[Psychometric quality of the "Eating Attitudes Test" (German version EAT-26D) for measuring disordered eating in pre-adolescents and proposal for a 13-item short version]

Abstract

To detect risky eating behavior questionnaires should be economic but at the same time they should fulfill the psychometric quality criteria. Available instruments are too long for the target group (e. g. EDE-Q, 28 items), restricted on primary symptoms (short version of EDI, 23 items) and with minor reliability (e. g. SCOFF and WC-Scale, 5 items each). Using the German version of the Eating Attitudes Test (EAT-26D, which comprises 26 items) in a community sample of 1 331 11-13 year old girls and 906 boys from Thuringia, Germany, we measured a internal consistency of Cronbachs' Alpha=0.85 for girls and 0.78 for boys. In a principal factor analysis, we could replicate the 6-factorial structure of previous studies. A confirmatory factor analysis verified the suitability of the EAT-26D for both, girls and boys. Reducing the EAT-26D on the 3 core-factors leads to an economic 13 item short version with an internal consistency of 0.87 for girls and 0.80 for boys.