empowered to achieve a healthy weight for their body type.

Both the behavior change tools and increased self-efficacy are presumed to contribute to reduced

body dissatisfaction, negative affect, and ED pathology. HW may not decrease thin-ideal

internalization to the degree that CD does (Stice, Trost, & Chase, 2003), however. This may be

because participants do not engage in dissonance inducing activities, and may still feel pressure

(from outside sources and/or themselves) to pursue the thin-ideal.

As noted above, research indicates that CD can be delivered by collegiate peer-

facilitators when they receive sufficient training and manuals are adjusted to their needs. Having

peers (i.e., endogenous providers who are natural to the social system) deliver interventions is

potentially one method for improving large-scale dissemination of efficacious programs given

that many social systems do not have the resources to administer interventions to a large number
Running Head: PEER-FACILITATED COGNITIVE DISSONANCE 6

of participants with doctoral or even masters-level clinicians. For instance, based on the research

supporting CD generally and peer-facilitated CD specifically, the Delta Delta Delta fraternity

(i.e., Tri Delta sorority) has begun large scale dissemination of peer-facilitated CD (i.e.,

Reflections: Body Image Program) with the aim of reaching 20,000 collegiate women over 5

years (Becker, Stice, Shaw & Woda, 2008). The Tri Delta executive office has noted that peer-

facilitation was a crucial factor in making the program financially feasible. Research also

indicates that administering CD can have positive implications for the peer-facilitators

themselves. Becker, Bull, Smith, & Ciao (2008) found that peer-facilitators experienced further

reductions in ED risk factors beyond those they experienced as participants in the same program.

The purpose of the present randomized effectiveness/dissemination study was to

determine whether or not a modified version of HW (MHW) could be effectively delivered by

endogenous providers (i.e., peer-facilitators), which would assist dissemination. Although CD

has been embraced by specific communities (e.g., sororities), other communities may prefer a

program that has a weight management/nutrition/physical activity focus. For instance, in a

related line of ongoing research, we have found that many athletes prefer the face validity of a

program that targets eating behavior versus body image. Similarly, schools that face pressure to

address obesity as well as eating disorders might prefer a program that not only reduces eating

disorder risk factors but may also reduce obesity onset; HW has been shown in one trial to

reduce risk for obesity onset 1 year following the intervention (Stice et al., 2006). In summary, it

makes sense to provide communities with as many effective prevention options as possible so as

to meet their needs.

In a pilot study (Becker, Schaumberg, Mallett, Hay, & Williams, 2007), we found that

peer-facilitators showed poor adherence to HW even though we already had modified structural
Running Head: PEER-FACILITATED COGNITIVE DISSONANCE 7

aspects of the HW protocol (e.g., having peer-facilitators guide participants through a self review

of food logs as opposed to having peer-facilitators personally review logs) so as to make the

intervention more amenable to peer-facilitation. Furthermore, a small but meaningful number of

participants actively provided negative feedback about HW. More specifically, participants

reported feeling that the original version of HW encouraged pursuit of the thin-ideal and made

them upset about weight gain that they had experienced since coming to college. Based on this,

we reviewed session audiotapes which suggested that because HW had a focus on managing

weight many participants had trouble distinguishing pursuit of the healthy-ideal, which we define

as however a woman’s body looks when she is doing everything she can to appropriately and

simultaneously maximize physical health, mental health, and quality of life, versus the thin-ideal

standard of female beauty. As such, participants interpreted peer-facilitated HW as promoting the

thin-ideal. This is likely because undergraduate peer-facilitators have limited clinical skills and

the manual was not sufficiently tailored for their needs.

To address these issues, we decided to further modify the healthy weight protocol and

rename it MHW so as to differentiate the version used in this study with that used by Stice et al.

(2006; 2008). Revisions were not made with a specific theory in mind. Rather, we viewed this as

an iterative process whereby we sought to alter the intervention in response to participants’

confusion about pursuit of the healthy- versus thin-ideal. Thus, whereas HW largely focuses on

tuning the caloric intake/output balance to obtain a healthy body weight, MHW also adds a

greater focus on increasing the nutrient density of one’s diet and exploration of the benefits of

the healthy-ideal and differences compared to the thin-ideal so that participants understand that

the goal of tuning energy intake and output is healthy weight management not pursuit of the thin-

ideal. We added the new components to provide peer facilitators with explicit tasks to help
Running Head: PEER-FACILITATED COGNITIVE DISSONANCE 8

participants better understand that the intervention was promoting healthy weight management.

Based on past experiences with CD, we have learned that giving peer-facilitators specific

dialogue works better than expecting them to spontaneously address complex issues. We also

added psychoeducation about the role lack of sleep can play in weight gain. Because this line of

participatory research involves delivering interventions to an entire social system, some

participants without marked eating/exercise concerns attend the interventions. Many college

students have poor sleep habits, however, and this gives students who have good eating/exercise

behaviors an additional option for improving a behavior that can impact weight via sleep’s

effects on leptin and ghrelin (Taheri, Lin, Austin, Young & Mignot, 2004).

This universal-selective prevention study compares peer-facilitated MHW to peer-

facilitated CD when administered to a mixed risk population of college women (see Becker et

al., 2005; Becker, Bull, Schaumberg et al., 2008 for evidence showing that sororities consist of

members who are both at low and high risk, which is why we refer to this approach as universal-

selective). Because CD has been shown to produce superior results to waitlist/assessment only

controls (e.g., Becker et al., 2005; Stice et al., 2006; 2008), alternative interventions (e.g., Becker

et al., 2006) and placebo control (e.g., Stice et al., 2006), we employed a comparative

intervention strategy (Kazdin, 2003) and chose to compare peer-facilitated MHW with the ED

prevention program that has amassed the greatest amount of empirical support (i.e., CD). We

also chose to adopt this strategy because the present study was conducted as part of a long-term

participatory research relationship with a sorority system. Participatory research methodology

involves insuring that all relevant stakeholders have a real voice in decision making (Israel, Eng,

Schulz, & Parker, 2005), and sorority leaders did not believe that the research benefits of a no-

intervention control group outweighed the clinical downsides to their members given the
Running Head: PEER-FACILITATED COGNITIVE DISSONANCE 9

extensive data showing that CD produces positive effects. The present study also adds a longer

follow-up than has been previously published with peer-facilitated CD. Consistent with the

findings of Stice et al. (2006), we predict that CD will reduce negative affect, thin-ideal

internalization, body dissatisfaction, dietary restraint, and bulimic pathology to a greater degree

than MHW at post-intervention, but that differences between the two interventions will dissipate

by 14-month follow-up (also consistent with Stice et al., 2006). We hypothesize that both CD

and MHW will produce reductions in negative affect, thin-ideal internalization, body

dissatisfaction, dietary restraint, and bulimic pathology long term when delivered by endogenous

providers (i.e., peers).

Method

Participants

New members who joined any of the seven local sororities at Trinity University in

February 2008 were required to participate in an annual body image program associated with

sorority orientation. The program consists of attending the group sessions. Participation in the

study, however, which consisted of filling out several questionnaires, was completely voluntary.

Out of the 114 new members who joined a sorority, 5 were excused from participation in the

program due to academic commitments or de-pledging. Ninety-seven percent of the remaining

members participating in the program agreed to participate in the study (N = 106). As in our

previous studies (Becker et al., 2005; 2006; 2008), participants (n = 4) who appeared to meet

criteria for an ED based on their responses to the Eating Disorder Examination-Questionnaire

(Fairburn & Beglin, 1994) were excluded from all analyses because the interventions being

studied are aimed at prevention not treatment. The remaining 102 participants ranged in age from

18 to 21 years (M = 18.73, SD = 0.72). Mean body mass index (BMI), based on self-reported
Running Head: PEER-FACILITATED COGNITIVE DISSONANCE 10

height and weight, was 22.07 (SD = 3.33). The majority of participants self-identified their

ethnicity as Caucasian (80%). The remainder endorsed African American (1%), Asian (2%),

more than one race (10%), or no response (7%). Participants received no compensation for their

participation.

Procedure

Overview. As appropriate, the program and study were reviewed and approved by the

Institutional Review Board (IRB), Greek Council, and Student Affairs at Trinity University (e.g.,

Greek Council approved both study and program whereas IRB approved only the study). A list

of all new sorority members was compiled, and new members were randomized by

undergraduate research assistants (RAs) into 12 groups of 8 to 10 members stratified by sorority.

Thus, each group had roughly equal representation of members from each sorority. After the

groups were created the RAs blindly randomized the groups to either CD or MHW. At the

beginning of the program, new sorority members from each of the seven campus sororities

attended a brief orientation session. Consenting participants completed baseline questionnaires

using a self-generated ID number to ensure confidentiality and then placed these in a large

envelope. All participants were informed that they could pretend to fill out the questionnaires

and return them in the envelope to reduce coercion. After completing questionnaires, participants

met with their assigned groups to begin the first session. The second session took place exactly

one week after the first session. Sessions were designed to last approximately 105 minutes plus

time to complete questionnaires. Consenting participants filled out post-intervention

questionnaires after the completion of the second session. Additional questionnaires were

completed at 8-week, 8-month, and 14-month follow-up during each respective sorority’s weekly

meeting.
Running Head: PEER-FACILITATED COGNITIVE DISSONANCE 11

All sessions were audio taped to evaluate peer-facilitator adherence to the program

manuals. Each tape was rated by two raters who determined to what degree each key step in the

highly scripted manuals was completed using an adherence rubric. Raters consisted of

undergraduate RAs who had participated in over 40 hours of training on the interventions.

Because kappa, which is commonly used to assess inter-rater reliability, can at times produce

surprisingly low scores even when rater agreement is very high (Gwet, 2002), we utilized the

alternative chance corrected AC1 statistic proposed by Gwet (2002) to assess inter-rater

reliability, which was excellent in this study (AC1 = .96). In contrast to Becker, Schaumberg et

al. (2007), in which peer-facilitator adherence to HW was poor, we found that peer-facilitators

had good adherence to both the CD and MHW protocols.

Peer-facilitator training. Research assistants (RAs) recruited interested sorority

members at weekly meetings for each respective sorority. All peer-facilitators were sophomores,

juniors, or seniors who had previously participated in an ED prevention intervention in one of

our earlier studies. Potential peer-facilitators were asked to refrain from becoming a peer-

facilitator if they were suffering from an ED or had significant body image concerns. Further we

emphasized the need for peer-facilitators to be able to be a positive role model for health and

body image even after the program was completed because the appearance of hypocrisy on the

part of peer-facilitators could damage the program. Peer-facilitators were told that if they chose

to withdraw we would assume it was because the time commitment was too high, so that there

was no social stigma involved in withdrawing. Peer-facilitators were randomly assigned to two

4.5 hour CD or MHW training sessions based on their availability to attend training. Experiential

training sessions were conducted by a licensed psychologist (CBB) and undergraduate RAs.

During training, peer-facilitators were given an overview of the respective intervention (i.e., CD
Running Head: PEER-FACILITATED COGNITIVE DISSONANCE 12

or MHW), and in teams of three, ran an abbreviated version of the intervention while the

remaining peer-facilitators acted as mock participants. Training sessions included nine peer-

facilitators and so, by the end of training, each peer-facilitator had experienced the intervention

once as a facilitator and twice as a participant. Peer-facilitators received detailed supervision

after each practice session and had the opportunity to ask questions or voice concerns about

leading the interventions. Thus, all peer-facilitators heard three rounds of supervision.

Intervention programs. CD and MHW consisted of two sessions. At the start of the first

session of both interventions, participants (a) commited to give the program a try and keep an

open mind, and (b) agreed to keep personal information brought up in the groups confidential.

Dissonance-based intervention. In session one, CD participants: (a) defined the thin-

ideal, (b) discussed the origin of the thin-ideal and how it is perpetuated, (c) brainstormed costs

of pursuing the thin-ideal, (d) participated in a verbal challenge activity (i.e., identified a time

when they felt pressure to pursue the thin-ideal and indicated how they could have responded to

counter the thin-ideal message) and (e) were given a mirror homework assignment. The mirror

assignment asked them to stand in front of a mirror wearing as little clothing as they felt

comfortable and list positive physical and emotional qualities about themselves. In session two,

participants (a) reviewed the mirror assignment, (b) engaged in role plays in which peer-

facilitators assumed the roles of women invested in the thin-ideal and participants tried to

discourage pursuit of the thin-ideal, (c) discussed ways to challenge and avoid common “fat talk”

statements, (d) listed ways to resist pressure to pursue the thin-ideal both individually and

collectively as sororities (called “body activism”) (e) discussed possible barriers to body activism

and ways to overcome those barriers, and (f) individually selected a self-affirmation exercise to

continue their practice of positive body talk.

Running Head: PEER-FACILITATED COGNITIVE DISSONANCE 13

Modified healthy weight. In session one, MHW participants (a) defined the thin-ideal,

(b) defined the healthy-ideal and contrasted it with the thin-ideal (c) discussed the importance of

eating nutrient dense foods in maintaining an intake/output balance (d) listed the benefits of

aspiring to a healthy-ideal, (e) discussed the importance of sleep in maintaining a healthy weight

and body, (f) listed reasons to pursue the healthy-ideal, and (g) were given two homework

assignments. The first assignment involved selecting a specific individual health goal to change

within the next week. The health goal could be eating, exercise, or sleep related. Because the

program is run on a semi-mandatory basis, it is possible that an occasional participant who is

randomized to MHW may report not needing to change exercise or eating1. In this situation, the

participant can pick sleep – which also is often poor in college students. Participants were

encouraged, however, to pick eating or exercise over sleep if either area needed improvement,

and to do so for health not pursuit of the thin-ideal. The second assignment asked participants to

keep a food log for two weekdays and one weekend day, and an exercise log for the whole week.

In session two, participants (a) reviewed the benefits of pursuing the healthy-ideal, (b) discussed

the difference between healthy dietary restriction (e.g., moderate, flexible, aimed at pursuing the

healthy-ideal in a obesegenic food culture) and unhealthy dietary restriction (rigid, overly

restrictive, extreme, typically aimed at pursuing the thin-ideal), (c) reviewed food and exercise

logs, (d) identified healthy changes they could make to improve their diet with respect to nutrient

density, along with barriers to such change, and strategies to overcome barriers (e) discussed

specific ways to make meals more nutrient dense (f) discussed the benefits of exercise (g)

identified healthy changes they could make to be more active, along with barriers to change, and

strategies to overcome barriers (h) discussed ways that sororities could promote a healthy-ideal
Running Head: PEER-FACILITATED COGNITIVE DISSONANCE 14

for their members, and (i) committed to specific goals to continue their pursuit of the healthy-

ideal.

Measures

Negative affect. Negative affect was assessed with the fear, guilt, and sadness subscales

from the Positive and Negative Affect Schedule-Revised (PANAS-X; Watson & Clark, 1992).

Participants indicated how much they had been feeling various emotional states (e.g., nervous,

scared, and lonely) over the past few weeks by providing a rating from 1 = very slightly or not at

all to 5 = extremely (scale range: 1-5). Scores from the 17 items were averaged. Past research

with this scale has demonstrated good internal consistency (α = .95), convergent validity with

affective measures, and predictive validity for bulimic symptom onset (Stice & Agras, 1998;

Watson & Clark, 1992). In the present study internal consistency was good (α = .89).

Thin-ideal internalization. Thin-ideal internalization was assessed with the Ideal Body

Stereotype Scale-Revised (IBSS-R; Stice, Ziemba, Margolis, & Flick, 1996). This scale consists

of 10 items, in which participants endorse how much they agree (1 = strongly disagree, 5 =

strongly agree; scale range: 1-5) with statements such as “thin women are more attractive.”

Scores from the items were averaged. In past studies, this scale has demonstrated adequate

internal consistency (α = .89) and test-retest reliability (r = .63) (Stice, 2001; Stice & Agras,

1998). Internal consistency in the present sample was consistent with past research (α = .87).

Body dissatisfaction. Body dissatisfaction was calculated using the shape concern and

weight concern subscales from the Eating Disorder Examination Questionnaire, (EDE-Q;

Fairburn & Beglin, 1994). The EDE-Q is a self report version of the Eating Disorder

Examination, which is currently considered the “gold standard” for assessing ED pathology

(EDE; Fairburn & Cooper, 1993). The EDE-Q, which has been extensively researched and tested
Running Head: PEER-FACILITATED COGNITIVE DISSONANCE 15

for its psychometric properties (e.g. see Luce & Crowther, 1999; Mond, Hay, Rodgers, Beumont,

& Owen, 2004; Peterson, et al., 2007), has been widely used. The shape concern and weight

concern EDE-Q subscales assess body disattisfaction over the past month and both subscales

have shown good internal consistency at baseline (shape concern, α = .93; weight concern, α =

.89) and 2-week test-retest reliability (shape concern, r = .94; weight concern, r = .92) (Luce &

Crowther, 1999). Body dissatisfaction scores were calculated by adding scores from the shape

concern subscale and weight concern subscale and dividing by two. The two scales were highly

correlated (r = .92), which provides support for combining them into one scale, and internal

consistency for this combined scale was excellent (α = .94) in the present study.

Dietary restraint. Although dietary restraint measures have not been shown to be good

measures of actual dietary intake (Stice, Fischer & Lowe, 2004), these measures have been

shown to predict bulimic pathology (Stice, Fischer & Lowe). Thus, we used the restraint

subscale from the EDE-Q (Fairburn & Beglin, 1994) to measure this bulimic predictive

construct. On this measure, participants report on how many days over the past month they have

engaged in dietary restraint (e.g., “Have you had a definite desire to have an empty stomach with

the aim of influencing your shape or weight?”). This 5-item subscale assessing restraint has

shown good internal consistency (α = .84) and 2-week test-retest reliability (r = .81) (Luce &

Crowther, 1999). Internal consistency in the present study was adequate (α = .78).

Bulimic pathology. As in our previous peer-facilitated studies (Becker et al., 2006;

2008) we generated a composite bulimic scale from the diagnostic items (e.g., “over the past 28

days how many times have you taken laxatives as a means of controlling your shape or weight?”)

of the EDE-Q (Fairburn & Beglin, 1994) to assess bulimic pathology. This 10-item measure
Running Head: PEER-FACILITATED COGNITIVE DISSONANCE 16

assesses to what degree participants have engaged in bulimic behaviors over the past month.

Internal consistency for the bulimic composite was adequate (α = .81) in the present sample.

Analysis

Despite the lack of differential dropout, all analyses were conducted on an intent-to-treat

basis and missing data points were filled in with maximum likelihood imputation procedures for

participants who did not complete all follow-ups. By 14-month follow-up, nine participants had

subsequently undergone training as peer-facilitators for the next year’s program. These

participants were evenly distributed between the two interventions (5 in CD and 4 in MHW).

Analyses including and excluding these participants did not yield significantly different results.

Therefore, all analyses included data from these participants.

In order to test our preliminary hypothesis that CD would produce greater decreases than

MHW on all dependent measures at post-intervention, ANOVAs were conducted with the

intervention (CD, MHW) as the between-subjects factor and time as a two level (pre, post)

within-subjects factor. For our main analyses, we used time (pre-, post-, 8-week, 8-month, 14-

month) as the within-subjects factor. Skewed EDE-Q data was normalized using a square root

transformation and skewed PANAS-X data was normalized using a logarithmic transformation.

A one-way ANOVA revealed no significance baseline differences between interventions for all

dependent measures, age, or BMI. Table 1 shows dependent variable means by intervention

group and time, as well as calculated Cohen’s d effect sizes for post-intervention and all follow-

ups. Table 1 also shows results from paired t-tests between baseline and all post-intervention

time points for each group. These analyses should not be used to infer differences between

groups beyond those supported by ANOVA analyses.

Running Head: PEER-FACILITATED COGNITIVE DISSONANCE 17

Results

Participant Flow

Of the 102 participants who were included in the analyses, 90 (88%) completed 8-week

follow-up, 83 (81%) completed 8-month follow-up, and 75 (74%) completed 14-month follow-

up (see Figure 1). Three participants dropped out after the first session and three dropped out

after session 2. There was no significant difference between conditions in dropout rates.

Preliminary Pre-Post Analyses

Negative affect. The repeated-measures ANOVA (pre-post) for negative affect revealed

a significant Time effect and a Time x Group interaction (Table 2). There was no effect for

Group. Consistent with our hypothesis, CD produced a larger effect size than MHW (Table 1).

Thin-ideal internalization. The repeated-measures ANOVA for thin-ideal

internalization revealed a significant effect of Time and a Time x Group interaction (Table 2).

CD yielded a significantly greater decrease in internalization as compared to MHW (Table 1).

There was no effect for Group.

Body dissatisfaction. The repeated-measures ANOVA for body dissatisfaction revealed

a significant effect of Time. There was no effect for Group or a Time x Group interaction (Table

2). Thus, our hypothesis was not supported.

Dietary restraint. The repeated-measures ANOVA for dietary restraint revealed a

significant effect of Time. There was no effect for Group or Time x Group interaction, which

indicates that results failed to support our hypothesis about CD being superior (Table 2).

Bulimic pathology. The repeated-measures ANOVA for bulimic pathology revealed a

significant effect of Time and a Time x Group interaction (Table 2). CD yielded larger effects as

compared to MHW (Table 1). There was no effect for Group.

Running Head: PEER-FACILITATED COGNITIVE DISSONANCE 18

Main Analyses

Negative Affect

The repeated-measures ANOVA for negative affect revealed a significant effect of Time,

but no effect of Group and no Time x Group interaction (see Table 3). Post hoc analyses

indicated that CD reduced negative affect at all post-intervention time points, whereas MHW

only showed significant reductions at 8 months and 14 months.

Thin-Ideal Internalization

The repeated-measures ANOVA for internalization also revealed a significant Time

effect. There was no effect for Group or Time x Group interaction (Table 3). Consistent with our

hypotheses, both groups showed reductions in thin-ideal internalization at 14 months. Table 1

displays additional results from post-hoc tests.

Body Dissatisfaction

The repeated-measures ANOVA for body dissatisfaction revealed a significant effect of

Time. There was no effect for Group or Time x Group interaction (Table 3). Post-hoc tests

indicated that both CD and MHW significantly reduced body dissatisfaction from baseline at all

post-intervention time points, including 14 months.

Dietary Restraint

The repeated-measures ANOVA for dietary restraint revealed a significant effect of

Time. There was no effect for Group or Time x Group interaction (Table 3). Both groups

decreased dietary restraint at 14 months, and all other post-intervention assessments with the

exception of 8 months for MHW.

Bulimic Pathology
Running Head: PEER-FACILITATED COGNITIVE DISSONANCE 19

The repeated-measures ANOVA for bulimic pathology revealed a significant effect of

Time. There was no effect for Group or Time x Group interaction (Table 3). Once again, both

groups showed comparable and significant reductions in bulimic pathology over 14 months (see

Table 1 for additional detail).

Discussion

This study sought to replicate and extend findings for two ED prevention programs that

have garnered empirical support, CD and MHW. As predicted, both CD and MHW reduced ED

risk factors at 14-month follow-up. At post-intervention, CD produced significantly greater

reductions than MHW in negative affect, thin-ideal internalization, and bulimic pathology. This

is an important and unusual finding in that it suggests that at post-intervention CD produces

larger effects than a credible and structurally equivalent alternate prevention intervention. This

outcome is rare, though not unheard of, in prevention studies. Indeed, this very specific result has

been found in a previous study comparing CD with HW (Stice, et al., 2006), despite differences

in participants and providers, which suggests that it may represent a real difference between CD

and MHW/HW. This finding also provides support for the notion that these interventions operate

via different mechanisms and suggests that CD may be a faster acting intervention than

MHW/HW, which makes sense given that the benefits of changing health behaviors may take

time to become evident. As noted above, however, differences between CD and MHW were not

present for any measures at longer follow-up (all p > .293), suggesting that differences between

these two interventions fade over time.

It is unclear why initial differences fade. For the most part, it appears that MHW largely

caught up to CD, although CD consistently produced larger – but not significantly larger – effect

sizes at 14 months. One option, as implicated above, is that MHW simply takes more time to
Running Head: PEER-FACILITATED COGNITIVE DISSONANCE 20

produce effects. Alternatively, the context in which these interventions are delivered may

partially explain the dissipation of initial differences. Participants are members of a social system

(group of local sororities) that is actively attempting to reject the thin-ideal and to reduce the use

of “fat talk” statements on a daily basis secondary to the annual use of CD for over 7 years.

Thus, MHW participants are eventually exposed to some of the components of CD. Also, peer-

facilitators, who have the greatest exposure to the program, are encouraged to push their fellow

sorority members to “live” the overarching message of the program (i.e., reject the thin-ideal and

embrace the healthy-ideal) on a daily basis even after the interventions are completed. This

tendency has become even more pronounced now that peer-facilitated CD, which was developed

with the local sororities at Trinity University, has been launched throughout North America (i.e.,

Reflections: Body Image Program) by Tri Delta. Campus sorority members take significant pride

in seeing what they perceive as “their program” expanded to a significant number of other

campuses (e.g., 41 in 2009-2010). In addition, sorority members report using what they learn

during the program to collaborate with other members as “body activists,” both individually and

within their sororities. Thus, continual sorority-wide collaboration in rejecting the thin-ideal may

cause the initial differences seen between CD and MHW to decrease over time. It should be

noted, however, that Stice et al. (2006) also found that post-intervention differences faded

between CD and HW at 1 year follow-up and they did not run their study in a structured social

system – which suggests that what was observed in this study is not solely due to this setting.

One alternative explanation is that sorority members naturally improve over time on ED

risk factors. This seems implausible for several reasons. First, although limited data address this

question, Allison and Park (2004) found that 57 sorority women showed no change, either

positive or negative, in drive for thinness, body dissatisfaction and bulimia over a three year
Running Head: PEER-FACILITATED COGNITIVE DISSONANCE 21

period. Second, in a previous study we found that CD significantly outperformed a waitlist

control group at one month (Becker et al., 2005) and that the control group showed virtually no

changes from pre-intervention to one-month follow-up (dietary restraint d = -.01, thin-ideal

internalization d = .14; body dissatisfaction d = .01; bulimic pathology d = -.11). Further, in

another study (Becker et al., 2006) we found that whereas CD produced significant reductions in

dietary restraint, thin-ideal internalization, body dissatisfaction and bulimic pathology at 8

months in sorority members, an alternate credible media advocacy intervention failed to do so in

3 out of the 4 dependent variables (d range -.03 to .14) suggesting no natural improvement over

that time period and an equivalent sample. Finally, in yet another study we found that whereas

low risk sorority members who participated in CD improved in risk status over 8 months, low-

risk members who participated in an alternate intervention showed small but worsening changes

over time (Becker, Bull, Schaumberg et al., 2008). Thus it seems unlikely that the significant

results that were found here resulted from natural improvement. Rather, given that Stice et al.

(2006) found that both CD and HW significantly outperformed assessment only at one year and

given that within group effect sizes in the present study at 14 months (CD d range = 0.48 - 0.97;

HW d range = 0.34-0.83) compare closely to those reported by Stice et al. for 12 months (CD d

range = 0.48 - 0.65; HW d range = 0.32-0.63), it appears that this study replicated previous

findings.

Despite espousing markedly different aims, both CD and MHW do contain certain

similarities that may bolster their effectiveness beyond the strict content of the sessions. Indeed,

Stice and colleagues have posited that such overlapping techniques (e.g., motivational exercises

and public commitments to change) may contribute to the concurrent effectiveness of CD and

HW (Stice, Shaw et al., 2008). Trials for both interventions also have been largely multi-session,
Running Head: PEER-FACILITATED COGNITIVE DISSONANCE 22

which allows participants to complete homework assignments reinforcing topics covered within

sessions. For instance, for homework in CD, participants are asked to stand in front of a mirror

and list positive qualities about themselves, both physical and emotional. In HW (and MHW),

participants are asked to identify and commit to specific changes in diet and exercise between

sessions. These assignments are individualized in that they require participants to apply what

they have discussed within sessions to their own personal body image and health concerns. Thus,

HW clearly utilizes some of the general principles that have been included in CD and only

research dismantling each intervention will be able to parcel out the exact contributions from

different components.

Results from the current study also suggest that MHW may be superior to HW in terms of

the viability of using peer-facilitators for dissemination given that adherence was poor in our

pilot study of peer-facilitated HW (Becker, Schaumberg et al., 2007) and good in the present

examination of MHW. We also had no student reports of distress with peer-facilitated MHW in

contrast to peer-facilitated HW. Thus, the additional modifications to MHW appear to have been

sufficient to allow peers to lead this intervention.

This study also extended follow-up as compared to our previous trials (Becker et al.,

2006; 2008) and indicated that effects for peer-led CD remain even at 14 months. Eight-month

effect sizes for CD were also fairly consistent with previous trials, providing additional

replication. Importantly, as noted above, review of means and standard deviations in Stice et al.

(2006), which is arguably one of the best controlled eating disorders prevention studies

conducted to date, indicate that the 14-month effect sizes for CD in the present study are

generally similar (e.g., negative affect, thin-ideal internalization) or larger (e.g., body

dissatisfaction, dietary restraint, bulimic pathology) than the one year effect sizes found by Stice
Running Head: PEER-FACILITATED COGNITIVE DISSONANCE 23

et al. This suggests that using peer-facilitators to deliver a two-session version of CD in a

supportive social system may produce similar effects to the 4 session non-peer-facilitated version

of CD. This is particularly exciting given that at 3 years Stice et al. (2008) found that CD

reduced onset of eating disorders by 60%. This is also impressive given that Stice et al.

employed a high risk sample with elevated body image concerns, whereas the present study

accepted all new sorority members and our previous research has shown that sorority members

comprise a mixed risk population with both lower and higher risk members (Becker et al., 2005;

2008). Thus, the present study is less likely to be influenced by regression to the mean and may

have had greater problems with floor effects.

There are several limitations to this study, the first being that assessment was limited to

self-report due to its utility in non-clinical settings. Secondly, there may have been spillover

effects between the two intervention groups because sororities members spend a significant

amount of time with one another. Third, although we used intent-to-treat analyses and

determined that dropout was not different between groups, at 14-month follow-up we only

retained three quarters of our sample, less than Stice and colleagues typically retain. It should be

noted, however, that this was an unfunded naturalistic study and that, in contrast to Stice et al.

(2006), participants were not paid for attending follow-up. Thus, the similarity in findings

between this study and the Stice et al. study support the generalizability of Stice et al.’s results.

Finally, in this study, we did not use a no intervention control group for comparison. The present

study is best viewed as a dissemination and implementation study that is aimed at determining

how well interventions perform in real world settings when implemented by cost-effective

endogenous providers. As noted above, past efficacy studies have shown that both CD and HW
Running Head: PEER-FACILITATED COGNITIVE DISSONANCE 24

outperform waitlist control (CD only: Becker et al., 2005; CD & HW: Stice et al., 2003) and

assessment only control groups (CD & HW; Stice et al., 2006; 2008).

This study suggests that both CD and MHW are effective when delivered by peer-

facilitators, which is encouraging in lieu of current and future dissemination efforts. It should be

noted that informal qualitative feedback from the sororities seemed to indicate that participants

preferred CD to MHW, although peer-facilitators who were exposed to both CD as a participant

in earlier studies and MHW as facilitators in this study felt that both were useful. Further, they

described MHW as a good follow-up to CD. It remains to be seen, however, if there truly is an

additive benefit to receiving both interventions.

This study also supports the longer-term effectiveness of peer-facilitated CD, which

extends past research. Future research would be helpful in order to ascertain the effective

components of each intervention and explore the use of message framing in ED prevention.

Replication of results, particularly with respect to MHW is also needed.

Running Head: PEER-FACILITATED COGNITIVE DISSONANCE 25

References

Allison, K. C., & Park, C. L. (2004). A prospective study of disordered eating among sorority

and nonsorority women. International Journal of Eating Disorders, 35, 354-362.

American Psychological Association Task Force on Psychological Intervention Guidelines.

(1995). Template for developing guidelines: Interventions for mental disorders and

psychological aspects of physical disorders. Washington, DC: American Psychological

Association.

Becker, C. B., Bull, S., Schaumberg, K., Cauble, A., & Franco, A. (2008). Effectiveness of peer-

led eating disorders prevention: A replication trial. Journal of Consulting and Clinical

Psychology, 76, 347-354.

Becker, C. B., Bull, S., Smith, L. M., & Ciao, A. C. (2008). Effects of being a peer-leader in an

eating disorder prevention program: Can we further reduce eating disorder risk factors?

Eating Disorders: The Journal of Treatment and Prevention, 16, 444-459.

Becker, C., Schaumberg, K., Mallett, J., Hay, L., & Williams, A. (2007, September). A

randomized pilot trial of peer-led healthy weight eating disorders prevention. Poster

presented at the 2007 Annual Meeting of the Eating Disorder Research Society, Pittsburgh,

PA.

Becker, C. B., Smith, L. M., & Ciao, A. C. (2005). Reducing eating disorder risk factors in

sorority members: A randomized trial. Behavior Therapy, 36, 245-253.

Becker, C. B., Smith, L. M., & Ciao, A. C. (2006). Peer-facilitated eating disorder prevention: A

randomized effectiveness trial of cognitive dissonance and media advocacy. Journal of

Counseling Psychology, 53, 550-555.

Running Head: PEER-FACILITATED COGNITIVE DISSONANCE 26

Becker, C. B., Stice, E., Shaw, H., & Woda, S. (2009). Use of empirically-supported interventions for

psychopathology: Can the participatory approach move us beyond the research-to-practice gap?

Behaviour Research and Therapy, 47, 265-274.

Fairburn, C. G., & Bèglin, S. J. (1994). Assessment of eating disorders: Interview or self-report

questionnaire? International Journal of Eating Disorders, 16, 363-370.

Fairburn, C.G., & Cooper, Z. (1993). The Eating Disorder Examination (12th ed.). In C. G.

Fairburn & G. T. Wilson (Eds.) Binge eating: Nature, assessment, and treatment. New

York: Guilford Press.

Fairburn, C. G., Cooper, Z., Doll, H. H., Norman, P., & O’Connor, M. (2000). The natural

course of bulimia nervosa and binge eating disorder in young women. Archives of General

Psychiatry, 57, 659-665.

Fairburn, C. G., Norman, P. A., Welch, S. L., O’Connor, M. E., Doll, H. A., & Peveler, R. C.

(1995). A prospective study of outcome in bulimia nervosa and the long-tern effects of three

psychological treatements. Archives of General Psychiatry, 52, 304-312.

Festinger, L (1962). A theory of cognitive dissonance. Stanford, CA: Stanford University Press.

Green, M., Scott, N., Diyankova, I., Gasser, C., & Pederson, E. (2005). Eating disorder

prevention: An experimental comparison of high level dissonance, low level dissonance,

and no-treatment control. Eating Disorders: The Journal of Treatment & Prevention, 13,

157-169.

Gwet, K. (2002). Kappa statistic is not satisfactory for assessing the extent of agreement between

raters. Statistical Methods for Inter-rater Reliability Assessment, 1, 1–6.

Running Head: PEER-FACILITATED COGNITIVE DISSONANCE 27

Herzog, D. B., Greenwood, D. N., Dorer, D. J., Flores, A. T., Ekeblad, E. R., Richards, A., et al.

(2000). Mortality in eating disorders: A descriptive study. International Journal of Eating

Disorders, 28, 20-26.

Israel, B. A., Eng, E., Schulz, A. J., & Parker, E. A. (2005). Introduction to methods in

community-based participatory research for health. In Israel, B. A., Eng, E., Shulz, A. J., &

Parker, E. A. (Eds.), Methods in community-based participatory research for health (pp. 3-

26). San Francisco: Jossey-Bass.

Kaplan, A. S., & Woodside, D. B. (1987). Biological aspects of anorexia nervosa and bulimia

nervosa. Journal of Consulting and Clinical Psychology, 55, 645-653.

Kazdin, A. E. (2003). Research design in clinical psychology (4th ed.). Boston, MA: Allyn and

Bacon.

Latimer, A. E., Rench, T. A., Rivers, S. E., Katulak, N. A., Materese, S. A., Cadmus, L., et al.

(2008). Promoting participation in physical activity using framed messages: An application

of prospect theory. British Journal of Health Psychology, 13, 659-681.

Luce, K. H., & Crowther, J. H. (1999). The reliability of the eating disorder Examination—Self-

report questionnaire version (EDE-Q). International Journal of Eating Disorders, 25, 349-

351.

Matusek, J. A., Wendt, S. J., & Wiseman, C. V. (2004). Dissonance thin-ideal and didactic

healthy behavior eating disorder prevention programs: Results from a controlled trial.

International Journal of Eating Disorders, 36, 376-388.

Mitchell, J. E., & Crow, S. (2006). Medical complications of anorexia nervosa and bulimia

nervosa. Current Opinion in Psychiatry, 19, 438-443.

Running Head: PEER-FACILITATED COGNITIVE DISSONANCE 28

Mitchell, K. S., Mazzeo, S. E., Rausch, S. M., & Cooke, K. L. (2007). Innovative interventions

for disordered eating: Evaluating dissonance-based and yoga interventions. International

Journal of Eating Disorders, 40, 120-128.

Mond, J. M., Hay, P. J., Rodgers, B., Beumont, P. J. V., & Owen, C. (2004). Temporal stability

of the eating disorder examination questionnaire. International Journal of Eating Disorders,

36, 195-203.

Pearson, J., Goldklang, D., & Striegel-Moore, R. H. (2002). Prevention of eating disorders:

Challenges and opportunities. International Journal of Eating Disorders, 31, 233-239.

Perez, M., Becker, C. B., Ramirez-Cash, A. (in press). Transportability of an empirically

supported dissonance-based prevention program for eating disorders. Body Image.

Peterson, C. B., Crosby, R. D., Wonderlich, S. A., Crow, S. J., Bardone-Cone, A., le Grange, D.,

et al. (2007). Psychometric properties of the eating disorder examination-questionnaire:

Factor structure and internal consistency. International Journal of Eating Disorders, 40,

386-389.

Stice, E. (1994). Review of the evidence for a sociocultural model of bulimia nervosa and an

exploration of the mechanisms of action. Clinical Psychology Review, 14, 633-661.

Stice, E. (2001). A prospective test of the dual-pathway model of bulimic pathology: Mediating

effects of dieting and negative affect. Journal of Abnormal Psychology, 110, 124-135.

Stice, E., & Agras, W. S. (1998). Predicting onset and cessation bulimic behaviors during

adolescence: A longitudinal grouping analysis. Behavior Therapy, 29, 257-276.

Stice, E., Chase, A., Stormer, S., & Appel, A. (2001). A randomized trial of a dissonance-based

eating disorder prevention program. International Journal of Eating Disorders, 29, 247-262.
Running Head: PEER-FACILITATED COGNITIVE DISSONANCE 29

Stice, E., Fisher, M., Lowe, M. R. (2004). Are dietary restraint scales valid measures of acute

dietary restriction? Unobtrusive observational data suggest not. Psychological Assessment,

16, 51-59.

Stice, E., Marti, C. N., Spoor, S., Presnell, K., & Shaw, H. (2008). Dissonance and healthy

weight eating disorder prevention programs: Long-term effects from a randomized efficacy

trial. Journal of Consulting and Clinical Psychology, 76, 329-340.

Stice, E., Mazotti, L., Weibel, D., & Agras, W. S. (2000). Dissonance prevention program

decreases thin-ideal internalization, body dissatisfaction, dieting, negative affect, and

bulimic symptoms: A preliminary experiment. International Journal of Eating Disorders,

27, 206-217.

Stice, E., & Shaw, H. (2004). Eating disorder prevention programs: A meta-analytic review.

Psychological Bulletin, 130, 206-227.

Stice, E., Shaw, H., Becker, C. B., & Rohde, P. (2008). Dissonance-based interventions for the

prevention of eating disorders: Using persuasion principles to promote health. Prevention

Science, 9, 114-128.

Stice, E., Shaw, H., Burton, E., & Wade, E. (2006). Dissonance and healthy weight eating

disorder prevention programs: A randomized efficacy trial. Journal of Consulting and

Clinical Psychology, 74, 263-275.

Stice, E., Shaw, H., & Marti, C. N. (2007). A meta-analytic review of eating disorder prevention

programs: Encouraging findings. Annual Review of Clinical Psychology, 3, 207-231.

Stice, E., Trost, A., & Chase, A. (2003). Healthy weight control and dissonance-based eating

disorder prevention programs: Results from a controlled trial. International Journal of

Eating Disorders, 33, 10-21.

Running Head: PEER-FACILITATED COGNITIVE DISSONANCE 30

Taheri, S., Lin, L., Austin, D., Young, T., & Mignot, E. (2004). Short Sleep Duration Is

Associated with Reduced Leptin, Elevated Ghrelin, and Increased Body Mass Index. PLoS

Medicine, 1, 210-217.

Stice, E., Ziemba, C., Margolis, J., & Flick, P. (1996). The dual pathway model differentiates

bulimics, subclinical bulimics, and controls: Testing the continuity hypothesis. Behavior

Therapy, 27, 531-549.

Running Head: PEER-FACILITATED COGNITIVE DISSONANCE 31

Footnotes

1. Based on peer-leader training, we estimate that over 95% of students can identify ways to
improve exercise or eating. The sleep option is available for the remaining 5%.
Running Head: PEER-FACILITATED COGNITIVE DISSONANCE 32

Table 1
Means and Standard Deviations for Dependent Measures
Post 8-wk. 8-mo. 14-mo. Post
Baseline Intervention Follow-up Follow-up Follow-up Intervention 8 wk. 8 mo. 14 mo.

Measures M (SD) M (SD) M (SD) M (SD) M (SD) d d d d

Negative
Affect
CD 1.76 (0.62) 1.48 (0.47) 1.60 (0.66) 1.54 (0.64) 1.47 (0.58) 0.51* 0.25* 0.35* 0.48*
MHW 1.58 (0.48) 1.48 (0.44) 1.51 (0.63) 1.34 (0.34) 1.41 (0.51) 0.22 0.12 0.58* 0.34*

Thin-Ideal.
Internalization
CD 3.37 (0.61) 2.89 (0.68) 3.30 (0.69) 3.18 (0.67) 3.05 (0.65) 0.74* 0.11 0.30* 0.51*
MHW 3.37 (0.62) 3.16 (0.56) 3.38 (0.70) 3.38 (0.51) 3.10 (0.71) 0.36* -0.02 0.00 0.41*

Body
Dissatisfaction
CD 2.30 (1.53) 1.60 (1.23) 1.65 (1.36) 1.46 (1.29) 1.03 (1.05) 0.50* 0.45* 0.59* 0.97*
MHW 1.73 (1.25) 1.39 (1.01) 1.19 (0.97) 1.34 (0.90) 0.84 (0.87) 0.30* 0.48* 0.36* 0.83*

Dietary
Restraint
CD 1.52 (1.22) 0.84 (0.88) 0.93 (1.21) 1.10 (1.17) 0.65 (0.99) 0.64* 0.49* 0.35* 0.78*
MHW 1.27 (1.16) 0.81 (0.73) 0.78 (0.88) 0.91 (0.96) 0.62 (0.99) 0.47* 0.48* 0.34 0.60*

Bulimic
Pathology
CD 12.25 (8.64) 8.18 (6.36) 9.06 (7.85) 7.90 (7.15) 5.93 (6.00) 0.54* 0.39* 0.55* 0.85*
MHW 9.45 (6.81) 7.94 (5.12) 6.68 (5.70) 6.87 (4.99) 4.92 (6.54) 0.25 0.44* 0.43* 0.68*

Note: Cognitive Dissonance (CD) n = 53, Modified Healthy Weight (MHW) n = 49. All analyses are intent to treat.
* indicates p < .05 for post-hoc t-test comparison with baseline
Running Head: PEER-FACILITATED COGNITIVE DISSONANCE 33

Table 2.

Results of Repeated Measures ANOVA from Pre-Intervention to Post-Intervention_

2
Analysis F p< pη

Negative Affect (df = 100)

Time 24.09 .001* 0.19

Group 0.88 .352 0.01

Time x Group 4.60 .034* 0.04

Thin Ideal Internalization (df = 100)

Time 33.35 .001* 0.25

Group 1.56 .214 0.02

Time x Group 5.33 .023* 0.05

Body Dissatisfaction (df = 100)

Time 34.70 .001* 0.26

Group 2.32 .131 0.02

Time x Group 3.74 .056 0.04

Restraint (df = 100)

Time 43.02 .001* 0.30

Group 0.54 .464 0.01

Time x Group 2.20 .141 0.02

Running Head: PEER-FACILITATED COGNITIVE DISSONANCE 34

Bulimic Pathology (df = 100)

Time 17.53 .001* 0.15

Group 0.91 .343 0.01

Time x Group 4.27 .041* 0.04

Note: * indicates significant effect with significance level set at p < .05.
Running Head: PEER-FACILITATED COGNITIVE DISSONANCE 35

Table 3.

Results of Repeated Measures ANOVA from Pre-Intervention to Post-Intervention_

2
Analysis F p< pη

Negative Affect (df = 100)

Time 18.23 .001* 0.15

Group 1.55 .217 0.02

Time x Group 0.02 .884 0.00

Thin Ideal Internalization (df = 100)

Time 4.46 .037* 0.04

Group 1.63 .205 0.02

Time x Group 0.01 .905 0.00

Body Dissatisfaction (df = 100)

Time 66.06 .001* 0.40

Group 2.27 .135 0.02

Time x Group 1.62 .206 0.02

Restraint (df = 100)

Time 32.18 .001* 0.24

Group 0.62 .433 0.01

Time x Group 0.32 .575 0.00

Running Head: PEER-FACILITATED COGNITIVE DISSONANCE 36

Bulimic Pathology (df = 100)

Time 57.60 .001* 0.37

Group 1.64 .203 0.02

Time x Group 0.07 .786 0.00

Note: * indicates significant effect with significance level set at p < .05.
 Running Head: PEER-FACILITATED COGNITIVE DISSONANCE 37

Figure 1.0. Sampling and Flow of Participants

Eligible Participants (n= 114)

Excluded:
Not meeting inclusion criteria
(n= 0)
Refused to participate (n= 3)
Enrollment Other reasons:
Did not continue to pursue
sorority membership or
Randomization (n= 106) excused absence (n= 5)

Allocated to intervention CD Allocated to intervention MHW

(n= 53) (n= 53)
Received complete intervention Received complete intervention
(n= 52) (n= 51)
Did not continue study past time 1 Did not continue study past time 1
Allocation
(n= 1) (n= 2)
Give reasons: excused absence by Give reasons: excused absence by
sorority from second session sorority from second session

Did not continue study past time 2 (n= 1) Did not continue study past time 2 (n= 2)
Did not continue study past follow-up #1 Did not continue study past follow-up #1
(n= 4) (n= 1)
Did not continue study past follow-up #2 Did not continue study past follow-up #2
(n= 11) (n= 7)
Give reasons: Not present during Give reasons: Not present during
collection of follow-up data at weekly Follow-Up collection of follow-up data at weekly
sorority meetings and did not attend other sorority meetings and did not attend other
sessions for follow-up data collection sessions for data collection

Completed follow-up #3, n= 36 (53-1-1-4-11) Completed follow-up #3, n = 41 (53-2-2-1-7)

Missing Data Analysis, n= 17 (1+ 1 + 4+11) Missing Data Analysis, n= 12 (2+2+1+7)
Excluded from analysis, n= 0 Excluded from analysis, n= 4
Give reasons: Met criteria for likely eating Analysis Give reasons: Met criteria for likely eating
disorder disorder
Total Analyzed, n= 53 (53 - 0) Total Analyzed, n= 49 (53-4)