Underwater weighing using the Hubbard tank vs the standard tank

The purpose of this study was to determine whether underwater weighing in the Hubbard tank is a valid and reliable method for estimating body composition. Thirty healthy subjects were weighed underwater for four separate trials; each trial consisted of five weighings. The first trial was completed in a standardized sit-in underwater weighing tank, and the last three trials were completed in a Hubbard tank. Validity was determined by calculating Pearson's correlation coefficient across the four underwater trials, which resulted in r = .077 or higher. The coefficient of variation was found to be 1.22 percent for the sit-in tank and 1.23 percent for the Hubbard tank. A paired t test between the two methods was significant at the .001 level. Reliability for the Hubbard tank method was determined by the repeated measures analysis of variance across the three trials and found not to be significant. The intraclass correlation for the three Hubbard tank trials resulted in R = .997. In addition, the interrater reliability between the two researchers who made independent observations of the weight scale resulted in r = .999 for the standardized sit-in tank and r = .998 for the Hubbard tank. Therefore, the Hubbard tank appears to be valid, reliable, and equally effective as the standardized sit-in underwater weighing tank for determining body composition. Because the paired t test was significant, however, the two methods should not be interchanged; one or the other method should be used, but not both.