Modeling individual choices in experiments : reply to Conte and Moffatt

Abstract

In a comment to Cappelen, Hole, Sørensen, and Tungodden (2007b), Conte and Moffatt (2009) challenge our use of a random utility model when studying individual choices in a fairness experiment. They propose an alternative approach, what we call the random behavioral model, and they show that the choice of modeling strategy has profound implications for our understanding of the observed behavior. In this note, we discuss how the two approaches differ, and we show that the random behavioral model of Conte and Moffatt (2009) fails to fit the data from our fairness experiment.