Appetitive traits associated with higher and lower body mass index: evaluating the validity of the adult eating behaviour questionnaire in an Australian sample

Kimberley M. Mallan, Alison Fildes, Xochitl de la Piedad Garcia, Jayne Drzezdzon, Matthew Sampson, Clare Llewellyn

The aims of this study were to evaluate the factor structure of the newly developed Adult Eating Behaviour Questionnaire (AEBQ) (Hunot et al., Appetite 105:356-63, 2016) in an Australian sample, and examine associations between the four food approach and four food avoidance appetitive traits with body mass index (BMI). Participants (N = 998) recruited between May and October 2016 via a university research participation scheme and online social network sites completed an online version of the AEBQ and self-reported demographic and anthropometric data. Of the sample, 84.8% were females, 29.6% had completed a university degree and the overall mean age was 24.32 years (SD = 8.32). Confirmatory factor analysis (CFA) was used to test three alternative factor structures (derived from issues raised in the original development study): the original 8 factor model, a 7 factor model with Food Responsiveness and Hunger scales combined, and a 7 factor model with the Hunger scale removed. The CFA revealed that the original 8 factor model was a better fit to the data than the 7 factor model in which Food Responsiveness and Hunger scales were combined. However, while reliability estimates for 7 of the 8 scales were good (Cronbach's α between 0.70-0.86), the reliability of the Hunger scale was modest (0.67) and dropping this factor resulted in a good fitting model. All food avoidance scales (except Food Fussiness) were negatively associated with body mass index (BMI) whereas Emotional Overeating was the only food approach scale positively associated with BMI. The study supports the use of the AEBQ as a reliable and valid measure of food approach and avoidance appetitive traits in adults. Longitudinal studies that examine continuity and stability of appetitive traits across the lifespan will be facilitated by the addition of this measurement tool to the literature.