对于构成不同的样本,直接比较它们的总率是不合适的。当标准分布未知时,比较样本的两组资料的率通常是采用对各构成组的率差加权平均进行的,一般选择权仅注意了人口构成的差异。本文提出一种新的加权方法,它既反映人口构成的差异,也反映了率的构成差异。 For composing different samples, it is inappropriate to directly compare their total rates. When the standard distribution is unknown, the rates of the two sets of data in the comparative sample are usually weighted averages of the rate differences of the constituent groups, and the general options only focus on differences in population composition. In this paper, a new weighting method is proposed, which not only reflects the difference of population composition but also reflects the compositional difference of rate.