While it is true that statistical differences (i.e., real differences) are
not always functionally different (e.g., a real difference of a couple of
IQ points has no functional (clinincal) significance), the issue can be
more complicated than just that. Many years ago I springboarded off Paul
Meehl's work which talked about some of the more complex issues of
statistical vs. clinical significance.
Ghiselli, W.B. (1983). Clinical and statistical issues related to
predicting therapeutic outcome.
J. Clin. Psych., 39, 651-657.
Meehl, P.E. (1959). A comparison of clinicans with five statistical
methods of identifying
psychotic MMPI profiles. J. Couns. Psych., 6, 102-109.
Meehl, P.E. (1965). Seer over sign: The first good example. J. Exp. Res.
in Personality, 1, 27-32.
Bill Ghiselli
Department of Psychology
University of Missouri at Kansas City
"One man's fish is another man's poisson"
