Comparative Evaluation of Narrative Reviews and Meta-Analyses

Abstract

Abstract. Summarizing and organizing research in narrative reviews is a classic procedure for cumulating research. In recent years narrative reviews have been increasingly, though not completely, replaced by meta-analyses. Using a case study of a prominent narrative review of the behavioral priming literature (Bargh, Schwader, Hailey, Dyer, & Boothby, 2012), we show that narrative reviews run the risk of drawing a picture that tends to be too good to be true, when the effect-sizes of the papers cited in the narrative review are compared to meta-analyses of the respective topic. We shortly discuss the reasons for this, emphasizing two sources of bias that may inflict narrative reviews to a larger degree than meta-analyses, namely bias in study selection, and bias in study aggregation.