Regression toward the mean in effectiveness studies: theoretically possible, not mathematically inevitable

Regression Toward the Mean (RTM) in extreme-group quasi-experiments is a thoroughly theoretical matter, not a simple mathematical necessity, so results of extreme-group before-and-after treatment effectiveness studies should not be doubted simply because of the theoretical possibility of RTM. The estimation of RTM, wherever one has a treatment group with a pretreatment mean off to one side of its population’s mean, requires knowledge of the population mean, standard deviation, and regression slope. RTM ought not be presumed to be operative in studies unless its several theoretical assumptions have been justified and its implications for sample data have been corroborated on that data. RTM can properly be adjusted for only when it has been adequately estimated from the relevant population data.