Consensus Decision-Making: Performance of Heuristics and Mental Models

Document Type

Peer-Reviewed Article

Publication Date



Consensus decision-making, found in settings ranging from formal institutions to ad hoc groups, represents a critical component of human social interaction. In such cases, decision-makers often must agree to a course of action with awareness of others' behavior (e.g., votes) but not group members' underlying motivations and strategies. How, then, should individual agents/decision-makers balance the payoffs of available options against the time it takes to reach consensus (ranging from quick consensus to stalemate)? In the current research, a novel, repeated consensus task was played by simulated agents. These agents differed in how they a) evaluated the available options and b) anticipated the choices of fellow group members. While intuition may suggest that intractably selfish agents will outcompete those willing to compromise, the data demonstrate that socially minded agents – i.e., one type that employed a simple heuristic and two types that employed sophisticated social cognition – were the only ones to exhibit evolutionarily stable strategies. However, sophisticated cognition did not guarantee good performance, suggesting that caution is warranted in trying to “outsmart” competing agents. While these agents were restricted to simple behaviors, the approach and models described here provide a potentially useful framework for studying consensus in humans.