Proceedings of the Annual Meeting of the Cognitive Science Society

Recent research (e.g., Burns & Krygier, 2015; Chi & Burns, 2022) demonstrated that people could exhibit a strong bias towards the smaller first digits, which is consistent with the pattern predicted by Benford's law. However, this psychological phenomenon was predominantly observed when generating meaningful numbers for decision-making. We investigated explanations rooted in the statistical acquisition of distributional information and the impact of anchoring during number estimation. Undergraduate students were asked to estimate the weight, lifespan and group-size of animals after learning different distributions of these variables, supplied with an anchored value, either an average or a starting point, for reference. The Benford bias reasonably emerged regardless of the variable distribution, yet was strongly influenced by the anchored information. Notably, showing average values significantly suppressed Benford bias. These findings offered insights into the cognitive process of number estimation in the presence of statistical evidence and anchored information.