*2.6. Data Aggregation and Analysis*

Participant responses were aggregated according to each questionnaire's specifications. For questionnaire total scores, univariate analyses of variance (ANOVAs) were conducted to assess differences between groups (OB-BED vs. OB vs. CO). To account for multicollinearity among questionnaire subscales, between-group differences on subscales were analyzed using one-way multivariate analysis of variance (MANOVAs). Anthropometric variables (age, BMI) were compared between groups using one-way ANOVA, and a Chi-squared test of independence (for sex ratio), with group as the independent variable. Chi-square tests were also used to analyze symptom severity in the YFAS 2.0.

For assessing the relative contribution of different sets of features in grouping and distinguishing OB-BED patients from OB and CO individuals, a linear discriminant analysis (LDA) was conducted with participant group as the criterion variable and questionnaire scores as predictor variables. The LDA's primary goal is to identify along how many and which dimensions (i.e., the discriminant functions) the participant groups can be distinguished from each other based on a set of predictor variables. Correlations (loadings) between predictors and discriminant functions can be used to indicate the relative value of each questionnaire to the discriminant function. Prior probabilities were adjusted to control for unequal group sizes. Box's *M* statistic was used to test for violations of the assumption of equal covariance matrices.

The significance level for all analyses was set at *p* ≤ 0.05. Effect sizes are reported as *η*2. Post hoc pairwise comparisons report Bonferroni-adjusted *p*-values for multiple comparisons. Variable values are reported as mean ± standard deviation. Z-standardized values of questionnaires are depicted in Figure 1. Unstandardized means ± standard deviation for questionnaire responses are summarized in Appendix A (Table A1). All data were analyzed with the Statistical Package for the Social Sciences (SPSS 25; IBM Corp., Armonk, NY, USA).

**Figure 1.** Mean questionnaire scores (z-standardized) and standard errors for the examined groups. CO = control; OB = obese; OB-BED = obese with co-morbid Binge Eating Disorder.
