Abstract / Synopsis
We model prejudice in the context of an iterated prisoner’s dilemma tournament. Prejudice here is defined as the inability to distinguish amongst members of an identifiable group. We run a computer simulation where agents of one of two groups with defined strategies are randomly matched against each other. The agents either cooperate with or defect against the agent they are matched with. Our focus is on agents who play an unprejudiced version of the Tit-for-Tat strategy (cooperate with any player in the first encounter, and then apply the same strategy [defect or cooperate] that the opposing player played against them in the last encounter) and a prejudiced strategy (cooperate with any member of the other group in the first encounter with a member of the other group, and then apply the same strategy [defect or cooperate] that any member of that group played against them in the last encounter with a member of that group). We find that even though agents are initially inclined to give members of another group the benefit of the doubt when interacting with them, the introduction of very minor levels of defection, real or perceived, can lead to universal defection against members of the other group.
Recommended Citation
Alan Lockard & Edwin Harcourt, "A Note on Modeling Prejudice," Journal of Humanistic Mathematics, Volume 16 Issue 1 (January 2026), pages 203-217. . Available at: https://scholarship.claremont.edu/jhm/vol16/iss1/12
