ABSTRACT
OBJECTIVE: To determine the internal consistency and criteria validity of the SCOFF questionnaire for screening adolescent eating behaviour disorders. DESIGN: Validation study with stochastic sampling. SETTING: Three schools at Bucaramanga, Colombia. PARTICIPANTS: 241 students selected at random from 1595 possible ones. MAIN INTERVENTIONS: SCOFF questionnaire and semi-structured CIDI interview, administered independently. MAIN MEASUREMENTS: Social and personal details, Cronbach's alpha, sensitivity analysis, specificity, forecast values, Cohen mean kappa test, and ROC curve. RESULTS: Sensitivity found was 81.9% (95% CI, 70.7-89.7) and specificity, 78.7% (95% CI, 71.7-84.6). The area under the ROC curve was 0.8596 (95% CI, 0.8108-0.9084); the Cronbach's alpha test, 0.436, and the Cohen mean kappa, 0.554 (95% CI, 0.425-0.683). CONCLUSIONS: The SCOFF questionnaire is an acceptable alternative for screening groups of adolescents because of its simplicity and speed of administration.
