Statistical Power in Rorschach Research
The Rorschach test has long remained the center of controversy, despite its widespread popularity in clinical settings (Luhin, Larsen, & Matarazzo. 1984). Nowhere is this more evident than in the long, often acrimonious, series of reviews of the test published in the Mental Measurements Yearbook CWJ\.1}B: d. Jensen, 1965). One long-standing controversy concerns whether the Rorschach is actually a "test" at all or whether it is more appropriately thought of as a clinical "technique" (Eron, 1965; Rabin, 1972; Zuhin, Eron, & Schumer. 1965). In the fourth MMYB, Sargent, a supporter of the test, stated that "the Rorschach test is a clinical technique, not a psychometric method" (Sargent. 1953, p. 218). As a test, the Rorschach has been assailed hy psychometrically minded psychologists as failing to meet many, if not most. of the standard criteria of test constnIction, including indices of internal consistency. interrater reliahility, and validity (Dana, 1%5; Jensen. 1965; McArthur, 1972) Further, as early as 1949, Cronhach expressed concerns about the quality of Rorschach research. In a quote used hy Eysenck in a scathing il1..\.llB review of the Rorschach (Eysenck, 1959), Cronhach declared that "perhaps ninety percent of the conclusions so far puhlished as a result of statistical studies are unsuhstantiated-not necessarily false-but hased on unsound analysis" (Cronhach, 1949. p. 425),
Cronhach continued, "One cannot attack the test merely because most Rorschach hypotheses are still in the pre-research stage" (p. 426) The critics of the test, alternately, wonder that "years of negative research have not cooled the ardor of the Rorschach supporter" (Knutson, 1972, p. '-140). The very nature of the Rorschach; the divergent systems of administration, scoring. and research; the nature of Rorschach scores and the shapes of score
distributions obtained; and the type of statistics commonly used (typically distribution free or nonparametric) seem to favor the views of the Rorschach's critics.