Twodifferent approaches have been used to derive measures of effect size. One approach is based on the comparison of treatment means, The standardized mean difference is an appropriate measure of effect size when one is merely comparing two treatments, but there is no satisfactory analogue for comparing more than two treatments. The second approach is based on the proportion of variance in the dependent variable that is explained by the independent variable. Estimates have been proposed for both fixed-factor and random-factor designs, but their sampling properties are not well understood. Nevertheless, measures of effect size can allow quantitative comparisons to be made across different studies, and they can be a useful adjunct to more traditional outcome measures such as test statistics and significance levels.Most psychological researchers appreciate in abstract terms at least that statements describing the outcomes of tests of statistical inference need to be distinguished from statements describing the importance of the relevant findings in theoretical or practical terms. The latter may have more to do with the magnitude of the effects in question than their level of statistical significance. Cohen (1965) remarked that in research concerned with comparisons among treatment means, investigators nonetheless typically confined themselves to reporting test statistics such as t or F and did not attempt to derive measures of effect size. More specifically, Craig, Eison, and Metze (1976) surveyed the articles in three different psychological journals that had employed Student's t test; none of these reported a measure of effect size, and in many instances the "significant" effects proved on inspection to be relatively slight in magnitude. Craig et al. concluded that "researchers and journal editors as a whole tend to (over)rely on 'significant' differences as the definition of meaningful research" (p. 282). This situation does not seem to have altered in the intervening time.This paper reviews research on the development and practical value of different measures of effect size. Classically, two different approaches have been taken in deriving such measures. One approach is based on the comparison of different treatment means, and the other approach evaluates the proportion of the variance in the dependent variable that is explained by the independent variable. Winer, Brown, and Michels (1991) noted that the first approach tends to be used in fixed-effects de-The author is grateful to Jacob Cohen, Richard Schweickert, and two anonymous reviewers for their comments on previous versions of this paper. Correspondence should be addressed to 1.T. E. Richardson, Department of Human Sciences, Brunei University, Uxbridge, Middlesex UB8 3PH, United Kingdom (e-mail: john.richardson@ brunel.ac.uk).Copyright 1996 Psychonomic Society, Inc. 12 signs, where the treatments employed exhaust the population of interest. The second approach is typically used in random-effects designs, in which the treatments are regarded as a sample f...