Monday, December 1, 2008

Is there a hole in medical statistics?

Toward Evidence-Based Medical Statistics. 1: The P Value Fallacy
Ann Int Med. 1999 Jun 15;130(12):995-1004.

Toward Evidence-Based Medical Statistics. 2: The Bayes Factor
Ann Int Med. 1999 Jun 15;130(12):1005-1013

This utterly fascinating (or positively sleep inducing depending on your perspective) and well written set of articles reveal some critical problems with P values and hypothesis testing as they are currently used and interpreted.

From the abstract:
An important problem exists in the interpretation of modern medical research data: Biological understanding and previous research play little formal role in the interpretation of quantitative results...The standard statistical approach has created this situation by promoting the illusion that conclusions can be produced with certain error rates, without consideration of information from outside the experiment. This statistical approach, the key components of which are P values and hypothesis tests, is widely perceived as a mathematically coherent approach to inference. There is little appreciation in the medical community that the methodology is an amalgam of incompatible elements, whose utility for scientific inference has been the subject of intense debate among statisticians for almost 70 years...


These are well worth the read (and reread), but for those who have time only for the distilled information, an introduction to the concepts will be posted on this blog in the near future.

No comments:

Post a Comment

Note: Only a member of this blog may post a comment.