A New Approach to the Study of Editing of Repeated Lotteries

We propose a new theoretical approach to the study of editing rules applied by decision makers when dealing with repeated lotteries. Under the assumption that decision makers detect statedominance among simply two-outcome lotteries and always prefer n draws of a dominating lottery to n draws of a dominated lottery, we study editing rules beyond the use of acceptance rates. We derive an appropriate experimental methodology based on loss and gain differences, which also allows us to quantify the strength of preferences. An experiment supports previous findings showing that editing might depend on the risk profile of the underlying lottery. Moreover, we provide evidence that acceptance rates could lead to different conclusions than our methodology, because they generally do not account for the strength of preferences.