A p-curve analysis of the emotional Stroop effect among women with eating disorders

OBJECTIVE

A recent meta-review of attentional bias research in eating disorders suggests that meta-analyses and systematic reviews include many low-quality and underpowered studies (Stott et al., 2021). As such, we examined whether published research examining the link between attentional bias, using the emotional Stroop task, and eating disorders among women with eating disorders has evidential value (ruling out selective reporting of a statistically significant effect) using a p-curve analysis. A p-curve analysis plots statisticall significant p-values onto a curve from .01 to .05 to examine its distribution. We hypothesized that the p-curve would be flat, indicating no true effect.

METHOD

The hypothesis, database search strategy, and data analytic approach were pre-registered. The inclusion criteria were reports that compared control and eating disorder groups, reported inferential statistics, and that used body shape/weight or general threat target words.

RESULTS

Fifty published reports were included in the p-curve analyses. Unexpectedly, the half and full p-curves were significantly right-skewed, indicating evidential value. However, the results were not robust to the exclusion of the seven lowest p-values and on average, reports were underpowered. There were also 18 reports with null results (they had a p-value greater than .05), which precluded their inclusion in the p-curve analyses.

DISCUSSION

The findings suggest that most of the evidence from research examining attentional biases using the emotional Stroop task among women with an eating disorder or with elevated eating disorder symptoms is underpowered and so should be interpreted with considerable caution.

PUBLIC SIGNIFICANCE

Concerns have been raised about the low quality of research examining attentional biases among women with eating disorders using the emotional Stroop task. In the current research, we observed that the evidential value of primary research reporting differences between women with and without eating disorders was equivocal and had low statistical power. These results can guide researchers towards conducting more rigorous research on attentional biases among people with eating disorders.

No Evidence That Exposure to Materialistic Advertisements Influence Appearance Overvaluation and Financial Success Overvaluation in the Self-concept

Theory and prior research indicate that placing overriding importance on a life domain (e.g., appearance, financial success, health, work, interpersonal relationships) can negatively influence mental and physical health. In particular, people who overvalue appearance have been shown to engage in maladaptive weight-control behaviours and to have eating disorders. Likewise, people who overvalue financial success have been shown to engage in risky gambling and to have disordered gambling. Although the consequences of overvaluing a life domain are palpable, much less is known about its antecedents, which we examined in the current research. According to the Consumer Culture Impact Model, exposure to sociocultural ideals regarding luxury, wealth, and appearance via advertisements influence appearance overvaluation. We proposed that exposure to such sociocultural ideals should also influence financial success overvaluation. We tested the hypothesis that appearance and financial success overvaluation increase in response to viewing materialistic advertisements. First and second year undergraduate students completed self-report measures of appearance and financial success overvaluation in September 2020 (N = 185). They were re-contacted in March 2021 to complete a consumer decision making task. In the task, participants were randomly assigned to a control condition wherein they viewed and compared several non-materialistic advertisements (e.g., pencils, coffee) or to an experimental condition wherein they viewed and compared materialistic advertisements (e.g., champagne, luxury vacation). Afterwards, all participants completed again the same measures of overvaluation. Unexpectedly, linear regression analyses showed that there were no pre-post changes in appearance and financial success overvaluation from before to after exposure to materialistic (relative to non-materialistic) advertisements. Exploratory Bayesian regression analyses revealed support for the null hypothesis. Findings do not support the Consumer Culture Impact Model. Holding materialistic values are discussed as a potential moderator.