Addressing replicability concerns via adversarial collaboration: Discovering hidden moderators of the minimal intergroup discrimination effect

In response to a series of studies failing to replicate the well-known ingroup favoritism effect in the minimal group paradigm, an adversarial collaboration was undertaken to identify possible moderators of the effect. The variables examined were prominent differences between studies undertaken by the co-authors that had vs. had not observed the effect. The resulting study found that the tendency to allocate more rewards to members of one's minimal group (vs. allocating equally to minimal in- and outgroup members) was stronger when participants were not socially isolated, when they received written (vs. verbal) instructions, and among Americans (vs. Australians). These results are consistent with certain theories of ingroup favoritism, and contradict others, particularly those that rely on some universal, context-independent psychological process (e.g., striving to enhance self-esteem following mere social categorization; evolved xenophobia). The study also illustrates the utility of adversarial collaborations, particularly when issues of replicability arise.