Cognitive processing of body and appearance words as a function of thin-ideal internalization and schematic activation

Abstract

To better understand how women at risk of body image disturbance respond when their body concerns are activated, we examined attentional and memory biases in undergraduate women with high thin-ideal internalization, an identified risk factor for eating disorders, following priming of body and appearance concerns. Female undergraduates (N = 186) viewed photos of either sports cars or attractive swimsuit models, then completed the Lexical Decision Test, a word recall test, and questionnaires assessing thin-ideal internalization and eating disorder symptomatology. High thin-ideal internalizers did not exhibit cognitive biases predicted by cognitive models of eating disorders, even when their body and appearance concerns were primed by exposure to attractive models. Converging evidence suggests that high-risk non-clinical samples rarely exhibit cognitive biases characteristic of individuals with eating disorders, and, in fact, may actually incorporate ideal appearance into their schemas and preferentially attend to attractive stimuli.

Introduction

Cognitive behavioural approaches for the treatment of body image disturbances including eating disorders help individuals to identify and modify maladaptive behaviours and cognitions (Cash & Strachan, 2002). According to Vitousek and Hollon's (1990) cognitive theory of eating disorders, individuals with eating disorders develop organized cognitive structures (self-schemas) around weight and its implications for the self. According to this theory, individuals with eating disorders selectively process and have enhanced recall for schema-congruent information (e.g., fat stimuli) and resist schema-incongruent information (e.g., thin stimuli)—this pattern of cognitive processing has been theorized to result in the persistence of eating disorder symptomatology. For example, these individuals may attend to body parts they are most dissatisfied with and ignore body parts they are relatively more satisfied with. An understanding of cognitive processes in individuals at risk of body image disturbance, such as how they respond when their body concerns are activated, can provide guidance as to where to intervene in treatment.

Attentional bias is one form of cognitive bias that has received much attention in both patient and non-patient samples. The assumption underlying attentional biases is that people who are preoccupied with their body weight, body shape, or food will selectively attend to such stimuli (Williamson, White, York-Crowe, & Stewart, 2004). A meta-analysis of attentional biases in eating disorders found evidence for an attentional bias towards body-related words in anorexia nervosa, and towards body and food-related words in bulimia nervosa (Dobson & Dozois, 2004). There was little consistent evidence for an attentional bias in high-risk non-clinical samples (e.g., those with high drive for thinness, weight preoccupation, body dissatisfaction, or dietary restraint) (Dobson & Dozois, 2004; Lee & Shafran, 2004).

With few exceptions (Rieger et al., 1998, Sackville et al., 1998), previous studies have focused exclusively on attention to negatively valenced information (e.g., “fat” words) and neglected positively valenced information (e.g., “thin” words), so it is uncertain whether individuals with eating disorders or eating disorder risk factors selectively process schema-congruent information and resist schema-incongruent information as Vitousek and Hollon's (1990) theory predicts, or attend to all body-related information regardless of valence. In addition, no studies have assessed the processing of information related to overall physical appearance (i.e., words pertaining to attractiveness and unattractiveness) so it is uncertain whether individuals with eating disorders or eating disorder risk factors preferentially process body information in particular, or appearance information more generally.

Memory bias has also been examined in eating disorder research, albeit to a lesser extent than attentional bias. The assumption underlying memory biases is that individuals who are preoccupied with body weight or shape allocate additional resources to the processing of body-related information, allowing for greater elaboration of stimuli during encoding and enhanced recall of body-related information (King, Polivy, & Herman, 1991; Sebastian, Williamson, & Blouin, 1996).

Individuals with anorexia nervosa have demonstrated a memory bias for body and food-related words (Hermans, Pieters, & Eelen, 1998), whereas those with bulimia nervosa have demonstrated a memory bias for body-related words (Hunt & Cooper, 2001; Sebastian et al., 1996), but not food-related words (Hunt & Cooper, 2001). It is noteworthy that this pattern of results is opposite to that reported in the attentional bias literature (Dobson & Dozois, 2004), suggesting that perhaps individuals with anorexia nervosa feel anxious about their bodies and are preoccupied with their bodies and food, whereas individuals with bulimia nervosa are preoccupied with their bodies and feel anxious about their bodies and food.

Examining memory biases in non-clinical populations with eating disorder risk factors have generated mixed results. Preliminary findings suggest that restrained eaters have better recall (King et al., 1991) and recognition (Boon, Vogelzang, & Jansen, 2000) of body-related words than unrestrained eaters, and that women who base their self-worth on weight and shape have greater false recognition of weight and shape words in a false alarm effect paradigm relative to women who base little of their self-worth on weight and shape (Geller, Johnson, & Madsen, 1997). In contrast, a study comparing weight preoccupied individuals to controls found no differences between groups in the recall of body-related words (Sebastian et al., 1996). A study comparing individuals with low and high body dysphoria found that those with high body dysphoria recalled significantly more fat words and fewer thin words than those with low body dysphoria; however, this finding was not replicated using a word recognition task (Baker, Williamson, & Sylve, 1995). Given the scarcity of studies examining memory biases in individuals with eating disorder risk factors, these findings await replication.

One possible explanation to account for the weak or lack of cognitive biases in high-risk non-clinical samples is that individuals’ body concerns have not been activated prior to engaging in the cognitive task. Lending support to this theory is a study which found that non-depressed controls performed in a similar manner to depressed patients on an attentional test, but only when a priming procedure was used to make their cognitive state more similar to depressed patients (Segal & Vella, 1990). Similarly, another study found that appearance-schematic women demonstrated attentional biases to a heterogeneous list of appearance words (e.g., lipstick, ugly, bikinis), but only when their appearance schemas were activated (Labarge, Cash, & Brown, 1998). According to Beck's cognitive theory (Beck, Rush, Shaw, & Emery, 1979), contextual events serve to activate or “prime” otherwise latent schemas, resulting in preferential processing of schema-congruent information. If schematic activation is required for the detection of cognitive biases, then the results of previous studies examining cognitive biases in high-risk non-clinical samples are inconclusive because the lack of significant findings may be attributed to lack of schematic activation. The current study explored this possibility by attempting to activate schemas to increase the likelihood that any cognitive biases would be detected.

Exposure to attractive models may serve as a schema-congruent trigger which activates body and appearance-related concerns (Altabe & Thompson, 1996), particularly for women with high thin-ideal internalization. Thin-ideal internalization, an identified risk factor for eating disorders, refers to the extent to which an individual accepts socially defined ideals of attractiveness, such as those portrayed by professional models in the visual media (Thompson & Stice, 2001). Social comparison may be the mechanism by which exposure to attractive models adversely affects body satisfaction (Hargreaves & Tiggemann, 2004; Tiggemann & McGill, 2004). Studies examining the effect of the media on body image found that women who were explicitly instructed to rate the attractiveness of models or to compare themselves to attractive models experienced greater body dissatisfaction (Cattarin, Thompson, Thomas, & Williams, 2000; Hargreaves & Tiggemann, 2004; Tiggemann & McGill, 2004) and negative schematic activation (Altabe & Thompson, 1996). Even a brief exposure to attractive models from the media results in a negative body image state, but only among women with high thin-ideal internalization and social comparision tendencies (Yamamiya, Cash, Melnyk, Posavac, & Posavac, 2005), suggesting that exposure to attractive models may be a schema-congruent trigger for women with high thin-ideal internalization.

The present study attempted to provide a more rigorous test of cognitive processing in a high-risk non-clinical sample by activating body and appearance-related concerns prior to assessing cognitive processing. Undergraduate women were divided into two groups based on their level of thin-ideal internalization (low vs. high). Next, they completed a photo-rating (priming) task, in which they were alternately assigned to view photos of either sports cars (control condition) or attractive swimsuit models (model exposure condition). Following the priming manipulation, participants completed the Lexical Decision Test and a word recall test to assess attentional and memory biases for negatively and positively valenced body (e.g., fat, thin) and appearance-related words (e.g., unattractive, attractive). Viewing attractive swimsuit models would be expected to activate body and appearance schemas among women with high thin-ideal internalization, and thus, bias their cognitive processing of body and appearance words.

We hypothesized that women with high thin-ideal internalization who were primed with photos of swimsuit models would be the only group to demonstrate attentional and memory biases. Specifically, we hypothesized that: (1) Women with high thin-ideal internalization who were primed with photos of swimsuit models would respond faster to and have better recall of fat and unattractive words relative to neutral words, and respond slower to and have poorer recall of thin and attractive words relative to neutral words; (2) Women with high thin-ideal internalization who were not primed with photos of swimsuit models and women with low thin-ideal internalization would not demonstrate attentional or memory biases for body or appearance words relative to neutral words. In addition to demonstrating a cognitive bias relative to neutral words, cognitive theories also predict that women with high thin-ideal internalization who are primed with photos of attractive models would attend to and have better recall of fat words than thin words, and of unattractive words than attractive words. Therefore, we hypothesized that there would be an interaction between internalization, priming condition, and word valence such that: (3) Women with high thin-ideal internalization who were primed with photos of swimsuit models would attend to and have better recall of fat than thin words and of unattractive than attractive words; (4) Women with high thin-ideal internalization who were not primed with photos of swimsuit models and women with low thin-ideal internalization would demonstrate similar attention to and memory for fat and thin words and for unattractive and attractive words.