ABSTRACT
For almost three quarters of a century, the received doctrine about appraising psychological theories has been to perform a statistical significance test. In the “soft” areas (clinical, counseling, developmental, personality, and social psychology), where the primitive state of theory only rarely permits strong conjectures as to the mathematical functions (let alone their parameters!), refutation of the null hypothesis has usually been the sole theory-testing procedure employed. In the 1960s, several psychologists (Bakan, 1966; Lykken, 1968; Meehl, 1967a; Rozeboom, 1960) came independently, for somewhat different reasons and hence with varied emphases, to entertain doubts as to the merits of null-hypothesis testing as a theoretical tool. (I set aside in this article the reliance on statistical significance in technology-e.g., benefit of a psychotropic drug, efficacy of an instructional method.) At the close of that decade, sociologists Morrison and Henkel (1970) edited a volume reprinting critical articles, and replies to them, by biologists, sociologists, psychologists, statisticians, and an economist. This excellent book should by rights be called “epoch-making” or “path-breaking,” but, regrettably, it was not. I do not know how well it sold, but it is rarely cited; and I find that the majority of psychology students in my department have never heard of it, let alone been urged by their professors to read it. Judging from published research in current soft psychology, the PhD orals I serve on, and colloquium lectures by job candidates, the book has had negligible influence.
