Abstract
In a recent meta-analysis, Krauss et al. found support for a reciprocal model of low self-esteem and eating disorders where, in a vicious circle, low self-esteem makes people more vulnerable to developing eating disorders and eating disorders, in turn, scars individuals' self-esteem. However, in the present reanalyses of the same meta-analytic data, we found that the prospective effects between self-esteem and eating disorders are likely spurious, meaning they do not reflect a true causal effect, but rather correlations with residuals and regression to the mean. Consequently, the claims by Krauss et al. can be challenged. To avoid statistical artifacts, we recommend researchers to fit, as we did in the present study, complementary models to their data in order to evaluate if prospective effects may be genuinely increasing or decreasing or if they appear to be spurious.