Meta-analysis

If I ever was inclined to conduct a meta-analysis, and I hope I never am, I'd be sure to read this paper first (presented at the Southern Economic Association Meetings):

The Use (and Abuse) of Meta-Analysis in Environmental and Natural Resource Economics: An Assessment [PDF]

Jon P. Nelson*, a and Peter E. Kennedy

Abstract
Motivated by the 2006 report of a Work Group appointed by the Environmental Protection Agency (EPA), this paper examines the present state of meta-analysis in environmental economics and offers recommendations for its future use. To this end we summarize and assess 130 meta-analyses from 115 published and unpublished studies, covering seventeen topical categories in environmental and resource economics. First, we provide several generic meta-analysis models as reference points and discuss major estimation issues. Five econometric issues are identified as part of a complete analysis: (1) sample selection criteria; (2) basic data summary; (3) primary data heterogeneity; (4) heteroskedasticity; and (5) non-independence of multiple observations from primary studies. Second, a tabular summary is presented for the 130 meta-analyses with respect to estimation methods. Third, a narrative summary is presented for 19 meta-analyses, including the three value-of-statistical-life studies examined by the EPA Work Group and one analysis from each of sixteen other categories. Fourth, we offer a set of “best practice” guidelines for future meta-analyses in this and other areas of economics. Last, the paper comments on the use of meta-analytic methods for benefit transfers of environmental values.

"Appendix to Meta-Analysis Survey: 115 Environmental Studies"

Comments

[Will]

Given that the review 115 meta-analyses in environmental economics (and have 10 more not yet in there), how many more opportunities could there be to write one?

It is an excellent paper, by the way, especially for their consideration of the econometrics necessary to handle the odd data characteristics of a meta-analysis.