RESEARCH ARTICLE
The Effect of Attentional Training on Body Dissatisfaction and Dietary Restriction Evelyn Smith* & Elizabeth Rieger University of Sydney, Sydney, Australia
Abstract The aim of the present study is to investigate the effect of attentional training towards shape, weight and food related information on body dissatisfaction and dietary restriction. A total of 98 female participants were trained to attend to negative shape/weight words, positive shape/weight words, negative (high calorie) food words, positive (low calorie) food words or neutral words. Subsequently, a body image challenge was presented and participants’ body dissatisfaction and dietary restriction were measured. Results indicated that negative shape/weight attentional biases exacerbated body dissatisfaction and a bias towards negative food words intensified dietary restriction. The study provides evidence for specificity in the effects of attentional biases and supports the notion that attentional training may be a useful component in interventions to improve body image and reduce dieting. Copyright # 2009 John Wiley & Sons, Ltd and Eating Disorders Association.
Keywords eating disorders; body dissatisfaction; attentional bias; selective attention; dot probe *Correspondence Evelyn Smith, School of Psychology, University of Sydney, Sydney, NSW 2006, Australia. Tel: þ61293517479. Fax: þ61290365223. Email:
[email protected]
Published online 31 March 2009 in Wiley InterScience (www.interscience.wiley.com) DOI: 10.1002/erv.921
Introduction Cognitive models of eating disorders propose that selective attention to shape-, weight- and food-related information are implicated in the maintenance of eating disorder symptoms (Vitousek & Hollon, 1990; Vitousek & Orimoto, 1993; Williamson, Muller, Reas, & Thaw, 1999). Consistent with this assertion, attentional biases in eating disorder patients have been found in research using the modified Stroop task (Ben-Tovim & Walker, 1991; Cooper, Anastasiades, & Fairburn, 1992; Cooper & Fairburn, 1992; Dobson & Dozois, 2004; Jones-Chesters, Monsell, & Cooper, 1998; Rofey, Corcoran, & Tran, 2004; Stormark & Torkildsen, 2004; Sackville, Schotte, Touyz, Griffiths, & Beumont, 1998), the dichotic listening task (Schotte, McNally, & Turner, 1990), and
the dot probe task (Rieger, Schotte, Touyz, Beumont, Griffiths, & Russell, 1998; Shafran, Lee, Cooper, Palmer, & Fairburn, 2007). These biases appear to be specific to samples with eating disorder symptoms of clinical severity (Boon, Vogelzang, & Jansen, 2000; Dobson & Dozois, 2004; Rieger et al., 1998). However, these correlational studies do not address the aetiological importance of attentional biases. Indeed, some have claimed that selective attention is merely a consequence of the body image and eating concerns of individuals with an eating disorder rather than a factor driving these concerns. Moreover, according to this view, selective attention does not require treatment in its own right but resolves as patients’ eating disorder symptoms improve (Ball, Mitchell, Touyz, Griffiths, & Beumont, 2004; Cooper &
Eur. Eat. Disorders Rev. 17 (2009) 169–176 ß 2009 John Wiley & Sons, Ltd and Eating Disorders Association.
169
Attentional Training and Body Dissatisfaction
Fairburn, 1994; Green, Wakeling, & Rogers, 1998). Yet a more recent experimental study provided initial support for the causal role of selective attention in exacerbating the body dissatisfaction characteristic of this patient group. Specifically, Smith and Rieger (2006) randomly allocated participants to undergo an attentional induction towards negative shape/weight words, negative emotion words or neutral words. Participants in the negative shape/weight condition subsequently reported significantly higher levels of body dissatisfaction when exposed to a body image challenge (i.e. encouraging comparison with media images of females representing the thin-ideal) than participants in the two control conditions. In providing initial support for the contention that an attentional bias towards shape- and weight-related information has a causal role in intensifying body dissatisfaction, Smith and Rieger (2006) suggested that attention training may prove to be a useful component in interventions designed to improve eating disorder symptoms such as body dissatisfaction. However, further research is needed to justify and inform the development of such an intervention. At the most basic level, the findings obtained by Smith and Rieger (2006) demonstrating the effect of an attentional bias towards negative shape/weight words on body dissatisfaction need to be replicated. Furthermore, attentional training towards other stimuli, such as positive shape/weight words or food stimuli, and its effects on body dissatisfaction would need to be tested. In addition, an examination of the impact of these biases on a broader range of eating disorder symptoms (beyond body dissatisfaction) is required. Perhaps most important in this regard is to explore the effect of selective attention on dietary restriction given the centrality accorded dieting in maintaining other eating disorder features such as binge eating and emaciation in dominant theoretical models (Fairburn, Cooper, & Shafran, 2003; Stice, Shaw, & Nemeroff, 1998). Indeed, longitudinal research indicates that both dieting and body dissatisfaction are strong predictors of other eating disorder symptoms (Jacobi, Hayward, de Zwaan, Kraemer, & Agras, 2004; Keel, Baxter, Heatherton, & Joiner, 2007). The inclusion of a broader range of constructs will also permit investigation of the specificity of the attentional bias. That is, to clarify whether different inductions of attentional bias, such as shape/weight versus food stimuli, have different effects on body dissatisfaction and dietary restriction. 170
E. Smith and E. Reiger
Firstly, attentional training to positive body words is hypothesized to protect or enhance body image. Rieger et al. (1998) found that, relative to dieting and nondieting controls, individuals with anorexia nervosa and bulimia nervosa were characterized by a tendency to direct attention towards negative shape/weight words (e.g. ‘fat’) but away from positive shape/weight words (e.g. ‘thin’). This differential allocation of attentional resources regarding negative versus positive bodyrelated information among individuals with high levels of body dissatisfaction suggests that while an attentional bias towards negative body words may exacerbate body image disturbance, an attentional bias towards positive body words may be protective against body image disturbance or even enhance body image. The identification of protective factors may help to inform the development of more effective prevention programs for eating disorders given the generally disappointing results produced to date (Stice & Shaw, 2004). Furthermore, the causal effect of attentional training towards negative high calorie food stimuli is hypothesized to exacerbate body dissatisfaction and dietary restriction, while attentional training towards positive low calorie food stimuli is hypothesized to protect or decrease body dissatisfaction and dietary restriction. Only one study has assessed attentional biases regarding food-related stimuli using a dot probe procedure in eating disorder patients (Shafran et al., 2007). In this study, the responses of a heterogeneous eating disorders group (comprised of individuals with anorexia nervosa, bulimia nervosa and eating disorder not otherwise specified (EDNOS)) were compared with various control groups (including individuals with high anxiety levels, and high, moderate and low shape concerns). Participants completed a dot probe task which included positive pictorial images of low calorie foods (e.g. salad) and negative pictorial images of high calorie foods (e.g. pizza). It was found that the eating disorder group was significantly more likely than most of the control groups (with the exception of the high shape-concern group) to selectively attend to negative food images, and was significantly more likely to direct attention away from positive food images compared to each of the control groups. The results thus correspond to those obtained by Rieger et al. (1998) for positive and negative shape/weight stimuli. Thus the primary aim of the current study is to provide further understanding regarding the role of inducing positive and negative shape-, weight- and
Eur. Eat. Disorders Rev. 17 (2009) 169–176 ß 2009 John Wiley & Sons, Ltd and Eating Disorders Association.
E. Smith and E. Reiger
food-related attentional biases and its effect on body dissatisfaction and dieting behaviour. To investigate the effects of these induced attentional biases, a non-clinical sample will by necessity be utilized given the findings from previous research that attentional biases have generally not already been established in this population. The main hypotheses are as follows: (1) Attentional training to negative shape/weight and negative (high calorie) food related stimuli will result in significantly higher body dissatisfaction and dieting behaviour compared to a control group trained to attend to neutral words. (2) In contrast, participants trained to attend to positive shape/weight and positive (low calorie) food related stimuli will not show these adverse effects.
Method Participants One hundred and seventy-two female first year undergraduate psychology students from the University of Sydney participated for course credit. Participants had a mean age of 19.1 years (SD ¼ 1.94; range ¼ 17– 26) and were required to have a body mass index (BMI ¼ kg/m2) within the normal range as defined by the World Health Organisation (1998) of 18.5–24.9 to exclude those with a weight suggestive of an eating disorder (23 participants were excluded on the basis of this criterion). In addition, participants were excluded if they reported symptoms suggestive of a past or current eating disorder as measured by a simple eating disorder screening questionnaire (which resulted in the exclusion of a further 37 participants). Potential participants were also excluded if they reported trait body dissatisfaction in the lower or higher range (i.e. 1.5 SD below and above the mean score on the body shape questionnaire (BSQ), respectively), which resulted in 4 participants being excluded from the lower range and 10 excluded from the higher range. The final sample thus consisted of 98 participants.
Measures The profile of mood states-adolescents (POMS-A) is a 24item self-report inventory used to assess six dimensions of the mood construct (anger, confusion, depression, fatigue, tension and vigour) as well as yielding a total
Attentional Training and Body Dissatisfaction
score of overall mood disturbance. The POMS-A questionnaire is a validated shortened version of the POMS that was originally developed for use with adolescents. However, Terry, Lane, and Fogarty (2003) demonstrated that the POMS-A is also a valid instrument for the assessment of mood disturbance among adults. Internal consistencies of the subscale and total scale scores, and content, criterion and construct validities have all been found to be acceptable (Terry et al., 2003). The physical appearance state and trait anxiety scale (PASTAS) is a measure of state and trait body image anxiety (Reed, Thompson, Brannick, & Sacco, 1991). Only the 16-item state subscale was used in this study to assess state body dissatisfaction. The PASTAS has good internal consistency and excellent test–retest reliabilities (Reed et al., 1991). The BSQ is a 34-item inventory which was used to assess trait body dissatisfaction (Cooper, Taylor, Cooper, & Fairburn, 1987). The BSQ has high internal consistency and its validity is supported in terms of significant correlations with other measures of body image disturbance and its ability to discriminate between women with bulimia nervosa and female controls (Cooper et al., 1987; Faith & Allison, 1996). A simple eating disorder screening questionnaire was used to screen out potential participants with symptoms suggestive of a past or current eating disorder since these individuals may have pre-existing attentional biases (Smith & Rieger, 2006). Participants were excluded if they reported (1) a BMI outside of the specified range; (2) binge eating at least twice a week; (3) excessive exercise defined as engaging in physical activity on 5 or more days a week solely for the purpose of weight control and/or (4) vomiting or using laxatives or other medications to avoid weight gain. The use of the measures above will be outlined in the ‘Procedure’ section.
Procedure Each experimental session consisted of seven consecutive phases: (1) Participants initially completed a mood-testing phase as measured by the POMS-A. (2) An attentional training phase was then undertaken which consisted of a modified dot probe task (Cognition and Emotion Laboratory, UWA; MacLeod, Rutherford, Campbell, Ebsworthy, & Holker, 2002) carried out via a computer using a
Eur. Eat. Disorders Rev. 17 (2009) 169–176 ß 2009 John Wiley & Sons, Ltd and Eating Disorders Association.
171
Attentional Training and Body Dissatisfaction
variety of different stimuli. Participants were randomly allocated to one of the following target word conditions: Trained to attend to positive shape/weight words such as ‘slim’ ( positive shape/ weight group); negative shape/weight words such as ‘fat’ (negative shape/weight group); low calorie food words such as ‘carrot’ ( positive food group) or high calorie food words such as ‘cake’ (negative food group). All target words were paired with neutral words (e.g. ‘car’) matched for word length and usage frequency (Kucera & Francis, 1967). Following Smith and Rieger (2006), the control group was directed to attend to neutral words (neutral group) paired with negative shape and weight related stimuli, as negative weight words have been found to induce body dissatisfaction. In the 240-trial dot probe task, a pair of words appeared (one neutral and one target) for 500millisecond on the computer screen, one above the other and was followed by the presentation of a probe (> or
