McCausland, William J. et al
Testing the random utility hypothesis directly - The Economic Journal: A Journal of the Royal Economic Society - 130(625), Jan, 2020: p.183-207
We test a set of inequalities in choice probabilities, shown to be necessary and sufficient for random utility by Falmagne (1978). We run an experiment in which each of 141 participants chooses six times from each doubleton or larger subset of a universe of five lotteries. We compute Bayes factors in favour of random utility, versus an alternative with unrestricted choice probabilities. There is strong evidence that a large majority of participants satisfy random utility; however, there is strong evidence against random utility for four participants. Results are fairly robust to the choice of prior. – Reproduced
Bayesian Analysis, Hypothesis Testing, Laboratory, Individual Behavior, Microeconomic Behavior: Underlying Principles
Testing the random utility hypothesis directly - The Economic Journal: A Journal of the Royal Economic Society - 130(625), Jan, 2020: p.183-207
We test a set of inequalities in choice probabilities, shown to be necessary and sufficient for random utility by Falmagne (1978). We run an experiment in which each of 141 participants chooses six times from each doubleton or larger subset of a universe of five lotteries. We compute Bayes factors in favour of random utility, versus an alternative with unrestricted choice probabilities. There is strong evidence that a large majority of participants satisfy random utility; however, there is strong evidence against random utility for four participants. Results are fairly robust to the choice of prior. – Reproduced
Bayesian Analysis, Hypothesis Testing, Laboratory, Individual Behavior, Microeconomic Behavior: Underlying Principles