• dance students;
  • eating disorders;
  • receiver operating characteristic curve;
  • screening;
  • two-phase survey;
  • validity



To compare the validity of the Eating Attitudes Test (EAT) and the Bulimic Investigatory Test Edinburgh (BITE) as screening tools for eating disorders (EDs), and to identify a new threshold for each questionnaire to detect ED cases among dance and nondance students.


Dance students enrolled in high schools with gifted dance programs and nondance students randomly chosen from the same or nearby schools were invited to participate in a 2-phase ED survey. Participants completed the EAT and BITE questionnaires in the first phase. All participants who screened positive and 10% of the participants who screened negative were interviewed blindly using the Structured Clinical Interview for DSM-IV-TR Axis I Disorders Patient Edition.


The BITE had better accuracy than the EAT in detecting ED in general among both dance and non-dance students. BITE scores of 19 and 16 were the optimal cutoff values for determining ED among dance and nondance students, respectively. The optimal cutoff value for the EAT to diagnose an ED was 19 for dance students and 12 for nondance students. Both questionnaires showed higher sensitivity and lower specificity in dance students than nondance students at the same cutoff points.


The BITE had better diagnostic performance than the EAT in this nonclinical population, although its ability to detect restrictive behaviors is likely as limited as that of the EAT. Plausible explanations for these results and limitations of this study are discussed in the text. © 2013 Wiley Periodicals, Inc. (Int J Eat Disord 2014; 47:105–111)