Davis C E
Am J Epidemiol. 1976 Nov;104(5):493-8. doi: 10.1093/oxfordjournals.aje.a112321.
In this paper, we have noted the ways in which regression to the mean can affect the measurement of treatment effects in clinical and epidemiologic studies. It is apparent that regression can have a sizeable effect and may lead to erroneous conclusions concerning treatment effects. Thus, the procedures outlined here should be useful in taking regression into account at the planning stage as well as at the time of analysis.
在本文中,我们已经指出了均值回归可能影响临床和流行病学研究中治疗效果测量的方式。显然,回归可能会产生相当大的影响,并可能导致有关治疗效果的错误结论。因此,这里概述的程序在规划阶段以及分析时考虑回归方面应该是有用的。