Statistics and ethics in medical research. VIII. Improving the quality of statistics in medical journals

Abstract

In these articles I have concentrated very much on one aspect of research. This is not meant to imply that statistics is of over-riding importance, but rather that it is an area where much improvement is both highly desirable and possible. By emphasising the ethical implications of carrying out research and publishing papers with incorrect statistics, I have argued that this is not just a matter for the individual researcher. There needs to be a wider appreciation of the importance of correct statistical thinking, and a great improvement in the standard of published research so that the sorts of errors discussed become very much the exception rather than commonplace. In the long term improved teaching and the greater involvement of statisticians will help; in the short term it is essential to have higher standards for published papers.