To examine the accuracy of different methods of bioelectrical impedance analyzers (BIA) for body composition measurement compared to air displacement plethysmography (ADP) in college athletes. A secondary purpose was to examine the impact of hydration status on the accuracy of BIA. Percent fat estimates from the following types of BIA models were compared to percent fat from ADP: hand -to- hand (HH), foot -to- foot (FF), commercially available hand -to- foot (CHF) model, and laboratory hand -to- foot (LHF). Participants were 53 college athletes (21 males and 32 females, aged 19.8 ± 1.5 years). Hydration status was measured with an optical urine specific gravity refractometer and participants were classified as adequately hydrated (≤ 1.020 g·mL−1; 42% of sample) or significantly dehydrated (> 1.020 g·mL−1; 58% of sample). Intraclass correlations were calculated between body composition methods. A two-way (method × gender) analysis of variance with repeated measures and Fisher's LSD tests were used to compare the means estimated by ADP and BIA models. Cohen's Delta (ES) was used to quantify the size of the differences between means. Simple regression analysis was used to examine the agreement between ADP and BIA models, and to quantify prediction accuracy for each BIA model. BIA estimates of percent fat were moderately to highly correlated with ADP (R = .74 to .90). No method × gender interaction affect was found (p > .05). Mean estimated percent fat from FF (21.5 ± 7.7% fat) did not differ from ADP (21.5 ± 7.1% fat). LHF (20.0 ± 6.0% fat), HH (19.6 ± 4.9% fat), and CHF (27.6 ± 6.9% fat) produced estimates of percent fat that were significantly different from ADP. Mean differences were small (ES ≤ 0.32) between ADP vs. HH, FF, and LHF, but the mean difference was large (ES = 0.87) between ADP and CHF. Regression analysis produced total errors of 4.2% fat for HH, 4.4% fat for FF, 6.9% fat for CHF, and 4.1% fat for LHF. Regression analysis was also used to examine the effect of hydration on BIA accuracy. Hydration status did not add significantly to the prediction of ADP from BIA models. Compared to ADP, the CHF model did not provide accurate estimates of percent fat. HH, FF, and LHF provided reasonably accurate estimates of percent fat. The effect of hydration status on the accuracy of BIA models was minimal. Accurate body composition assessment is an important consideration in collegiate athletics from both a safety and performance standpoint. Athletic trainers and strength and conditioning specialists often do not have access to laboratory methods to assess body composition so practical methods should be easy to use, inexpensive, reliable, and accurate. The currently recommended BIA pre-assessment procedures (i.e., no food or drink for 4 hours prior to testing) may contribute to a high prevalence of dehydration among college athletes. In the current study the CHF model produced less accurate estimates of percent fat than the other BIA models tested. While proper hydration is important for safety and athletic performance it appears to have little influence on the accuracy of BIA models.