H. Stumpf, GENDER DIFFERENCES IN PERFORMANCE ON TESTS OF COGNITIVE-ABILITIES - EXPERIMENTAL-DESIGN ISSUES AND EMPIRICAL RESULTS, Learning and individual differences, 7(4), 1995, pp. 275-287
Experimental design variables take on particular importance when infer
ences are made about gender differences in cognitive abilities, especi
ally when conclusions are derived from standardized tests administered
to very large samples. Problems in the representativeness of the samp
les of subjects and test items, in test reliability and factor purity,
in the interpretation of effect sizes, and heterogeneity of variance
can inadvertently bias conclusions drawn from scores of females and ma
les on psychometric tests. Empirical research in this highly politiciz
ed area is reviewed with an emphasis on those factors that could cause
gender-related differences to be over- or underestimated.