To examine the concordance of a questionnaire-based categorization as ‘at-risk for exercise dependence’ and an interview-based diagnosis of exercise dependence.One hundred thirty four subjects answered the German version of the Exercise Dependence Scale-21 (EDS-G). They were also assessed with a structured clinical interview for exercise dependence.The congruence between the questionnaire-based categorizations of ‘at-risk for exercise dependence’ and the diagnosis of exercise dependence based on the interview was examined using κ-coefficients.The agreement between questionnaire-based and interview-based diagnoses was fair to moderate with more false positive categorization based on the EDS-G.Assuming that a structured clinical interview allows a more accurate diagnostic categorization, the EDS-G might overestimate the prevalence of exercise dependence.