In Reply.—Dr Powers is correct in every regard. In fact, well before receipt of his letter, the Editorial Board and I were equally concerned about such discrepancies between our statistical articles and the original articles we have published. The problem is extensive, given the large number of manuscripts that we handle. Our solution to the problem Dr Powers highlights is as follows.
Dr Brown and his colleagues have been "commissioned" to undertake a review of all articles published since I became editor from the standpoint of soundness of research design and statistical treatment and interpretation of the data. We will publish this review when completed and also use it to implement the next phase.
Dr Brown has just been appointed to the editorial board, and together with members of the board and me, will assist in the development of policies and procedures to ensure adequate review of