摘要：El objetivo de este estudio ha sido presentar una version reducida del Eating Attitudes Test (EAT-26) para su utilizacion en el contexto espanol. Se busco especificamente conocer su fiabilidad y estructura factorial para, finalmente, comprobar el modelo teorico propuesto. Para ello, contamos con una muestra de 212 estudiantes universitarios con una edad media de 19.4 anos, la mayoria mujeres (61.8%). Estos respondieron al EAT-26 y a preguntas sociodemograficas. Inicialmente, a traves de un analisis paralelo se observo la coherencia de la estructura factorial. A continuacion se seleccionaron, a partir del poder discriminativo (p < 0.05), cuatro items para cada factor. Mediante el analisis de Escalamiento Multidimensional, se observo que los 12 items resultantes se distribuian segun las dimensiones teoricas esperadas (S-Stress = 0.07; RSQ = 0.97), aspecto corroborado a traves del analisis factorial confirmatorio (x2 / g.l = 2.08; GFI = 0.92; AGFI = 0.88; CFI = 0.94; e RMSEA = 0.071). Los Alfas de Cronbach para los tres factores pueden ser considerados satisfactorios: Hacer dieta (0.87), Bulimia y preocupacion por la alimentacion (0.60) y Control oral (0.52). Se concluye que la version propuesta presenta parametros psicometricos satisfactorios, pudiendo ser utilizada como medida de screening en la deteccion precoz de casos de anorexia nerviosa en la poblacion general.
其他摘要：The aim of this study was presenting a reduced version of the Eating Attitudes Test (EAT-26), for use in a Spanish context. It also intended to know its consistency and factor structure in order to confirm the proposed theoretical model. The sample was constituted by 212 undergraduate students with a mean age of 19.4 years, mostly females (61.8%). The participants were asked to answer the EAT-26 and socialdemographic questions. Initially, based on the results of a parallel analysis, the coherence of the factor structure was observed. Following that analysis, four items per factor were selected according to its discrimination power (p < .05). Using a Multidimensional Scaling Analysis, it was observed that the final 12 items were distributed according to the expected theoretical dimensions (S - Stress = 0.07; RSQ = 0.97); these results were corroborated by a Confirmatory Factor Analysis (÷2 / g.l = 2.08; GFI = 0.92; AGFI = 0.88; CFI = 0.94; e RMSEA = 0.071). The Cronbach's Alpha of the three factors can be considered adequate: Dieting (.87), Bulimia and food preoccupation (.60) and Oral control (.52). It can be concluded that the proposed version presents satisfactory psychometric parameters, and it may be used as a measure to evaluate and help in the previous detection of anorexia nervosa cases in general population.
