Authors' reply

BMJ 1996; 313 doi: https://doi.org/10.1136/bmj.313.7050.166b (Published 20 July 1996) Cite this as: BMJ 1996;313:166
  1. Martin Prince,
  2. Anthony Mann,
  3. Robert Blizard,
  4. Anne Bird
  1. Lecturer in psychiatry Professor of epidemiological psychiatry Institute of Psychiatry, London SE5 8AF
  2. Lecturer in statistics Consultant psychiatrist Royal Free Hospital, London NW3 2QG

    EDITOR,—Strictly speaking, it is never possible to prove a null hypothesis. However, lack of support for rejecting a null hypothesis is an important finding in a study in which treatment has been randomly allocated to over 2500 subjects. This sample size would have given us 99.9% power to detect a difference in the rate of change in the scores obtained in the paired associate learning test of >/=0.1 point a year between treatments.

    Although it is desirable to test multiple cognitive domains, there are …

    View Full Text

    Log in

    Log in through your institution


    * For online subscription