Abstract: | Mohr's (Evaluation and Program Planning 23(4) (2000)) provocative comments are examined. I agree wholeheartedly with Mohr's admonition that researchers not be too quick to assume that measurement error is random. However, I remain unconvinced by Mohr's arguments that the notion of regression toward the mean “has been a costly distraction,” that measurements of “artificial constructs” be treated as if they are free from measurement error, and that change-scores analysis generally be avoided in favor of regression analysis. |