ABSTRACT
Extensive research exists on the association between self-reported emotional eating (EE) and disordered eating (DE) behaviors. Heterogeneity exists by type (e.g., unidimensional vs. multidimensional) and valence (e.g., negative vs. positive) of self-reported EE, and no previous meta-analyses have examined the association between self-reported EE and DE behaviors. A total of 67 studies (N = 26,289; 43 reporting relations in one model, and 24 reporting relations in more than one model) met inclusion criteria; ranges for age and publication date were 18.0-61.8 years old and 1995 to 2022. Five models quantified relations between DE behaviors and 1) broad negative EE, 2) EE in response to depression, 3) EE in response to anger and anxiety, 4) EE in response to boredom, and 5) EE in response to positive emotions. Using random-effects models, pooled Cohen's d effect sizes suggested small, positive relations between DE behaviors and self-reported broad negative EE (d = 0.40, p < 0.001), EE-depression (d = 0.41, p < 0.001), EE-anger/anxiety (d = 0.35, p < 0.001), and EE-boredom (d = 0.38, p < 0.001). A significant, but very small, positive relation was observed between DE behaviors and self-reported EE-positive (d = 0.08, p = 0.01). Subgroup analyses suggested a medium, positive relation between self-reported broad negative EE and binge eating (d = 0.53, p < 0.001) and a small, positive relation between self-reported broad negative EE and dietary restraint (d = 0.20, p < 0.001). Significant heterogeneity was identified across all models except for the EE-boredom and DE behaviors model. Higher BMI, but not age, clinical status, or type of DE behavior strengthened the positive relation between self-reported broad negative EE and DE behaviors. Findings support previous research suggesting that negative and positive EE are distinct constructs, with negatively valenced EE being more closely associated with DE behaviors, especially binge eating.
