Pike M C, Anderson J, Day N
Epidemiol Community Health. 1979 Mar;33(1):104-6. doi: 10.1136/jech.33.1.104.
We have studied Miettinen's multivariate confounder score method of controlling confounding in case-control studies both theoretically and by simulation. The main conclusion to be drawn from out results is that the method will in many practical situations seriously exaggerate the statistical significance achieved, and its use is not to be recommended.
我们已经从理论和模拟两方面研究了米耶蒂宁在病例对照研究中控制混杂因素的多变量混杂因素评分方法。从我们的结果中得出的主要结论是,在许多实际情况下,该方法会严重夸大所达到的统计显著性,因此不建议使用。