The purpose of this study was to compare the Diagnostic and Statistical Manual of Mental Disorders, fourth edition (DSM-IV) and the proposed Diagnostic and Statistical Manual of Mental Disorders, fifth edition (DSM-5) diagnostic criteria in terms of the number of cases of eating disorder not otherwise specified (EDNOS) and to see which diagnostic system can effectively capture variance in psychiatric symptoms in a Japanese sample. One thousand and twenty-nine women with an eating disorder (ED) participated in this study. Assessment methods included structured clinical interviews and administration of the Eating Attitudes Test and the Eating Disorder Inventory. Relaxing the diagnostic criteria for anorexia nervosa and bulimia nervosa and recognizing binge ED decreased the proportion of EDNOS (from 45.1% to 26.1%). The DSM-5 categorization of patients was better able to capture variance in psychopathology scales. The proposed revisions to EDs in the DSM-5 partially reduced reliance on EDNOS. The DSM-5 may differentiate ED groups more effectively than the DSM-IV.