>From: "DAP Louw (Sielkunde)" <[EMAIL PROTECTED]>
>Can someone please refer my to a source that will shed some light
>on what precisely the "difference" between clinical and statistical
>significance is. One often hears somebody saying: "Yes, there is
>no statistical significance, but there is definitely a clinical
>significance."
Clinical significance is sometimes advocated as a non-statistical criterion
for evaluation of change in small-n designs. But I think you are probably
thinking of the distinction Rosenthal & Rosnow make between statistically
"significant effects" and "important" effects. On the one hand, it is
possible to obtain statistically signicant differences between groups based
on an effect size that has little parctical impact (e.g., with a large
enough sample, I might find that a treatment program produces a
statistically signifianct increase of 1.5 points on an IQ test). On the
other hand, effects that might not attain statistical significance and/or
might produce small effect sizes can produce effects that are considered
important. Rosenthal & Rosnow advocate using binomial effect-size displays
to get at these important effects. They give the following research as
examples:
A comparison of alcohol abuse/dependence and depression among Vietnam
veterans and veterans of the same cohort who did not serve in Vietnam found
correlations between type of service and later problems with depression and
alcohol abuse/dependence to be r = .06 and r = .07, respectively.
Statistically these are small. But the rates of depression and alcohol
abuse in the Vietnam veterans were twice as large as the rates in the
non-Vietnam veteran population, an important contrast.
The relation between use of aspirin and heart attack in a group of 22,000
physicians was r = .034 but the study was terminated early and all
participants were encouraged to begin taking aspirin because nearly twice
the number of physicians in the placebo group were having heart attacks
than in the aspirin group.
Rosenthal, R., & Rosnow, R. L. (1991). Essentials of behavioral research:
Methods and data analysis (2nd ed.). New York: McGraw-Hill.
Rosenthal, R. (1990). How are we doing in soft psychology? American
Psychologist, 45, 775-777.
For critiques of this approach, see:
Crow, E. L. (1991). Response to Rosenthal's comment "How are we doing in
soft psychology?" American Psychologist, 46, 1083.
Strahan, R. F. (1991). Remarks on the binomial effect size display.
American Psychologist, 46, 1083-1084.
McGraw, K. O. (1991). Problems with the BESD: A comment on Rosenthal's
"How are we doing in soft psychology?" American Psychologist, 46, 1084-1086.
And a rejoinder by Rosenthal:
Rosenthal, R. (1991). Effect sizes: Pearson's corelation, its display
via the BESD, and alternative indices. American Psychologist, 46, 1086-1087.
________________________________________________________
Claudia J. Stanny, Ph.D. e-mail: [EMAIL PROTECTED]
Department of Psychology Phone: (850) 474 - 3163
University of West Florida FAX: (850) 857 - 6060
Pensacola, FL 32514 - 5751
Web: http://www.uwf.edu/psych/stanny.html
