<i>p</i> &lt; .05 — But Is It <i>Clinically</i> Significant? Practical Examples for Clinicians

Many researchers report the p value of an analysis to communicate whether findings are significant, but a better understanding is gained by then calculating the effect size (d), which enables effects and strengths of association to be compared among studies.

Abstract

Many researchers report the p value of an analysis to communicate whether findings are significant. This may be misinterpreted to imply that if p &lt; .05 the finding is also of clinical significance. Researchers should also report the magnitude of the association, an easily calculated statistic, to allow a better understanding of results when p &lt; .05. Methods for doing this for t , χ 2 , and F values are given. The clinical significance of an intervention is also commonly determined by whether p &lt; .05. A better understanding, however, is gained by then calculating the effect size ( d ). This measure enables effects and strengths of association to be compared among studies. Methods for doing this, for the same statistics, are given.