The present research examined three factors that were thought to affect cooperation in the replicated prisoner’s dilemma game: (1) motivation for participating, (2) understanding of the game, and (3) reward level. Results indicated that both high-understanding instructions and recruiting Ss to participate for money (rather than research credit) promoted cooperation, while reward level had no significant effect. These findings were used to argue that past studies, which tended to show low levels of cooperation for less than 50 replications, generated results that do not validly model real-life conflict.
Messé, Lawrence A.; Bolt, Martin; and Sawyer, Jack, "Nonstructural determinants of behavior in the replicated prisoner’s dilemma game" (1971). University Faculty Publications and Creative Works. 516.