Thinking and Deciding

444 SOCIAL DILEMMAS: COOPERATION VERSUS DEFECTION
Table 18.2: Effects of cooperation (C) or defection (D) on subjects’ pay
Bob chooses C Bob chooses D
Art’s pay Bob’s pay Art’s pay Bob’s pay
Art chooses C $8 $8 $2 $10
Art chooses D $10 $2 $4 $4
Effects of repetition
In the typical prisoner’s dilemma laboratory task (and in some real-world analogues
of it, such as arms races) the game is played repeatedly by the same two subjects.
This leads to various attempts on the part of each player to make the other player
cooperate. Players adopt various strategies to induce cooperation: A strategy is a rule
that determines which choice a player makes on each play. Axelrod (1984) argues
that one of the most effective strategies, both in theory and in fact, is the “tit for tat”
approach. You begin by cooperating, but after that you imitate the other player’s
choice on each successive play, thereby “punishing” uncooperative behavior.
If such a strategy is effective in getting the other player to cooperate, the game
is not really a social dilemma at all, for the strategy itself becomes an action that is
best for each player. When games are repeated, it is helpful to think not of individual
plays but rather of strategies as the choices at issue. When we study performance
in laboratory games, then, we must be aware of the fact that repeated games may
not actually involve social dilemmas at all. The laboratory studies I shall review,
however, use nonrepeated games and are therefore true social dilemmas.
N-person prisoner’s dilemma
The basic prisoner’s dilemma can be extended to several people, becoming the “Nperson
prisoner’s dilemma.” (N stands for the number of people.) In one form of this
game, each choice of Option C — as opposed to Option D — leads to a small loss
for the player who makes the choice and a large gain for everyone else overall. In the
simplest version of this, imagine a class with twenty students. Each student writes C
or D on a piece of paper without knowing what the other students have written. Each
student who writes C causes each of the other students to receive $1. Each student
who writes D gets $1 in addition, but this has no effect on the payoff for others. If
everyone writes C, each gets $19. If everyone writes D, each gets $1. But writing D
always leads to $1 more than writing C. If you write D and everyone else writes C,
you get $20 and everyone else gets $18.
Table 18.3 and Figure 18.1 illustrate a four-person game in which the benefit of
cooperation is $4 for each other player and the benefit of defection is $2 for the self.
The slope of the lines represents the benefits that accrue to others from each person’s
decision to cooperate rather than defect; each person’s decision to cooperate moves