The normative force of Savage’s axiom and the Allais paradoxes: more evidence

Abstract We report the results of an experiment designed to rigorously investigate the normative force of the sure-thing principle. In a baseline treatment, we replicate our recent finding that common consequence effect violations of the axiom persist, following an opportunity to revise choices in light of a normative argument supporting it (Humphrey and Kruse in Theory and Decision 96:1–17, 2024). In a second treatment, we study the robustness of that result, by applying the same test to the common ratio effect version of the Allais paradox. We observe that whilst the result replicates, behaviour is influenced by decision task parameters. Two further treatments control for alternative explanations of observed behaviour; specifically, an experimenter demand effect and a status quo effect. We find that the original result is robust. Our conclusions are threefold. Firstly, neither version of the Allais paradox is the result of choice errors. Secondly, this finding is not explained by an experimenter demand effect or a status quo bias. Thirdly, under plausible assumptions, for which support can be found in our data, we observe a common consequence effect that can be interpreted as a rejection of the normative justification of the sure-thing principle.