Lies, damned lies and statistics (relating to medical publications)
A blog in MedPage Today this week, F. Perry Wilson (an assistant professor of medicine at the Yale School of Medicine) provides a very interesting overview of how statistics can be falsely manipulated in publications and how to spot the manipulations. The author also provides some useful solutions, including encouraging training in statistical ethics, limiting the current paradigm of positive study = publication, and pressurising researchers to prespecify the statistical tests that they are going to use. The full blog can be found here.
Leave a Reply