Studies of life span development in everyday problem solving suggest two trajectories of change in adulthood: individuals become less effective at solving well-defined instrumental problems but more effective at managing ill-defined interpersonal problems. Two experiments were conducted to examine the ability of young and older adults to effectively manage an interpersonal problem that has a well-defined measure of instrumental success. Participants played an iterated Prisoner's Dilemma game with same-age, computer-simulated strangers (Experiment 1) and friends (Experiment 2). Success was dependent upon one's ability to put aside self-interest and cooperate with a partner. Computer-simulated partners reciprocated the participants' decisions 100% of the time or behaved in a more self-interested manner. Young and older adults' tendencies to create conflict with the reciprocating partner and their defensive reactions to the selfish partner were examined. Although young adults outperformed older adults when playing the game on their own, they did not carry this performance advantage into the interactive rounds. In fact, despite their success when playing alone, young adults were no more successful than older adults when interacting with others. Young and older adults both cooperated more with friends than with strangers and more with the reciprocating partner than the selfish partner. However, when the participants' first interaction was with a selfish stranger, older adults were more cooperative than young adults and consequently accrued more reward. This is consistent with previous research demonstrating that older adults use more passive interpersonal problem solving strategies than young adults, and it also partially supports the prediction that advancing age leads to more effective strategy implementation when solving interpersonal problems.