Meta-analysis: the glass eye of evidence-based practice?

Meta-analysis was developed as a technique for combining the results of many different quantitative studies: it is often used to produce quantitative estimates of causal relations and/or association between variables. Meta-analysis is sometimes regarded as a central component of evidence-based practice. We draw attention to an incompatibility in the epistemology and methods of reasoning in quantitative meta-analysis and the epistemology and reasoning implicit in expert practice. We argue that this may be because the common perception of meta-analysis appeals to truth as correspondence; we suggest that rejecting the naive realism that underpins truth as correspondence allows meta-analysis to be understood in terms of truth as coherence. We can then develop an account of meta-analysis that does not depend upon reduction to a mathematical procedure but is an attempt to maximise coherence in beliefs about what works that is consistent with clinical reasoning in expert practice.