1920–2003 · clinical psychologist and philosopher of science, Minnesota

aul Meehl was a practising psychotherapist who spent his career at the University of Minnesota asking awkward questions about his own field. He was president of the American Psychological Association in 1962.

what he did

In Clinical versus Statistical Prediction (Meehl, 1954) he compared clinicians' forecasts about patients with forecasts from simple actuarial rules, in the studies then published, and found the rules as good or better almost every time [paraphrase, not a quotation]. Clinicians did not thank him. With Lee Cronbach he wrote the classic paper on construct validity, on what it means for a test to measure something you cannot see directly.

His methodological papers are still assigned. Meehl (1967) pointed out a paradox: in physics, better measurement makes a theory's test harder to pass; in much of psychology, where the test is only whether a difference is non-zero, better measurement makes it easier [paraphrase, not a quotation]. Meehl (1978) blamed the slow progress of "soft" psychology partly on that habit of significance testing [paraphrase, not a quotation].

why he is here

His 1954 book is the starting point of algorithm aversion, and his methodology papers sit at the start of the replication timeline.

in the tour: the forking corridor has a placard called "Meehl: clinical and statistical" ···

Primary works and sources

1. Meehl (1954). Clinical versus Statistical Prediction: A Theoretical Analysis and a Review of the Evidence. Minneapolis: University of Minnesota Press.
2. Cronbach & Meehl (1955). Construct validity in psychological tests. Psychological Bulletin. doi:10.1037/h0040957
3. Meehl (1967). Theory-Testing in Psychology and Physics: A Methodological Paradox. Philosophy of Science. doi:10.1086/288135
4. Meehl (1978). Theoretical risks and tabular asterisks: Sir Karl, Sir Ronald, and the slow progress of soft psychology. Journal of Consulting and Clinical Psychology. doi:10.1037/0022-006x.46.4.806

see also

	algorithm aversion — the human–AI version of his question
	timeline: replication — from 1959 on
	annotated: metascience — the shelf