Is the Association Between Appearance Overvaluation and Dietary Restraint Robust to Substantive and Methodological Specifications? A Specification Curve Analysis

In the cognitive-behavioural model, dietary restraint follows from appearance overvaluation (i.e., the core psychopathology of disordered eating). However, little research has examined the association between appearance overvaluation and dietary restraint when accounting for shared variance with other factors. Moreover, results generally hinge on researchers' decision-making, including addressing outliers and covariates. Herein, specification curve analysis was used to examine the association between appearance overvaluation and dietary restraint under 80 unique regression models based on different combinations of substantive factors and methodological decisions. Results indicated a positive association between appearance overvaluation and dietary restraint among university women (N=569; mean β=0.26), however, the association was not statistically significant when all factors were included in the model, and removal of outliers made results unstable. Hence, it is possible that sociocultural and cognitive-behavioural factors may be mediating or confounding the association. Findings also highlight the importance of limiting researcher degrees of freedom during data analysis.