Williams D, Anderson T, Currier D
Phys Ther. 1984 May;64(5):658-64. doi: 10.1093/ptj/64.5.658.
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.
本研究的目的是确定在哈伯德水槽中进行水下称重是否是一种有效且可靠的估计身体成分的方法。30名健康受试者在水下进行了四次独立试验；每次试验包括五次称重。第一次试验在标准化的坐式水下称重水槽中完成，后三次试验在哈伯德水槽中完成。通过计算四次水下试验的皮尔逊相关系数来确定有效性，结果r = 0.077或更高。发现坐式水槽的变异系数为1.22%，哈伯德水槽的变异系数为1.23%。两种方法之间的配对t检验在0.001水平上具有显著性。哈伯德水槽法的可靠性通过对三次试验的重复测量方差分析来确定，结果发现不具有显著性。三次哈伯德水槽试验的组内相关系数R = 0.997。此外，对体重秤进行独立观察的两位研究人员之间的评分者间信度，标准化坐式水槽的r = 0.999，哈伯德水槽的r = 0.998。因此，对于确定身体成分，哈伯德水槽似乎是有效、可靠的，并且与标准化坐式水下称重水槽同样有效。然而，由于配对t检验具有显著性，两种方法不应互换使用；应使用其中一种方法，但不能同时使用两种方法。
