This study was designed to examine the reliability and validity of the bioelectrical impedance method (BIA) of measuring body composition and compare its accuracy with the results obtained by standard anthropometric methods BIA, skinfold fat, and hydrostatically measured percent fat (% fat) were obtained on 44 women and 24 men. Each subject was tested four times by two testers on two different days. Generalizability theory was used to estimate reliability and measurement error that considered both day-to-day and intertester error. The BIA, skinfold fat, and hydrostatic methods were all found to be reliable (Rxx = 0.957-0.987) with standard errors ranging from 0.9 to 1.5% fat. An additional 26 men (n = 50) and 38 women (n = 82) were tested once and combined with the data used for the reliability analysis to cross-validate BIA estimates of % fat with hydrostatically determined % fat. The cross-validation correlations for the BIA determinations of % fat ranged from 0.71 to 0.76, which were significantly lower than that obtained with the sum of seven (sigma 7) skinfolds equations (rxy = 0.92 for men and 0.88 for women). The correlations between the weight-to-height ratio body mass index (BMI) and hydrostatically determined % fat were 0.75 and 0.74 for men and women, respectively. The standard errors of estimate for the two BIA models ranged from 4.6 to 6.4% fat compared with 2.6 and 3.6% fat for the sigma 7 equations. The BIA method for measuring body composition was comparable to the BMI method, with height and weight accounting for most of the variance in the BIA equation.