Meta-analysis of Advertising Elasticity Estimates: How to Correct for the Sampling Error Bias

Abstract

Meta-analysis is a powerful statistical technique that allows social scientists to cumulate research findings across studies. Marketing researchers have used meta-analysis to measure the effectiveness of marketing mix strategies⁄ e.g., advertising, pricing, and promotion. Typically, researchers determine an average elasticity as a measure of the strategy to influence sales, and try to demonstrate causal relationship between the elasticity and some moderator variables. However, extreme caution should be exercised before making any causal inference from meta-analytic results. For example, the variance of an elasticity may be biased due to the sampling error in each estimate. Hence, the sampling error variance must be measured and accounted for before attempting to unravel any causal relationship between the elasticity and moderator variables. In this study by Bandyopadhyay⁄ a meta-analysis of advertising elasticity is done to demonstrate how to correct for the sampling error variance and measure the effect of moderator variables on elasticities. The results, according to the author, can help brand managers make useful inferences about the overall advertising effectiveness.