Portal de Periódicos da CAPES

Consequences of prejudice against the null hypothesis.

1975; American Psychological Association; Volume: 82; Issue: 1 Linguagem: Inglês

10.1037/h0076157

1939-1455

Anthony G. Greenwald,

Decision-Making and Behavioral Economics

The consequences of prejudice against accepting the null hypothesis were examined through (a) a mathematical model intended to stimulate the research-publication process and (b) case studies of apparent erroneous rejections of the null hypothesis in published psychological research.The input parameters for the model characterize investigators' probabilities of selecting a problem for which the null hypothesis is true, of reporting, following up on, or abandoning research when data do or do not reject the null hypothesis, and they characterize editors' probabilities of publishing manuscripts concluding in favor of or against the null hypothesis.With estimates of the input parameters based on a questionnaire survey of a sample of social psychologists, the model output indicates a dysfunctional research-publication system.Particularly, the model indicates that there may be relatively few publications on problems for which the null hypothesis is (at least to a reasonable approximation) true, and of these, a high proportion will erroneously reject the null hypothesis.The case studies provide additional support for this conclusion.Accordingly, it is concluded that research traditions and customs of discrimination against accepting the null hypothesis may be very detrimental to research progress.Some procedures that can help eliminate this bias are prescribed.In a standard college dictionary (Webster's New World, College Edition, 1960), null is defined as "invalid; amounting to nought; of no value, effect, or consequence; insignificant."In statistical hypothesis testing, the null hypothesis most often refers to the hypothesis of no difference between treatment effects or of no association between variables.Interestingly, in the behavioral sciences, researchers' null hypotheses frequently satisfy the nonstatistical definition of null, being "of