A questionnaire to assess eating behavior: Structure, validity and responsiveness of a new German eating behavior scale (SEV).

Numerous weight-loss interventions promoting healthy and weight-reducing eating behavior have been developed over the past years. To evaluate the effectiveness of such interventions for eating behavior change, short, validated and sensitive instruments are needed. In this study series, we developed and validated a new outcome measure to assess health-conscious and weight-controlling eating behavior for the evaluation of weight-loss interventions. The preliminary version of the German eating behavior scale (Skala zumErnährungsverhalten [SEV]) included 40 items. Three studies were conducted to (a) reduce the preliminary item pool, (b) investigate structural validity and internal consistency using exploratory and confirmatory factor analyses (EFA, CFA), as well as McDonald's ω, and (c) test construct validity with physiological and behavioral parameters. Responsiveness to change was also assessed after a 12-week weight-loss intervention. EFA indicated a two-factor solution with health-conscious (hc-EB) and weight-controlling eating behavior (wc-EB) as subscales, CFA confirmed the two-factor solution with acceptable model fit. Internal consistencies of both subscales were also acceptable to good (hc-EB: ω = 0.88; wc-EB: ω = 0.78). Significantly small to moderate correlations to the Healthy Eating Index (r = 0.51) as well as blood glucose (r = 0.31), blood lipids (r = 0.23), and vascular age (r = 0.31) were found, supporting the construct validity of the SEV and its subscales. Both subscales detected intervention-related changes in eating behavior among subjects of a weight-loss intervention with Standardized Response Means of 0.52 and 0.67, indicating good responsiveness of the SEV. In sum, findings provide evidence that the SEV is a valid and responsive measure to assess health-conscious and weight-controlling eating behavior.