Adjustment for Publication Bias and Quality Bias in Bayesian Meta-analysis

There has been an enormous recent increase in the use of meta-analysis as a statistical technique for combining the results of many individual analyses (1-3). While the combined analysis may have increased inferential power over any individual study, there are several drawbacks to meta-analysis (4-6). One well-documented concern is the need to collect all studies, both published and unpublished, relevant to the meta-analysis if the subsequent inferences are to be valid (7-9). A meta-analysis based on only a subset of all relevant studies may result in biased conclusions.