|
Field
Primary Outcomes (Explanation)
|
Before
For each participant, game, and game round, a cooperation rate is computed as the share of prosocial choices out of all decisions made. In Game Rounds 3 and 4, where group identity is visible, cooperation rates are computed separately for decisions matched with in-group member decisions and decisions matched with out-group member decisions.
The in-group cooperation advantage is defined as the difference between the in-group cooperation rate and the out-group cooperation rate. Mean cooperation is defined as the average of the two. Both outcomes are averaged across the three games to obtain pooled measures, and also analyzed separately for each game.
|
After
For each participant, game, and game round, a cooperation rate is computed as the share of prosocial choices out of all decisions made. In Game Rounds 3 and 4, where group identity is visible, cooperation rates are computed separately for decisions matched with in-group member decisions and decisions matched with out-group member decisions.
The in-group cooperation advantage is defined as the difference between the in-group cooperation rate and the out-group cooperation rate. Mean cooperation is defined as the average of the two. Both outcomes are averaged across the three games to obtain pooled measures, and also analyzed separately for each game.
Starting with session 17, generosity (dictator game) and bid intensity (all-pay auction) are computed as the share of the endowment/prize given or bid, separately for each of the five information conditions. In all-pay auctions higher bids reflect competitive rather than prosocial behavior. The one-sided conditions serve as the primary test of H1/H2 in these games, given their structural similarity to Game Rounds 3 and 4.
|
|
Field
Intervention (Hidden)
|
Before
In each Game Round, participants make 20 decisions per game, resulting in 60 decisions per Game Round. In each game, one of the two options is the prosocial choice. In Game Rounds 3 and 4, the 20 decisions per game are split into 10 decisions matched with in-group member decisions and 10 decisions matched with out-group member decisions.
The three games cover qualitatively distinct classes of 2-player strategic interactions: the Prisoner's Dilemma features a dominant strategy to defect, the Stag Hunt is a coordination game with two Nash equilibria, and the Mini Trust Game captures asymmetric payoff structures. Together, they span the key strategic constellations that arise in social interactions between two players.
Treatment assignment is subject to a no-deception constraint: the Good treatment is only available in a given session if at least one Klee group member and one Kandinsky group member actually chose the prosocial option in the Prisoner's Dilemma during Game Round 2, and analogously for the Bad treatment. Treatment assignment is therefore determined after Game Round 3, once the relevant Game Round 2 data are available.
In the Good treatment, participants are informed that one Klee group member and one Kandinsky group member chose the prosocial option in the Prisoner's Dilemma in Game Round 2. In the Bad treatment, participants are informed that one Klee group member and one Kandinsky group member chose the non-prosocial option. Since this refers to a single decision out of 20 made in the Prisoner's Dilemma in Game Round 2, the information is statistically irrelevant.
To eliminate second-order strategic reasoning, decisions in Game Rounds 3 and 4 are matched against recorded decisions from Game Round 2, in which group identity was not yet revealed to participants. Any behavioral change between Game Round 3 and Game Round 4 therefore reflects the participant's own belief updating in response to the anecdote.
|
After
In each Game Round, participants make 20 decisions per game, resulting in 60 decisions per Game Round. In each game, one of the two options is the prosocial choice. In Game Rounds 3 and 4, the 20 decisions per game are split into 10 decisions matched with in-group member decisions and 10 decisions matched with out-group member decisions.
The three games cover qualitatively distinct classes of 2-player strategic interactions: the Prisoner's Dilemma features a dominant strategy to defect, the Stag Hunt is a coordination game with two Nash equilibria, and the Mini Trust Game captures asymmetric payoff structures. Together, they span the key strategic constellations that arise in social interactions between two players.
Treatment assignment is subject to a no-deception constraint: the Good treatment is only available in a given session if at least one Klee group member and one Kandinsky group member actually chose the prosocial option in the Prisoner's Dilemma during Game Round 2, and analogously for the Bad treatment. Treatment assignment is therefore determined after Game Round 3, once the relevant Game Round 2 data are available.
In the Good treatment, participants are informed that one Klee group member and one Kandinsky group member chose the prosocial option in the Prisoner's Dilemma in Game Round 2. In the Bad treatment, participants are informed that one Klee group member and one Kandinsky group member chose the non-prosocial option. Since this refers to a single decision out of 20 made in the Prisoner's Dilemma in Game Round 2, the information is statistically irrelevant.
To eliminate second-order strategic reasoning, decisions in Game Rounds 3 and 4 are matched against recorded decisions from Game Round 2, in which group identity was not yet revealed to participants. Any behavioral change between Game Round 3 and Game Round 4 therefore reflects the participant's own belief updating in response to the anecdote.
To explore the effect of a large (i.e., almost continuous) strategy space, we include two additional games after game round 4, starting with session 17 (n = 361 collected): an all-pay auction as a strategic game with a continuous strategy space and a non-strategic dictator game with a continuous strategy space.
We elicit participants' cooperative and competitive behavior towards the in- and out-group in an all-pay auction with no information (i.e., neither player is informed about the group memberships), one-sided information (i.e., the decision-maker is informed, but the opponent is not informed about the group memberships) and two-sided information (i.e., both are informed about the group memberships and this is common knowledge).
We elicit participants' generosity towards in- and out-group members with no information (i.e., neither player is informed about the group memberships), one-sided information (i.e., the decision-maker is informed, but the recipient is not informed about the group memberships) and two-sided information (i.e., both are informed about the group memberships and this is common knowledge; the recipient is informed through a message shown at the end of the study).
|