On the significance of effects and the effects of significance
Discusses 3 issues in applying and interpreting effect-size (ES) estimation in psychological research. First, criteria for choosing an appropriate ES metric and the advisability of adopting a single ES (PV, or the percent of shared variance) for all research is examined. A common metric would (a) forfeit information about study design and (b) permit imprecise problem definitions. Further, the different design-related metrics suggested by J. Cohen (1977) neither enhance nor trivialize the relations they describe. Ways are suggested to substantively evaluate magnitudes of effect within specific topic areas. The interpretive yardsticks include (a) multiple choices of contrasting effect sizes, (b) practical significance, and (c) research methodology. The most informative ES interpretation occurs when ES is compared to other ESs involving the same or similar variables. Finally, the value of reporting significance tests along with effect sizes is scrutinized. The two statistics are not completely redundant since ES cannot indicate that an effect size of zero has been effectively ruled out. (19 ref) (PsycINFO Database Record (c) 2006 APA, all rights reserved). © 1981 American Psychological Association.
Volume / Issue
Start / End Page
International Standard Serial Number (ISSN)
Digital Object Identifier (DOI)