A statistical note on "differential misdiagnosis of blacks and whites by the MMPI"

This report presents a critique of the conclusion by Strauss et al, that MMPI diagnostic decision-making rules derived from whites are inapplicable to blacks. It is shown that for the data from two of their three MMPI indices, a multiple discriminant function and the Goldberg Index, there was no clear evidence for greater misdiagnosis among blacks than whites when the hit rates among psychotics, behavior disorders or both are examined. For the third measure, semiactuarial profile analysis, the results show less diagnostic accuracy for blacks psychotics, but this one result does not warrant a broad generalization. A preferable method of analysis and a source of potential bias in one of the measures are pointed out.