The Recent GSK J4 Is Twice The Fun

Shared gambles were like standard gambles except if the subject took the gamble and lost, half the loss amount was ostensibly inflicted on the partner instead of the subject. Thus, shared gambles entailed moral hazard. Catch gambles were like standard gambles, but the amount to be won or lost was $10. Thus, the sure gain of $15 strictly dominated all catch gambles. Selection of a catch gamble was taken as evidence of thoughtless responding. Subjects were offered every combination of probability, amount, and gamble type twice (except that catch gambles were presented only once each). This arrangement yielded a total of 84 trials, which were presented in a random order. When a gamble was selected, the outcome (whether the ErbB subject won or lost) was not shown during the task, because we did not want outcomes to influence subsequent choices. 2.1.3. Procedure The experimenter met both subjects of each dyad together and explained the gambling task orally and with a handout. (See Appendix A in Supplementary Material for the full text of the handout.) Subjects were told the task had two roles, Player A and Player B. Player B would be offered shared gambles (which they could take to benefit themselves at the chance of causing loss to Player A), whereas Player A would not be offered any shared gambles. Thus, Player B could causes losses to Player A, but not vice versa. Subjects were told they would be randomly assigned to roles. In reality, all subjects were assigned to the role of Player B. We used this deception (and a similar one in Study 2) because our interest was in the behavior of the most powerful player, so having real subjects take on less powerful roles would have wasted subjects. After this meeting with the experimenter, the two subjects of each dyad went to adjacent private rooms to complete the task on computers. The computer told each subject that they were Player B and then displayed a message depending on a randomly assigned experimental condition. In the Anonymous condition, the message was ��Your partner will not be told what choices you made or whether you affected their chances of winning $5.�� In the Visible condition, the message was ��After the task is complete, your partner will see a complete list of the choices you made and how you affected their chances of winning $5. The two of you will then have the opportunity to discuss your choices.�� This message was the only manipulation that depended on condition. After subjects completed the gambling trials, their total earnings from the gambling task were displayed, and they were paid $5 of real money. (This $5 was the same $5 mentioned above as compensation.) The behavior of subjects in the Visible condition was not actually shown to other subjects. We think it would have been unethical to do so, since showing a subject the shared gambles taken by her partner could be upsetting to both subjects, and we had no plan to collect useful data from such an exchange. 2.2.