ABSTRACT
Osterlind asserts that Jensen’s statistical techniques for assessing the presence of bias are entirely correct and accurate; thus Jensen’s conclusion of no bias in standardized tests of mental ability is defensible. Osterlind also identifies the major weakness of his own argument, confessing disappointment that Jensen relied so heavily on the psychometrically flawed groups-by-items interaction (primarily ANOVA) methods for detecting bias and ignored more appropriate item response theory (IRT) methods. Finally, Osterlind offers two excuses on Jensen’s behalf: (1) ‘at the time Jensen was conducting studies on test bias he was following accepted methods’ (p. 195), and (2) regardless, the findings from more valid IRT methods lead to the same conclusions.
