APPROXIMATE BAYESIAN-INFERENCE FOR RANDOM EFFECTS METAANALYSIS

Whilst meta-analysis is becoming a more commonplace statistical techni que, Bayesian inference in metaanalysis requires complex computational techniques to be routinely applied. We consider simple approximations for the first and second moments of the parameters of a Bayesian rand om effects model for meta-analysis. These computationally inexpensive methods are based on simple analytical formulae that provide an effici ent tool for a qualitative analysis and a quick numerical estimation o f posterior quantities. They are shown to lead to sensible approximati ons in two examples of meta-analyses and to be in broad agreement with the more computationally intensive Gibbs sampling. (C) 1998 John Wile y & Sons, Ltd.