Hammitt, James K. and Robinson, Lisa A. (2015) Introduction to the Special Series on Research Synthesis: A Cross-Disciplinary Approach. Risk Analysis, 35 (6). pp. 963-970.

Abstract

To estimate the effects of a policy change, analysts must often rely on available data as time and resource constraints limit their ability to commission new primary research. Research synthesis methods—including systematic review, meta-analysis, and expert elicitation—play an important role in ensuring that this evidence is appropriately weighed and considered.We present the conclusions of a multidisciplinary Harvard Center for Risk Analysis project that evaluated and applied these methods, and introduce the resulting series of articles. The first step in any analysis is to clearly define the problem to be addressed; the second is a systematic review of the literature. Whether additional analysis is needed depends on the quality and relevance of the available data to the policy question, and the likely effect of uncertainty on the policy decision. Meta-analysis promotes understanding the variation between studies and may be used to combine the estimates to develop values for application in policy analysis. Formal, structured expert elicitation promotes careful consideration of the evidence when data are limited or inconsistent, and aids in extrapolating to the policy context. Regardless of the methods used, clear communication of the approach, assumptions, and uncertainty is essential.