9825_Evaluating the potential roles of body dissatisfaction in exercise avoidance

Graduate Theses and Dissertations
Iowa State University Capstones, Theses and
Dissertations
2017
Evaluating the potential roles of body
dissatisfaction in exercise avoidance
Kimberly Rae More
Iowa State University
Follow this and additional works at: https://lib.dr.iastate.edu/etd
Part of the Psychology Commons
This Thesis is brought to you for free and open access by the Iowa State University Capstones, Theses and Dissertations at Iowa State University Digital
Repository. It has been accepted for inclusion in Graduate Theses and Dissertations by an authorized administrator of Iowa State University Digital
Repository. For more information, please contact digirep@iastate.edu.
Recommended Citation
More, Kimberly Rae, “Evaluating the potential roles of body dissatisfaction in exercise avoidance” (2017). Graduate Theses and
Dissertations. 15380.
https://lib.dr.iastate.edu/etd/15380

Evaluating the potential roles of body dissatisfaction in exercise avoidance

by

Kimberly Rae More

A thesis submitted to the graduate faculty

in partial fulfillment of the requirements for the degree of

MASTER OF SCIENCE

Major: Psychology (Health and Social Psychology)

Program of Study Committee:
L. Alison Phillips, Major Professor
Marcus Crede
Laura Ellingson

The student author and the program of study committee are solely responsible for the content
of this thesis. The Graduate College will ensure this thesis is globally accessible and will not
permit alterations after a degree is conferred.

Iowa State University

Ames, Iowa

2017

Copyright © Kimberly Rae More, 2017. All rights reserved.
ii

TABLE OF CONTENTS

Page
LIST OF FIGURES ………………………………………………………………………………………
iii
LIST OF TABLES
………………………………………………………………………………………..
vi
ABSTRACT……………………………….
……………………………………………………..
v
CHAPTER 1
INTRODUCTION ………………………………………………………………..
1
Links Between Body Dissatisfaction and Health
………………………………………….
1

Tested Interventions to Increase Exercise and Body Satisfaction] ………………….
2

Body Dissatisfaction Prevents Engagement in Exercise ……………………………….
3

The Current Study
……………………………………………………………………………………
4
CHAPTER 2
METHOD ……………………………………………………………………………
6

Participants C ………………………………………………………………………………………….
6

Procedure Ch
…………………………………………………………………………………………..
6

Measures Cha
………………………………………………………………………………………….
7
CHAPTER 3
RESULTS …………………………………………………………………………..
12

Hypothesis 1……………………………………………………………………………………………
14

Hypothesis 2……………………………………………………………………………………………
15

Hypothesis 3……………………………………………………………………………………………
15

CHAPTER 4
DISCUSSION ………………………………………………………………………
17
REFERENCES …………………………………………………………………………………………….
22
APPENDIX A
: SELF-REPORT MEASURES …………………………………………………
32
APPENDIX B : SCATTERPLOTS …………………………………………………………………
35
APPENDIX C : RESULTS WITH EXCLUSION OF OUTLIERS………………………
44
APPENDIX D
: RESULTS WITHOUT MEAN IMPUTATION
…………………………
48
APPENDIX E : RESULTS WITH LOG10 TRANSFORMATIONS …………………..
46

iii

LIST OF FIGURES

Page

Figure 1. Mediation Model to Test Hypothesis 1 and 2
……………………………………..
30

Figure 2. Mediation Model to Test Hypothesis 3
………………………………………………
31

iv

LIST OF TABLES

Page
Table 1. Descriptive Statistics and Correlations ………………………………………………
25
Table 2. Exploratory Factor Analysis of Body Satisfaction

Scale and Body Appreciation Scale
…………………………………………………..
27
Table 3. Mediation Analysis Results for Hypothesis 1
……………………………………..
28
Table 4. Mediation Analysis Results for Hypothesis 2
……………………………………..
28
Table 5. Mediation Analysis Results for Hypothesis 3
……………………………………..
29

v

ABSTRACT

Body dissatisfaction is experienced by individuals in all weight classes and has been
linked with poor mental and physical health outcomes in both women and men. Exercise
interventions are a common tool used to improve body dissatisfaction, but their impact is
relatively small. Reasons for this small impact might include high rates of attrition and
difficulty in recruiting those who are most sedentary in the first place, or who avoid exercise
(at most high-risk/high-need). The present study evaluates the extent to which exercise
avoidance mediates the association of body dissatisfaction with exercise frequency and
whether perceived embarrassment, exercise fatigue, and exercise self-efficacy explain the
association of body dissatisfaction with exercise avoidance. Participants were 110 students
and staff from an urban, private US university. Body dissatisfaction, exercise avoidance, and
hypothesized mediators were measured at baseline; objective exercise was measured with
accelerometers for one month. Exercise avoidance mediated the relation between body
dissatisfaction and exercise frequency (B = -.02 (SE =.01) [95% CI: -.04 to -.01]).
Additionally, the relation between body dissatisfaction and exercise avoidance was fully
mediated by embarrassment (B = .24 (SE .10) [95% CI: .08 to .47])and fatigue (B = .10 (SE
.06) [95% CI: .01 to .28]) but not by self-efficacy (B = -.00 (SE .02) [95% CI: -.06 to .01]).
Thus, exercise interventions may not effectively target individuals who are dissatisfied with
their body because they may be avoiding exercise due to perceived embarrassment and
fatigue.
1

CHAPTER 1. INTRODUCTION

Both women and men are susceptible to feeling dissatisfied with their bodies; 61% of
women and 41% of men were found to rate themselves as ‘too heavy’, and 21% of women and
11% of men reported that they feel they are unattractive (Frederick, Peplau & Lever, 2006).
Perceptions of being overweight are common, even among those who are not considered to be
medically overweight (Frederick et al., 2006). Perceptions of being overweight and body
dissatisfaction in general are linked with poor mental and physical well-being, even after
controlling for actual markers of mental and physical health (Bucchianeri & Neumark-Sztainer,
2014; Černelič-Bizjak & Jenko-Pražnikar, 2014; Vartanian & Novak, 2011; Wilson, Latner &
Hayashi, 2013). We briefly review that literature here and suggest that targeting body
dissatisfaction may be a key first step in improving individuals’ mental and physical health. In
particular, we evaluate the possible roles of body dissatisfaction in preventing involvement in
regular exercise, which is an important behavior for mental and physical health (Penedo & Dahn,
2005).
Links Between Body Dissatisfaction and Health
First, body dissatisfaction is associated with risky health behaviors and poor mental and
physical well-being. With regards to mental health, body dissatisfaction has been found to
mediate the relationship between BMI and psychological health, such that higher BMI leads to
poorer mental-health outcomes at least in part due to body dissatisfaction (Bucchianeri &
Neumark-Sztainer, 2014). Specifically, body dissatisfaction mediates the relationship between
BMI and self-esteem and depressed mood (Mond, van den Berg, Boutelle, Hannan, & Neumark-
Sztainer, 2011). With regards to physical health, body dissatisfaction has been shown to mediate
2

the relationship between BMI and physical health-related quality of life, such that higher BMI
was associated with poorer physical health at least partially due to body dissatisfaction (Wilson,
Latner & Hayashi, 2013). This may be due to the influence of body dissatisfaction on health-
related behaviors as well as directly on biological health processes: regarding behaviors, body
dissatisfaction is related to dieting which increases the risk of disordered eating (Stice & Shaw,
2002), and individuals who subscribe to anti-fat attitudes and experience weight-related stigma
avoid exercise (Vartanian & Novak, 2011). For women and men, dissatisfaction with the body is
linked to an increase of inflammatory biomarkers (i.e., C-reactive protein)
even after controlling for weight and other known predictors, such as sleep quantity, alcohol
consumption, gender, and age (Černelič-Bizjak & Jenko-Pražnikar, 2014). Thus, body
dissatisfaction appears to be a potential independent contributor to heart disease.
Second, positive body image, or body satisfaction, is associated with mental and physical
health benefits. Regardless of actual body shape, both women and men who have a positive body
image were less likely to adopt unhealthy diets and more likely to make an effort to protect their
bodies from sun damage (Gillen, 2015). Further, those with a positive body image are more
likely to have higher self-esteem and were less likely to be depressed (Gillien, 2015). This
association between body satisfaction and mental health may be due to individual differences in
optimism and due to more proactive coping strategies (e.g., exercise) among those who report
positive levels of body satisfaction (Avalos, Tylka, & Wood-Barcalow, 2005).
Tested Interventions to Increase Exercise and Body Satisfaction
Interventions to improve/enhance body satisfaction have primarily focused on increasing
participants’ physical activity; these efforts have had some success, across different types of
exercise and groups of individuals (Campbell & Hausenblas, 2009; Hausenblas & Fallon 2006).
3

One meta-analysis of 57 studies examining the impact of exercise interventions on body
satisfaction found that individuals in intervention groups had a small but significant
improvement in body satisfaction between baseline and follow-up relative to control groups
(Campbell & Hausenblas, 2009), and this effect did not differ by participants’ overweight status
or change in fitness level or BMI during the intervention. These results indicate that both
medically healthy and medically overweight individuals can improve their body satisfaction by
exercising and that these changes may not require improvements in actual fitness or BMI level.
Though exercise interventions have been shown to improve body satisfaction, these
effects have been relatively small (Campbell & Hausenblas, 2009). Further, these interventions
likely only worked for those individuals who remained in the studies, and attrition is a large
problem for exercise interventions (Linke, Gallo, & Norman, 2011). Therefore, the effectiveness
of exercise interventions (on behavior, as well as mental and physical health outcomes) may be
increased by first addressing reasons for participant attrition and the intention-behavior gap (e.g.,
participants’ perceived behavioral control/self-efficacy and barriers to exercise; Ajzen, 1991;
2013; Sniehotta, Scholz & Schwarzer, 2005).
Body Dissatisfaction Prevents Engagement in Exercise
We propose that exercise interventions may not effectively target individuals who are
dissatisfied with their bodies, because individuals low in body satisfaction may be more likely to
avoid signing up for an intervention advertising exercise, or to drop out or not adhere to the
intervention. That is, we propose that body dissatisfaction may contribute to active avoidance of
exercise, which is known to decrease engagement in moderate or vigorous exercise (Vartanian &
Shaprow, 2008), thereby limiting the effectiveness of exercise interventions for those at-risk
individuals (those high in body dissatisfaction).
4

Body dissatisfaction may contribute to exercise avoidance due to several possible factors:
first, Schmalz (2010) found that individuals who perceive that weight stigmatization is common
are more likely to believe that they are not competent enough to engage in physical activity and
that perceived exercise competence was explained by body satisfaction but not by actual weight
status (i.e., BMI). Therefore, body dissatisfaction may lead to exercise avoidance due to low
perceived competence or self-efficacy.
Second, individuals who are dissatisfied with their bodies may perceive greater barriers
to exercise, and perceiving barriers to exercise can prevent an individual from adopting and
maintaining engagement in regular exercise (Booth, Bauman, Owen & Gorge, 1997; Grubbs &
Carter, 2002)—regardless of whether the perceived barriers are real (Simonavice & Wiggins,
2008). Grubbs and Carter (2002) found that individuals who do not engage in regular exercise
were more likely to perceive exercising as embarrassing than individuals who are regular
exercisers. Importantly, non-exercising individuals are not only more likely to perceive barriers
to exercise, such as feeling embarrassed or fatigued when exercising, but are also less likely to
perceive benefits to exercise (Grubbs & Carter, 2002). Thus, individuals who are dissatisfied
with their bodies may not only avoid exercise due to experiencing (or perceiving) more of these
barriers, but they may also be less likely to find the appeal in joining an exercise intervention.
The Current Study
The present study has two purposes. The first is to empirically evaluate a link between
body dissatisfaction and exercise frequency via (i.e. mediated by) exercise avoidance. The
second is to evaluate whether body dissatisfaction predicts psychological well-being via exercise
avoidance. These relationships have not been explicitly tested or proposed in the literature, to our
knowledge, but the links between body dissatisfaction, health behaviors (including exercise), and
5

mental and physical health outcomes suggest that these relationships exist. Assuming body
dissatisfaction will be significantly related to exercise avoidance, the third objective is to
evaluate specific, potentially-changeable factors that could account for the relationship between
body dissatisfaction and exercise avoidance. Specifically, we test the hypothesis that the
relationships between body dissatisfaction and exercise avoidance will be mediated by
individuals’ exercise self-efficacy and perceived exercise barriers—embarrassment and fatigue
from exercise, controlling for participants’ reported BMI so that we isolate the effect of body
dissatisfaction on exercise frequency and avoidance from the effect of being overweight on
exercise frequency and avoidance.
If results suggest that exercise self-efficacy, perceived embarrassment, and perceived
fatigue mediate the relationship between body dissatisfaction and exercise avoidance, they could
potentially be targeted in those with body dissatisfaction prior to commencing any exercise
program in order to boost individuals’ engagement in (and potential benefits from) the exercise
intervention. By targeting body satisfaction and therefore exercise avoidance first, it is possible
that exercise interventions could increase participation of and effectiveness for individuals at
high risk for poor mental and physical health.

6

CHAPTER 2. METHODS

Participants
The sample consisted of 123 adults (students and N = 36 staff) recruited through the
psychology subject pool and departmental e-mails from an urban, private university in the US.
Participants ranged from 18 to 73 years of age, with a mean age of 24.7 (11.24) years. The
majority of the sample was female (72%). Most participants identified as Caucasian (76%); other
identified ethnicities were South Asian (9.1%), Black (6.6%), East Asian (3.3%), and Middle
Eastern (0.8%). The final sample excluded participants who failed random response checks
and/or who reported being NCAA athletes (N = 110).
Procedure

The data obtained for the present study was part of a larger, month-long observational
study, and the hypotheses evaluated in the current analyses have not been evaluated or published
elsewhere. Prospective participants were recruited if they were willing to try to be active at least
two times per week for 20 consecutive minutes for the duration of the study. At the first time
point in the study, participants answered questions regarding their body satisfaction and
demographic information. Next, participants attended an in-person session where they were
given their Fitbit. Researchers helped the participants create a ‘Fitbit action plan’ that was
intended to help participants remember to wear their Fitbits for the entire study. Additionally,
participants downloaded the Fitbit app that allowed them to sync their Fitbit activity to the app.
Participants were asked to begin wearing their Fitbit the next day.

Within 48 hours of the in-person session participants received a link to complete the
second online questionnaire. The questionnaire included questions to assess self-efficacy and
7

perceived embarrassment and fatigue while exercising. Participants returned to the lab four-
weeks after the initial in-person session to return their assigned Fitbit, to get weighed, and to
answer self-report questionnaires assessing exercise avoidance and psychological wellbeing.
Measures
Body dissatisfaction
Body dissatisfaction was measured with two scales from the published literature. First,
the Body Appreciation Scale (Avalos, Tylka & Wood-Barcalow, 2005) is a 13-item measure that
assesses four aspects of body acceptance: having positive opinions about one’s body, body
acceptance regardless of perceived imperfections, respecting one’s body by engaging in health
behaviors, and maintaining positive body satisfaction by not idealizing thin body types typically
displayed in the media. The measure has a 5-point-likert- type scale and response options range
from ‘Not at all’ to ‘Extremely’, with higher scores being indicative of higher body appreciation.
The variable was scored as instructed and then re-scored so that higher scores indicate
greater body dissatisfaction for analyses of the current hypotheses (which are stated in terms of
body dissatisfaction). Scores on the Body Appreciation Scale has been shown to have high
internal consistency (α = .94) and have convergent validity (See Avalos, Tylka & Wood-
Barcalow, 2005; e.g., body surveillance -.55 and body shame r = -.77). The reliability coefficient
for scores on the Body Appreciation Scale, measured at baseline in the present sample, was 0.92,
and test re-test reliability with follow-up body appreciation scores was found (r = .83, paired
t(112) = -.66, p = .51). Second, the Body Shape Satisfaction Scale (Pingitore, Spring, Garfield,
1997) is a 10-item measure that assesses body satisfaction with specific parts of the body (e.g.,
waist, stomach, thighs). Items are rated on a 5-point-likert-type scale ranging from ‘Very
dissatisfied’ to ‘Very satisfied’, the items were reverse coded so that higher scores indicated high
8

levels of body dissatisfaction. Scores on the Body Shape Satisfaction Scale has been shown to
have high internal consistency (α = .88) (Pingitore, Spring & Garfield, 1997) and convergent
validity (See Petrie, Tripp & Harvey, 2002; bodily shame r = -.63 and appearance evaluation r =
-.75). The reliability coefficient for the Body Shape Satisfaction Scale in the present study was
0.88. Finally, baseline and follow up scores from the Body Shape Satisfaction Scale
demonstrated test re-test reliability in the present sample (r = .82, paired t(114) = .79, p = .43).
Recently, distinctions have been made between body dissatisfaction/satisfaction and body
appreciation (e.g., Tylka & Wood-Barlow, 2015; Tiggerman & McCourt, 2013). Researchers
have proposed that negative body-image (e.g., body dissatisfaction) and positive body-image
(e.g., body appreciation) do not represent equivalent dimensions (e.g., Tylka & Woof-Barlow,
2015). Therefore, exploratory factor analysis (maximum likelihood extraction – direct oblimin)
along with parallel analysis (which verifies how many factors should be attained above chance
level; Hayton, Allen, & Scarpello, 2004) is used in the current study to determine whether body
dissatisfaction and body appreciation should be used as combined or as separate predictors for
the main hypotheses.
Exercise frequency
The Fitbit (Fitbit.com, “Zip” model) counts movement as steps using accelerometry
technology. Past studies have found high convergent validity between counted steps by Fitbits
and manually counted steps (Evenson, Goto & Furberg, 2015). Additionally, Fitbits have been
shown to have high inter-device reliability for counting steps. Even sedentary individuals should
show light activity on a day-to-day basis (e.g., walking between classes). Lack of activity is
therefore an indication that the participant was not wearing the Fitbit. Participants were excluded
from the final dataset if they had no recorded activity on 25% or more of the intervention days;
9

further, days where participants wore the device for less than 10 hours were excluded from
analyses. In the present study, exercise frequency was the proportion of days that individuals
engaged in at least one 20-minute exercise session (i.e., 20 or more consecutive minutes of
moderate or vigorous exercise activity).
Exercise avoidance
Exercise avoidance was measured using two items developed by Vartanian and Shaprow
(2008): ‘I avoid engaging in physical activity when others might be around’, and ‘I feel
uncomfortable going to a gym’. These two items are part of a 3-item scale used by Vartanian and
Novak (2011) to represent exercise avoidance. One of the original items was not used in the
present study because it assessed exercise avoidance due to embarrassment. Response options
were displayed as a 7-point-likert-type scale with higher scores reflecting higher levels of
exercise avoidance. Observed reliability was α = .86 and test re-test reliability was assessed with
follow-up scores (r= .78, paired t(115) = -1.54, p = .13).

Subjective psychological well-being
The flourishing scale was used to assess participants’ subjective psychological well-being
(Diener et al., 2009). The flourishing scale consists of 8 items (e.g., ‘I lead a purposeful and
meaningful life’) that are rated on a 7-point-likert-type scale ranging from ‘Strongly disagree’ to
‘Strongly agree’. Higher scores correspond to higher levels of general psychological well-being.
The flourishing scale has been shown to have temporal stability and high internal validity (α =
.87). Additionally, the flourishing scale has convergent validity with other measures of
psychological well-being (e.g., satisfaction with life r = .62 and optimism r = -.59, where low
scores reflect optimism; Diener et al., 2010). The reliability coefficient for the flourishing scale,
10

measured at follow-up, was 0.95, and test re-test reliability was evaluated with baseline scores r
= .50, paired t(116) = .41, p = .68).
Perceived barriers to exercise
The Benefits and Barriers to Exercise Scale was used to assess perceived barriers to
exercise (Sechrist, Walker, & Pender, 1987). Two barriers are proposed to mediate the relation
between body dissatisfaction and exercise avoidance in the present study. Embarrassment was
measured using the item ‘I am too embarrassed to exercise’. Fatigue was measured using the
items ‘I am fatigued by exercise’ and ‘exercise is hard work’ (α = .76). Both of the items were
rated on a 5-point-likert-type scale ranging from ‘Strongly disagree’ to ‘Strongly agree’. For both
variables, higher scores indicate greater experience of barriers to exercise. The evaluated barriers
in the current study were chosen due to the barriers related to body dissatisfaction in the existing
literature (Booth, Bauman, Owen & Gorge, 1997; Grubbs & Carter, 2002; Schmalz, 2010); the
other barriers in the scale are not conceptually related to body dissatisfaction and so were not
included as tested mediators of the relationship between body dissatisfaction and exercise
avoidance (e.g., ‘Inconvenient facility schedules’ and ‘Costs too much to exercise’).
Self-efficacy
Self-efficacy was measured using the item ‘I am confident that I can exercise for at least
20 minutes, three times per week for the next month’. This item was derived from the Theory of
Planned Behavior Questionnaire and is thought to measure the capacity component of perceived
behavioral control, which is behavior-specific self-efficacy (Ajzen, 2013). The item is measured
on a 7-point-likert-type scale ranging from ‘False’ to ‘True’. Scores on the self-efficacy item
measured at baseline and follow-up did significantly differ (r = .24, paired t(115) = 4.66, p < .001) suggesting that self-efficacy may not be a stable construct. 11 Demographics Self-report questions were used to identify participants’ age, gender identification, race and ethnicity identifications, year of schooling, height, and weight. Random response check Two items were included in the survey to check for random responding. The first item required participants to answer “mostly untrue” and the second item required participants to select the value “4” to help identify individuals who randomly responded to items. Random responses have been shown to drastically alter effect sizes (Credé, 2010). We took a conservative approach to eliminating data by excluding only those individuals who missed both checks. 12 CHAPTER 4. RESULTS Mean imputation was used to correct the missing self-report data that was present in the dataset. Missing data was not imputed for one-item scales (i.e., self efficacy, missing N = 1) or for scales in which a participant answered zero of multiple items (i.e., six participants did not complete any Time 2 self-report questionnaires, due to dropping out of the study). Therefore, in total, six participants had missing data, with six total cells missing. Data was examined for multivariate outliers using Mahalonobis Distance values. Skewness and kurtosis was examined using z-scores, variables with a score greater than 3.3 were transformed using a log10 transformation (i.e., exercise avoidance, embarrassment, self-efficacy, psychological well-being, and exercise frequency). To demonstrate the robustness of the results, all hypotheses were analyzed with and without the inclusion of imputed data, the multivariate outlier, and the transformed variables (see Steegen, Tuerlinckx, Gelman, & Vanpaemel, 2016). There were no differences between the main analyses and the aforementioned alternative analyses. Thus, all results are reported with mean imputation, the inclusion of the multivariate outlier, and using the non-transformed variables. Table 1 contains descriptive statistics and correlations. It should be noted that Proportion of Exercise Days and BMI report the observed minimum and maximum values. One participant had an estimated BMI value of 0.75, which was determined to be a mistake in his/her self- reported height; accordingly this value was replaced with the mean. Combined body dissatisfaction descriptive statistics were derived from raw scores The exploratory factor analysis of the Body Appreciation Scale and the Body Shape Satisfaction Scale, along with the solution from the parallel analysis, resulted in a single-factor 13 solution. The item “height” from the Body Shape Satisfaction Scale was removed after initial analyses as it loaded poorly onto the factor (.25) and had a low extracted communality (0.06). After removal of the item the single factor accounted for 47.58% of the variance in scores and all of the remaining items from both the Body Shape Satisfaction Scale and the Body Appreciation Scale loaded onto the single factor (See Table 3). Thus, individual z-scores for the Body Shape Satisfaction Scale and the Body Appreciation Scale were calculated and combined to form an overall measure of body dissatisfaction that was used as the predictor variable in all subsequent analyses (referred to as body dissatisfaction). There is no theoretical reason or reason identified in previous research to suggest that gender or the category of student versus staff would influence the relationships between the tested variables. However, to statistically test whether these categories should be included as moderators of the main analyses a series of regression analyses and independent samples t-test were conducted. To test whether gender should be included as a moderator of the mediated relationship the interaction between gender and body image (mean centered) was compared against gender and body image as individual predictors of exercise avoidance. Similarly, the interaction between gender and exercise avoidance (mean centered) was compared against gender and exercise avoidance as individual predictors of exercise frequency and psychological well-being. Participants’ status as a student or staff member was also tested as a possible moderator using the aforementioned method. For all linear regression analyses examining gender as a possible moderator, the interaction term was non-significant. Additionally, for all independent samples t-tests gender did not significantly predict any of the variables. Thus, gender will not be included as a moderator in any tests of hypotheses. With regards to the category of student versus staff, the interaction term with body dissatisfaction was 14 not a significant predictor of exercise avoidance. Additionally, the interaction term with exercise avoidance was not a significant predictor of psychological well-being. However, the interaction was a significant predictor of exercise frequency. Likewise, there was a significant relationship between the category of student versus staff and exercise avoidance and exercise in the independent samples t-tests. Thus, analyses to test the research hypothesis were conducted with and without the inclusion of staff to determine whether the inclusion of staff members changed the results. No differences were found between the direction and significance of the indirect effect in the analyses, thus only the results with the inclusion of both student and staff members are reported. Each hypothesis test was conducted using Hayes’ PROCESS procedures for bootstrapped mediation analyses using 1000 bootstrapped samples and 95% confidence intervals (Hayes, 2013). In each analysis using participants’ estimated BMI as a covariate controlled for the effect of BMI on the results. Hypothesis 1 As hypothesized, the relationship between body dissatisfaction and exercise frequency was mediated by exercise avoidance, even after controlling for BMI [Indirect effect: B = -.02 (SE =.01) (95% CI: -.04 to -.01]). That is, a one-unit increase in body dissatisfaction results in a -.02 unit decrease in exercise frequency through the mediating variable of exercise avoidance. Additionally, the relationship between body dissatisfaction and exercise frequency depended on exercise avoidance since the direct effect of body dissatisfaction on exercise frequency was non- significant (see path c’ Figure 1 and Table 4). That is, individuals who are more dissatisfied with their body exercise less because they are more likely to be avoiding exercise. Importantly, this relationship does not depend on BMI. 15 Hypothesis 2 Counter to the hypothesis, the relationship between body dissatisfaction and psychological well-being was not mediated by exercise avoidance after controlling for BMI [B = .05 (SE =.19) (95% CI: -.31 to .41]). The path between body dissatisfaction and exercise avoidance (path a) was significant, as in the test of Hypothesis 1 (See Figure 1; see path coefficients in Table 5). However, the path between exercise avoidance and psychological well- being was not significant. The direct effect and total effect were significant, even after controlling for BMI. That is, higher levels of body dissatisfaction predicted lower levels of psychological well-being above and beyond the effect of the mediator (exercise avoidance) and BMI. Hypothesis 2 was partially supported in that body dissatisfaction was predictive of psychological well-being, but this relationship was not mediated by exercise avoidance. Hypothesis 3 As hypothesized, the relationship between body dissatisfaction and exercise avoidance was mediated by embarrassment and fatigue as barriers to exercise, even after controlling for BMI (B = .24 (SE .10) [95% CI: .08 to .47]) and (B = .10 (SE .06) [95% CI: .01 to .28]), respectively— but not by self-efficacy (B = -.00 (SE .02) [95% CI: -.06 to .01]). Thus, a one-unit increase change in body dissatisfaction leads to a .24 or .10 increase in exercise avoidance due to increased perceived embarrassment and fatigue, respectively. The direct effect of body dissatisfaction on exercise avoidance was non-significant, suggesting that the relationship between body dissatisfaction and exercise avoidance was accounted for by perceived embarrassment and fatigue as barriers to exercise (see tested model in Figure 2; see path coefficients in Table 6). Thus, Hypothesis 3 was partially confirmed, in that the relationship 16 between body dissatisfaction and exercise avoidance was explained by embarrassment and fatigue as barriers to exercise, but not by self-efficacy. 17 CHAPTER 4. DISCUSSION Body dissatisfaction has been linked to poor mental and physical health outcomes including depressed mood and increased risk of heart disease (e.g., Bucchianeri & Neumark- Sztainer, 2014; Černelič-Bizjak & Jenko-Pražnikar, 2014). Exercise interventions have been used as a common tool to improve body image in participants (Campbell & Hausenblas, 2009; Hausenblas & Fallon 2006). Although these interventions have had some success, it is possible that they do not effectively target individuals experiencing body dissatisfaction. That is, individuals who are dissatisfied with their bodies may avoid exercise (and interventions to promote exercise) because they feel that they are unable to exercise or that exercise would make them feel fatigued or embarrassed (Booth, Bauman, Owen & Gorge, 1997; Grubbs & Carter, 2002; Schmalz, 2010). The present study is the first to examine the link between body dissatisfaction and exercise avoidance, as well as specific, potentially-changeable factors, such as perceived embarrassment and fatigue from exercise, that may account for the aforementioned relationship. The present study found evidence of the mediating role of exercise avoidance in the relationship between body dissatisfaction and exercise frequency and identified two factors that account for this relationship (i.e., perceived embarrassment and fatigue). The results therefore indicate that, by first targeting embarrassment and fatigue as perceived barriers to exercise, interventions may influence individuals who are dissatisfied with their bodies to be more likely to join and adhere to an exercise intervention. Furthermore, since perceived embarrassment and fatigue from exercise mediated the relationship between body dissatisfaction and exercise avoidance it is possible that interventions that target these two factors alone may have very large effects. It is important to note that mediation can exist without a significant bivariate relationship between the 18 predictor variable and the outcome variable (e.g., Shrout & Bolger, 2002; Hayes & Rockwood, 2016). In this instance, a non-significant direct effect signifies either that the predictor and outcome variable are non-linearly related or that the relationship depends on a third variable (mediator) (Hayes & Rockwood, 2006). In light of this, the insignificant direct effect between body dissatisfaction and exercise frequency (hypothesis 1) and the between body dissatisfaction and exercise avoidance (hypothesis 2) should not detract from the mediation (indirect) effect found in both cases. Although a low level of body satisfaction is predictive of lower levels of psychological well-being, we did not find that this relationship is due to individuals’ exercise avoidance. Thus, Hypothesis 2 was rejected. It may be that the relationship between body dissatisfaction and psychological well-being is due to a general negative mood or that poor well-being causes poor estimates of body image. If a more domain-specific measure of well-being, or exercise-related affect, were measured in place of the very general flourishing scale, it is possible that body dissatisfaction would predict this exercise-related affect/experience via exercise avoidance. A second unexpected finding was that exercise self-efficacy did not mediate the relationship between body dissatisfaction and exercise avoidance. That is, perceived ability to exercise did not explain why individuals who have low body satisfaction avoid exercise. This null finding is surprising, as a meta-analysis of the components of the Theory of Planned Behavior found that self-efficacy is a unique and useful predictor of both intention and behavior (Armitage & Conner, 2001). Additionally, Schmalz (2010) found that body esteem was predictive of perceived competence to engage in physical activity. It is possible that the null finding can be 19 explained by the use of a limited 1-item measure used to assess self-efficacy. However, others have successfully used single-item measures of self-efficacy to assess health outcomes (e.g., Hoeppner, Kelly, Urbanoski, & Slaymaker, 2001). It is also possible that the ceiling effect observed for the self-efficacy item (i.e., 71.8% of participants reported the highest level of self- efficacy) compromised the mediation analysis of the effect of body dissatisfaction on exercise avoidance through perceived ability. That is, it is possible that the lack of variation in scores for self-efficacy impacted the results of the mediation analysis. To this end, it may be helpful for future studies to set a higher threshold for self-efficacy that better reflect the current physical activity guidelines for adults (e.g., Office of Disease Prevention and Health Promotion, 2017). There are several limitations to the present study that must be addressed. First, the scale used to measure exercise avoidance focused solely on social exercise avoidance (e.g., I avoid engaging in physical activity when others might be around). Therefore, it is possible that there are other facets of exercise avoidance that relate differentially to body dissatisfaction (e.g., private exercise avoidance or exercise avoidance due to physical reactions). It would be beneficial to expand measurement of exercise avoidance to include additional facets in order to further explore the relationship between body dissatisfaction and exercise avoidance. A second limitation to the present study is the range restriction with regards to exercise frequency and body dissatisfaction. That is, the procedure stated that participants must be willing to engage in exercise for a minimum of 20 minutes at least two times per week. Although some participants reported being currently sedentary (at baseline), it is possible that individuals who have the worst body image (and would not be willing to do minimal activity for a study) did not volunteer for the present study. Beneficial future research would find alternative methods for recruiting sedentary individuals and individuals who are the most dissatisfied with their bodies.