When assessing the efficacy of a new antidepressant in comparison with a standard treatment, most clinical trials have come to the conclusion that the nullhypothesis (equal efficacy) cannot be rejected, and have not been reformulated with respect to (at least) equivalent studies. However, it cannot be concluded from this that the compared treatments have the same (or similar) efficacy, because in many of the studies the statistical power is not sufficient. Using the effect-size formula described by Glass et al. (1981), a meta-analysis were performed combining the results of comparative trials of maprotiline, mianserin, viloxazine, trazodone, nomifensine, fluvoxamine, and fluoxetine, performed according to similar objectives and designs (similar patient selection, double-blind, randomized, etc.) and with imipramine as reference compound. Together with the results of a former meta-analysis of amitriptyline-controlled studies (Möller and Haug, 1988) the present investigation indicates differences in efficacy, which in the case of most of the new generation antidepressants is similar to the reference compounds imipramine and amitriptyline.