Assessing the Importance of Experimental Outcomes

Abstract

This article discusses a common statistical error that frequently appears in efforts to assess the importance of experimental outcomes in the social work practice and research literature, failure to provide proportions of explained variance in a manner which permits judgements as to their substantive importance or clinical significance. It then briefly describes a partial remedy and presents practical statistical formulas which research consumers can use to calculate proportions of explained variance when published reports fail to provide this information.