- Email: [email protected]
Accounting for fluctuations in body dissatisfaction
Body Image 8 (2011) 315–321
Contents lists available at ScienceDirect
Body Image journal homepage: www.elsevier.com/locate/bodyimage
Accounting for fluctuations in body dissatisfaction Lauren A. Colautti, Matthew Fuller-Tyszkiewicz ∗ , Helen Skouteris, Marita McCabe, Stephen Blackburn, Elise Wyett School of Psychology, Deakin University, 221 Burwood Highway, Burwood, Victoria 3125, Australia
a r t i c l e
i n f o
a b s t r a c t
Article history: Received 11 January 2011 Received in revised form 27 May 2011 Accepted 3 July 2011 Keywords: Body dissatisfaction State body image Context Mood states Experience sampling method
The present study evaluated whether the strength of relationship between contextual cues (presence of company and mood) and state body dissatisfaction varied as a function of individual differences in key trait measures (body shame, body surveillance tendencies, internalization of appearance standards, and trait affect) which have been linked to trait body dissatisfaction. Fifty-five undergraduate women completed a questionnaire containing the trait-based measures and then carried a Personal Digital Assistant (PDA) for a 7-day period. The PDA prompted participants six times daily to self-report their current mood and state body dissatisfaction. Multi-level modeling revealed that individual differences in body shame predicted inter-individual variability in the strength of the relationships between presence of company and state body dissatisfaction, and positive mood and state body dissatisfaction. Trait positive affect also explained variance in the positive mood state-body dissatisfaction relationship. The implications of the findings for prevention of body image disturbances are discussed. Crown Copyright © 2011 Published by Elsevier Ltd. All rights reserved.
Introduction Body image is a multidimensional construct defined as an individual’s perception of, and attitudes towards, his or her body and appearance (Cash, Fleming, Alindogan, Steadman, & Whitehead, 2002). It encompasses a range of cognitive, affective, and perceptual phenomena (Banfield & McCabe, 2002; Thompson, Heinberg, Altabe, & Tantleff-Dunn, 1999). Body dissatisfaction is one aspect of body image, relating to an individual’s degree of dissatisfaction with particular parts of the body (Cook-Cottone & Phelps, 2003). It is argued that body dissatisfaction consists of both state and trait aspects (Cash et al., 2002; Thompson, 2004; Vocks, Hechler, Rohrig, & Legenbauer, 2009). Trait body dissatisfaction is considered to be a stable and unchanging characteristic that is transferable across a wide range of contexts. Trait body dissatisfaction has been linked to personality traits (e.g., perfectionism, trait affect, and selfesteem), appearance-related factors (e.g., shame felt about one’s appearance, body surveillance, internalization of the thin ideal), and socio-cultural influences (e.g., media, interpersonal relations) (Anschutza, Engelsa, Van Leeuwea, & van Strie, 2009; Stice, 2002; Tissot & Crowther, 2008; van den Berg & Thompson, 2007). In contrast, state body dissatisfaction is thought to fluctuate on a moment-by-moment basis, and these fluctuations have
∗ Corresponding author. Tel.: +61 3 9251 7344; fax: +61 3 9244 6858. E-mail address: [email protected] (M. Fuller-Tyszkiewicz).
been associated with contextual factors (e.g., social settings, exercise, and body exposing situations, such as looking in a mirror), current mood state, and individual differences in personality dispositions and disordered eating symptomatology (Hausenblas & Fallon, 2006; LePage & Crowther, 2010; Melnyk, Cash, & Janda, 2004; Paquette & Raine, 2004; Rudiger, Cash, Roehrig, & Thompson, 2007). The relationship between body dissatisfaction and its purported contributors has most commonly been measured using trait measurements and within cross-sectional designs. However, as this approach is insensitive to the state-dependent aspects of body dissatisfaction ratings, the results of these studies are likely to conflate state- and trait-based components of body dissatisfaction (Melnyk et al., 2004). State-based approaches to measure body dissatisfaction have typically involved the use of simple pre–post induction based methods to elicit state-like changes in body dissatisfaction (Cash et al., 2002; Haimovitz, Lansky, & O’Reilly, 1993; Tiggemann, 2001; Vocks, Legenbauer, & Heil, 2007). For example, Krones, Stice, Batres, and Orjada (2004) compared changes in self-reported body dissatisfaction for women who interacted with a confederate who either conformed to the thin-ideal or who had body dimensions representative of the average woman in the general population. They found that body dissatisfaction increased significantly for participants exposed to the thin-ideal confederate but not for the normal weight confederate. While induction-based studies demonstrate the malleability of body dissatisfaction, they are limited in several ways. First, such methods may have limited ecological validity as the event that
1740-1445/$ – see front matter. Crown Copyright © 2011 Published by Elsevier Ltd. All rights reserved. doi:10.1016/j.bodyim.2011.07.001
316
L.A. Colautti et al. / Body Image 8 (2011) 315–321
induces changes in body dissatisfaction may be more (or less) likely than everyday situations to elicit changes in body dissatisfaction. This may inaccurately represent the extent to which body dissatisfaction levels actually vary from moment to moment (i.e., the level of intra-individual variability). Second, pre–post assessments only measure two time points and, as such, do not allow for measurement error in these ratings. More extensive, repeated measurement is preferable as atypical and outlying responses can then be averaged out. Moreover, the averaged effect is likely to be more representative of the actual effect size than a single instance of testing (Nesselroade & Ram, 2004). Third, despite evidence of within-group variability in pre–post change in body dissatisfaction ratings in these induction-based studies, the possibility that the relationship between state variables may vary as a function of individual difference factors (inter-individual variability) has been largely ignored. Recent methodological advances have led to techniques of data collection that are cost effective, less invasive than experimental designs, and suitable for capturing moment-by-moment data (Beal & Weiss, 2003). The experience sampling method (ESM; Csikszentmihalyi & Larson, 1987), which involves repeated measurement of behaviors, cognitions, and emotions in situ, permits researchers to address more sophisticated questions based on the intra-individual variability of these state variables and the dynamic relationship between state- and trait-based variables than has previously been possible. This has led to empirical evaluations of whether the predictors of trait-based body dissatisfaction may also be predictive of state body dissatisfaction. For instance, it has been shown that personality factors (notably, perfectionistic self-presentation) and appearance-relevant trait measures (psychological investment in appearance, disturbed eating attitudes, and appearance-fixing coping strategies) are predictive of intra-individual variability in state body dissatisfaction ratings (Lattimore & Hutchinson, 2010; Melnyk et al., 2004; Rudiger et al., 2007). While these studies demonstrate that trait measures are predictive of variability in state body dissatisfaction, it is likely that contextual (or state-dependent) cues may also provide opportunities for an individual to feel more (or less) dissatisfied with her/his appearance. To date, only two studies have used ESM to evaluate the extent to which contextual factors influence state body dissatisfaction. Lattimore and Hutchinson (2010) monitored postmeal mood, body dissatisfaction, and perceived caloric intake in a group of women enrolled in a weight loss program. They found that intra-individual variability in state body dissatisfaction was linked with intra-individual variability in both mood and perceived caloric intake. LePage and Crowther (2010) took a different approach by examining the extent to which individual differences at the trait level could account for inter-individual variation in the strength of relationship between instances of exercise and state body dissatisfaction. They found that exercise was associated with decreases in state body dissatisfaction and that the relationship between motives for exercise and state body dissatisfaction was moderated by trait body dissatisfaction. While the results of these state-based body dissatisfaction studies demonstrate that contextual variables and a number of key trait variables are related to variability in state body dissatisfaction, there are still many important questions that need to be addressed. First, the role that social context (company) plays in fluctuations in state body dissatisfaction ratings throughout a person’s day-to-day life has yet to be empirically evaluated. This is despite theoretical arguments that state that the presence of others provides an opportunity for body-based comparisons with others and the possibility to have one’s body visually scrutinized; either of which could result in increased body dissatisfaction (Fredrickson & Roberts, 1997). Body-based comparisons are most likely to occur in the company
of same-sex individuals of similar age (such as friends), although appearance concerns may also be salient when in the company of one’s romantic partner or a potential mate. Second, there has been limited research into the influence of mood state on fluctuations in body dissatisfaction. Although Lattimore and Hutchinson (2010) found an association between intra-individual variability in mood and state body dissatisfaction, this finding does not ensure that changes in body dissatisfaction and mood occurred synchronously. Third, the possibility that interindividual differences exist in the strength of association between context and body dissatisfaction has been neglected. Trait-based variables which have been linked to the etiology and maintenance of trait body dissatisfaction may also make individuals more susceptible to increases in state body dissatisfaction in the presence of others and/or when one’s mood state becomes increasingly negative. For instance, it is possible that the influence of company on state body dissatisfaction is greatest for individuals with traitlevel body image concerns, such as internalization of appearance standards, body shame, and the tendency to engage in bodysurveillance behaviors. The present study used an ESM approach to determine whether: (a) state body dissatisfaction can be predicted by current mood state and/or presence of others and (b) inter-individual differences in the strength of the relationship between state body dissatisfaction and these contextual factors are influenced by key trait measures (body shame, body surveillance, negative or positive trait affect, internalization of the thin ideal tendencies) that, as noted above, have been shown to relate to trait body dissatisfaction. ESM was chosen in preference to a simple induction design in order to capture intra-individual variability in body dissatisfaction states in a more ecologically valid way. By re-sampling the relationship between context and body dissatisfaction six times daily for a period of 1 week, it was expected that estimates would better reflect daily experience and be less influenced by atypical responses and/or extraneous influences (Beal & Weiss, 2003). Based on previous findings, it was hypothesised that decreased positive mood, increased negative mood, and presence of company would be predictive of higher state body dissatisfaction. It was additionally predicted that a stronger relationship between state body dissatisfaction and presence of company would be evidenced for women with higher body shame, body surveillance tendencies, and who were more strongly invested in the sociocultural value placed on physical appearance (internalization of the thin ideal). To further explore associations found previously between mood states and body dissatisfaction (LePage & Crowther, 2010), trait negative and positive affects were also included (in addition to the appearance-based measures) to explain interindividual variation in the relationship between mood states and state body dissatisfaction. It was anticipated that trait negative affect would moderate the relationship between negative mood and body dissatisfaction, whereas trait positive affect would moderate the relationship between positive mood and state body dissatisfaction.
Method Participants A total of 57 women were recruited from advertisements made during undergraduate lectures and tutorials offered at a large metropolitan university in Melbourne, Australia. Two women were excluded due to incomplete data. The 55 remaining participants ranged in age from 19 to 51 years with a mean age of 29.69 (SD = 9.62) years. Self-reported body mass indices (BMI = kg/m2 ) ranged from 17.96 to 36.29 (M = 24.03, SD = 4.38).
L.A. Colautti et al. / Body Image 8 (2011) 315–321
Materials Pre-test measures. Demographics. This questionnaire obtained information concerning the participants’ age, height and weight. Trait affect. Five items from the Trait Affect Scale (TAS; Blore, 2008) were used to measure positive and negative trait affect. Prior research into the structure of affect has shown that these five items adequately capture these two affective traits (Davern, 2004; Yik, Russell, & Feldman-Barrett, 1999). The positive items ask participants to rate how happy, satisfied, and content they generally feel, whereas the negative affect items ask how unhappy and discontent participants feel in general. Items are presented with 5-point Likert-style response scales, ranging from 1 (Never) to 5 (Always). Item responses are summed, resulting in separate scores for each of trait positive and negative affect. Blore (2008) demonstrated the convergent validity of this measure. Cronbach’s alpha estimates in the present study were .78 for positive affect and .82 for negative affect. Internalization of appearance standards. The nine-item internalization-general subscale of the Sociocultural Attitudes Towards Appearance Questionnaire-Version 3 (SATAQ-3; Thompson, van den Berg, Roehrig, Guarda, & Heinberg, 2004) was used to assess the extent to which participants endorse and accept cultural ideals of physical appearance (e.g., “I would like my body to look like the people who are in the movies”). Items were rated on a 5-point Likert scale from 1 (Definitely Disagree) to 5 (Definitely Agree). This subscale has been shown to be reliable and valid (Thompson et al., 2004). Cronbach’s alpha in the current study for Internalization-general scores was .93. Objectified body consciousness. The body surveillance and body shame subscales of the Objectified Body Consciousness Scale (OBCS; McKinley & Hyde, 1996) were used to measure the extent to which participants think of their bodies in terms of appearance (e.g., “I often worry about whether the clothes I am wearing make me look good”) and the extent to which the participant believed they were a bad person for not attaining cultural standards of appearance (e.g., “When I can’t control my weight, I feel like something must be wrong with me”), respectively. Participants responded on a 7-point Likert scale ranging from 1 (Disagree Strongly) to 7 (Agree Strongly) with higher scores reflecting greater body surveillance or body shame. Scale reliability and validity has been demonstrated (McKinley & Hyde, 1996). Cronbach’s alpha in the present study was .84 for body surveillance scores and .81 for body shame scores. PDA-based measures. State body dissatisfaction. The Body Image States Scale (BISS; Cash et al., 2002) is a six-item measure of current body image (dis)satisfaction at a given time point. Participants were asked to rate how satisfied they felt “right now at this moment” in regards to their (a) physical appearance, (b) body size and shape, (c) weight, and (d) physical attractiveness. Participants were additionally asked to rate how their current feelings regarding their looks relative to (e) how they typically felt, and (f) how the average person looks. Items were rated on 9-point Likert scale ranging from 1 (Extremely Dissatisfied) to 9 (Extremely Satisfied). Item scores were averaged and then subtracted from the maximum possible score (9) so that high scores reflected greater body dissatisfaction. Reliability and internal consistency of the BISS has been reported (Rudiger et al., 2007). Cronbach’s alpha in the present study was .93.
317
Mood states. The five items of the TAS (see above) were modified so that participants indicated how they ‘felt right’ now instead of ‘in general’. Items were rated on an 11-point scale ranging from 0 (Not at all) to 10 (Extremely). This measure has been shown to be sensitive to moment-by-moment fluctuations in mood (Blore, 2008). In the present study, Cronbach’s alpha was .94 for both positive and negative mood. Social context. Participants were asked to indicate who they were with at the time of completing the survey on the PDA by selecting one or more of the following options: (a) alone, (b) with partner, (c) with friends, (d) with work colleagues, or (e) with others (unspecified). Although the authors intended to compare state body dissatisfaction ratings across each of the five social contexts, few participants endorsed all of the response options throughout the duration of the PDA testing period. Accordingly, response options two to five were collapsed into a single category representing ‘with others’ and were compared against the ‘alone’ response option. The limitations of this re-categorization are covered in Discussion section. Daily diary apparatus. The PDAs used in the present study were Palm Z22s System running the Palm 5.4 Operating System (©Palm Inc., USA). Questionnaires were constructed using the Purdue Momentary Assessment Tool (PMAT; Weiss et al., 2004). Procedure Ethics approval for the study was provided by the university’s Human Research Ethics Committee. The study was advertised in undergraduate lectures and tutorials. Interested participants met with a member of the research team, at which time they received the Plain Language Statement and Consent Form. After informed consent was obtained, participants completed the pre-test questionnaires listed above. Participants were then given a PDA, charger kit, and instructions for the study. Participants were asked to complete a PDA survey to practice and familiarize themselves with the task, and were encouraged to ask questions if they felt there were any ambiguities with the process or items. Participants carried a PDA with them for a 7-day period. The PDA emitted an audible tone six times per day at random intervals between 10:00 a.m. and 8:00 p.m. This tone prompted participants to complete the questionnaires on the PDA. The PDAs were programmed to signal at intervals of no less than 90 min in order to allow for sampling across the entire day. Moreover, for each time interval, participants had a 15-min window in which to complete the survey before it was counted as missing data by the PDA. Participants were instructed to complete the momentary assessments whenever the PDA signaled, provided it was safe or appropriate to do so. Driving a car or being in an important meeting were given as examples of times where it may be inappropriate to respond to the PDA. The order of presentation of items was held constant across all testing intervals. After 7 days, participants returned the PDA to the research team, where they were debriefed and received a $10 gift voucher as an honorarium. Data Analytic Strategy Given that participants provided multiple PDA responses over time, the data were hierarchical in nature, with responses to PDA surveys (Level 1 data) nested within individuals (Level 2). As a consequence of ignoring the clustered nature of repeated measures data (and violating the assumption of independence of errors), traditional approaches such as multiple regression produce prediction models which under-estimate standard error terms, in turn leading
318
L.A. Colautti et al. / Body Image 8 (2011) 315–321
to Type I error inflation for estimates of the relative contributions of IVs for predicting the DV (Hox, 2002). Repeated measures designs (e.g., ANOVA and MANOVA) are equally problematic as they aggregate data to Level 2 (i.e., inter-individual differences), effectively ignoring differences at Level 1 (intra-individual variability). The present study utilized multilevel modeling (MLM), which is a technique that overcomes any potential violations of the assumptions of independence that can occur when data is clustered within groups (Jackson, 2010). MLM treats each participant as a separate group and thus provides estimates of intra-individual and inter-individual variability in the relationship between state body dissatisfaction and the contextual factors. In the present study, Level 1 variables consisted of state body dissatisfaction, negative mood, positive mood, and company estimates as based on the PDA survey (N = 1530). Level 2 represented estimates of trait variables (body shame, body surveillance, internalization tendencies, and negative and positive trait affect) for each of the participants in the data set (N = 55). In order to assess the relationship between contextual variables and state body dissatisfaction, three aspects of the MLM output were relevant: (a) the Level 1 fixed components, which assessed the strength of the relationship between contextual factors and state body dissatisfaction when averaged across all individuals (i.e., ignoring clustering effects); (b) the Level 1 variance components which assessed the extent to which the strength of these Level 1 relationships varied from person to person (i.e., inter-individual variability); and (c) the Level 2 intercept-and-slopes-as-outcomes model which allowed for evaluation of the extent to which the Level 2 trait variables (internalization tendencies, body shame, body surveillance, and negative and positive trait affect) explained interindividual variability in the strength of the relationship between Level 1 variables. Because BMI has been shown to correlate with these trait variables (Stice, 2002), the Level 2 intercept-and-slopesas-outcomes model was evaluated twice; once with unadjusted estimates of effects, and a second run with effects adjusted for individual differences in BMI.
Results Data Preparation While missing data does not present statistical problems for multi-level modeling, the researcher must decide how much missing data to allow and determine whether the pattern of missingness has implications for interpretability of findings (Hox, 2002; Tabachnick & Fidel, 2006). Two participants were eliminated from the study as they each answered only three of the possible 42 PDA surveys. No other participant missed more than 50% of the surveys and, given this still allowed for at least 21 estimates of state body dissatisfaction across the week, the decision was made to retain all participants who met this 50% cut-off. While this cut-off of 50% may appear low, it must be noted that the number of times the survey was answered throughout the testing period (M = 30.75, SD = 8.66) was still more than those provided in previous daily diary studies of state body dissatisfaction. A related concern is that procedural non-compliance may potentially undermine the generalizability of findings. Therefore, several preliminary analyses were conducted to evaluate whether the likelihood of missing PDA data was related to day of the week or individual difference variables (age, BMI, body dissatisfaction, body shame, internalization, positive affect, and negative affect). A one-way ANOVA revealed that the effect of day of the week on compliance rates was non-significant, F(6, 49) = 1.24, p > .05. Furthermore, bivariate correlations between compliance rates for PDA
surveys and the trait measures used in this study were all nonsignificant (all ps > .05). Prior to the main analysis, Phase 1 data were screened to ensure that they met the assumptions of multivariate analysis (Tabachnick & Fidel, 2006). Less than 2% missing data was found overall, and dealt with using expectation maximization. There were no outliers or evidence of non-normality in any variables. Table 1 presents the means, standard deviations, and possible range of scores for the Level 1 and Level 2 variables in the current study. MLM Analyses When initially modeled as fixed effects, positive and negative mood each made significant contributions to prediction of state body dissatisfaction, t(54) = −15.50, p < .001 and t(54) = 2.06, p < .05, respectively. Presence of company was not reliably associated with state-based body dissatisfaction ratings, t(54) = 0.10, p = .93. However, as shown in Table 2, significant inter-individual variation in the strength of relationships between body dissatisfaction and the three contextual variables was observed once the Level 1 relationships were allowed to vary. Thus, the non-significant fixed effect for the company-body dissatisfaction relationship is not representative of all individuals in the data set. Likewise, that the other two Level 1 relationships differ also suggests that the fixed effect for mood states and body dissatisfaction may under-estimate the true strength of relationship for a subset of individuals in this sample. Finally, the Level 2 intercepts-and-slopes-as-outcomes model showed the extent to which individual difference factors (trait affect, internalization, body shame, and body surveillance) explained variance in the relationships between Level 1 variables. As revealed in Table 3, inter-individual variation in the relationship between company and state body dissatisfaction was significantly predicted by body shame in the unadjusted model, t(50) = 2.73, p < .01. This effect reduced to non-significance after controlling for individual differences in BMI, t(49) = 1.56, p > .05. Thus, although individuals with heightened body shame were most likely to report body dissatisfaction in the company of others, it seems that this effect may, in part, be attributable to the influence of BMI. Individual differences in body surveillance tendencies and level of acceptance of appearance standards were not predictive of inter-individual variations in the relationship between company and body dissatisfaction in either the adjusted or unadjusted models. Inter-individual variation in the relationship between positive mood and state body dissatisfaction was predicted by body shame and trait positive affect, t(50) = −1.66, p < .05 and t(50) = 1.89, p < .05, respectively. Heightened body shame served to weaken the relationship between positive mood and body dissatisfaction whereas trait positive affect strengthened this association. These relationships remained significant after controlling for BMI. None of the trait variables reliably accounted for variance in the strength of the relationship between negative mood and body dissatisfaction, although there was a trend for trait negative affect to be predictive of a stronger relationship between negative mood and body dissatisfaction, t(50) = 1.42, p = .07. Discussion To date, there has been a dearth of research examining the predictors of state body dissatisfaction in everyday life. In particular, there has been insufficient attention given to the possibility that proximal factors (such as mood and social setting) may interact with trait factors (such as appearance-related variables and personality dispositions) to produce momentary fluctuations in body dissatisfaction. Thus, the aim of the present study was to evaluate: (a) the extent to which current mood states and presence of com-
L.A. Colautti et al. / Body Image 8 (2011) 315–321
319
Table 1 Means, standard deviations, and possible score ranges for Level 1 and Level 2 variables. Level
Variable
Mean
SD
Possible range
Level 1
Positive mood Negative mood Body dissatisfaction Internalization Body shame Body surveillance Negative affect Positive affect
17.95 5.11 4.61 28.05 3.42 4.76 5.11 11.11
5.55 4.44 1.73 8.85 1.18 0.97 1.21 1.64
0–30 0–20 1–9 9–45 1–7 1–7 2–10 3–15
Level 2
Notes: The Level 1 variables are state measures, whereas the Level 2 variables are trait-based measures. Table 2 Variance components between individuals for the Level 1 random-intercept-and-slopes model predicting state body dissatisfaction. Variance components between individuals
Variance estimate
SD
df
2
Positive mood Negative mood Company
0.01 0.01 0.39
0.08 0.09 0.62
52 52 52
168.78*** 113.28*** 159.63***
***
p < .001.
pany were predictive of state body dissatisfaction ratings and (b) whether trait affect and key body image measures (internalization of the thin ideal, body shame, and body surveillance tendencies) accounted for inter-individual variation in the strength of relationship between contextual factors and state body dissatisfaction. Present findings supported the hypothesis that contextual factors influence body dissatisfaction ratings. While positive and negative mood, but not company, were predictive of state body dissatisfaction as fixed effects (i.e., when averaged across all participants), there was considerable inter-individual variability in the strength of association between body dissatisfaction and these contextual factors. Thus, even for the non-significant relationship between company and body dissatisfaction, the fixed effect underestimated the strength of this relationship for some participants in the present study. This latter finding with the random effects model illustrates the utility of MLM over more traditional analytic approaches to repeated measures designs (such as ANOVA and multiple regression) which ignore inter-individual variability and, in so doing, can lead to inaccurate or misleading conclusions. Reliance upon these traditional analytic strategies for repeatedmeasure designs may, in part, explain why the influence of mood states on body dissatisfaction has received mixed support in simple pre–post induction designs (e.g., Cash et al., 2002; Haimovitz et al., 1993; Tiggemann, 2001; Vocks et al., 2007). The finding that presence of company is predictive of state body dissatisfaction ratings for at least some individuals is consistent with prior theoretical speculation and empirical literature.
Fredrickson and Roberts (1997) argued that the presence of others (whether they are strangers, friends, partner, etc.) provides an opportunity for one’s body to be appraised. For some individuals (particularly those with negative body image, elevated BMI, and who view appearance as important for their sense of self-worth), the potential for public scrutiny can result in feelings of anxiety and dissatisfaction towards their appearance. In the present study, the relationship between presence of company and state body dissatisfaction was particularly strong for individuals with heightened body shame. Importantly, when individual differences in BMI were controlled for, the influence of body shame on the company-state body dissatisfaction relationship reduced to non-significance. Thus, the influence that presence of others exerts on state body dissatisfaction may derive from one’s awareness or perception of her/his actual body size. However, neither internalization nor body surveillance tendencies were predictive of inter-individual variability in the relationship between company and body dissatisfaction. Perhaps within the context of body shame and elevated BMI, these variables simply do not offer unique predictive value for the relationship between these state variables. Body shame was also predictive of the relationship between positive mood and state body dissatisfaction. Specifically, it was found that the relationship between positive mood and (reduced) body dissatisfaction was weaker for individuals with heightened body shame, suggesting that trait components of body dissatisfaction may adversely affect the positive influence of contextual cues on state body dissatisfaction. Such an effect is consistent
Table 3 MLM results indicating the role of trait affect and appearance-related variables for explaining variability in the strength of the relationships between company and state body dissatisfaction, and mood and state body dissatisfaction. Between individual effects
Predictors
Unadjusted Coefficient
Company
Positive mood
Negative mood
* **
p < .05. p < .01.
Body shame Body surveillance Internalization Body shame Body surveillance Positive affect Internalization Body shame Body surveillance Negative affect Internalization
0.19 0.12 −0.01 −0.02 0.01 0.01 0.01 0.01 −0.01 0.01 0.01
Adjusted for BMI SE 0.07 0.13 0.01 0.01 0.01 0.01 0.01 0.01 0.02 0.01 0.01
t **
2.73 0.93 −0.77 −1.66* 0.28 1.89* 0.32 0.76 −0.55 1.42 0.06
Coefficient
SE
t
0.16 0.13 −0.01 −0.02 0.01 0.01 0.01 0.02 −0.01 0.02 −0.01
0.11 0.12 0.01 0.01 0.01 0.01 0.01 0.02 0.02 0.01 0.01
1.56 1.13 −0.27 −1.88* 0.42 1.92* 0.64 0.93 −0.70 1.49 −0.30
320
L.A. Colautti et al. / Body Image 8 (2011) 315–321
with prior research which has shown that positive appearancerelated comments may have adverse consequences for individuals with heightened trait body dissatisfaction (Herbozo & Thompson, 2006). Similarly, LePage and Crowther (2010) found that trait body dissatisfaction moderated the relationship between fitness/health motivations for exercise and state body dissatisfaction following an exercise session. While exercise for health reasons resulted in lower levels of state body dissatisfaction for individuals with minimal trait body dissatisfaction, the opposite pattern was observed for individuals with high trait body dissatisfaction. Taken together, these findings are consistent with the notion that trait body dissatisfaction and/or trait body shame are particularly strong determinants of state body dissatisfaction levels and that these trait variables may, in some way, block the positive effects of contextual cues from influencing one’s body dissatisfaction states (LePage & Crowther, 2010; Rudiger et al., 2007). Clearly, this is an area with potential clinical implications as present findings seem to suggest that the benefits of contextual factors such as exercise, presence of others, and positive mood states on how an individual feels about her/his body from moment to moment will only obtain once trait body image disturbances are treated. This potential hierarchy of effects warrants further empirical investigation. The present findings suggest that trait affect may also influence the strength of association between mood states and body dissatisfaction ratings. It was found that the effect of positive mood states on state body dissatisfaction was greatest for individuals with heightened trait positive affect. Although non-significant, there was also a trend for negative affect to predict inter-individual variation in the strength of association between negative mood and body dissatisfaction. That is, general satisfaction/dissatisfaction may be more likely to extend to body (dis)satisfaction for these individuals. These observations extend prior findings which show that mood induction can produce increases in body (dis)satisfaction (Baker, Williamson, & Sylve, 1995; Plies & Florin, 1992), and suggest that the extent of this effect may depend on trait affect. Limitations and Implications for Future Research A limitation of the current study is the potentially burdensome nature of the ESM component of the study. Whereas previous ESMbased studies have used fewer testing points per day (usually one or two, often outside of work hours), the PDA was set in the present study to signal at random intervals, six times per day in order to obtain data that more accurately reflects everyday fluctuations in body dissatisfaction. It is likely that this added burden contributed to non-compliance (the average response rate per participant was approximately 75%). However, it is important to note that although individuals differed in the number of PDA-based surveys they completed, this did not appear to diminish the integrity of obtained data. Completion rate was not reliably associated with age, BMI, or any of the trait measures used in the present study. Moreover, there was no indication of increased non-compliance towards the end of the 7-day testing period. Thus, for the present study at least, there is less concern that results may have been biased by noncompliance. Nevertheless, the need to balance data collection goals against burden of the ESM task for participants is an ongoing issue that warrants further investigation. Another potential limitation of the PDA component of this study is that the repeated questioning about mood and body dissatisfaction may have made participants more attentive to their appearance and mood than they otherwise would be. Given the novelty of the ESM approach, there is currently little knowledge on the effects of daily diaries on participants’ experiences and behaviors (Bolger, Davis, & Rafaeli, 2003), and it is possible that results obtained in the present study were influenced by study design. Attempts were made to reduce habituation to response sets by set-
ting the PDA to signal randomly instead of at fixed intervals (Bolger et al., 2003). Further research is needed to evaluate the efficacy of this and other approaches to reduce method-related confounds. Finally, dichotomization of the ‘presence of company’ variable may have impacted results obtained in the present study. Although this dichotomization was necessary due to low endorsement of several response options for this variable, reducing the variable to presence versus absence of others does not allow for evaluation of how type of company (e.g., partner versus stranger) affects state body dissatisfaction. Future research is therefore needed to probe the potential relationship between presence of company and body dissatisfaction. It may also be useful to evaluate how key characteristics of the companion (e.g., relative age, gender, suitability as a potential mate) affect this relationship. Despite these limitations, the present study demonstrates the need to jointly consider the influences of trait and contextual variables on state body dissatisfaction. The ESM approach may further develop theoretical understanding of how transient body image disturbances manifest and the implications of these disturbances for other problematic behaviors, such as disordered eating. This approach allows researchers to identify potential triggers in the environment that may promote negative or positive feelings about one’s appearance. Furthermore, by modeling interindividual variability in the relationship between context and state body dissatisfaction, researchers may develop more comprehensive models which identifies at-risk populations (based on trait variables) for the influence of context on body image. Present findings provide evidence to suggest that current mood state and presence of others influence state body dissatisfaction. It was also shown that trait variables influence the strength of the relationships between contextual cues and state body dissatisfaction. While positive mood was associated with reduced body dissatisfaction for the sample as a whole, this positive effect was less pronounced for individuals with heightened trait body shame. Similarly, presence of others exerted greatest influence on state body dissatisfaction for individuals with heightened body shame. Taken together, these findings suggest that improvements in state body satisfaction may be enhanced by treatment of trait-level body image disturbances. However, to more fully understand whether the distinctions between state and trait level body image disturbances has clinical utility, future research which evaluates the health implications of heightened state body dissatisfaction is needed.
References Anschutza, D., Engelsa, R., Van Leeuwea, J., & van Strie, T. (2009). Watching your weight? The relations between watching soaps and music television and body dissatisfaction and restrained eating in young girls. Psychology and Health, 24, 1035–1050. Baker, J. D., Williamson, D. A., & Sylve, C. (1995). Body image disturbance: Memory bias, and body dysphoria: Effects of negative mood induction. Behavior Therapy, 26, 747–759. Banfield, S. S., & McCabe, M. P. (2002). An evaluation of the construct of body image. Adolescence, 37, 373–393. Beal, D. J., & Weiss, H. M. (2003). Methods of ecological momentary assessment in organizational research. Organizational Research Methods, 6, 440–464. Blore, J. D. (2008). Subjective wellbeing: An assessment of competing theories. Unpublished manuscript, Deakin University, Melbourne, Australia. Retrieved on September 22, 2010, from Deakin University, Australian Centre of Quality of Life Website: http://www.deakin.edu.au/research/acqol/publications/membertheses.php. Bolger, N., Davis, A., & Rafaeli, E. (2003). Diary methods: Capturing life as it is lived. Annual Review of Psychology, 54, 579–616. Cash, T. F., Fleming, E. C., Alindogan, J., Steadman, L., & Whitehead, A. (2002). Beyond body image as a trait: The development and validation of the Body Image States Scale. Eating Disorders, 10, 103–113. Cook-Cottone, C., & Phelps, L. A. (2003). Body dissatisfaction in college women: Identification of risk and protective factors to guide college counselling practices. Journal of College Counselling, 6, 80–89. Csikszentmihalyi, M., & Larson, R. (1987). Validity and reliability of the experiencesampling method. Journal of Nervous & Mental Disease, 175, 526–537.
L.A. Colautti et al. / Body Image 8 (2011) 315–321 Davern, M. T. (2004). Subjective wellbeing as an affective construct. Unpublished manuscript, Deakin University, Melbourne. Retrieved on September 22, 2010, from Deakin University, Australian Centre of Quality of Life Website: http://www.deakin.edu.au/research/acqol/publications/member-theses.php. Fredrickson, B. L., & Roberts, T. (1997). Objectification theory: Toward understanding women’s lived experiences and mental health risks. Psychology of Women Quarterly, 21, 173–206. Haimovitz, D., Lansky, L. M., & O’Reilly, P. (1993). Fluctuations in body image across situations. International Journal of Eating Disorders, 13, 77–84. Hausenblas, H. A., & Fallon, E. A. (2006). Exercise and body image: A meta analysis. Psychology and Health, 21, 33–47. Herbozo, S., & Thompson, J. K. (2006). Appearance-related commentary, body image, and self-esteem: Does the distress associated with the commentary matter? Body Image, 3, 255–262. Hox, J. (2002). Multilevel analysis: Techniques and applications. Mahwah, NJ: Erlbaum. Jackson, D. L. (2010). Reporting results of latent growth modelling and multilevel modelling analyses: Some recommendations for rehabilitation psychology. Rehabilitation Psychology, 55, 272–285. Krones, P., Stice, E., Batres, C., & Orjada, K. (2004). In vivo social comparison to a thin-ideal peer promotes body dissatisfaction: A randomized experiment. International Journal of Eating Disorders, 38, 134–142. Lattimore, P., & Hutchinson, R. (2010). Perceived calorie intake and state body image satisfaction in women attempting weight loss: A preliminary investigation. Body Image, 7, 15–21. LePage, M. L., & Crowther, J. H. (2010). The effects of exercise on body satisfaction and affect. Body Image, 7, 124–130. Melnyk, S. E., Cash, T. F., & Janda, L. H. (2004). Body image ups and downs: Prediction of intra-individual level and variability of women’s daily body image experiences. Body Image, 1, 225–235. McKinley, N. M., & Hyde, J. S. (1996). The Objectified Body Consciousness Scale: Development and validation. Psychology of Women Quarterly, 20, 181–215. Nesselroade, J. R., & Ram, N. (2004). Studying intraindividual variability: What we have learned that will help us understand lives in context. Research in Human Development, 1, 9–29. Paquette, M., & Raine, K. (2004). Sociocultural context of women’s body image. Social Science and Medicine, 59, 1047–1058. Plies, K., & Florin, I. (1992). Effects of negative mood induction on the body image of restrained eaters. Psychology and Health, 7, 235–242.
321
Rudiger, J. A., Cash, T. F., Roehrig, M., & Thompson, J. K. (2007). Day-to-day body image states: Prospective predictors of intra-individual level and variability. Body Image, 4, 1–9. Stice, E. (2002). Risk and maintenance factors for eating pathology: A meta-analytic review. Psychological Bulletin, 128, 825–848. Tabachnick, B. G., & Fidel, L. S. (2006). Using multivariate statistics (5th ed.). Boston: Allyn and Bacon. Thompson, J. K., Heinberg, L. J., Altabe, M., & Tantleff-Dunn, S. (1999). An introduction to the concept of body image disturbance: History, definitions and descriptions. In J. Thompson, L. Heinberg, M. Altabe & S. Tantleff-Dunn (Eds.), Exacting beauty: Theory, assessment, and treatment of body image disturbance (pp. 3–15). Washington: American Psychological Association. Thompson, J. K. (2004). The mis(measurement) of body image: Ten strategies to improve assessment for applied and research purposes. Body Image, 1, 7–14. Thompson, J. K., van den Berg, P., Roehrig, M., Guarda, A., & Heinberg, L. (2004). The Sociocultural Attitudes Towards Appearance Scale-3 (SATAQ-3): Development and validation. International Journal of Eating Disorders, 35, 293–304. Tiggemann, M. (2001). Person × Situation interactions in body dissatisfaction. International Journal of Eating Disorders, 29, 65–70. Tissot, A. M., & Crowther, J. H. (2008). Self-oriented and socially prescribed perfectionism: Risk factors within an integrative model for bulimic symptomatology. Journal of Social & Clinical Psychology, 27, 734–755. van den Berg, P., & Thompson, J. K. (2007). Self-schema and social comparison explanations of body dissatisfaction: A laboratory experiment. Body Image, 4, 29–38. Vocks, S., Hechler, T., Rohrig, S., & Legenbauer, T. (2009). Effects of a physical exercise session on state body image: The influence of pre-experimental body dissatisfaction and concerns about weight and shape. Psychology and Health, 24, 713–728. Vocks, S., Legenbauer, T., & Heil, A. (2007). Food intake affects state body image: Impact of restrained eating patterns and concerns about eating, weight and shape. Appetite, 49, 467–475. Weiss, H. M., Beal, D. J., Lucy, S. L., & MacDermid, S. M. (2004). Constructing EMA studies with PMAT: The Purdue Momentary Assessment Tool user’s manual. Software manual available at http://www.ruf.rice.edu/∼dbeal/PMAT.html Yik, M., Russell, J. A., & Feldman-Barrett, L. (1999). Structure of self-reported current affect: Integration and beyond. Journal of Personality and Social Psychology, 77, 600–619.
Contents lists available at ScienceDirect
Body Image journal homepage: www.elsevier.com/locate/bodyimage
Accounting for fluctuations in body dissatisfaction Lauren A. Colautti, Matthew Fuller-Tyszkiewicz ∗ , Helen Skouteris, Marita McCabe, Stephen Blackburn, Elise Wyett School of Psychology, Deakin University, 221 Burwood Highway, Burwood, Victoria 3125, Australia
a r t i c l e
i n f o
a b s t r a c t
Article history: Received 11 January 2011 Received in revised form 27 May 2011 Accepted 3 July 2011 Keywords: Body dissatisfaction State body image Context Mood states Experience sampling method
The present study evaluated whether the strength of relationship between contextual cues (presence of company and mood) and state body dissatisfaction varied as a function of individual differences in key trait measures (body shame, body surveillance tendencies, internalization of appearance standards, and trait affect) which have been linked to trait body dissatisfaction. Fifty-five undergraduate women completed a questionnaire containing the trait-based measures and then carried a Personal Digital Assistant (PDA) for a 7-day period. The PDA prompted participants six times daily to self-report their current mood and state body dissatisfaction. Multi-level modeling revealed that individual differences in body shame predicted inter-individual variability in the strength of the relationships between presence of company and state body dissatisfaction, and positive mood and state body dissatisfaction. Trait positive affect also explained variance in the positive mood state-body dissatisfaction relationship. The implications of the findings for prevention of body image disturbances are discussed. Crown Copyright © 2011 Published by Elsevier Ltd. All rights reserved.
Introduction Body image is a multidimensional construct defined as an individual’s perception of, and attitudes towards, his or her body and appearance (Cash, Fleming, Alindogan, Steadman, & Whitehead, 2002). It encompasses a range of cognitive, affective, and perceptual phenomena (Banfield & McCabe, 2002; Thompson, Heinberg, Altabe, & Tantleff-Dunn, 1999). Body dissatisfaction is one aspect of body image, relating to an individual’s degree of dissatisfaction with particular parts of the body (Cook-Cottone & Phelps, 2003). It is argued that body dissatisfaction consists of both state and trait aspects (Cash et al., 2002; Thompson, 2004; Vocks, Hechler, Rohrig, & Legenbauer, 2009). Trait body dissatisfaction is considered to be a stable and unchanging characteristic that is transferable across a wide range of contexts. Trait body dissatisfaction has been linked to personality traits (e.g., perfectionism, trait affect, and selfesteem), appearance-related factors (e.g., shame felt about one’s appearance, body surveillance, internalization of the thin ideal), and socio-cultural influences (e.g., media, interpersonal relations) (Anschutza, Engelsa, Van Leeuwea, & van Strie, 2009; Stice, 2002; Tissot & Crowther, 2008; van den Berg & Thompson, 2007). In contrast, state body dissatisfaction is thought to fluctuate on a moment-by-moment basis, and these fluctuations have
∗ Corresponding author. Tel.: +61 3 9251 7344; fax: +61 3 9244 6858. E-mail address: [email protected] (M. Fuller-Tyszkiewicz).
been associated with contextual factors (e.g., social settings, exercise, and body exposing situations, such as looking in a mirror), current mood state, and individual differences in personality dispositions and disordered eating symptomatology (Hausenblas & Fallon, 2006; LePage & Crowther, 2010; Melnyk, Cash, & Janda, 2004; Paquette & Raine, 2004; Rudiger, Cash, Roehrig, & Thompson, 2007). The relationship between body dissatisfaction and its purported contributors has most commonly been measured using trait measurements and within cross-sectional designs. However, as this approach is insensitive to the state-dependent aspects of body dissatisfaction ratings, the results of these studies are likely to conflate state- and trait-based components of body dissatisfaction (Melnyk et al., 2004). State-based approaches to measure body dissatisfaction have typically involved the use of simple pre–post induction based methods to elicit state-like changes in body dissatisfaction (Cash et al., 2002; Haimovitz, Lansky, & O’Reilly, 1993; Tiggemann, 2001; Vocks, Legenbauer, & Heil, 2007). For example, Krones, Stice, Batres, and Orjada (2004) compared changes in self-reported body dissatisfaction for women who interacted with a confederate who either conformed to the thin-ideal or who had body dimensions representative of the average woman in the general population. They found that body dissatisfaction increased significantly for participants exposed to the thin-ideal confederate but not for the normal weight confederate. While induction-based studies demonstrate the malleability of body dissatisfaction, they are limited in several ways. First, such methods may have limited ecological validity as the event that
1740-1445/$ – see front matter. Crown Copyright © 2011 Published by Elsevier Ltd. All rights reserved. doi:10.1016/j.bodyim.2011.07.001
316
L.A. Colautti et al. / Body Image 8 (2011) 315–321
induces changes in body dissatisfaction may be more (or less) likely than everyday situations to elicit changes in body dissatisfaction. This may inaccurately represent the extent to which body dissatisfaction levels actually vary from moment to moment (i.e., the level of intra-individual variability). Second, pre–post assessments only measure two time points and, as such, do not allow for measurement error in these ratings. More extensive, repeated measurement is preferable as atypical and outlying responses can then be averaged out. Moreover, the averaged effect is likely to be more representative of the actual effect size than a single instance of testing (Nesselroade & Ram, 2004). Third, despite evidence of within-group variability in pre–post change in body dissatisfaction ratings in these induction-based studies, the possibility that the relationship between state variables may vary as a function of individual difference factors (inter-individual variability) has been largely ignored. Recent methodological advances have led to techniques of data collection that are cost effective, less invasive than experimental designs, and suitable for capturing moment-by-moment data (Beal & Weiss, 2003). The experience sampling method (ESM; Csikszentmihalyi & Larson, 1987), which involves repeated measurement of behaviors, cognitions, and emotions in situ, permits researchers to address more sophisticated questions based on the intra-individual variability of these state variables and the dynamic relationship between state- and trait-based variables than has previously been possible. This has led to empirical evaluations of whether the predictors of trait-based body dissatisfaction may also be predictive of state body dissatisfaction. For instance, it has been shown that personality factors (notably, perfectionistic self-presentation) and appearance-relevant trait measures (psychological investment in appearance, disturbed eating attitudes, and appearance-fixing coping strategies) are predictive of intra-individual variability in state body dissatisfaction ratings (Lattimore & Hutchinson, 2010; Melnyk et al., 2004; Rudiger et al., 2007). While these studies demonstrate that trait measures are predictive of variability in state body dissatisfaction, it is likely that contextual (or state-dependent) cues may also provide opportunities for an individual to feel more (or less) dissatisfied with her/his appearance. To date, only two studies have used ESM to evaluate the extent to which contextual factors influence state body dissatisfaction. Lattimore and Hutchinson (2010) monitored postmeal mood, body dissatisfaction, and perceived caloric intake in a group of women enrolled in a weight loss program. They found that intra-individual variability in state body dissatisfaction was linked with intra-individual variability in both mood and perceived caloric intake. LePage and Crowther (2010) took a different approach by examining the extent to which individual differences at the trait level could account for inter-individual variation in the strength of relationship between instances of exercise and state body dissatisfaction. They found that exercise was associated with decreases in state body dissatisfaction and that the relationship between motives for exercise and state body dissatisfaction was moderated by trait body dissatisfaction. While the results of these state-based body dissatisfaction studies demonstrate that contextual variables and a number of key trait variables are related to variability in state body dissatisfaction, there are still many important questions that need to be addressed. First, the role that social context (company) plays in fluctuations in state body dissatisfaction ratings throughout a person’s day-to-day life has yet to be empirically evaluated. This is despite theoretical arguments that state that the presence of others provides an opportunity for body-based comparisons with others and the possibility to have one’s body visually scrutinized; either of which could result in increased body dissatisfaction (Fredrickson & Roberts, 1997). Body-based comparisons are most likely to occur in the company
of same-sex individuals of similar age (such as friends), although appearance concerns may also be salient when in the company of one’s romantic partner or a potential mate. Second, there has been limited research into the influence of mood state on fluctuations in body dissatisfaction. Although Lattimore and Hutchinson (2010) found an association between intra-individual variability in mood and state body dissatisfaction, this finding does not ensure that changes in body dissatisfaction and mood occurred synchronously. Third, the possibility that interindividual differences exist in the strength of association between context and body dissatisfaction has been neglected. Trait-based variables which have been linked to the etiology and maintenance of trait body dissatisfaction may also make individuals more susceptible to increases in state body dissatisfaction in the presence of others and/or when one’s mood state becomes increasingly negative. For instance, it is possible that the influence of company on state body dissatisfaction is greatest for individuals with traitlevel body image concerns, such as internalization of appearance standards, body shame, and the tendency to engage in bodysurveillance behaviors. The present study used an ESM approach to determine whether: (a) state body dissatisfaction can be predicted by current mood state and/or presence of others and (b) inter-individual differences in the strength of the relationship between state body dissatisfaction and these contextual factors are influenced by key trait measures (body shame, body surveillance, negative or positive trait affect, internalization of the thin ideal tendencies) that, as noted above, have been shown to relate to trait body dissatisfaction. ESM was chosen in preference to a simple induction design in order to capture intra-individual variability in body dissatisfaction states in a more ecologically valid way. By re-sampling the relationship between context and body dissatisfaction six times daily for a period of 1 week, it was expected that estimates would better reflect daily experience and be less influenced by atypical responses and/or extraneous influences (Beal & Weiss, 2003). Based on previous findings, it was hypothesised that decreased positive mood, increased negative mood, and presence of company would be predictive of higher state body dissatisfaction. It was additionally predicted that a stronger relationship between state body dissatisfaction and presence of company would be evidenced for women with higher body shame, body surveillance tendencies, and who were more strongly invested in the sociocultural value placed on physical appearance (internalization of the thin ideal). To further explore associations found previously between mood states and body dissatisfaction (LePage & Crowther, 2010), trait negative and positive affects were also included (in addition to the appearance-based measures) to explain interindividual variation in the relationship between mood states and state body dissatisfaction. It was anticipated that trait negative affect would moderate the relationship between negative mood and body dissatisfaction, whereas trait positive affect would moderate the relationship between positive mood and state body dissatisfaction.
Method Participants A total of 57 women were recruited from advertisements made during undergraduate lectures and tutorials offered at a large metropolitan university in Melbourne, Australia. Two women were excluded due to incomplete data. The 55 remaining participants ranged in age from 19 to 51 years with a mean age of 29.69 (SD = 9.62) years. Self-reported body mass indices (BMI = kg/m2 ) ranged from 17.96 to 36.29 (M = 24.03, SD = 4.38).
L.A. Colautti et al. / Body Image 8 (2011) 315–321
Materials Pre-test measures. Demographics. This questionnaire obtained information concerning the participants’ age, height and weight. Trait affect. Five items from the Trait Affect Scale (TAS; Blore, 2008) were used to measure positive and negative trait affect. Prior research into the structure of affect has shown that these five items adequately capture these two affective traits (Davern, 2004; Yik, Russell, & Feldman-Barrett, 1999). The positive items ask participants to rate how happy, satisfied, and content they generally feel, whereas the negative affect items ask how unhappy and discontent participants feel in general. Items are presented with 5-point Likert-style response scales, ranging from 1 (Never) to 5 (Always). Item responses are summed, resulting in separate scores for each of trait positive and negative affect. Blore (2008) demonstrated the convergent validity of this measure. Cronbach’s alpha estimates in the present study were .78 for positive affect and .82 for negative affect. Internalization of appearance standards. The nine-item internalization-general subscale of the Sociocultural Attitudes Towards Appearance Questionnaire-Version 3 (SATAQ-3; Thompson, van den Berg, Roehrig, Guarda, & Heinberg, 2004) was used to assess the extent to which participants endorse and accept cultural ideals of physical appearance (e.g., “I would like my body to look like the people who are in the movies”). Items were rated on a 5-point Likert scale from 1 (Definitely Disagree) to 5 (Definitely Agree). This subscale has been shown to be reliable and valid (Thompson et al., 2004). Cronbach’s alpha in the current study for Internalization-general scores was .93. Objectified body consciousness. The body surveillance and body shame subscales of the Objectified Body Consciousness Scale (OBCS; McKinley & Hyde, 1996) were used to measure the extent to which participants think of their bodies in terms of appearance (e.g., “I often worry about whether the clothes I am wearing make me look good”) and the extent to which the participant believed they were a bad person for not attaining cultural standards of appearance (e.g., “When I can’t control my weight, I feel like something must be wrong with me”), respectively. Participants responded on a 7-point Likert scale ranging from 1 (Disagree Strongly) to 7 (Agree Strongly) with higher scores reflecting greater body surveillance or body shame. Scale reliability and validity has been demonstrated (McKinley & Hyde, 1996). Cronbach’s alpha in the present study was .84 for body surveillance scores and .81 for body shame scores. PDA-based measures. State body dissatisfaction. The Body Image States Scale (BISS; Cash et al., 2002) is a six-item measure of current body image (dis)satisfaction at a given time point. Participants were asked to rate how satisfied they felt “right now at this moment” in regards to their (a) physical appearance, (b) body size and shape, (c) weight, and (d) physical attractiveness. Participants were additionally asked to rate how their current feelings regarding their looks relative to (e) how they typically felt, and (f) how the average person looks. Items were rated on 9-point Likert scale ranging from 1 (Extremely Dissatisfied) to 9 (Extremely Satisfied). Item scores were averaged and then subtracted from the maximum possible score (9) so that high scores reflected greater body dissatisfaction. Reliability and internal consistency of the BISS has been reported (Rudiger et al., 2007). Cronbach’s alpha in the present study was .93.
317
Mood states. The five items of the TAS (see above) were modified so that participants indicated how they ‘felt right’ now instead of ‘in general’. Items were rated on an 11-point scale ranging from 0 (Not at all) to 10 (Extremely). This measure has been shown to be sensitive to moment-by-moment fluctuations in mood (Blore, 2008). In the present study, Cronbach’s alpha was .94 for both positive and negative mood. Social context. Participants were asked to indicate who they were with at the time of completing the survey on the PDA by selecting one or more of the following options: (a) alone, (b) with partner, (c) with friends, (d) with work colleagues, or (e) with others (unspecified). Although the authors intended to compare state body dissatisfaction ratings across each of the five social contexts, few participants endorsed all of the response options throughout the duration of the PDA testing period. Accordingly, response options two to five were collapsed into a single category representing ‘with others’ and were compared against the ‘alone’ response option. The limitations of this re-categorization are covered in Discussion section. Daily diary apparatus. The PDAs used in the present study were Palm Z22s System running the Palm 5.4 Operating System (©Palm Inc., USA). Questionnaires were constructed using the Purdue Momentary Assessment Tool (PMAT; Weiss et al., 2004). Procedure Ethics approval for the study was provided by the university’s Human Research Ethics Committee. The study was advertised in undergraduate lectures and tutorials. Interested participants met with a member of the research team, at which time they received the Plain Language Statement and Consent Form. After informed consent was obtained, participants completed the pre-test questionnaires listed above. Participants were then given a PDA, charger kit, and instructions for the study. Participants were asked to complete a PDA survey to practice and familiarize themselves with the task, and were encouraged to ask questions if they felt there were any ambiguities with the process or items. Participants carried a PDA with them for a 7-day period. The PDA emitted an audible tone six times per day at random intervals between 10:00 a.m. and 8:00 p.m. This tone prompted participants to complete the questionnaires on the PDA. The PDAs were programmed to signal at intervals of no less than 90 min in order to allow for sampling across the entire day. Moreover, for each time interval, participants had a 15-min window in which to complete the survey before it was counted as missing data by the PDA. Participants were instructed to complete the momentary assessments whenever the PDA signaled, provided it was safe or appropriate to do so. Driving a car or being in an important meeting were given as examples of times where it may be inappropriate to respond to the PDA. The order of presentation of items was held constant across all testing intervals. After 7 days, participants returned the PDA to the research team, where they were debriefed and received a $10 gift voucher as an honorarium. Data Analytic Strategy Given that participants provided multiple PDA responses over time, the data were hierarchical in nature, with responses to PDA surveys (Level 1 data) nested within individuals (Level 2). As a consequence of ignoring the clustered nature of repeated measures data (and violating the assumption of independence of errors), traditional approaches such as multiple regression produce prediction models which under-estimate standard error terms, in turn leading
318
L.A. Colautti et al. / Body Image 8 (2011) 315–321
to Type I error inflation for estimates of the relative contributions of IVs for predicting the DV (Hox, 2002). Repeated measures designs (e.g., ANOVA and MANOVA) are equally problematic as they aggregate data to Level 2 (i.e., inter-individual differences), effectively ignoring differences at Level 1 (intra-individual variability). The present study utilized multilevel modeling (MLM), which is a technique that overcomes any potential violations of the assumptions of independence that can occur when data is clustered within groups (Jackson, 2010). MLM treats each participant as a separate group and thus provides estimates of intra-individual and inter-individual variability in the relationship between state body dissatisfaction and the contextual factors. In the present study, Level 1 variables consisted of state body dissatisfaction, negative mood, positive mood, and company estimates as based on the PDA survey (N = 1530). Level 2 represented estimates of trait variables (body shame, body surveillance, internalization tendencies, and negative and positive trait affect) for each of the participants in the data set (N = 55). In order to assess the relationship between contextual variables and state body dissatisfaction, three aspects of the MLM output were relevant: (a) the Level 1 fixed components, which assessed the strength of the relationship between contextual factors and state body dissatisfaction when averaged across all individuals (i.e., ignoring clustering effects); (b) the Level 1 variance components which assessed the extent to which the strength of these Level 1 relationships varied from person to person (i.e., inter-individual variability); and (c) the Level 2 intercept-and-slopes-as-outcomes model which allowed for evaluation of the extent to which the Level 2 trait variables (internalization tendencies, body shame, body surveillance, and negative and positive trait affect) explained interindividual variability in the strength of the relationship between Level 1 variables. Because BMI has been shown to correlate with these trait variables (Stice, 2002), the Level 2 intercept-and-slopesas-outcomes model was evaluated twice; once with unadjusted estimates of effects, and a second run with effects adjusted for individual differences in BMI.
Results Data Preparation While missing data does not present statistical problems for multi-level modeling, the researcher must decide how much missing data to allow and determine whether the pattern of missingness has implications for interpretability of findings (Hox, 2002; Tabachnick & Fidel, 2006). Two participants were eliminated from the study as they each answered only three of the possible 42 PDA surveys. No other participant missed more than 50% of the surveys and, given this still allowed for at least 21 estimates of state body dissatisfaction across the week, the decision was made to retain all participants who met this 50% cut-off. While this cut-off of 50% may appear low, it must be noted that the number of times the survey was answered throughout the testing period (M = 30.75, SD = 8.66) was still more than those provided in previous daily diary studies of state body dissatisfaction. A related concern is that procedural non-compliance may potentially undermine the generalizability of findings. Therefore, several preliminary analyses were conducted to evaluate whether the likelihood of missing PDA data was related to day of the week or individual difference variables (age, BMI, body dissatisfaction, body shame, internalization, positive affect, and negative affect). A one-way ANOVA revealed that the effect of day of the week on compliance rates was non-significant, F(6, 49) = 1.24, p > .05. Furthermore, bivariate correlations between compliance rates for PDA
surveys and the trait measures used in this study were all nonsignificant (all ps > .05). Prior to the main analysis, Phase 1 data were screened to ensure that they met the assumptions of multivariate analysis (Tabachnick & Fidel, 2006). Less than 2% missing data was found overall, and dealt with using expectation maximization. There were no outliers or evidence of non-normality in any variables. Table 1 presents the means, standard deviations, and possible range of scores for the Level 1 and Level 2 variables in the current study. MLM Analyses When initially modeled as fixed effects, positive and negative mood each made significant contributions to prediction of state body dissatisfaction, t(54) = −15.50, p < .001 and t(54) = 2.06, p < .05, respectively. Presence of company was not reliably associated with state-based body dissatisfaction ratings, t(54) = 0.10, p = .93. However, as shown in Table 2, significant inter-individual variation in the strength of relationships between body dissatisfaction and the three contextual variables was observed once the Level 1 relationships were allowed to vary. Thus, the non-significant fixed effect for the company-body dissatisfaction relationship is not representative of all individuals in the data set. Likewise, that the other two Level 1 relationships differ also suggests that the fixed effect for mood states and body dissatisfaction may under-estimate the true strength of relationship for a subset of individuals in this sample. Finally, the Level 2 intercepts-and-slopes-as-outcomes model showed the extent to which individual difference factors (trait affect, internalization, body shame, and body surveillance) explained variance in the relationships between Level 1 variables. As revealed in Table 3, inter-individual variation in the relationship between company and state body dissatisfaction was significantly predicted by body shame in the unadjusted model, t(50) = 2.73, p < .01. This effect reduced to non-significance after controlling for individual differences in BMI, t(49) = 1.56, p > .05. Thus, although individuals with heightened body shame were most likely to report body dissatisfaction in the company of others, it seems that this effect may, in part, be attributable to the influence of BMI. Individual differences in body surveillance tendencies and level of acceptance of appearance standards were not predictive of inter-individual variations in the relationship between company and body dissatisfaction in either the adjusted or unadjusted models. Inter-individual variation in the relationship between positive mood and state body dissatisfaction was predicted by body shame and trait positive affect, t(50) = −1.66, p < .05 and t(50) = 1.89, p < .05, respectively. Heightened body shame served to weaken the relationship between positive mood and body dissatisfaction whereas trait positive affect strengthened this association. These relationships remained significant after controlling for BMI. None of the trait variables reliably accounted for variance in the strength of the relationship between negative mood and body dissatisfaction, although there was a trend for trait negative affect to be predictive of a stronger relationship between negative mood and body dissatisfaction, t(50) = 1.42, p = .07. Discussion To date, there has been a dearth of research examining the predictors of state body dissatisfaction in everyday life. In particular, there has been insufficient attention given to the possibility that proximal factors (such as mood and social setting) may interact with trait factors (such as appearance-related variables and personality dispositions) to produce momentary fluctuations in body dissatisfaction. Thus, the aim of the present study was to evaluate: (a) the extent to which current mood states and presence of com-
L.A. Colautti et al. / Body Image 8 (2011) 315–321
319
Table 1 Means, standard deviations, and possible score ranges for Level 1 and Level 2 variables. Level
Variable
Mean
SD
Possible range
Level 1
Positive mood Negative mood Body dissatisfaction Internalization Body shame Body surveillance Negative affect Positive affect
17.95 5.11 4.61 28.05 3.42 4.76 5.11 11.11
5.55 4.44 1.73 8.85 1.18 0.97 1.21 1.64
0–30 0–20 1–9 9–45 1–7 1–7 2–10 3–15
Level 2
Notes: The Level 1 variables are state measures, whereas the Level 2 variables are trait-based measures. Table 2 Variance components between individuals for the Level 1 random-intercept-and-slopes model predicting state body dissatisfaction. Variance components between individuals
Variance estimate
SD
df
2
Positive mood Negative mood Company
0.01 0.01 0.39
0.08 0.09 0.62
52 52 52
168.78*** 113.28*** 159.63***
***
p < .001.
pany were predictive of state body dissatisfaction ratings and (b) whether trait affect and key body image measures (internalization of the thin ideal, body shame, and body surveillance tendencies) accounted for inter-individual variation in the strength of relationship between contextual factors and state body dissatisfaction. Present findings supported the hypothesis that contextual factors influence body dissatisfaction ratings. While positive and negative mood, but not company, were predictive of state body dissatisfaction as fixed effects (i.e., when averaged across all participants), there was considerable inter-individual variability in the strength of association between body dissatisfaction and these contextual factors. Thus, even for the non-significant relationship between company and body dissatisfaction, the fixed effect underestimated the strength of this relationship for some participants in the present study. This latter finding with the random effects model illustrates the utility of MLM over more traditional analytic approaches to repeated measures designs (such as ANOVA and multiple regression) which ignore inter-individual variability and, in so doing, can lead to inaccurate or misleading conclusions. Reliance upon these traditional analytic strategies for repeatedmeasure designs may, in part, explain why the influence of mood states on body dissatisfaction has received mixed support in simple pre–post induction designs (e.g., Cash et al., 2002; Haimovitz et al., 1993; Tiggemann, 2001; Vocks et al., 2007). The finding that presence of company is predictive of state body dissatisfaction ratings for at least some individuals is consistent with prior theoretical speculation and empirical literature.
Fredrickson and Roberts (1997) argued that the presence of others (whether they are strangers, friends, partner, etc.) provides an opportunity for one’s body to be appraised. For some individuals (particularly those with negative body image, elevated BMI, and who view appearance as important for their sense of self-worth), the potential for public scrutiny can result in feelings of anxiety and dissatisfaction towards their appearance. In the present study, the relationship between presence of company and state body dissatisfaction was particularly strong for individuals with heightened body shame. Importantly, when individual differences in BMI were controlled for, the influence of body shame on the company-state body dissatisfaction relationship reduced to non-significance. Thus, the influence that presence of others exerts on state body dissatisfaction may derive from one’s awareness or perception of her/his actual body size. However, neither internalization nor body surveillance tendencies were predictive of inter-individual variability in the relationship between company and body dissatisfaction. Perhaps within the context of body shame and elevated BMI, these variables simply do not offer unique predictive value for the relationship between these state variables. Body shame was also predictive of the relationship between positive mood and state body dissatisfaction. Specifically, it was found that the relationship between positive mood and (reduced) body dissatisfaction was weaker for individuals with heightened body shame, suggesting that trait components of body dissatisfaction may adversely affect the positive influence of contextual cues on state body dissatisfaction. Such an effect is consistent
Table 3 MLM results indicating the role of trait affect and appearance-related variables for explaining variability in the strength of the relationships between company and state body dissatisfaction, and mood and state body dissatisfaction. Between individual effects
Predictors
Unadjusted Coefficient
Company
Positive mood
Negative mood
* **
p < .05. p < .01.
Body shame Body surveillance Internalization Body shame Body surveillance Positive affect Internalization Body shame Body surveillance Negative affect Internalization
0.19 0.12 −0.01 −0.02 0.01 0.01 0.01 0.01 −0.01 0.01 0.01
Adjusted for BMI SE 0.07 0.13 0.01 0.01 0.01 0.01 0.01 0.01 0.02 0.01 0.01
t **
2.73 0.93 −0.77 −1.66* 0.28 1.89* 0.32 0.76 −0.55 1.42 0.06
Coefficient
SE
t
0.16 0.13 −0.01 −0.02 0.01 0.01 0.01 0.02 −0.01 0.02 −0.01
0.11 0.12 0.01 0.01 0.01 0.01 0.01 0.02 0.02 0.01 0.01
1.56 1.13 −0.27 −1.88* 0.42 1.92* 0.64 0.93 −0.70 1.49 −0.30
320
L.A. Colautti et al. / Body Image 8 (2011) 315–321
with prior research which has shown that positive appearancerelated comments may have adverse consequences for individuals with heightened trait body dissatisfaction (Herbozo & Thompson, 2006). Similarly, LePage and Crowther (2010) found that trait body dissatisfaction moderated the relationship between fitness/health motivations for exercise and state body dissatisfaction following an exercise session. While exercise for health reasons resulted in lower levels of state body dissatisfaction for individuals with minimal trait body dissatisfaction, the opposite pattern was observed for individuals with high trait body dissatisfaction. Taken together, these findings are consistent with the notion that trait body dissatisfaction and/or trait body shame are particularly strong determinants of state body dissatisfaction levels and that these trait variables may, in some way, block the positive effects of contextual cues from influencing one’s body dissatisfaction states (LePage & Crowther, 2010; Rudiger et al., 2007). Clearly, this is an area with potential clinical implications as present findings seem to suggest that the benefits of contextual factors such as exercise, presence of others, and positive mood states on how an individual feels about her/his body from moment to moment will only obtain once trait body image disturbances are treated. This potential hierarchy of effects warrants further empirical investigation. The present findings suggest that trait affect may also influence the strength of association between mood states and body dissatisfaction ratings. It was found that the effect of positive mood states on state body dissatisfaction was greatest for individuals with heightened trait positive affect. Although non-significant, there was also a trend for negative affect to predict inter-individual variation in the strength of association between negative mood and body dissatisfaction. That is, general satisfaction/dissatisfaction may be more likely to extend to body (dis)satisfaction for these individuals. These observations extend prior findings which show that mood induction can produce increases in body (dis)satisfaction (Baker, Williamson, & Sylve, 1995; Plies & Florin, 1992), and suggest that the extent of this effect may depend on trait affect. Limitations and Implications for Future Research A limitation of the current study is the potentially burdensome nature of the ESM component of the study. Whereas previous ESMbased studies have used fewer testing points per day (usually one or two, often outside of work hours), the PDA was set in the present study to signal at random intervals, six times per day in order to obtain data that more accurately reflects everyday fluctuations in body dissatisfaction. It is likely that this added burden contributed to non-compliance (the average response rate per participant was approximately 75%). However, it is important to note that although individuals differed in the number of PDA-based surveys they completed, this did not appear to diminish the integrity of obtained data. Completion rate was not reliably associated with age, BMI, or any of the trait measures used in the present study. Moreover, there was no indication of increased non-compliance towards the end of the 7-day testing period. Thus, for the present study at least, there is less concern that results may have been biased by noncompliance. Nevertheless, the need to balance data collection goals against burden of the ESM task for participants is an ongoing issue that warrants further investigation. Another potential limitation of the PDA component of this study is that the repeated questioning about mood and body dissatisfaction may have made participants more attentive to their appearance and mood than they otherwise would be. Given the novelty of the ESM approach, there is currently little knowledge on the effects of daily diaries on participants’ experiences and behaviors (Bolger, Davis, & Rafaeli, 2003), and it is possible that results obtained in the present study were influenced by study design. Attempts were made to reduce habituation to response sets by set-
ting the PDA to signal randomly instead of at fixed intervals (Bolger et al., 2003). Further research is needed to evaluate the efficacy of this and other approaches to reduce method-related confounds. Finally, dichotomization of the ‘presence of company’ variable may have impacted results obtained in the present study. Although this dichotomization was necessary due to low endorsement of several response options for this variable, reducing the variable to presence versus absence of others does not allow for evaluation of how type of company (e.g., partner versus stranger) affects state body dissatisfaction. Future research is therefore needed to probe the potential relationship between presence of company and body dissatisfaction. It may also be useful to evaluate how key characteristics of the companion (e.g., relative age, gender, suitability as a potential mate) affect this relationship. Despite these limitations, the present study demonstrates the need to jointly consider the influences of trait and contextual variables on state body dissatisfaction. The ESM approach may further develop theoretical understanding of how transient body image disturbances manifest and the implications of these disturbances for other problematic behaviors, such as disordered eating. This approach allows researchers to identify potential triggers in the environment that may promote negative or positive feelings about one’s appearance. Furthermore, by modeling interindividual variability in the relationship between context and state body dissatisfaction, researchers may develop more comprehensive models which identifies at-risk populations (based on trait variables) for the influence of context on body image. Present findings provide evidence to suggest that current mood state and presence of others influence state body dissatisfaction. It was also shown that trait variables influence the strength of the relationships between contextual cues and state body dissatisfaction. While positive mood was associated with reduced body dissatisfaction for the sample as a whole, this positive effect was less pronounced for individuals with heightened trait body shame. Similarly, presence of others exerted greatest influence on state body dissatisfaction for individuals with heightened body shame. Taken together, these findings suggest that improvements in state body satisfaction may be enhanced by treatment of trait-level body image disturbances. However, to more fully understand whether the distinctions between state and trait level body image disturbances has clinical utility, future research which evaluates the health implications of heightened state body dissatisfaction is needed.
References Anschutza, D., Engelsa, R., Van Leeuwea, J., & van Strie, T. (2009). Watching your weight? The relations between watching soaps and music television and body dissatisfaction and restrained eating in young girls. Psychology and Health, 24, 1035–1050. Baker, J. D., Williamson, D. A., & Sylve, C. (1995). Body image disturbance: Memory bias, and body dysphoria: Effects of negative mood induction. Behavior Therapy, 26, 747–759. Banfield, S. S., & McCabe, M. P. (2002). An evaluation of the construct of body image. Adolescence, 37, 373–393. Beal, D. J., & Weiss, H. M. (2003). Methods of ecological momentary assessment in organizational research. Organizational Research Methods, 6, 440–464. Blore, J. D. (2008). Subjective wellbeing: An assessment of competing theories. Unpublished manuscript, Deakin University, Melbourne, Australia. Retrieved on September 22, 2010, from Deakin University, Australian Centre of Quality of Life Website: http://www.deakin.edu.au/research/acqol/publications/membertheses.php. Bolger, N., Davis, A., & Rafaeli, E. (2003). Diary methods: Capturing life as it is lived. Annual Review of Psychology, 54, 579–616. Cash, T. F., Fleming, E. C., Alindogan, J., Steadman, L., & Whitehead, A. (2002). Beyond body image as a trait: The development and validation of the Body Image States Scale. Eating Disorders, 10, 103–113. Cook-Cottone, C., & Phelps, L. A. (2003). Body dissatisfaction in college women: Identification of risk and protective factors to guide college counselling practices. Journal of College Counselling, 6, 80–89. Csikszentmihalyi, M., & Larson, R. (1987). Validity and reliability of the experiencesampling method. Journal of Nervous & Mental Disease, 175, 526–537.
L.A. Colautti et al. / Body Image 8 (2011) 315–321 Davern, M. T. (2004). Subjective wellbeing as an affective construct. Unpublished manuscript, Deakin University, Melbourne. Retrieved on September 22, 2010, from Deakin University, Australian Centre of Quality of Life Website: http://www.deakin.edu.au/research/acqol/publications/member-theses.php. Fredrickson, B. L., & Roberts, T. (1997). Objectification theory: Toward understanding women’s lived experiences and mental health risks. Psychology of Women Quarterly, 21, 173–206. Haimovitz, D., Lansky, L. M., & O’Reilly, P. (1993). Fluctuations in body image across situations. International Journal of Eating Disorders, 13, 77–84. Hausenblas, H. A., & Fallon, E. A. (2006). Exercise and body image: A meta analysis. Psychology and Health, 21, 33–47. Herbozo, S., & Thompson, J. K. (2006). Appearance-related commentary, body image, and self-esteem: Does the distress associated with the commentary matter? Body Image, 3, 255–262. Hox, J. (2002). Multilevel analysis: Techniques and applications. Mahwah, NJ: Erlbaum. Jackson, D. L. (2010). Reporting results of latent growth modelling and multilevel modelling analyses: Some recommendations for rehabilitation psychology. Rehabilitation Psychology, 55, 272–285. Krones, P., Stice, E., Batres, C., & Orjada, K. (2004). In vivo social comparison to a thin-ideal peer promotes body dissatisfaction: A randomized experiment. International Journal of Eating Disorders, 38, 134–142. Lattimore, P., & Hutchinson, R. (2010). Perceived calorie intake and state body image satisfaction in women attempting weight loss: A preliminary investigation. Body Image, 7, 15–21. LePage, M. L., & Crowther, J. H. (2010). The effects of exercise on body satisfaction and affect. Body Image, 7, 124–130. Melnyk, S. E., Cash, T. F., & Janda, L. H. (2004). Body image ups and downs: Prediction of intra-individual level and variability of women’s daily body image experiences. Body Image, 1, 225–235. McKinley, N. M., & Hyde, J. S. (1996). The Objectified Body Consciousness Scale: Development and validation. Psychology of Women Quarterly, 20, 181–215. Nesselroade, J. R., & Ram, N. (2004). Studying intraindividual variability: What we have learned that will help us understand lives in context. Research in Human Development, 1, 9–29. Paquette, M., & Raine, K. (2004). Sociocultural context of women’s body image. Social Science and Medicine, 59, 1047–1058. Plies, K., & Florin, I. (1992). Effects of negative mood induction on the body image of restrained eaters. Psychology and Health, 7, 235–242.
321
Rudiger, J. A., Cash, T. F., Roehrig, M., & Thompson, J. K. (2007). Day-to-day body image states: Prospective predictors of intra-individual level and variability. Body Image, 4, 1–9. Stice, E. (2002). Risk and maintenance factors for eating pathology: A meta-analytic review. Psychological Bulletin, 128, 825–848. Tabachnick, B. G., & Fidel, L. S. (2006). Using multivariate statistics (5th ed.). Boston: Allyn and Bacon. Thompson, J. K., Heinberg, L. J., Altabe, M., & Tantleff-Dunn, S. (1999). An introduction to the concept of body image disturbance: History, definitions and descriptions. In J. Thompson, L. Heinberg, M. Altabe & S. Tantleff-Dunn (Eds.), Exacting beauty: Theory, assessment, and treatment of body image disturbance (pp. 3–15). Washington: American Psychological Association. Thompson, J. K. (2004). The mis(measurement) of body image: Ten strategies to improve assessment for applied and research purposes. Body Image, 1, 7–14. Thompson, J. K., van den Berg, P., Roehrig, M., Guarda, A., & Heinberg, L. (2004). The Sociocultural Attitudes Towards Appearance Scale-3 (SATAQ-3): Development and validation. International Journal of Eating Disorders, 35, 293–304. Tiggemann, M. (2001). Person × Situation interactions in body dissatisfaction. International Journal of Eating Disorders, 29, 65–70. Tissot, A. M., & Crowther, J. H. (2008). Self-oriented and socially prescribed perfectionism: Risk factors within an integrative model for bulimic symptomatology. Journal of Social & Clinical Psychology, 27, 734–755. van den Berg, P., & Thompson, J. K. (2007). Self-schema and social comparison explanations of body dissatisfaction: A laboratory experiment. Body Image, 4, 29–38. Vocks, S., Hechler, T., Rohrig, S., & Legenbauer, T. (2009). Effects of a physical exercise session on state body image: The influence of pre-experimental body dissatisfaction and concerns about weight and shape. Psychology and Health, 24, 713–728. Vocks, S., Legenbauer, T., & Heil, A. (2007). Food intake affects state body image: Impact of restrained eating patterns and concerns about eating, weight and shape. Appetite, 49, 467–475. Weiss, H. M., Beal, D. J., Lucy, S. L., & MacDermid, S. M. (2004). Constructing EMA studies with PMAT: The Purdue Momentary Assessment Tool user’s manual. Software manual available at http://www.ruf.rice.edu/∼dbeal/PMAT.html Yik, M., Russell, J. A., & Feldman-Barrett, L. (1999). Structure of self-reported current affect: Integration and beyond. Journal of Personality and Social Psychology, 77, 600–619.