My title question, "When should we use our heads instead of the formula?" is not rhetorical. I am sincerely asking what I see as an important question. I find the two extreme answers to this question, namely, "Always" and "Never," equally unacceptable. But to formulate a satisfactory answer upon the present evidence seems extraordinarily difficult. I put the question in the practical clinical context. This is where Sarbin put it in his pioneering study 14 years ago, and this is where it belongs. Some critics of my book (5) have repudiated the whole question by saying that, always and necessarily, we use both our heads and the formula. No, we do not. In research, we use both; the best clinical research involves a shuttling back and forth between clever, creative speculation and subsequent statistical testing of empirical derivations therefrom. So far as I am aware, nobody has ever denied this. Even the arch-actuary George Lundberg approved of the clinician as hypothesis-maker. In research one cannot design experiments or concoct theories without using his head, and he cannot test them rigorously without using a formula. This is so obvious that I am surprised to find that people will waste time in discussing it. The clinical-statistical issue can hardly be stated so as to make sense in the research context, and I should have thought it clear that a meaningful issue can be raised only in the context of daily clinical activity.
[1]
Paul E. Meehl,et al.
Clinical Versus Statistical Prediction: A Theoretical Analysis and a Review of the Evidence
,
1996
.
[2]
L. Cronbach,et al.
Construct validity in psychological tests.
,
1955,
Psychological bulletin.
[3]
Paul E. Meehl,et al.
Wanted—a good cook-book.
,
1956
.
[4]
Symposium on Clinical and Statistical Prediction: The Tie That Binds.
,
1956
.
[5]
P. Meehl,et al.
Antecedent probability and the efficiency of psychometric signs, patterns, or cutting scores.
,
1955,
Psychological bulletin.
[6]
C. McArthur.
Analyzing the clinical process.
,
1954
.