Dados Bibliográficos

Resumo

Rosenthal and Rubin (1985) pointed out that in research on extreme situations (e.g., new treatments for terninally ill patients) any noticeable statistical trend in the desirable direction is valuable. It should be published even if it is of low magnitude and fails to meet our traditional criteria of statistical significance. Their approach is now being misused by those defending Rushton's (1988) 'theory' about American Blacks (based on weak trends in excessively suspect data sets). Hasty and eager acceptance of weak, biased, and unrepresentative data as scientific evidence of genetically based and relatively immutable racial differences in human potential amounts to psychological warfare on oppressed racial groups. Similar defamation of vulnerable minorities by Nazi pseudoscientists led to the loss of millions of human lives in the past. Statistical theory classifies similar endeavors as a Type I error (a misleading rejection of the null hypothesis).

Referências Citadas