人体测量数据误差的评估

The estimation of anthropometric measurement errors

作者将人体测量数据的方差表示为σ_x^2=σ_a^2+σ_a^2+σ_c^2,σ_a^2、σ_b^2、σ_c^2分别代表个体差异、测试者误差和仪器及其它随机误差。用交叉分组或系统分组的试验方法,可求得样本估计量S_a^2、S_b^2、S_c^2,从而可计算可靠性度量R=S_r^2/S_x^2(S_r^2=S_b^2+S_c^2)、精确性度量I=S_c^2/S_r^2和各误差波动范围±1.96S_a,±1.96S_b,±1.96S_c与国外同类型的研究报告相比,作者纠正了其中的不合理情况。

In this paper , the variance of anthropometric measures is expressed as and σ2c being the variances from subjects , measurers , and instruments or other random effects , respectively . By the experiment of two-way or nested classification , Sa2, Sb2and Sc2 can be estimated . So the measure of reliability can be estimated as and the measure of precision as The corresponding estimated error limits can be expressed as ±1.96Sa, ±1.96 Sband ± 1.96 Sc. This paper has also corrected a mistake which appears in a related report from the external literature .