Validity of the Eating Attitudes Test: A study of Mexican eating disorders patients

Abstract

To evaluate the psychometric properties of the Mexican version of the Eating Attitudes Test (EAT-40) in clinical and control populations in Mexico City. 276 female patients with eating disorders [52 with anorexia nervosa (AN), 102 with bulimia nervosa (BN) and 122 with eating disorders not otherwise specified (EDNOS)] and a comparison group of 280 normal control female subjects completed the EAT. The EAT had an adequate level of internal consistency in the clinical sample (Cronbach's alpha=0.90). Total score was significantly correlated with criterion group membership (r=0.77, p< or =0.0001), suggesting a high level of concurrent validity. There was a small overlap in the frequency distribution of the 2 groups (eating disorders and control) and only 6% of the normal controls scored as high as lowest anorexic and bulimic patients. The factorial analysis yielded five factors accounting for 46.6% of total variance: 1) Dietary restraint, 2) Bulimia, 3) Drive of thinness, 4) Food preoccupation and 5) Perceived social pressure. This study provides evidence that the Mexican version of the EAT is an economical, reliable and potentially useful instrument for research in this field.