Experimental Design
The experiment is a within-subject design with two games: dictator games and public goods games. The dictator games and public goods games are presented in random order at the individual level.
The dictator games consist of two tasks: a decision task and a prediction task. The decision task will elicit subjects' excuse-driven risk behavior, while the prediction task will evaluate their predictions of others' excuse-driven behavior under the same risk scenarios. To control for possible ordering effects, we will randomize the order of the decision and prediction tasks at the individual level. In each task, subjects are presented with the same five sets of multiple price lists, comprising a normalization price list and four valuation price lists. Except for the normalization price list, the rest of four price lists will be presented in a random order for each subject.
The public good games, framed as an irrigation provision problem, contain five different games. Subjects in the same experimental session participate in all five games. The experiment always begins with the baseline scenario (G0) in which each subject must choose between keeping one token with some private returns or investing it in a group fund with some collective returns to each participant for each token invested by the group. The rest of four games present varying degrees of risk to private and collective returns. Here, we conduct a between-subjects design with two treatments: a control group without partial insurance and a treatment group with partial insurance. In the control group, subjects face private investment alone (G1), private investment with a reduced risk (G3), public investment alone (G2), and public investment with a reduced risk (G4). In the treatment group, the private-risk game (G1) and the collective-risk game (G2) are the same as those in the control group. Instead, games of G3 and G4 are changed to private investment alongside a partial insurance option (G3’) and public investment alongside a partial insurance option (G4’). To control for possible ordering effects, we randomize the order of the five different games. With the purpose of reducing subjects’ cognitive load and subsequent noisy responses and maintaining task consistency, the experiment always starts with G0, then G1 or G2, and then G3 or G4 (G3’ or G4’).