This study examined the prevalence of eating disorders (ED) and the possible difference between ED symptoms and true ED by using questionnaires as compared with an interview and clinical evaluation in Norwegian elite female athletes (n = 522) from 35 sports and nonathletic controls (n = 448). In addition to the 117 athletes classified as "at risk" to develop ED, 90 subjects were randomly chosen, comprising 30 athletic controls, 30 at-risk nonathletes, and 30 nonathletic controls. All were interviewed and clinically examined. A significantly higher number of athletes (18%) than controls (5%) were found to actually suffer from ED, particularly athletes competing in sports in which leanness or a specific weight were considered important. When results from the screening study were compared to those from the interviews and clinical examinations, a significant underreporting of ED among athletes was demonstrated. The athletes also reported the use of other pathogenic methods in the screening study compared to what they reported in the interview. Nonathletes more correctly reported the use of pathogenic methods but overreported the prevalence of ED. Thus the issue of using questionnaires alone or in combination with personal interview/clinical examination merits further investigation.