The Effect of Minimal Group Framing in a Dictator Game Experiment

Abstract

Our study examines the impact of framing of tasks on dictator game choices. We specifically examine the effect of using two different framing of instructions, one of which we deem the charity frame and the other the group frame. The former frames the dictator choice in terms of giving to a randomly matched participant, whereas the latter frames the dictator choice in terms of allocating between the dictator and a randomly matched participant who is referred to as partner. We find that the group framing is associated with significantly higher transfers from the proposer to the recipient in the experiment. Dictators in the group frame transfer on average INR 40 (8 percent of their endowment) more than those in the charity frame. Furthermore, approximately 16 percent of subjects in the charity frame provide equal splits as compared to 42 percent in the group frame. Finally, approximately 8 percent of dictators gave nothing to the recipient in the group frame as compared to almost 18 percent who were purely selfish in the charity frame. The difference due to framing is robust and remains even after controlling for demographic characteristics such as age, gender, family income and educational background. A hurdle regression shows us that the group frame significantly affects the decision to give the recipient a positive amount, but not the amount actually given conditional on the decision to give. Finally, results from logit regressions indicate that the group frame significantly increases the probability of making 50:50 (or higher) offers to the recipient.