Are People Bayesian? Uncovering Behavioral Strategies

Abstract

Economists and psychologists have recently been developing new theories of decision making under uncertainty that can accommodate the observed violations of standard statistical decision theoretic axioms by experimental subjects. We propose a procedure which finds a collection of decision rules that best explain the behavior of experimental subjects. The procedure is a combination of maximum likelihood estimation of the rules together with an implicit classification of subjects to the various rules, and a penalty for having too many rules. We apply our procedure to data on probabilistic updating by subjects in four different universities. We get remarkably robust results which show that the most important rules used by the subjects (in order of importance) are Bayes ' s rule, a representativeness rule (ignoring the prior), and to a lesser extent, conservatism (over-weighting the prior).

Expected utility theory and prospect theory: one wedding and a decent funeral

“…Mixture models are also virtually identical to "latent class models" used in many areas of statistics, marketing and econometrics, even though the applications often make them seem quite different (e.g., Goodman 1974aGoodman , 1974bVermunt and Magidson 2003). In experimental economics, El-Gamal and Grether (1995) estimate a finite mixture model of Bayesian updating behavior, and contrast it to a related approach in which individual subject behavior is classified completely as one type of the other. Stahl and Wilson (1995) develop a finite mixture model to explain behavior in a normal form game, differentiating between five types of boundedly rational players.…”

“…One could alternatively define a grand likelihood in which observations or subjects are classified as following one model or the other on the basis of the latent probabilities π EUT and π PT . El-Gamal andGrether (1995) illustrate this approach in the context of identifying behavioral strategies in Bayesian updating experiments.…”

An Experimental Comparison of Induced and Elicited Beliefs

“…Furthermore, some subjects might use one rule, while others use a different one (El Gamal and Grether, 1995). Grether (1992) also argues that subjects may change rules depending on the decision situation (e.g.…”

“…1 We begin by interpreting predictions with an ordered probit statistical model assuming strict rationality, but are unable to recover the induced probability. In the spirit of El Gamal and Grether (1995), we then use a mixture type of ordered probit that assumes subjects are bounded rational and heterogeneous-subjects use one of four decision rules to make their predictions or just choose randomly. The first rule assumes strict rationality, while the others are heuristic rules of thumb based on the expected outcome.…”

Learning About Computers: An Analysis of Information Search and Technology Choice

“…8 Finally, we note that that our estimation procedure does not actually impose that consumers update in a strict Bayesian manner. El-Gamal and Grether (1995), in their experiment on Bayesian learning, found that, while Bayes' rule was the most commonly used rule, subjects often used a "representativeness heuristic." This means they update too much when they receive new information.…”