ABSTRACT
Decisions regarding the uses of statistical levels of significance have typically been rendered by social scientists with specialized training and considerable expertise. Journals of practicing professions and occupations often reach an audience who is relatively unsophisticated in scientific methodology and techniques of statistical analysis. Hence, the statistical worker finds welcome relief in the.05 convention, which prescribes that the null hypothesis, is automatically to be rejected whenever the probability of being wrong in that decision is 5 percent or less. The choice of a statistical level of significance, that is, establishing the probability of rejecting the null hypothesis when in fact it is true, apparently demands little psychic energy on the part of researchers. Casual examination of the literature discloses that the common, arbitrary, and virtually sacred levels of.05,.01, and.001 are almost universally selected regardless of the nature and type of problem. Of these three,.05 is perhaps most sacred.
