Experimental Design
This project examines the decisions of human participants to opt into or out of conflict in a game with private information. Each of two players will privately observe a "secret number," which are iid random draws from a uniform distribution between 0 and 100. The players then choose simultaneously "Red" or "Green." Depending on the treatment, if both (either) player chooses Green, then both players get a payoff of 60, but if either (both) players choose Red, then the player with the higher secret number gets a payoff of 100, while the other gets 0. We compare treatments in which conflict is unilateral or mutual between subjects. Within subjects, we also compare these treatments to modified treatments in which each player also has an unobserved "hidden number," also iid random draws from a uniform distribution between 0 and 100. If conflict occurs (by unilateral or mutual choice) then each player's secret number is compared with the same player's hidden number. If a player's secret number is higher than the same player's hidden number, the player gets a payoff of 100, otherwise 0. Thus, in this case it is possible that both players get 100 or 0. These treatments remove the adverse selection effects of the main treatments so that Bayesian Nash equilibrium changes, but fully cursed equilibrium does not.
Condition UNILATERAL is the game in which conflict occurs if either player chooses Red and secret numbers are compared with each other to determine the outcome.
Condition MUTUAL is the game in which conflict occurs if both player choose Red and secret numbers are compared with each other to determine the outcome.
Condition UNILATERAL-HIDDEN is the game in which conflict occurs if either player chooses Red and secret numbers are compared with hidden numbers to determine the outcome.
Condition MUTUAL-HIDDEN is the game in which conflict occurs if both player choose Red and secret numbers are compared with hidden numbers to determine the outcome.
Each session will start by playing either the UNILATERAL or MUTUAL condition repeatedly for 10 periods, matched randomly into pairs of 2 within a matching group of 8 in each period. Then participants will play another 10 periods of the corresponding HIDDEN condition, again matched randomly into pairs of 2 within the same matching group of 8 in each period.
Then, participants will play a single round of the UNILATERAL or MUTUAL condition (corresponding to the condition played at the start of the session) in teams of 2. Each participants will be matched with a teammate from the other matching group, so that teammates have not previously interacted in the session. Teammates will simultaneously send free-form messages to each other before observing their team's secret number. After observing each other's messages and the team's secret number, each teammate will simultaneously choose Red or Green, with disagreement resolved randomly by the computer. If conflict occurs (by the UNILATERAL or MUTUAL rule corresponding to the treatment), the team's secret number will be compared to the secret number of the other team with whom they are randomly matched. For the team with the higher secret number, both teammates get a payoff of 100, while both members of the other team get 0. If conflict does not occur, both members of both teams get payoffs of 60.
Finally, all participants will complete several secondary tasks, including social value orientation, cognitive reflection test, multiple price list risk preference elicitation, and a demographic questionnaire.
Predictions:
UNILATEAL condition: Bayesian Nash equilibrium predicts choosing conflict (Red) regardless of the secret number.
MUTUAL condition: Bayesian Nash equilibrium predicts choosing resolution (Green) regardless of the secret number.
UNILATERAL-HIDDEN and MUTUAL-HIDDEN conditions: Bayesian Nash equilibrium predicts choosing conflict (Red) if the secret number is at least 60, otherwise choose resolution (Green).
In all conditions, Fully Cursed equilibrium predicts choosing conflict (Red) if the secret number is at least 60, otherwise choose resolution (Green).