Q Aim of present study is to re-evaluate predictive value of body dissatisfaction for accounting changes in self-esteem Home, - The Relationship between Body Dissatisfaction and Self-esteem Body dissatisfaction refers to an individual’s degree of dissatisfaction with particular parts of the body (such as arms, thighs, etc.) and/or with global aspects of appearance (such as shape and weight) (Cook-Cottone& Phelps, 2003). Although body dissatisfaction is a key feature of eating disorders (American Psychological Association; APA, 2000), dissatisfaction with appearance is also common in non-clinical populations (Fuller-Tyszkiewicz, Ricciardelli, McCabe, & Lavelle, 2012). It is estimated that up to 50% of adolescents and adults are at least moderately dissatisfied with their appearance (Fuller-Tyszkiewicz, Krawczyk, Thompson, & Ricciardelli, 2012). These prevalence rates have motivated researchers to evaluate the causes and consequences of body dissatisfaction. One variable that is often linked with body dissatisfaction is self-esteem. Researchers have consistently demonstrated that individuals who are dissatisfied with their appearance also tend to have low self-esteem (McCauley, Mintz& Glenn, 1998; Miller & Downey, 1999; Webster &Tiggemann, 2003). This effect appears robust with respect to differences across samples in terms of age and gender (Wilcox, 1997). Whereas the majority of this research has relied on correlational designs, findings from longitudinal research (Tiggemann, 2005) suggests that the two variables may be causally related. Tiggemann (2005) demonstrated that changes over time in self-esteem can be predicted by body dissatisfaction, but that self-esteem is not predictive of changes in body dissatisfaction. Tiggemann’s findings are consistent with the contingencies of self-worth model (Crocker, Luhtanen, Cooper, &Bouvrette, 2003), which states that one’s global assessment of self-worth derives from satisfaction with various domains that s/he feels are important to her/his self-concept. Mellor, Fuller-Tyszkiewicz, McCabe and Ricciardelli (2010) identified several problems with Tiggemann’s (2005) findings. First, body dissatisfaction only accounted for 2.5% of the change in self-esteem scores over time, suggesting that other variables are also responsible for fluctuations in self-esteem.Second, by only evaluating one domain of satisfaction, Tiggemann’s (2005) findings may over-state the influence of body dissatisfaction on self-esteem. They argue that other domains of satisfaction, such as health, living standards, relationships, and achievements, are also relevant determinants of global self-worth. To date, a longitudinal examination of the relative influences of each domain on global self-esteem has not been conducted. The aim of the present study is to re-evaluate the predictive value of body dissatisfactionfor accounting for changes in self-esteem. It is predicted that body dissatisfaction will significantly predict self-esteem in a bivariate regression (Hypothesis 1), but that this effect will reduce to non-significance after controlling for other domains of satisfaction (Hypothesis 2). Method Participants The sample consisted of 350 women aged 18 to 40 years (M= 25.06, SD= 6.24) who were recruited from undergraduate courses offered at a large metropolitan university in Melbourne, Australia. Self-reported height and weight were used to calculate BMI (weight (kg)/height (m2)) for participants. National Institute of Health (NIH, 1998) guidelines were applied to interpret body mass index (BMI) scores: 8% of this sample were “underweight” (BMI<18.5), 75% were “normal weight” (BMI 18.5-24.9), 16% were “overweight” (BMI 25.0-29.9), and 1% were “obese” (BMI?30.0). For the overall sample, average BMI was 22.03 (SD=2.97). Materials Body dissatisfaction. Dissatisfaction with appearance was measured using the seven item Body Image Dissatisfaction subscale of the Body Change Inventory (BCI; Ricciardelli & McCabe, 2002). The first four items assess dissatisfaction with weight, shape, size, and muscles (e.g., ‘How dissatisfied are you with your weight?’). The remaining three items assess dissatisfaction with specific body regions: hips, thighs, and legs. Participants respond on a 5-point scale ranging from 0 (not at all) to 4 (extremely dissatisfied). Item scores are summed, such that higher total scores reflect greater body dissatisfaction. Researchers have demonstrated that the Body Image Dissatisfaction subscale exhibits strong psychometric properties (Ricciardelli& McCabe, 2002). In the current study, Cronbach’s alpha was .92. Self-esteem. Self-esteem was measured with the 10 item Rosenberg Self-Esteem Scale (RSES; Rosenberg, 1965). Five items are positively worded and five negatively worded. Item examples are ‘I feel that I have a number of good qualities’ and ‘At times I feel that I am no good at all’. Participants respond on a 5-point scale ranging from 0 (strongly disagree) to 4 (strongly agree). Self-esteem scores were obtained by adding all item responses, after the negatively worded item responses had been reverse-coded. High scores indicated a strong sense of self-esteem. In the present study, Time 2 estimates of self-esteem were subtracted from Time 1 scores to provide a measure of change in self-esteem across the testing period. Positive change scores indicate that a person’s self-esteem reduced, whereas a negative score reflects improvement in self-esteem at Time 2. Domains of satisfaction. Domains of satisfaction were measured using the items of the Personal Wellbeing Index (PWI; International Wellbeing Group [IWG], 2006). The PWI consists of seven items, each prefaced with ‘How satisfied are you with…’ The stems are: ‘your standard of living’, ‘your health’, ‘your achievements in life’, ‘your personal relationships’, ‘how safe you feel’, ‘feeling part of your community’, and ‘your future security’. Participants respond on an 11-point scale ranging from 0 (very dissatisfied) to 10 (very satisfied), with 5 (neutral) as a midpoint. Each item represents a separate domain as prior research has demonstrated their unique predictive value for key outcome variables, including overall wellbeing (IWG, 2006). A total wellbeing score was not required in this study; each item was used as a separate domain measure. Procedure Participants were recruited from Deakin University undergraduate courses. Announcements were made at the beginning of lectures to advertise the study to students. Flyers were left at the entrance to the lecture theatre and students who were interested in participating were invited to pick up a flyer at the end of class. The flyers contained a web link which students could use to access and fill out the online survey. Participants entered an alphanumeric code (mother’s maiden name plus participant’s year of birth; e.g., Smith1982) so that Time 1 and 2 data could be linked. The Time 1 survey contained all measures listed above, whereas the Time 2 survey only measured self-esteem. All statistical analyses were carried out using SPSS for windows. Results Descriptive Statistics The means, standard deviations, and range of scores for each variable used in the regression analyses are presented. On average, these women are moderately dissatisfied with their appearance, while also being moderately satisfied with all other domains of satisfaction. Regression Analyses Hierarchical regression analysis was conducted to evaluate the influence of Time 1 body dissatisfaction on change in self-esteem scores by itself (at Step 1), and within the context of other domains of satisfaction (at Step 2). At Step 1, body dissatisfaction scores reliably predict change in self-esteem, accounting for approximately 10% of variance in the DV; F(1, 348) = 51.76, p< .001. The inclusion of other domains of satisfaction explained an additional 8% variance in the DV; ΔR2 = .08, Fchange(7, 341) = 10.69, p = .01. Body dissatisfaction does not make a significant unique contribution to prediction after controlling for other domains of satisfaction (t = -1.36, p = .18). The only predictor to significantly contribute to prediction of change in self-esteem at Step 2 is satisfaction with achievement, which uniquely explains 5% variance in the DV (sr2 = .05). Discussion According to the Contingencies of Self-Worth model (Crocker et al., 2004), a domain of satisfaction is likely to affect global self-esteem if an individual deems that domain to be important to her/his self-concept. Tiggemann (2005) argues that, given the prevalence of body dissatisfaction and the sociocultural importance placed on physical appearance, body dissatisfaction is likely to be an important predictor of self-esteem. Thus, the present studyre-evaluated the influence of body dissatisfaction on self-esteem after controlling for other domains of satisfaction. While body dissatisfaction was predictive of change in self-esteem when entered individually, inclusion of other domains of satisfaction reduced this relationship to non-significance and showed satisfaction with life achievements to be a more important predictor of self-esteem. This pattern of findings supports Mellor et al.’s (2010) contention that previous estimates of the influence of body dissatisfaction on self-esteem (e.g., Tiggemann, 2005) have been over-stated due to failure to jointly consider the influence of other predictors of self-esteem. Interestingly, the strength of the bivariate association between body dissatisfaction and self-esteem(Step 1 of the hierarchical regression) is weaker than previous estimates (e.g., McCauley et al., 1998; Miller & Downey, 1999; Webster &Tiggemann, 2003). Two chief reasons for this difference are that the present study tested this relationship longitudinally in a large sample, whereas prior studies have typically relied on cross-sectional designs with small samples. Howell (2009) demonstrates that small sample sizes tend to inflate correlation size as they are more susceptible to biases due to sampling fluctuations. Similarly, Field and Hole (2003) argue that correlations over time tend to be weaker than cross-sectional equivalents as the former is subject to a range of time-related biases, including maturation and history effects. When these time-related threats are controlled, longitudinal associations provide a more stringent test of the possibility of a causal relationship as they evaluate potential temporal precedence of the IV, and its ability to influence change in the DV (Field & Hole, 2003). The present findings should be interpreted in light of key study limitations. First, although domains of satisfaction are measured, the present design did not measure whether individuals felt that these domains were important to their self-concept. It is, therefore, possible that body dissatisfaction failed to uniquely predict self-esteem because the current sample valued physical appearance less than previous studies in which body dissatisfaction was predictive of self-esteem. Second, the time frame (2 years) used to evaluate change in self-esteem may have been too long an interval to meaningfully detect changes in self-esteem that are due to body dissatisfaction. Recent findings have demonstrated meaningful changes in body dissatisfaction and self-esteem within hours and days (e.g., Colautti et al., 2011; Kernis, 2005), suggesting that an individual’s self-esteem level may change many times between the 2 year interval used in the present study. As such, it may have been more appropriate to correlate the body dissatisfaction levels measured at baseline against self-esteem levels at a more proximal distance to this initial assessment. In conclusion, present findings suggest that body dissatisfaction is longitudinally related to self-esteem, but this relationship is weaker than previously found from cross-sectional designs and appears dependent on whether other domains of satisfaction are also included in the predictive model.Further research is needed to replicate these findings in light of identified study limitations. In particular, researchers are encouraged to evaluate the strength of the longitudinal association between body dissatisfaction and self-esteem using shorter intervals between testing periods. Evaluation of the importance that participants place on appearance is also necessary given the possibility that this may moderate the body dissatisfaction-self esteem relationship. Nevertheless, present findings caution against considering this relationship in isolation of other domains of satisfaction, as the bivariate association is likely to overstate the influence that body dissatisfaction exerts on self-esteem.