Abstract : Based on experimental Dictator Games with probabilistic prospects, Brock, Lange, and Ozbay (2013) conclude that neither ex post nor ex ante comparisons can fully account for observed behavior. We argue that their conclusion that ex ante comparisons cannot explain the data is at best weakly supported by their results, and do so on three grounds: the absence of significant differences between the most relevant treatments, the implicit assumption of subjects' risk neutrality, and the asymmetry of treatments regarding the disclosure of dictators' choice.
