Inadequate data in, attractive theory out.

Contends that R. J. Sternberg (see record 1981-06882-001) presented correlations based on independent samples of about 25 cases each, made the usual statistical comparison of each independently with a population parameter of zero, and proceeded to interpret his results as if he had made the appropriate statistical tests. In one set of comparisons, he needed to use a nonzero population parameter representative of previous research findings to define his statistical hypothesis. It is concluded that these errors are endemic in reports of psychological research. (3 ref) (PsycINFO Database Record (c) 2010 APA, all rights reserved)