Advancing Strategic Decision Science Since 2014
The Behavioral and Experimental Lab at the Nevada Institute of Game Theory is a hub of activity where the pristine assumptions of classical game theory meet the messy reality of human psychology. Equipped with dozens of isolated computer terminals and sophisticated software for running interactive games, the lab recruits participants from diverse backgrounds to engage in classic paradigms like the Prisoner's Dilemma, the Ultimatum Game, and the Trust Game. The primary goal is to collect robust data on how people actually make decisions in strategic settings, data which consistently shows that notions of strict self-interest and hyper-rationality are poor predictors of behavior. Instead, actions are heavily influenced by social preferences, emotions, and cognitive biases.
One of the lab's most replicated findings concerns inequity aversion. In the Ultimatum Game, where one player proposes a split of a sum of money and the other can accept or reject (causing both to get nothing), classical theory predicts the proposer will offer the smallest possible amount and the responder will accept any positive offer. NIGT experiments, like many others, show this is false. Responders routinely reject low offers, punishing unfairness even at a personal cost, and proposers, anticipating this, typically make fair offers around 40-50%. This demonstrates that people care about both their own payoff and the fairness of the distribution. Other experiments have quantified preferences for reciprocity (a tendency to reward kindness and punish unkindness) and conditional cooperation (a willingness to contribute to a public good if others do as well).
Beyond social preferences, NIGT researchers study cognitive limits. People do not perform infinite backward induction or complex probability calculations. Instead, they use heuristics and learn from experience. Lab experiments track how strategies evolve over repeated rounds of a game. Researchers model this using adaptive learning rules, such as reinforcement learning or belief-based learning, where players update their strategies based on past outcomes. This work has led to the development of behavioral game theory models that incorporate these realistic learning processes and cognitive constraints. These models are far better at predicting outcomes in settings like auctions, market entry games, and price competition than their fully rational counterparts.
In a specialized wing of the lab, studies incorporate physiological measurements like skin conductance, heart rate variability, and eye-tracking, and some collaborate with neuroscientists for fMRI studies. These experiments aim to uncover the neural and physiological correlates of strategic decision-making. For example, researchers have investigated which brain regions activate when a player experiences betrayal in a trust game or when calculating a best response in a complex game. This interdisciplinary approach helps build a more complete picture of decision-making, linking psychological constructs, biological states, and observable behavior.
The insights from NIGT's experimental work are not confined to academia. They directly inform the design of mechanisms and institutions. For instance, understanding that people are conditional cooperators can improve the design of community-based resource management systems. Knowing that people are vengeful can guide the design of contract enforcement and dispute resolution mechanisms. The lab also tests new market designs and auction formats for robustness to behavioral 'irregularities' before they are deployed in the real world. By rigorously documenting how human psychology shapes strategic play, the Nevada Institute's experimental program ensures that game theory remains a vital and accurate tool for understanding and designing real-world interactions, bridging the gap between the ideal rational actor and the wonderfully complex human being.