Meta-analysis in behavioural and experimental economics

Lane, T

doi:10.4337/9781802207736.00094

Meta-analysis in behavioural and experimental economics

Lookup NU author(s): Dr Tom Lane

Downloads

Full text for this publication is not currently held within this repository. Alternative links are provided below where available.

Abstract

A meta-analysis is an analysis of analyses. The method entails identifying all suitable previous studies on a particular topic and combining the results into one dataset to reveal the quantitative patterns in the literature. Applications include identifying a literature-wide weighted average for some variable of interest (such as behaviour in a type of experimental game), as well as using meta-regression to investigate how this variable is influenced by study-level factors (such as the game's parameters, subjects’ demographics). When the variable of interest represents the effect of one factor on another, meta-analysis may be afflicted by publication bias, unreliably indicating that large effect sizes are the norm, as smaller effects remain hidden from the public domain. Meta-regression may suffer from omitted variable bias. Findings from meta-analyses should therefore be taken as suggestive rather than definitive but may serve as a useful impetus for future research.