On the independence of history: experience spill-overs between experiments

Abstract

A central understanding in experimental economics is that subjects’ decisions in the lab are independent of history. We test whether this assumption of between-experiment independence is indeed justified. We analyze experiments with an allocation decision (like a dictator or ultimatum game) and find that participation in previous experiments tends to increase the amount subjects allocate to themselves. Hence, independence between experiments cannot be presumed if subjects participate repeatedly. The finding has implications for the interpretation of previous allocation decision results and deserves attention when running future experiments.

Publication
Theory and decision