Conducting Meta-Regression Analyses

As part of the ‘Systematic Review of International Development’ initiative funded by DFID, Aus Aid and 3ie, MAER-Net will evaluate funded review protocols, provide technical assistance, and offer a methods workshop. We expect that the review of these funded review protocols will take several weeks. Below we offer some very broad guidance to novice review teams. Research teams may contact Tom Stanley for more information.

Registered Review Protocols:

Guidelines for the Meta-Analysis of Economics Research

 

  • Be Systematic
    • Search strategies and results should be independently replicable.
  • Be Comprehensive
    • Err on the side of inclusion but be sure that selected research results are comparable.
  • Be Cautious
    • Accommodate likely misspecification biases, potential publication selection biases, trends and dependencies in the original research literature.
  • Be Rigorous
    • Use state-of-the-art statistical and econometric methods that make proper allowance for heteroskedasticity, heterogeneity, and potential dependence among the reported estimates.
  • Be Robust
    • Make sure that your central findings are not the result of some unique selection of models, methods, or subsets of research.