The effect of social network structure on group anchoring bias

Abstract

Decisions—whether made by individuals or groups—often involve estimating quantities, a process that is subject to anchoring bias (Tversky and Kahneman in Science 185: 1124–1131, 1974). Differences in susceptibility to anchoring bias between individuals and groups have been recently explored with the result that groups appear less biased than individuals (Meub and Proeger in Theor Decis 85:117–150, 2018). However, existing studies treat groups monolithically without taking into account their network structure—the pattern of relationships among members. The present paper investigates the effects of group social network structure on anchoring bias. Using a structured survey instrument, we gathered data on competence-based trust relationships among 264 students enrolled in a university degree program. An anchoring experiment was conducted in which some of the students made estimates as individuals, while others did so in groups of different structures. The findings provide initial evidence of differences in bias levels across variously structured groups as well as relative to individuals. Groups with highly centralized trust networks (where a single person owned everyone’s trust) showed more anchoring bias than dense groups (where everyone trusted everyone else) and sparse groups (where no one trusted any other member of the group) showed more bias than dense groups. In addition, despite previous research demonstrating groups are less susceptible than individuals to anchoring bias, this study shows a higher presence of bias in both our centralized groups and sparse groups when compared to individuals, suggesting that group structure might moderate the mitigating effect of groups on anchoring bias. The research has implications for organizational behavior and social network literature. Specifically, this study contributes to the debate on anchoring bias for group decisions by highlighting the significant role of social network structure. At the same time, it contributes to the literature on network structure and performance by providing initial evidence of how network structure affects anchoring bias susceptibility. Moreover, our study contributes to management practice by alerting managers to the dangers of centralized networks, suggesting that competence-based trust plays a vital role in the resistance to anchoring bias.