Questionnaire measures of negative emotional eating (NEE) have been associated with elevated body mass index (BMI). Relatively fewer studies have examined positive emotional eating (PEE), which report that PEE is associated with lower BMI or is not associated with BMI. To examine whether NEE and PEE are linked to BMI, we conducted a systematic review and meta-analysis of studies that used the Emotional Appetite Questionnaire (EMAQ) and the Salzburg Emotional Eating Scale (SEES), which assess change in eating associated with positive and negative emotions, and measured BMI in adults with a range of BMIs. A search of databases (CINAHL Medline, and PsycINFO), citations (Google Scholar) and dissertations (Proquest), and a preprint registry (Open Science Framework, OSF) was conducted independently by three screeners. Forty-three cross-sectional studies were eligible for inclusion. Correlations between BMI and NEE scales and PEE scales were extracted. Age and sex were examined as potential moderators. We found statistically significant (ps<
.001) and small mean effect sizes with random-effects models. Higher EMAQ NEE (d=0.152 [0.11, 0.19], N=18,576) and SEES NEE scales (sadness, d=.209 [.168, .250]
angry, d=.096 [.047, .144]
anxiety, d=.169 [.124, .211], N=4,141) were associated with higher BMI. The EMAQ PEE (d=-.073 [-0.106, -0.041], N=18,806) and the SEES happy (d=-0.157 [-0.100, -0.114], N=4,141) scales were associated with lower BMI. There was significant heterogeneity in effect sizes for PEE and NEE
however, there was no statistically significant moderation by age or sex. There was also no evidence for publication bias except for SEES sadness. This analysis is limited to cross-sectional questionnaire-based studies. NEE may be associated more strongly than PEE with emotional regulation difficulties and overeating of energy-dense foods which may be why it is associated with higher BMI.