A Meta-Analysis of Effect Sizes in Consumer Behavior Experiments

Abstract

An investigation was undertaken to empirically document effect sizes (i.e., the strength of a relationship or the magnitude of a difference between variables) in consumer behavior experiments reported in the literature during the period 1970–1982. A total of 118 experiments and 1,036 experimental effects was analyzed. Using ω 2 as a measure of effect size, the investigation's results reveal that, on average, approximately 11 percent of the variance in a response variable was explained or accounted for by a statistically significant ( p ≤ 0.05) effect. This percentage, however, differed as a function of the methodological characteristics of the experiment.