Thorny Issues and Prickly Solutions: Publication Bias in Meta-Analytic Reviews in the Social Sciences

Abstract

Two main issues that involve publication bias are the types of studies that get published and the lack of published studies reflecting null results (i.e., the file drawer problem).  These issues are of central importance to researchers who conduct systematic (meta-analytic and narrative) reviews of a body of literature because they may result in misleading conclusions being drawn that, in turn, can adversely impact the decisions of policymakers and practitioners who rely on these reviews.  Yet without a thorough unbiased critique, a body of literature may not be as evidence-based as the reviewers believe it to be from the extant research analyzed.  Therefore, the purpose of this article is to consider issues involved in reporting and evaluating results of meta-analytic reviews, provide an overview of statistical techniques for addressing the file drawer problem (i.e., lack of journals publishing studies with null results), discuss the extent to which the file drawer is actually a “problem,” and offer some solutions.