Meta analysis and the curse of publication bias
5 December 2017
Speaker: Professor John Copas, Department of Statistics, University of Warwick
Publication bias is a major threat to the validity of the meta analysis of literature reviews. A common conjecture is that research studies reporting a significant result are more likely to be selected for review than studies whose results are inconclusive. It is impossible to correct for publication bias but, at least under this conjecture, we can assess how sensitively review results are likely to depend on different degrees of selection. We will look at examples taken from different areas of the social and medical sciences.