To the Editor.—
We read with interest the reanalysis by Andrews and Harvey (Archives 1981;38:1203-1208) of the data by Smith et al1 thought by the authors to answer satisfactorily the question "Does psychotherapy benefit the neurotic patient?"It is not our purpose to comment generally on the appropriateness of the meta-analytic approach to psycho therapy outcome research, troubled as it is by the méthodologie and conceptual heterogeneity of the studies reviewed and the inadequacy of a school- based approach to the assessment of treatment effects. The report itself is so fraught with méthodologie short comings, however, that in our view it contributes little to answering the important question outlined in its title. Its major problem lies in the meth