Regression toward the mean: its etiology, diagnosis, and treatment.

This paper explores the phenomenon of "regression toward the mean." The primary effect of this is to affect scores on retesting so that they are closer to the population mean. Thus, people who are selected for inclusion in a study because their scores on some measure are above (or below) some criterion have values on retesting that are less extreme. This may make it appear that the study participants have improved; this will occur even in the absence of an effective intervention. We explore the reasons for regression toward the mean and how it can be detected and discuss some methods that may minimize its effects.