Correcting for Primary Study Misspecifications in Meta-Analysis

Abstract

Misspecifications and differences in operational definitions of elasticities in primary studies carry over to meta-analysis results. We show that the current practice of accounting for such primary study aberrations in a meta-analysis by means of dummy variables goes a long way in mitigating their negative effects on the bias and mean squared error of the estimator, and the size and the power of the statistical tests on the meta-estimate. Controlling for omitted variable bias has a bigger beneficial impact on the meta-analysis results than the concomitant procedure for point versus double-log elasticities. However, the impact of mixing different types of elasticities on the results of a meta-analysis is smaller in any case.

Keywords: Meta-analysis, Monte Carlo simulation, Omitted variable bias, Elasticities, Model Misspecification

JEL Classification: C12, C15, C40