Objectives: To assess the validity of the Eating Attitude Test (EAT-26) in Arabic as a screening instrument in nonclinical populations. Methods: A representative sample of Grade 7–12 female students in Riyadh, Saudi Arabia, was selected randomly but proportional to various social classes. The girls were independently assessed by the EAT-26 and a structured clinical interview. Results: One hundred twenty-nine subjects were included. Twenty-five were identified by EAT-26 as having abnormal eating attitudes. One case was identified as anorexia nervosa by the interview and no cases of bulimia were found. Discussion: EAT-26 was found to be highly sensitive and reasonably specific. Like some other studies in non-Western populations, it yielded a high false positive rate and a low positive predictive value. Because of its low cost and practicality, EAT-26 might be a useful tool in screening large populations for eating disorders. © 1996 by John Wiley & Sons, Inc.