Do Individuals Learn To Maximise Expected Utility

Steven J. Humphrey
2002-08-29
Abstract:Violations of expected utility theory are sometimes attributed to imprecise preferences interacting with a lack of learning opportunity in the experimental laboratory. This paper reports a test of whether conditions which facilitate objective probability learning yield decisions better described by expected utility theory than is the case in experiments devoid of learning opportunity. The data show that expected utility maximising behaviour increases with the learning opportunity, but so too do systematic violations. Learning, therefore, may exacerbate choice anomalies.
What problem does this paper attempt to address?