Velema J P, Blettner M, Restrepo M, Muñoz N
International Agency for Research on Cancer, Lyon, France.
Epidemiology. 1991 Mar;2(2):107-15.
We used data from Colombia to compare responses from husbands and wives concerning the reproductive history of the women. We analyzed agreement in two ways: First, we compared kappa statistics between independent subgroups. Second, we evaluated agreement by means of log-linear models. Men underreported the total number of pregnancies and the number of abortions of their wives. Agreement on the term of the pregnancy was dependent on the ages of the partners. Log-linear modeling provides an attractive alternative to more usual ways of studying agreement.
我们使用来自哥伦比亚的数据，比较丈夫和妻子对女性生育史的回答。我们通过两种方式分析一致性：第一，我们比较独立亚组之间的kappa统计量。第二，我们通过对数线性模型评估一致性。男性少报了其妻子的怀孕总数和流产次数。关于怀孕期限的一致性取决于伴侣的年龄。对数线性模型为研究一致性的更常用方法提供了一种有吸引力的替代方法。
