Validation of the Dutch Eating Behavior Questionnaire in Russian young people

• The sample included 530 students (147 men), whose age ranged from 18 to 45 years (M = 20.8, SD = 3.5), and their BMI ranged from 15.6 to 43.5 kg/m2 (M = 22.0, SD = 3.8).
• The sample was randomly divided into two subsamples for a split-half factor analysis approach.
• Exploratory factor analysis was conducted on the first subsample (n = 270) to identify the factor structure of the Dutch Eating Behavior Questionnaire (DEBQ). The analysis was performed using parallel factor analysis (diagonally weighted least squares) as the method of extraction factors and Promin as the rotation method. The polychoric correlation matrix was used to obtain factor solutions.
• Exploratory factor analysis showed that six factors should be selected based on Kaiser's criteria (eigenvalue > 1), while parallel analyses suggested only three factors. Three factors explained 60.8% of the variance and were fully consistent with the three DEBQ scales.
• Confirmatory factor analysis was carried out on the second subsample (n = 260) to test the three-factor structure of the DEBQ and the uniqueness of four items on the emotional scale.
• Acceptable values of fit indexes (cutoff for CFI and TLI above .90, for SRMR and RMSEA lower .08) were obtained for the fourth model, in which items of diffuse emotions (items 3, 8, 10, and 28) allow to correlate and problem items (items 15, 24, 27, and 33) are excluded.