Fit for Purpose, Psychometric Assessment of the Eating Attitudes Test-26 in an Irish Adolescent Sample

Abstract

The present study examined the psychometric properties of the Eating Attitudes Test-26 (EAT-26) in a sample of Irish adolescents (N=2444). Consistent with previous research, in adolescents, confirmatory factor analysis (CFA) failed to replicate the original three-factor structure of the EAT-26. Goodness-of-fit indices provided support for a recently published six-factor EAT-18 model. As the EAT-26 is typically used as a unitary measure by clinicians, a second-order factor model was investigated, which supported a general concept of eating problems. Convergent validity of the EAT-18 was assessed using the Eating Disorder Inventory-3 (EDI-3). Using the 90th percentile, a cut-off score of 13 was identified on the EAT-18 that discriminated between those indicative and non-indicative of disordered eating attitudes and behaviors. Findings suggest that the revised factor structures may be more suitable for the general adolescent population than the original three-factor EAT-26. Clinical implications of the EAT-18 and future research recommendations are addressed.