Adaptation, confirmatory factor analysis, and psychometric properties of the Brazilian version of the Repetitive Eating Questionnaire

Abstract

Objective: The Repetitive Eating Questionnaire (Rep(eat)-Q) is a brief self-report measure developed to assess grazing behavior—an understudied problematic eating behavior associated with eating disorder psychopathology and poor weight management. This study aimed to adapt and test the psychometric properties and factor structure of the Rep(eat)-Q in a Brazilian community sample. Method: We used the International Test Commission guidelines for the adaptation of psychological measures between cultures. Participants were recruited from the community and on a university campus and completed a set of questionnaires including the Rep(eat)-Q, psychological measures, and sociodemographic data. Results: A total of 718 (86.4% female) completed the Rep(eat)-Q and were eligible for the confirmatory factor analysis (CFA). Only 542 completed the entire survey (Rep(eat)-Q and other psychological measures) and were eligible for the analysis of the validity evidence based on the relations to other variables. CFA revealed support for a two first-order model as originally proposed, however, a hierarchical (i.e., second-order) model is proposed. We found support for second-order scalar invariance across sex. Both first- and second-order internal consistency estimates presented very good evidence. Analysis (n = 542) revealed moderate to strong correlations (≥.58 to.77) between grazing (total score and subscales) and binge eating and other psychological measures, presenting good convergent evidence. Discussion: This study reports good psychometric properties of the Rep(eat)-Q for the Brazilian community sample. Public Significance: Grazing is a problematic eating behavior and the Rep(eat)-Q is one of the measures to assess grazing. The article aimed to adapt and test the psychometric properties of the Rep(eat)-Q in a Brazilian community sample. The results support a two-factor structure and suggest that the Rep(eat)-Q possesses good psychometric properties to be used with the Brazilian populations.