50 Years of Anchoring: A Meta-Analysis and Meta-Study of Anchoring Effects

Abstract

One of the most robust phenomena studied across the social behavioral sciences is numeric anchoring, in which a comparison against a presumed-to-be arbitrary number preceding a judgment can influence a myriad of real-world relevant judgments. The authors meta-analyze this expansive literature containing 2,603 total effect sizes (1,283 comparing high anchors against low anchors), finding a large (d = 0.824, 95% CI[0.765, 0.883], I2 = 93.64%) effect with only a small reduction from publication-bias corrections. Evidence suggests reduced (or null) effects associated with incidental anchoring (i.e., numeric priming), anchors from different dimensions or from random numbers, the presence of incentives or debiasing interventions, and knowledge. The authors supplement the meta-analysis with a pre-registered meta-study (N = 1,968) comparing high against low anchors and find similar moderation by anchors from different dimensions, extremity, cognitive load, knowledge, and debiasing but not presence of incentives. The authors provide a comprehensive review of both the empirical and theoretical landscape, and offer recommendations for consolidating the literature, improving theory testing, future development of theory and methods related to anchoring, and guidance for managers attempting to use anchoring effects in strategic and policy decisions.

Keywords: Anchoring, Judgment and Decision Making, Meta-Analysis, Meta-Study