How to deal with inconsistent choices on multiple price lists

Abstract

Economic experiments are often based on the claim that some heterogeneous behavioural trait affects response to treatment. This trait is measured in another part of the experiment, using a multiple price list. Frequently, choices on such a list are not perfectly consistent.In this paper we argue that this inconsistency is a resource. It informs the researcher about the precision with which the trait is measured. Precision can be estimated with a measurement error model.With data from an experiment that studies the relation between risk aversion and punishment we illustrate the approach, and show that it matters.