| Abstract: | Comments on A. Feingold's (see record 1995-15953-001) claim that one cannot have an accurate image of group comparisons by looking only at mean difference and effect size, d. This is because, if the comparison groups have differenct variances, the differences in the left and right tails are not symmetrical and must be calculated separately using the variances of each distribution to estimate the relative tail weights at any given z. Feingold is commended for recognizing that tail differences must be examined separatetly, however, there is an assumption that it is possible to establish a standard formula for all comparisons. Although Feingold's contribution is not without merit, it hs the double flaw of resting on a questionable assumption of normality, and of, remaining insenstive to important aspects of data. (PsycINFO Database Record (c) 2010 APA, all rights reserved) |
