Even when the average performance and variability of a group are identical on a pretest and a posttest, the scores of some subjects are likely to change substantially. The percentage of subjects whose scores may be expected to change by any given amount is a simple function of the correlation between the pretest and posttest scores. Two examples were given in which the results obtained agreed closely with those “predicted” on the basis of the pretest-posttest correlation. Understanding regression effects will help investigators guard against misinterpretation of their data.
References
1.
CampbellD. T.StanleyJ. C.Experimental and quasi-experimental designs for research. Chicago: Rand McNally, 1966.
2.
FurbyL.Interpreting regression toward the mean in developmental research. Developmental Psychology, 1973, 8, 172–179.
