Journal of Contextual Behavioral Science 2 (2013) 39–48

Contents lists available at SciVerse ScienceDirect

Journal of Contextual Behavioral Science

2212-14http://d

n CorrE-m

journal homepage: www.elsevier.com/locate/jcbs

Research—Basic Empirical Research

Assessment of body image flexibility: The Body Image-Acceptanceand Action Questionnaire

Emily K. Sandoz a,n, Kelly G. Wilson b, Rhonda M. Merwin c, Karen Kate Kellum b

a University of Louisiana at Lafayette, Psychology Department, Girard 202-A, P.O. Box 43131, Lafayette, LA 70504, USAb University of Mississippi, USAc Duke University Medical Center, USA

a r t i c l e i n f o

Article history:Received 1 August 2012Received in revised form8 February 2013Accepted 17 March 2013

Keywords:Body imagePsychological flexibilityAcceptanceMindfulnessEatingAssessment

47/$ - see front matter & 2013 Published by Ex.doi.org/10.1016/j.jcbs.2013.03.002

esponding author. Tel.: +1 3373715440.ail address: emilysandoz@louisiana.edu (E.K. S

a b s t r a c t

Acceptance and mindfulness components are increasingly incorporated into treatment for eatingdisorders with promising results. The development of measures of proposed change processes wouldfacilitate ongoing scientific progress. The current series of studies evaluated one such instrument, theBody Image-Acceptance and Action Questionnaire (BI-AAQ), which was designed to measure body imageflexibility. Study one focused on the generation and reduction of items for the BI-AAQ and ademonstration of construct validity. Body image flexibility was associated with increased psychologicalflexibility, decreased body image dissatisfaction, and less disordered eating. Study two demonstratedadequate internal consistency and test–retest reliability of BI-AAQ. Study three extended findings relatedto structural and construct validity, and demonstrated an indirect effect of body image dissatisfaction ondisordered eating via body image flexibility. Research and clinical utility of the BI-AAQ are discussed. TheBI-AAQ is proposed as a measure of body image flexibility, a potential change process in acceptance-oriented treatments of eating disorders.

& 2013 Published by Elsevier Inc. on behalf of Association for Contextual Behavioral Science.

1. Introduction

A number of cognitive behavior therapies (CBTs) focus onteaching clients to behave effectively in the presence of disturbingthoughts and feelings (Hayes, 2004). This is particularly wellsuited for clinical populations which tend to be cognitively rigid(and therefore have thoughts that are difficult to change) or forwhom cognitions are consistent with a preferred self-image(despite being harmful), such as in eating disorders like anorexianervosa. Acceptance and Commitment Therapy (ACT; Sandoz,Wilson, & DuFrene, 2011), Dialectical Behavior Therapy (DBT; e.g., Safer, Telch, & Chen, 2009), Mindfulness Based CognitiveTherapy (MBCT; e.g., Baer, Fischer, & Huss, 2005b), and Mind-fulness Based Stress Reduction (MBSR; e.g., Kristeller & Hallet,1999) have all been adapted for treatment of disordered eating.Common to each of these approaches is an emphasis on buildingawareness and openness toward experience (for example, cues forhunger and satiation and dissatisfaction with the body) whilereplacing automatic or reactive eating behavior with more inten-tional, committed behaviors.

This emphasis seems to be particularly well suited for treatingdisordered eating (see Baer, Fischer, and Huss (2005a), Kristeller,

lsevier Inc. on behalf of Associatio

andoz).

Quillian-Wolever, and Baer (2006) for reviews), with recentevidence that symptoms are maintained, at least in part, by non-acceptance or experiential avoidance. ACT has recently gatheredpreliminary support in the treatment of eating disorders. Arandomized controlled trial comparing ACT to traditional cognitivetherapy (CT) for treatment of comorbid eating pathology foundthat ACT produced larger reductions in eating pathology than CT(Juarascio, Forman, & Herbert, 2010). A small case series investi-gating ACT as a treatment for unremitting anorexia nervosashowed clinically significant improvement in disordered eatingin all three clients (Berman, Boutelle, & Crow, 2009). A randomizedcontrolled trial investigating ACT as a treatment for obesity andobesity-related stigma showed significantly greater improvementsin obesity stigma, quality of life, psychological distress, and bodymass (Lillis, Hayes, Bunting, & Masuda, 2009).

2. Psychological flexibility and disordered eating

In ACT, this combination of openness to experience withconstructive behavior is called psychological flexibility. Psycholo-gical flexibility is the ability to fully experience the presentmoment while engaging in behavior that is consistent with one'schosen values even when the present moment includes difficultemotions, thoughts, memories or body sensations (Hayes, Luoma,Bond, Masuda, & Lillis, 2006). From this perspective, healthy

n for Contextual Behavioral Science.

E.K. Sandoz et al. / Journal of Contextual Behavioral Science 2 (2013) 39–4840

psychological functioning involves psychological flexibility (Hayeset al., 2006). A growing body of evidence supports psychologicalflexibility as central to psychological health (Kashdan &Rottenberg, 2010), and psychological inflexibility as central topsychological vulnerability (Kashdan, Barrios, Forsyth, & Steger,2006). Psychological flexibility seems to be important in under-standing the role of disordered eating-related cognitions in poorpsychological health (Masuda, Price, Anderson, & Wendell, 2010;Merwin et al., 2011). Inflexibility with distressing thoughts andfeelings has been associated with binge-eating (Barnes & Tantleff-Dunn, 2010), anorectic (Merwin, Zucker, Lacy, & Elliott, 2010) andbulimic symptoms (Lavender, Jardin, & Anderson, 2009).

Among the thoughts and feelings most strongly associated withdisordered eating is body image dissatisfaction (e.g., Polivy &Herman, 2002). Conceptualizations of eating disorders have longpointed to body image dissatisfaction (also called “weight con-cerns)” as a common precursor to disordered eating (e.g., Polivy &Herman, 2002; Rosen, 1992; Shisslak & Crago, 2001; Slade, 1982;Stice, 2002; Thompson, 1992). Meta-analytic reviews supportthese theories, demonstrating the important role of body imagedissatisfaction in both the development and maintenance ofdisordered eating (Jacobi, Hayward, de Zwaan, Kraemer & Agras,2004; Stice & Shaw, 2002). Body image dissatisfaction seemsnecessary to the development of disordered eating but does notexplain how some can remain dissatisfied without it impactingtheir eating behavior (e.g., Polivy & Herman, 2002).

3. Psychological flexibility and body image dissatisfaction

Body dissatisfaction is so pervasive that it has been describedas “normative discontent” (Rodin, Silberstein, & Streigel-Moore,1985, p. 267). Thus, it cannot only be the presence of thedisturbing thoughts and feelings about the body that is of issue.Rather, the key skill may be the individual's ability to have adisturbing thought/feeling about the body and not let it impacthealth and well being (Masuda et al., 2010). For example, a womanmay, upon regarding her reflection, be filled with disgust andthink, “I've gained so much weight. I look awful.” If she is low inpsychological flexibility, it is likely that she will engage in somebehaviors to manage her thoughts and feelings about her bodysuch as canceling social plans, changing her clothes, having adrink, binge eating, or skipping a meal, even if these are contraryto how she would like to live her life. If she is high in psychologicalflexibility, however, she is more likely to notice her thoughts andthe disgust she feels without making choices that are inconsistentwith her personal values. If her relationships are important to her,it is likely she will keep her social plans, even on the days when itmakes her feel worse about her body to be around people.

The emerging acceptance and mindfulness-based treatmentsfor disordered eating are targeting precisely this kind of flexibilitywith promising results (see Baer et al. (2005), Kristeller et al.(2006) for reviews). Continued progress in the development ofthese treatments of eating disorders will require the demonstra-tion of both positive outcomes (e.g., improvements to quality oflife and eating behavior) and the processes by which theseoutcomes occur. Although processes of change have not beendemonstrated with respect to traditional cognitive-interventionsfor body image disturbance, the assessments exist that wouldallow for such research (e.g., Appearance Schemas Inventory—Revised; Cash, Melnyk, & Hrabosky, 2004; Assessment of Body-Image Cognitive Distortions; Jakatdar, Cash, & Engle, 2006). Bodyimage inflexibility could not be assessed using existing measuresof body image as it is functionally defined while these measuresfocus on formal properties of behavior. Body image inflexibility isnot characterized by the contents of thoughts (e.g., “What I look

like is an important part of who I am;” Cash et al., 2004) ortopography of behaviors (e.g., “I wear baggy clothes;” Rosen,Srebnik, Saltzberg, & Wendt, 1991), but on how body image isimpacting life (e.g., “When I start thinking about the size andshape of my body, it's hard to do anything else;” Sandoz & Wilson,2006).

4. Assessing body image flexibility

Psychological flexibility has typically been assessed using abrief self-report measure, the Acceptance and Action Question-naire (AAQ; Hayes et al., 2004; Bond et al., 2011) or one of itsvariations. Mediation of treatment outcomes may be betterassessed by measures of psychological flexibility that targetspecific areas of difficulty. Domain-specific measures have beenfound to be sensitive to changes in psychological flexibility withsmoking urges (Gifford et al., 2004), diabetes-related thoughts andfeelings (Gregg, Callaghan, Hayes, & Glenn-Lawson, 2007), obesitystigma (Lillis & Hayes, 2008), auditory hallucinations (Shawyeret al., 2007), and pain-related thoughts and feelings (McCracken,Vowles, & Eccleston, 2004). In some cases, the domain-specificmeasures mediated treatment outcomes where the general AAQdid not (e.g., Gifford et al., 2004; Gregg et al., 2007).

To date, two studies in this area have assessed body imageflexibility as a proposed process of change (Berman et al., 2009;Lillis et al., 2009). Berman et al., 2009 conducted a small caseseries investigating ACT with previously treated, unremittedanorexia nervosa, and demonstrated improvement in body imageflexibility in all three participants. The design did not, however,allow for a formal mediation analysis, and used an unpublishedversion of a measure the authors developed (Sandoz & Wilson,2006). Lillis et al., 2009 demonstrated weight-specific psychologi-cal flexibility as a mediator of weight control and quality of lifefollowing an ACT-based workshop. Their measure, the Acceptanceand Action Questionnaire for Weight-Related Difficulties (AAQ-W;Lillis & Hayes, 2008), may not be sensitive to body image inflex-ibility in the form of controlled eating (e.g., avoidant restriction).Although an item like, “I am in control of my eating behavior”might indicate greater flexibility for an individual struggling withbinge eating, it would likely indicate inflexibility in an individualengaging in harmful dietary restriction.

Further investigation of the role of body image flexibility in theprevention or treatment of disordered eating depends on thedevelopment of a brief measure of body image flexibility that ispsychometrically sound, easy to administer, and appropriate forindividuals with a range of weight concerns and a variety ofdisordered eating behaviors. For example, disordered eating andeating-related behaviors (e.g., purging and excessive exercise) areexhibited by both male (e.g., Drummond, 2002; Lavender, DeYoung, & Anderson, 2010; O'Dea & Abraham, 2002; Petrie,Greenleaf, Reel, & Carter, 2008) and female (e.g., Berg, Frazier, &Sherr, 2009; Napolitano & Himes, 2011; Reinking & Alexander,2005) college students. In many cases, subclinical levels of dis-ordered eating can be just as disruptive to functioning as eatingpatterns meeting diagnostic criteria (Fairburn & Bohn, 2005).However, measures developed in clinical populations may not beappropriate for these populations, as they might be insensitive tovariations in the subclinical range.

5. The Body Image-Acceptance and Action Questionnaire

The current series of studies present an attempt to develop ameasure of body image flexibility that could allow for researchthat examines the role of body image flexibility in preventing or

E.K. Sandoz et al. / Journal of Contextual Behavioral Science 2 (2013) 39–48 41

reducing disordered eating in a nonclinical population. The BodyImage-Acceptance and Action Questionnaire (BI-AAQ) wasadapted from existing measures of psychological flexibility tospecifically assess flexible responding to body-related thoughtsand feelings. Three studies were conducted in a college studentsample so as to include individuals with varying degrees ofdisordered eating. Pilot data suggested significant disorderedeating in around ten to fifteen percent of this particularpopulation.

Study one focused on the generation and selection of items forthe BI-AAQ, and an initial evaluation of its validity. Concurrentvalidity was examined in terms of the relationship between bodyimage flexibility and overall psychological flexibility, body imagedissatisfaction, and disordered eating. Study two focused on anevaluation of the reliability of the BI-AAQ. Study three focused onreplicating the results of study one.

6. Study one: initial validation

6.1. Participants

One hundred eighty-two (182) undergraduates were recruited forparticipation in return for extra credit in their psychology classes.Participants had an average age of 19.65 with 92% being between theages of 18 and 22. The sample was 70% female. They were 68.2%Caucasian and 15.3% African American. Only one individual in thesample reported having been diagnosed with an eating disorder atone time. Sixty-five percent of the sample had a calculated bodymass index (BMI) (from self-reported weight and height) withinnormal range (18.5 kg/m2oBMIo25 kg/m2), with 4.9% classified as“underweight,” 18% classified as “overweight,” and 10.4% classified as“obese” (WHO, 1995).

6.2. Measures

6.2.1. Body Image-Acceptance and Action QuestionnaireThe BI-AAQ (see Appendix) was designed to measure body

image flexibility. Body image flexibility is defined as the capacityto experience the ongoing perceptions, sensations, feelings,thoughts, and beliefs associated with one's body fully and inten-tionally while pursuing chosen values. The original BI-AAQ itempool included 46 items adapted from (1) the three publishedversions of the Acceptance and Action Questionnaire (Hayes et al.,2004; Bond & Bunce, 2003; Bond et al., 2011) and (2) the itempools from which the published AAQ items were selected. Theauthors rewrote these items to focus specifically on body-relatedcontent as opposed to psychological experiences in general. Itemsworded in the direction of inflexibility are reverse scored such thathigher summed scores are indicative of greater body imageflexibility.

6.2.2. Acceptance and Action Questionnaire-II (AAQ-II; Bond et al.,2011)

Psychological flexibility involves both the ability to acceptdifficult thoughts and feelings and to engage in valued activity intheir presence. This study employed the 10-item version of theAAQ-II (Bond et al., 2011), an improved version of the earlier AAQ(Hayes et al., 2004). Although the 7-item version has been recom-mended by the authors (Bond et al., 2011), the 10-item version waswhat was available and in common use at the time that the studywas conducted. Higher scores are indicative of greater psychologicalflexibility. Scores on the AAQ-II have been shown to have goodreliability and construct validity (Bond et al., 2011). The internalconsistency of the AAQ-II in this sample was good, with a Cron-bach's alpha of 0.84.

6.2.3. Body Shape Questionnaire (BSQ; Cooper, Taylor, Cooper,& Fairburn, 1986)

Body image dissatisfaction was assessed using the BSQ (Cooperet al., 1986), a commonly used 34-item measure. Participantsrespond on a five-point scale from never to always, and responsesare scored such that higher scores indicated higher body imagedissatisfaction. The BSQ has demonstrated good test–retest relia-bility, discriminant validity, concurrent validity, and criterion-related validity (Cooper et al., 1986; Rosen, Jones, Ramirez, &Waxman, 1996). The internal consistency of the BSQ in this samplewas excellent, with a Cronbach's alpha of 0.97.

6.2.4. Eating Attitudes Test-26 (EAT-26; Garner, Olmsted, Bohr,& Garfinkel, 1982)

Disordered eating behavior was assessed using the EAT-26(Garner et al., 1982). Participants respond to 26 items on a six-point scale from never to always. Responses were recorded suchthat “Always,” “Usually,” or “Often,” are scored as 3, 2, or 1,respectively. All other responses (i.e., “Sometimes,” “Rarely,” or“Never”) were scored as 0. Responses result in scores on threesubscales: Dieting, Food Preoccupation, and Oral Control. TheDieting subscale includes items that describe the extent to whichindividuals restrict intake, obsess about thinness, and experienceeating related discomfort. The Food Preoccupation subscaleincludes items that describe the extent to which individuals feelcontrolled by food and food-related thoughts and engage inbulimic behavior. The Oral Control subscale includes items thatdescribe the extent to which individuals exert highly-controlledeating behavior and feel pressure to gain weight or eat more. Onall three subscales, higher scores indicate more disordered eatingbehavior.

The EAT-26 has been shown to have good criterion-related,discriminant, concurrent validity (Garner et al., 1982; Mintz &O'Halloran, 2000; Koslowsky et al., 1992), internal consistency(Garner et al., 1982; Mazzeo, 1999), and test–retest reliability(Mazzeo, 1999). It has also been adapted as a valid measuredisordered eating behavior in a number of populations (seeGarfinkel & Newman, 2001 for a thorough review of the EAT).Internal consistency of the EAT-26 in this sample was good, with aCronbach's alpha of 0.86.

6.3. Procedures

After consenting to participation, participants completed thebattery of measures including the BI-AAQ item pool, the AAQ, theEAT-26, the BSQ, and a demographics questionnaire, in that order.Participants received a debriefing form upon completion, provid-ing more detailed information regarding the study's hypothesesand references for supplementary information and mental healthservices.

6.4. Study one results

6.4.1. Factor analysisBody image flexibility was defined to be actively contacting

perceptions, thoughts, beliefs, and feelings about the body withoutattempts to change their intensity, frequency, or form. It wasexpected to demonstrate a unidimensional structure. Seventeenitems were omitted based on low or negative item-total correla-tions (o0.30). Omission followed an iterative procedure. All itemswith negative item-total correlations were omitted. Then the itemwith the lowest item-total correlation was omitted, and theremaining item-total correlations examined. This was repeateduntil all items had item-total correlations over 0.30. Internal

E.K. Sandoz et al. / Journal of Contextual Behavioral Science 2 (2013) 39–4842

consistency of the remaining 29 items was good, with a Cronbach'salpha of 0.93.

The remaining data were then examined to further determinefactorability. First, the Kaiser–Meyer–Olkin (KMO) Measure of Sam-pling Adequacy was 0.91, above the recommended 0.6 (Kaiser, 1970;Kaiser & Rice, 1974) and indicating enough common variabilityamongst items to indicate factorability. Second, Bartlett's test ofSphericity was significant (χ2 (406)¼2596.45, po0.001), meaningthat the correlation matrix diverged significantly from an identitymatrix in which the items would be uncorrelated. Third, thediagonals of the anti-image correlation matrix (KMO for individualitems) were all over 0.8, supporting the inclusion of all 29 items inthe factor analysis based on partial correlations other items. Fourth,the initial communalities ranged from 0.33 to 0.72, suggesting asatisfactory amount of shared variance for each item with otheritems. Taken together, these indices suggested that despite themoderate sample size, factor analysis was appropriate for these data(MacCallum, Widaman, Zhang, & Hong, 1999).

Principal factor analysis was conducted to determine the latentstructure of the data with the intention of determining which itemsbest measure body image flexibility. Direct oblimin rotation (with adelta of 0) was employed so that multiple factors, if found, would beallowed to correlate. Parallel analysis (Horn, 1965), a Monte Carlo testfor retention of factors based on size of eigenvalues, was conductedto determine the number of factors corresponding to eigenvaluessignificantly greater than parallel random eigenvalues (i.e., eigenva-lues generated from completely independent items). One hundredsets of eigenvalues reflecting sampling were generated based onrandom data matrices representing 29 variables and 175 cases torepresent the effects of sampling error. Because of the tendency ofparallel analysis to overfactor, the 95th percentiles of the randomeigenvalues were used, instead of the means (Glorfeld, 1995;Harshman & Reddon, 1983). As seen in Fig. 1, a comparison of theactual eigenvalues obtained to those calculated from the randomdata, suggested retention of two factors would be appropriate. Thefirst factor accounted for 34.6% of the variance and the second factoraccounted for an additional 7.4% of the variance.

Examination of the rotated factor loadings revealed that all 29items loaded above 0.30 on one factor. Eleven of these items alsoloaded above 0.30 on both factors. Additionally, all of the itemsloading on the second factor were those that were worded in thedirection of flexibility, suggesting that this factor may be anartifact of item wording. There were no other cross loadings.Finally, the two factors were correlated 0.45. For these reasons,the analysis was repeated and one factor extracted, as would betheoretically consistent.

The single extracted factor accounted for 34.4% of the variance.Twenty-six items had factor loadings above 0.40, suggesting amore conservative criterion for retention could be used in order to

Fig. 1. Actual vs. randomly generated eigenvalues in study 1.

produce a measure of shorter length. Twelve items with factorloadings above 0.60 were retained for further analyses.

6.4.2. Calculation and descriptive statisticsBody image flexibility scores were calculated by reverse-

scoring and summing responses on the 12-item BI-AAQ (seeAppendix). The distribution of scores was examined for centraltendency and spread. BI-AAQ scores ranged from 14 to 84, with amean of 66.56 and a standard deviation of 15.11. Thus, thedistribution was slightly negatively skewed, indicating a trendtoward body image flexibility.

6.4.3. Internal consistencyThe twelve-item version of the BI-AAQ was examined for

internal consistency. All item-total correlations were over 0.57.Cronbach's alpha was 0.92, and decreased if any one item wasdeleted.

6.4.4. CovariatesCorrelational analyses were conducted to examine the relation-

ships between body image flexibility and theoretically relevantcriterion variables. The relationship between BMI and body imageflexibility approached significance, such that higher body massindex (BMI) was associated with lower body image flexibility(r (176)¼−0.15, p¼0.054). This suggests that it may be importantto control for BMI when examining the relationship between bodyimage flexibility and disordered eating.

Significant gender differences were also noted with the dis-tribution of scores being significantly less variable for males(p¼0.034) and males reporting significantly higher body imageflexibility than females (po0.001). Examination of correlationmatrices of body image flexibility with other covariates anddependent variables revealed that this difference did not extendto the multivariate level, meaning that the pattern of relationshipsbetween body image flexibility and other variables of interest wassimilar enough between genders to allow for subsequent analysesto be conducted on males and females as one group.

The pattern of correlations was wholly similar with one notableexception. The correlation between body image flexibility and BMI,although nonsignificant in the female sample and in the sample asa whole, was significant and of moderate size in the male sample,(r (53)¼−0.366, p¼0.007). For this reason, BMI was controlled forin subsequent analyses.

6.5. Construct validity

6.5.1. Concurrent validityBody image flexibility would be expected to relate negatively to

body dissatisfaction and disordered eating. As seen in Table 1,partial correlational analysis supported these relationships.BI-AAQ scores were highly negatively related to body imagedissatisfaction as assessed by the BSQ and disordered eating asassessed by the EAT-26 and the BULIT-R. The sole exception wereresponses on the Oral Control subscale, which were unrelated toBI-AAQ scores, suggesting that body image flexibility does notpredict oral control when BMI is controlled for. Body imageflexibility was low when body image dissatisfaction and disor-dered eating were high.

If the BI-AAQ measures psychological flexibility specific to thebody, BI-AAQ scores would be expected to have a moderatelypositive correlation with measures of overall psychological flex-ibility. As seen in Table 1, partial correlation coefficients providedinitial support for the hypothesized relationship. BI-AAQ scoreswere moderately positively related to psychological flexibility.

E.K. Sandoz et al. / Journal of Contextual Behavioral Science 2 (2013) 39–48 43

6.5.2. Incremental validityIf body image flexibility is important in predicting disordered

eating, then body image flexibility would be expected to (a) improveprediction of disordered eating behavior above and beyond bodyshape dissatisfaction and (b) moderate the predictive value of bodyimage dissatisfaction for disordered eating. Hierarchical regressionanalyses were conducted predicting disordered eating behavior fromBSQ and BI-AAQ scores after controlling for BMI, and with BI-AAQentered last. As indicated in Table 2, BI-AAQ scores remained asignificant predictor of EAT-26 scores even after controlling for bodydissatisfaction and actual body size (BMI). In addition, the modelincluding both BSQ and BI-AAQ accounted for over half of thevariance in disordered eating behavior scores.

Next, hierarchical regression analyses were repeated, includingthe BSQ�BI-AAQ interaction term to test for moderation effects ofbody image flexibility on the relationship between body shapedissatisfaction and disordered eating behavior. The interactionbetween body shape dissatisfaction and body image flexibilitywas a significant predictor of disordered eating behavior, indicat-ing that body image flexibility moderated the relationshipbetween body image dissatisfaction and disordered eating.(po0.001). As indicated in Fig. 2, the slope of the regression linefor predicted disordered eating at different levels of body imagedissatisfaction was significantly greater when body image flex-ibility was low (po0.001). This suggests that body image dis-satisfaction had a decreased impact on disordered eating whenbody image flexibility was high.

Additionally, if body image flexibility is to be a useful construct,it would be expected to improve prediction of disordered eating

Table 1Correlations between body image flexibility, psychological flexibility, and eating-related difficulties.

Body image flexibility

Study 1 Study 3rpartial r

Psychological flexibility (AAQ) 0.48nnn 0.30nnn

Body image dissatisfaction (BSQ) −0.73nnn −0.80nnn

Bulimic symptoms (BULIT-R) −0.70nnn

Dieting (EAT-26) −0.69nnn −0.70nnn

Food preoccupation (EAT-26) −0.60nnn −0.61nnn

Oral control (EAT-26) −0.09 −0.27nnn

Note: Study 1 correlations were calculated after controlling for Body Mass Index.npo0.05, nnpo0.01, nnnpo0.001..

Table 2Hierarchical regression model predicting disordered eating (EAT-26 total) from body sh

Std. beta

BMI −0.030

BMI −0.184Body image dissatisfaction 0.761

BMI −0.184Body image dissatisfaction 0.537Body image flexibility −0.309

BMI −0.172Body image dissatisfaction 1.121Body image flexibility 0.107Body image dissatisfaction�body image flexibility −0.434

Note: Values in parentheses are standard errors.npo0.05, nnpo0.01, nnnpo0.005, nnnnpo0.001.

behavior over existing measures of general flexibility. Hierarchicalregression analyses were conducted predicting disordered eatingbehavior from BSQ, AAQ, and BI-AAQ scores after controlling forBMI and with BI-AAQ scores entered last. The BI-AAQ remained asignificant predictor of EAT-26 scores even after controlling forBMI, body shape dissatisfaction, and general flexibility (po0.001).Even when general flexibility was a significant predictor ofdisordered eating behavior, BI-AAQ scores still contributedsignificantly to the overall predictive power of the model.

7. Study two: test–retest reliability

7.1. Participants

Two hundred thirty-four (234) undergraduates were recruitedfor participation in return for extra credit in their psychologyclasses. Participants were unique to this study (i.e., potentialparticipants were restricted from participation if they had parti-cipated in study one). Participants had an average age of 19.42years with 97% being between the ages of 18 and 22. The samplewas 68.4% female. The sample was 84.2% Caucasian and 12.8%African American.

ape dissatisfaction and body image flexibility after controlling for BMI.

t R2 ΔR2

−0.391 (0.149) 0.001 0.001

−3.571 (0.102) 0.556 0.555nnnn

14.750 (0.015)nnnn

−3.743 (0.097)nnnn 0.601 0.045nnnn

7.633 (0.020)nnnn

−4.437 (0.041)nnnn

−3.591 (0.094)nnnn 0.626 0.025nnn

5.996 (0.054)nnnn

0.758 (0.083)−3.357 (0.001)nnn

Fig. 2. Predicated disordered eating from body image dissatisfaction by body imageflexibility.

Table 3Hierarchical regression equation predicting disordered eating (EAT-26 scores) from body image dissatisfaction and body image flexibility.

Std. beta t R2 ΔR2

Body image dissatisfaction 0.712 16.834 (0.013)nnn 0.507 0.507nnn

Body image dissatisfaction 0.392 5.972 (0.021)nnn 0.565 0.059nnn

Body image flexibility −0.401 −6.104 (0.046)nnn

Body image dissatisfaction 1.417 7.910 (0.057)nnn 0.617 0.052nnn

Body image flexibility 0.320 2.399 (0.094)n

Body image dissatisfaction�body image flexibility −0.667 −6.093 (0.001)nnn

Note: Values in parentheses are standard errors.n po0.05, nnnpo0.001.

E.K. Sandoz et al. / Journal of Contextual Behavioral Science 2 (2013) 39–4844

7.2. Measures

In study two, the Demographic Questionnaire and the twelve-item BI-AAQ developed in study one were included in a packet ofmeasures of psychological flexibility and related constructs (e.g.,mindfulness). None of the other measures are of interest to thereliability of the BI-AAQ and will not be described further.

7.3. Procedure

After consenting to participation, participants completed thebattery of 12 brief questionnaires including the DemographicQuestionnaire and 12-item BI-AAQ. This took participants nolonger than 60 mins. Participants returned between two and threeweeks later and completed a shorter battery of three question-naires including the BI-AAQ. This took participants no longer than20 min. Participants received a debriefing form upon completion,providing more detailed information regarding the study's hypoth-eses and references for supplementary information and mentalhealth services.

7.4. Study two results

7.4.1. Calculation and descriptive statisticsThe distributions of scores were examined for central tendency

and spread. For the first administration, BI-AAQ scores rangedfrom 15 to 84, with a mean of 59.9 and a standard deviation of17.4. For the second administration, BI-AAQ scores ranged from 22to 84, with a mean of 63.3 and a standard deviation of 16.1.Consistent with study one, both distributions were negativelyskewed, indicating a trend toward body image flexibility.

7.5. Reliability

7.5.1. Internal consistencyThe 12-item version of the BI-AAQ was examined for internal

consistency in both administrations. For the first administration,all item-total correlations were over 0.54. Cronbach's alpha was0.92, and decreased if any one item was deleted. For the secondadministration, most item-total correlations were over 0.62, withone item (“If I start to feel fat, I try to think of something else”)correlating with the total 0.39. Cronbach's alpha was 0.93. Cron-bach's alpha decreased if any item except this item was deleted, inwhich case it stayed the same.

7.2.2. Test–retest reliabilityThe Pearson product–moment correlation between the two

administrations of the BI-AAQ was 0.80 (po0.01). Correlationsbetween administrations for individual items ranged from 0.30 forone item (“I will have better control over my life if I can control

negative thoughts about my body”) and 0.68 for another item(“I care too much about my body weight and shape”), but all weresignificant at 0.01 level.

8. Study three: validation, replication and extension

8.1. Participants

Two hundred eighty-eight (288) undergraduates wererecruited for participation in return for extra credit in theirpsychology classes. Participants were unique to this study (i.e.,potential participants were restricted from participation if theyhad participated in study one or two). Participants had an averageage of 19.5 years with 95.2% being between the ages of 18 and 22.The sample was 59.9% female. The sample was 77.4% Caucasianand 17.5% African American.

8.2. Measures

In study three, the AAQ-2, the EAT-26, BSQ and the BI-AAQdescribed above were included in a packet of measures of depres-sion, anxiety, and education-related difficulties. Other measureswere not of interest to the validity of the BI-AAQ and will not bedescribed further.

8.3. Procedure

After consenting to participation, participants completed thebattery of 16 brief questionnaires. This took participants no longerthan 60 min. Participants received a debriefing form upon com-pletion, providing more detailed information regarding the study'shypotheses and references for supplementary information andmental health services.

8.4. Study three results

8.4.1. Calculation and descriptive statisticsThe distribution of scores was examined for central tendency

and spread. BI-AAQ scores ranged from 15 to 84, with a mean of66.2 and a standard deviation of 15.8. Consistent with studies oneand two, the distribution was negatively skewed, indicating atrend toward body image flexibility.

7.4.2. Internal consistency and structural validityThe 12-item version of the BI-AAQ was examined for internal

consistency. The scale was found to have good internal consis-tency. All item-total correlations were over 0.52. Cronbach's alphawas 0.93, and decreased or stayed the same if any one item wasdeleted. The principal factor analysis and the parallel analysisconducted in study one were replicated, providing additional

Table 4Logistic regression equation predicting eating disorder risk (EAT-26≥11) from body image flexibility and body shape dissatisfaction.

Predictor Std. beta SE Wald's χ2 Odds ratio

Constant −0.672 0.127 28.066 N/A

Constant 5.483 0.775 50.057 N/ABody image flexibility −0.096 0.012 64.125nnn 0.908

Constant −2.021 1.565 1.668 N/ABody image flexibility −0.036 0.016 5.384nn 0.964Body image dissatisfaction 0.040 0.008 25.158nnn 1.041

Constant 3.087 3.199 0.931 N/ABody image flexibility −0.113 0.046 6.018nn 0.893Body image dissatisfaction −0.007 0.026 0.069 0.993Body image flexibility�body image dissatisfaction 0.001 0.000 3.307n 1.001

Evaluation of final model 128.254nnnn

Likelihood ratio test 131.179nnnn

Hosmer–Lemeshow goodness-of-fit test 6.392

n po0.10.nn po0.05.nnn po0.01.nnnn po0.001.

E.K. Sandoz et al. / Journal of Contextual Behavioral Science 2 (2013) 39–48 45

support for the stability of the unidimensional model. This singlefactor accounted for 54% of the variance in responses.

8.4.3. Concurrent validityBody image flexibility was related to overall psychological

flexibility and disordered eating in study one. As seen in Table 1,these relationships were replicated in the current study. BI-AAQscores were positively related to psychological flexibility. Theywere also negatively related to body shape dissatisfaction anddisordered eating.

8.4.4. Incremental validityIn study one, body image flexibility improved prediction of

disordered eating behavior above and beyond body shape dis-satisfaction and moderated the predictive value of body imagedissatisfaction for disordered eating. As seen in Table 3, hierarch-ical regression analyses replicated this finding. Notably, a full 61%of the variance in disordered eating was predictable from bodyimage dissatisfaction, body image flexibility, and the interactionterm. As in study 1, body image dissatisfaction had less of animpact on disordered eating when body image flexibility was high.

8.4.5. Criterion-related validityNinety-six participants (33.3%) were classified as at-risk for

eating disorders using the recommended cutoff for nonclinicalsamples of 11 on the EAT-26 (Orbitello et al., 2006). If body imageflexibility is important in predicting disordered eating, BI-AAQscores should discriminate those at-risk for eating disorders fromthe rest of the sample. Logistic regression was performed toexplore the extent to which increases body image flexibilitycorrespond to an increase in the likelihood of eating disorder risk.Body image dissatisfaction and the interaction term were includedas predictors in Blocks 2 and 3, respectively.

As seen in Table 4, all three models were significant. Bodyimage flexibility improved prediction over the null model, χ2(1)¼99.204, po0.001. Body image flexibility alone correctly classified91.5% of those with eating disorder risk, and 54.3% without. Bodyimage dissatisfaction improved prediction when included alongwith body image flexibility, χ2(2)¼128.254, po0.001. Body imageflexibility and body image dissatisfaction correctly classified 92.6%of those with eating disorder risk, and 63.6% without. The final

model, including the body image flexibility�body image dissa-tisfaction interaction, was also significant, χ2(2)¼128.254,po0.001. The full factorial model correctly classified 92.6% ofthose with eating disorder risk and 63.6% without, for a total of72.7% correct classification. Interestingly, body image dissatisfac-tion was not a significant predictor once the interaction term wasincluded.

9. Discussion

The current series of studies attempted to develop and provideinitial psychometric support for a brief self-report measure ofbody image flexibility, the Body Image-Acceptance and ActionQuestionnaire (BI-AAQ). To be useful, such an instrument wouldneed to demonstrate both that it does, indeed, measure bodyimage flexibility (validity), and that it does so consistently(reliability).

9.1. Reliability

The twelve-item BIAAQ was found to assess body imageflexibility with good reliability between items and administra-tions. Internal consistency was high (0.92–0.93) in all four admin-istrations across three different samples. Temporal stability wasgood (0.80) for overall BI-AAQ scores.

8.2. Validity

Body image flexibility was defined as the capacity to experiencethe perceptions, sensations, feelings, thoughts, and beliefs about thebody fully and intentionally while pursuing effective action in otherlife domains. Although conceptually, body image flexibility could bedistinguished into different capacities (e.g., capacity to experiencebody image vs. capacity to pursue effective action), each item wasgenerated to refer to both acceptance and action. In other words,the BI-AAQ was intended to measure body image flexibility as aunidimensional construct. Results of principal factor analysis in thefirst study supported the retention of 12 items that loaded on asingle factor with a conservative retention criterion. Studies 2 and3 replicated a single, highly homogenous factor. Since data werefirst collected for these studies, the item selection and

E.K. Sandoz et al. / Journal of Contextual Behavioral Science 2 (2013) 39–4846

unidimensional structure of the BI-AAQ has been replicated usingconfirmatory factor analysis in a large sample with more diversebody sizes (Ferreira, Pinto-Gouveia, & Duarte, in press). Together,these findings suggest that the unidimensional structure is stable.

Body image flexibility was the focus of these studies because ofits relationship with disordered eating, and purported role as aprocess of change in prevention and treatment of disorderedeating. Study one and study three both supported a strongnegative relationship between body image flexibility and a rangeof self-reported disordered eating behaviors. Dieting, bulimicsymptoms, and preoccupation with food were all strongly asso-ciated with inflexibility with body image after controlling for BMI.Oral control, however, was only associated with body imageflexibility when BMI was not controlled for. Consistent with theextant literature, this may reflect more variability in the functionsof oral control than with other facets of disordered eating(Masuda, Price, & Latzman, 2012). For some individuals, control-ling dietary intake may actually reflect increased flexibility. Thisalso may be an artifact of the homogenous samples common inthis literature.

Body image flexibility also proved important in correct identi-fication of individuals at risk for eating disorders. Nearly 73% ofparticipants' risk status was accurately predicted by body imageflexibility alone, and body image dissatisfaction did not contributesignificantly to the predictive model once the interaction termwasadded. Body image flexibility seems to be more important thanbody image dissatisfaction when it comes to predicting who is atrisk for development of eating disorders.

Construct validity depends, however, not only on relationshipswith the outcome of interest, but also on relationships with otherpredictors. Body image flexibility was positively related to overallpsychological flexibility. Body image flexibility improved predic-tion of disordered eating, however, above and beyond overallpsychological flexibility even after controlling for body imagedissatisfaction and BMI. This suggests that the AAQ might be lesssensitive to differences in the kind of psychological inflexibilityimportant in eating disorders. This also provides additional sup-port for the continued development of measures of psychologicalflexibility that are domain specific.

Body image flexibility was strongly negatively related to bodyimage dissatisfaction. This was not surprising, as many items onthe final BI-AAQ included specific reference to how people relateto aversive experiences with the body (e.g., “feeling fat,” or“negative thoughts about the body”). Notably, these items wereempirically selected. In study one, the items in the pool that didnot refer specifically to dissatisfaction (e.g., “When I think aboutmy body…”) did not meet the criterion for retention. This suggeststhat body image flexibility may be more important when peoplefeel badly about their bodies than when they feel good or neutral.

Despite the strong relationship between body image flexibilityand body image dissatisfaction, body image flexibility predicteddisordered eating above and beyond body image dissatisfaction.Consistent with the theoretical model (e.g., Sandoz, Wilson, &Kellum, 2009) and with extant literature (see Hayes et al., 2006),body image flexibility moderated the relationship between bodyimage dissatisfaction and disordered eating. Body image dissatis-faction had less of an impact on disordered eating when bodyimage flexibility was high. This was consistent in linear andcategorical models (i.e., when predicting EAT-26 scores and whenpredicting risk status).

These data complement previous research indicating thatmindfulness moderates the relationship between disordered eat-ing cognitions and disordered eating behaviors (Masuda et al.,2012). Masuda et al. (2012) have posited that overall psychologicalflexibility is not as important to predicting disordered eating as itis general psychological health. Flexibility specific to body image,

however, seems to make a considerable contribution. It is likelythat the moderation effect of body image flexibility extends to therelationship between disordered eating cognitions (of whichbody image dissatisfaction is an aspect) and disordered eatingbehaviors.

9.3. Methodological limitations

There were methodological issues that limit the conclusionsthat can be drawn from this series of studies. For one, all threesamples were drawn from the same population of undergraduatestudents. The participants were mostly Caucasian females aroundnineteen years of age. The homogenous samples limited range, andthus, statistical power in such a way as to prevent meaningfulanalyses examining individual differences in body image, disor-dered eating, or the relationships between body image anddisordered eating. Generalizability of these findings will dependon the BI-AAQ's validation with diverse genders, ethnicities, agegroups, cultures and clinical groups. This work is already under-way, and with promising initial findings (e.g., Ferrerira et al., inpress; Merwin, France, & Zucker, 2009; Rabitti, Manduchi, Miselli,Presti, & Moderato, 2010).

In addition, the order of administration of the questionnaireswas not randomized between subjects. This means that responseson the BI-AAQ could be affected by the participants' experience ofhaving just completed other questionnaires. Order effects seemunlikely, however, as the BI-AAQ was administered followingdifferent questionnaires in all three studies with similar univariateand multivariate results.

Finally, all data were collected via concurrent self-report.Anonymity reduces the likelihood of dishonesty. It does notensure, however, that the participants understood the questionsand successfully discriminated the correct answers. The validity ofthese findings also rests on the assumption that these relation-ships do not tend to vary significantly without intervention. Theseconclusions would be strengthened if BI-AAQ scores were found topredict performance on behavioral tasks or trends in behaviorover time.

9.4. Future directions

In addition to replications of this work in various samples andwith suggested methodological improvements, future work shouldfocus on further specification of body image flexibility as aconstruct. For example, efforts should be made to examine howbody image flexibility might contribute to the overall multidimen-sional construct of body image. Traditional cognitive behavioralapproaches have focused on body image as including threedimensions: perceptions, attitudes, and overt behavior (Cash &Pruzinsky, 2002). The current studies addressed only the relation-ship between body image flexibility and body image dissatisfac-tion, and in a limited way. Future research might integrate currentfindings with existing research by examining how body imageflexibility relates to traditional conceptualizations of body image.Of particular interest might be longitudinal studies that examinebody image flexibility and other body image variables over time,along with associated difficulties. It may be that body imageflexibility changes an individual's susceptibility to body imagedisturbances in perception, attitudes, and overt behavior.

Future research might also focus on outcomes of body imageflexibility interventions on different components of body image.Interventions that target body image flexibility may effect someimprovement of body image distortion, body image dissatisfaction,body image investment and body image avoidance. For example,cognitive interventions have traditionally focused not only onbody image dissatisfaction but also on body image investment

E.K. Sandoz et al. / Journal of Contextual Behavioral Science 2 (2013) 39–48 47

(e.g., Cash, 1997; Rosen, 1997). This is an important area ofdevelopment as body image investment predicts treatmentresponse, treatment outcome, and long-term remission (e.g.,Fairburn, Cooper, & Shafran, 2003; Fairburn, Peveler, Jones, Hope,& Doll, 1993; Masheb & Grilo, 2008; McFarlane, Olmsted, &Trottier, 2008; Spoor, Stice, Burton, & Bohon, 2007). There hasbeen some evidence, however, that targeting flexibility withthoughts can produce bigger reductions in believability ofthoughts than direct cognitive disputation (e.g., Zettle & Hayes,1986; Bach & Hayes, 2002). It could be that body image dissatis-faction and investment may be better targeted through interven-tions on body image flexibility.

Finally, the current studies provided support for body imageflexibility as a moderator of the relationship between body imagedissatisfaction and disordered eating. Body image dissatisfactionseems to have less of an impact on disordered eating if body imageflexibility is high than when it is low. Other approaches havesupported body image flexibility as a mediator of the relationshipbetween disordered eating cognitions, such as body image dis-satisfaction, and disordered eating behavior (Wendell, Masuda, &Le, 2012). This slightly different model suggests that body imageflexibility may account for the association between body imagedissatisfaction and disordered eating. This does not, however,speak specifically to the role of body image flexibility as a potentialmediator of changes to disordered eating or risk of disorderedeating following mindfulness-based prevention or treatmentefforts. Now that the BI-AAQ has demonstrated reliability andvalidity, future research should include direct examination ofvalidity and utility in the context of response to interventionsover time in both clinical and nonclinical samples.

The BI-AAQ was developed with the ultimate goal of facilitatingcontinued prevention and treatment development in the area ofdisordered eating. Mindfulness-based treatments of eating disor-ders are quickly gaining support for their effectiveness (see Baeret al. (2005), Kristeller et al. (2006) for reviews). To date, however,there has been no way to demonstrate that they are effective bychanging the processes they target. The findings of these studiesprovide not only psychometric support for a measure that could beused in precisely this way but also provide further support for theimportance of including body image flexibility in the conceptua-lization and treatment of disordered eating. The current findingssuggest that interventions targeting body image flexibility mayprevent or reduce disordered eating, and offers the BI-AAQ as anassessment of purported mechanisms of change in suchinterventions.

Appendix. BI-AAQ

Directions: Below you will find a list of statements. Please rate thetruth of each statement as it applies to you. Use the following ratingscale to make your choices. For instance, if you believe a statement is‘Always True,’ you would write a 7 next to that statement.

Nevertrue

Veryseldomtrue

Seldomtrue

Sometimestrue

Frequentlytrue

Almostalwaystrue

Alwaystrue

1

2 3 4 5 6 7

1.

Worrying about my weight makes it difficult for me to live alife that I value.

2.

I care too much about my weight and body shape. 3. I shut down when I feel bad about my body shape or weight. 4. My thoughts and feelings about my body weight and shape

must change before I can take important steps in my life.

5.

Worrying about my body takes up too much of my time. 6. If I start to feel fat, I try to think about something else. 7. Before I can make any serious plans, I have to feel better about

my body.

8. I will have better control over my life if I can control my

negative thoughts about my body.

9. To control my life, I need to control my weight.

10.

Feeling fat causes problems in my life. 11. When I start thinking about the size and shape of my body, it's

hard to do anything else.

12. My relationships would be better if my body weight and/or

shape did not bother me.

References

Bach, P., & Hayes, S. C. (2002). The use of Acceptance and Commitment Therapy toprevent the rehospitalization of psychotic patients: a randomized controlledtrial. Journal of Consulting and Clinical Psychology, 70, 1129–1139.

Baer, R. A., Fischer, S., & Huss, D. B. (2005a). Mindfulness and acceptance in thetreatment of disordered eating. Journal of Rational Emotive and CognitiveBehavioral Therapy, 23, 281–300.

Baer, R. A., Fischer, S., & Huss, D. B. (2005b). Mindfulness-based cognitive therapyapplied to binge eating disorder: a case study. Cognitive and Behavioral Practice,12, 351–358.

Barnes, R. D., & Tantleff-Dunn, S. (2010). Food for thought: examining the relation-ship between food thought suppression and weight-related outcomes. EatingBehaviors, 11, 175–179.

Berg, K. C., Frazier, P., & Sherr, L. (2009). Change in eating disorder attitudes andbehavior in college women: prevalence and predictors. Eating Behaviors, 10,137–142.

Berman, M. I., Boutelle, K. N., & Crow, S. J. (2009). A case series investigatingAcceptance and Commitment Therapy as a treatment for previously treated,unremitted patients with anorexia nervosa. European Eating Disorders Review,17, 426–434.

Bond, F. W., & Bunce, D. (2003). The role of acceptance and job control in mentalhealth, job satisfaction, and work performance. Journal of Applied Psychology,88, 1057–1067.

Bond, F. W., Hayes, S. C., Baer, R., Carpenter, K., Guenole, N., Orcutt, H., Waltz, T., &Zettle, R. (2011). Preliminary psychometric properties of the Acceptance andAction Questionnaire II: a revised measure of psychological flexibility andexperiential avoidance. Behavior Therapy, 42, 676–688.

Cash, T. F. (1997). The body image workbook. Oakland, CA: New Harbinger.Cash, T. F., & Pruzinsky, T. (2002). Body image: a handbook of theory, research, and

clinical practice. New York: Guilford Press.Cash, T. F., Melnyk, S. E., & Hrabosky, J. I. (2004). The assessment of body image

investment: an extensive revision of appearance schemas inventory. Interna-tional Journal of Eating Disorders, 35, 305–316.

Cooper, P. J., Taylor, M. J., Cooper, Z., & Fairburn, C. G. (1986). The development andvalidation of the Body Shape Questionnaire. International Journal of EatingDisorders, 6, 485–494.

Drummond, M. (2002). Men, body image and eating disorders. International Journalof Men's Health, 1, 79–93.

Fairburn, C. G., & Bohn, K. (2005). Eating disorder NOS (EDNOS): an example of thetroublesome “not otherwise specified” (NOS) category in DSM-IV. BehaviorResearch and Therapy, 43, 691–701.

Fairburn, C. G., Peveler, R. C., Jones, R., Hope, R. A., & Doll, H. A. (1993). Predictors of12-month outcome in bulimia nervosa and the influence of attitudes to shapeand weight. Journal of Consulting and Clinical Psychology, 61, 696–698.

Fairburn, C. G., Cooper, Z., & Shafran, R. (2003). Cognitive-behavior therapy foreating disorders: a “transdiagnostic” theory and treatment. Behavior Researchand Therapy, 41, 509–528.

Ferreira, C., Pinto-Gouveia, J., & Duarte, C. (2011). The validation of the Body ImageAcceptance and Action Questionnaire: Exploring the moderator effect ofacceptance on disordered eating. International Journal of Psychology & Psycho-logical Therapy, 11, 327–345.

Garfinkel, P. E., & Newman, A. (2001). The Eating Attitudes Test: twenty-five yearslater. Eating and Weight Disorders, 6, 1–24.

Garner, D. M., Olmsted, M. P., Bohr, Y., & Garfinkel, P. E. (1982). The Eating AttitudesTest: psychometric features and clinical correlates. Psychological Medicine, 12,871–878.

Gifford, E. V., Kohlenberg, B. S., Hayes, S. C., Antonuccio, D. O., Piasecki, M. M.,Rasmussen-Hall, M. L., et al. (2004). Applying a functional acceptance basedmodel to smoking cessation: an initial trial of Acceptance and CommitmentTherapy. Behavior Therapy, 35, 689–705.

Glorfeld, L. W. (1995). An improvement on Horn's parallel analysis methodology forselecting the correct number of factors to retain. Educational and PsychologicalMeasurement, 55, 377–393.

Gregg, J. A., Callaghan, G. M., Hayes, S. C., & Glenn-Lawson, J. L. (2007). Improvingdiabetes self-management through acceptance, mindfulness, and values: a

E.K. Sandoz et al. / Journal of Contextual Behavioral Science 2 (2013) 39–4848

randomized controlled trial. Journal of Consulting and Clinical Psychology, 75,336–343.

Harshman R. A., & Reddon J. R. (1983). Determining the number of factors bycomparing real with random data: a serious flaw and some possible correc-tions. In Proceedings of the classification society of North America at Philadelphia(pp. 14–15).

Hayes, S. C. (2004). Acceptance and commitment therapy and the new behaviortherapies: mindfulness, acceptance, and relationship. In: S. C. Hayes, V.M. Follette, & M. M. Linehan (Eds.), Mindfulness and acceptance: expanding thecognitive-behavioral tradition. New York: Guilford.

Hayes, S. C., Strosahl, K. D., Wilson, K. G., Bissett, R. T., Pistorello, J., Toarmino, D., &Zettle, R. (2004). Measuring experiential avoidance: a preliminary test of aworking model. The Psychological Record, 54, 553–578.

Hayes, S. C., Luoma, J., Bond, F., Masuda, A., & Lillis, J. (2006). Acceptance andCommitment Therapy: model, processes, and outcomes. Behaviour Research andTherapy, 44, 1–25.

Horn, J. L. (1965). A rationale and test for the number of factors in factor analysis.Psychometrika, 30, 179–185.

Jacobi, C., Hayward, C., de Zwaan, M., Kraemer, H. C., & Agras, W. S. (2004). Comingto terms with risk factors for eating disorders: application of risk terminologyand suggestions for a general taxonomy. Psychology Bulletin, 130, 19–65.

Jakatdar, T. A., Cash, T. F., & Engle, E. K. (2006). Body-image thought processes: thedevelopment and initial validation of the Assessment of Body-Image CognitiveDistortions. Body Image: An International Journal of Research, 3, 325–333.

Juarascio, A. S., Forman, E. M., & Herbert, J. D. (2010). Acceptance and commitmenttherapy versus cognitive therapy for the treatment of comorbid eatingpathology. Behavior Modification, 34, 175–190.

Kaiser, H. F. (1970). A second generation Little Jiffy. Psychometrika, 35, 401–413.Kaiser, H. F., & Rice, J. (1974). Little Jiffy, Mark IV. Educational and Psychological

Measurement, 34, 111–117.Kashdan, T. B., & Rottenberg, J. (2010). Psychological flexibility as a fundamental

aspect of health. Clinical Psychological Review, 30, 865–878.Kashdan, T. B., Barrios, V., Forsyth, J. P., & Steger, M. F. (2006). Experiential

avoidance as a generalized psychological vulnerability: comparisons withcoping and emotion regulation strategies. Behaviour Research and Therapy, 44,1301–1320.

Koslowsky, M., Scheinberg, Z., Bleich, A., Mordechai, M., Apter, A., Danon, Y., &Solomon, Z. (1992). The factor structure and criterion validity of the short formof the Eating Attitudes Test. Journal of Personality Assessment, 58, 27–35.

Kristeller, J., & Hallet, B. (1999). An exploratory study of a meditation-basedintervention for binge eating disorder. Journal of Health Psychology, 4, 357–363.

Kristeller, J., Quillian-Wolever, R., & Baer, R. A. (2006). Mindfulness-basedapproaches to eating disorders. In R. A. Baer (Ed.), Mindfulness-based treat-ment approaches: Clinician’s guide to evidence base and applications (pp. 75-91). San Diego, CA: Elsevier.

Lavender, J. M., Jardin, B. F., & Anderson, D. A. (2009). Bulimic symptoms inundergraduate men and women: contributions of mindfulness and thoughtsuppression. Eating Behaviors, 10, 228–231.

Lavender, J. M., De Young, K. P., & Anderson, D. A. (2010). Eating DisorderExamination Questionnaire (EDE-Q): norms for undergraduate men. EatingBehaviors, 11, 121–191.

Lillis, J., & Hayes, S. C. (2008). Measuring avoidance and inflexibility in weightrelated problems. International Journal of Behavioral Consultation and Therapy, 4,372–378.

Lillis, J., Hayes, S., Bunting, K., & Masuda, A. (2009). Teaching acceptance andmindfulness to improve the lives of the obese: preliminary test of a theoreticalmodel. Annals of Behavioral Medicine, 37, 58–69.

MacCallum, R. C., Widaman, K. F., Zhang, S., & Hong, S. (1999). Sample size in factoranalysis. Psychological Methods, 4, 84–99.

Masheb, R. M., & Grilo, C. M. (2008). Prognostic significance of two sub-categorization methods for treatment of binge eating disorder: negative affectand overvaluation predict, but do not moderate, specific outcomes. BehaviourResearch and Therapy, 46, 429–437.

Masuda, A., Price, M., Anderson, P. L., & Wendell, J. W. (2010). Disordered eating-related cognition and psychological flexibility as predictors of psychologicalhealth among college students. Behavior Modification, 34, 3–15.

Masuda, A., Price, M., & Latzman, R. D. (2012). Mindfulness moderates therelationship between disordered eating cognitions and disordered eatingbehaviors in a non-clinical college sample. Journal of Psychopathology andBehavioral Assessment, 34, 107–115.

Mazzeo, S. E. (1999). Modification of an existing measure of body shape concernand its relationship to disordered eating in female college students. Journal ofCounseling Psychology, 46, 42–50.

McCracken, L. M., Vowles, K. E., & Eccleston, C. (2004). Acceptance of chronic pain:component analysis and a revised assessment method. Pain, 107, 159–166.

McFarlane, T., Olmsted, M. P., & Trottier, K. (2008). Timing and prediction of relapsein a transdiagnostic eating disorder sample. International Journal of EatingDisorders, 41, 587–593.

Merwin R. M., France B. E., & Zucker N. L. (2009). Eating disorders and the newwave of cognitive-behavioral therapy: does acceptance matter. In Presentationat the annual meeting for the Association for Behavioral and Cognitive Therapy.New York, NY.

Merwin, R. M., Zucker, N. L., Lacy, J. L., & Elliott, C. A. (2010). Interoceptiveawareness in eating disorders: distinguishing lack of clarity from non-acceptance of internal experience. Cognition and Emotion, 24, 892–902.

Merwin, R. M., Timko, C. A., Moskovich, A. A., Ingle, K., Bulik, C. M., & Zucker, N. L.(2011). Psychological inflexibility and symptom expression in anorexia nervosa.Eating Disorders: The Journal of Treatment and Prevention, 19, 62–82.

Mintz, L. B., & O'Halloran, M. S. (2000). The Eating Attitudes Test: validation withDSM-IV eating disorder criteria. Journal of Personality Assessment, 74, 489–503.

Napolitano, M. A., & Himes, S. (2011). Race, weight, and correlates of binge eating infemale college students. Eating Behaviors, 12, 29–36.

O'Dea, J. A., & Abraham, S. (2002). Eating and exercise disorders in young collegemen. Journal of American College Health, 50, 273–278.

Orbitello, B., Ciano, R., Corsaro, M., Rocco, P. L., Taboga, C., Tonutti, L., & Balestrieri,M. (2006). The EAT-26 as screening instrument for clinical nutrition unitattenders. International Journal of Obesity, 30, 977–981.

Petrie, T. A., Greenleaf, C., Reel, J., & Carter, J. (2008). Prevalence of eating disordersand disordered eating behaviors among male collegiate athletes. Psychology ofMen and Masculinity, 9, 267–277.

Polivy, J., & Herman, C. P. (2002). Causes of eating disorders. Annual Review ofPsychology, 2002, 187–213.

Rabitti E., Manduchi K., Miselli G., Presti G.,& Moderato P. (2010). The validation ofthe Italian version of the Body Image Acceptance and Action Questionnaire. InPresentation at the annual meeting for the Association for Contextual BehavioralScience. Reno, NV.

Reinking, M. F., & Alexander, L. E. (2005). Prevalence of disordered-eating behaviorsin undergraduate female collegiate athletes and nonathletes. Journal of AthleticTraining, 40, 47–51.

Rodin, J., Silberstein, L. R., & Streigel-Moore, R. H. (1985). Women and weight: anormative discontent. In: T. B. Sondregger (Ed.), Psychology and gender:Nebraska symposium on motivation, 1984 (pp. 267–307). Lincoln: University ofNebraska Press.

Rosen, J. C. (1992). Body image disorder: definition, development and contributionto eating disorders. In: J. H. Crowther, D. L. Tennenbaum, S. E. Hobfoll, & M. A.P. Stephens (Eds.), The ideology of bulimia nervosa: the individual and familialcontext. Washington, DC: Hemisphere Publishing Corp.

Rosen, J. C. (1997). Cognitive-behavioral body image therapy. In: D. M. Garner, & P.E. Garfinkel (Eds.), Handbook of treatment for eating disorders (2nd ed.). NewYork: The Guilford Press.

Rosen, J. C., Srebnik, D., Saltzberg, E., & Wendt, S. (1991). Development of a bodyimage avoidance questionnaire. Psychological Assessment, 3, 32–37.

Rosen, J.C, Jones, A., Ramirez, E., & Waxman, S. (1996). Body Shape Questionnaire:studies of validity and reliability. International Journal of Eating Disorders, 20,315–319.

Safer, D. L., Telch, C. F., & Chen, E. Y. (2009). Dialectical behavior therapy for bingeeating and bulimia. New York: Guilford.

Sandoz E. K., & Wilson K. G. (July, 2006). Body Image Acceptance Questionnaire:embracing the ‘Normative Discontent.’ In Paper presented at the Association forContextual Behavioral Science ACT/RFT World Conference. London, England.

Sandoz E. K., Wilson K. G., & Kellum K. K. (2009). The effects of ACT for body imagedisturbance on eating behavior and valued living. In Presentation at the meetingfor the Association for Behavior Analysis International. Phoenix, AZ.

Sandoz, E. K., Wilson, K. G., & DuFrene, T. (2011). Acceptance and commitmenttherapy for eating disorders: a process-focused guide to treating anorexia andbulimia. Oakland, CA: New Harbinger.

Shawyer, F., Ratcliff, K., Mackinnon, A., Farhall, J., Hayes, S. C., & Copolov, D. (2007).The voices acceptance and action scale (VAAS): pilot data. Journal of ClinicalPsychology, 63, 593–606.

Shisslak, C. M., & Crago, M. (2001). Risk and protective factors in the developmentof eating disorders. In: J. K. Thompson, & L. Smolak (Eds.), Body image, eatingdisorders, and obesity in youth. Washington, DC: American PsychologicalAssociation.

Slade, P. D. (1982). Towards a functional analysis of anorexia nervosa and bulimianervosa. British Journal of Clinical Psychology, 21, 167–179.

Spoor, S. T. P., Stice, E., Burton, E., & Bohon, C. (2007). Relations of bulimic symptomfrequency and intensity to psychosocial impairment and health care utilizationin a community-recruited sample. International Journal of Eating Disorders, 40,14–505.

Stice, E. (2002). Risk and maintenance factors for eating pathology: a meta-analyticreview. Psychological Bulletin, 128, 825–848.

Stice, E., & Shaw, H. E. (2002). Role of body dissatisfaction in the onset andmaintenance of eating pathology: a synthesis of research findings. Journal ofPsychosomatic Research, 53, 985–993.

Thompson, J. K. (1992). Body image: extent of disturbance, associated features,theoretical models, assessment methodologies, intervention strategies, and aproposal for a new DSM-IV category-body image disorder. In: M. Hersen, R.M. Eisler, & P. M. Miller (Eds.), Progress in behavior modification, Vol. 28 (pp. 3–54). Sycamore, IL: Sycamore Publishing Inc.

Wendell, J. W., Masuda, A., & Le, J. K. (2012). The role of body image flexibility in therelationship between disordered eating cognitions and disordered eatingsymptoms among non-clinical college students. Eating Behaviors, 13, 240–245.

World Health Organization (1995). Physical status: the use and interpretation ofanthropometry. Report of a WHO Expert Committee. Geneva: World HealthOrganization WHO Technical Report Series 854.

Zettle, R. D., & Hayes, S. C. (1986). Dysfunctional control by client verbal behavior:the context of reason-giving. The Analysis of Verbal Behavior, 4, 30–38.