ABSTRACT
Sociologists have been formally warned that not only is.05 not sacred, but the selection of a significance level is a complex process. Note that none of the criteria should be considered in isolation—each should constitute just one of several guide lines in selecting a significance level. The authors give three suggestions pertaining to significance levels and how to report them: think and reflect on the arbitrary nature of conventional levels of significance, report the actual level obtained, and regardless of the level obtained, give an opinion on whether or not it supports the hypothesis. If these suggestions are followed extensively in research reporting, some of the problems of interpreting significance tests should diminish. To digress on tests of hypotheses, a large significance level makes it easier to reject the null hypothesis and accept the original hypothesis set up by the researcher.
