How to Conduct and Interpret Meta-Analyses

Abstract

Research questions may be answered with empirical studies or by aggregating existing research . In contrast to single studies, meta-analyses combine the quantitative findings of multiple primary studies to estimate an overall effect size, for example the mean predictive validity of spatial tests used for pilot selection . This aggregation allows researchers to make more reliable conclusions about the size of a specific relationship, and how much it varies between studies . Moderator analyses can be applied to examine the influence of different factors on the estimated effect size, such as whether the predictive validity varies with the type of aircraft (fixed wing or rotary) or sample (ab initio vs experienced pilots) . The aim of this chapter is to present the different steps taken to conduct a meta-analysis and explain different meta-analytic approaches for estimating the mean effect sizes and variation between studies . The chapter also presents different types of software available for both screening studies and meta-analysis calculations, and uses examples from aviation and human factors explaining how the results may be interpreted and used in applied settings .