In study 1, 21 females provided both honest and dishonest answers to the Eating Disorders Inventory-2 (EDI-2). It was found that the EDI-2 can be easily faked. The fake profile was used to screen subjects in a second study, in which 25 gymnasts and 25 matched controls were assessed on symptoms of eating disorders, energy intake, menstrual history, and bone mineral density (BMD). A Hotelling's T2 test (Wilks' lambda = 0.70) revealed that the gymnast and control groups did not differ significantly (P > 0.05) on the EDI-2 subscales; however, both groups exhibited scores on the Drive For Thinness (DFT) subscale of the EDI-2 that were higher than the published average for college women. More gymnasts (61%) than controls (24%) reported an absence of their menstrual cycle of 3 months or more. A higher percentage (8/11, 73%; chi 2 = 4.7, P < 0.05) of the subgroup with elevated DFT scores (i.e., > 14) reported having this disruption of their menstrual cycle compared with those with lower DFT scores (13/33, 39%). DFT scores were negatively (P < 0.05) related to energy intake (r = -0.48) and whole body BMD (r = -0.47). It is concluded that (a) DFT scores may be useful in identifying gymnasts at risk for problems associated with eating disorders, and (b) response distortion must be considered in future research using the EDI-2.