The interplay of social group biases in social threat learning

Abstract

Learning from other individuals (e.g. social learning) is subjected to biases affecting whom to learn from. Consistent with research in animals, showing similarity-based learning biases and a general tendency to display pro-social responses to in-group individuals, we recently demonstrated that social learning of both fear and safety was enhanced when information was transmitted between same-race individuals. Here, we addressed how two different social group categories jointly affect the transmission of fears by investigating the interplay between racial and supporter group membership. We demonstrate that supporter group membership differentially influenced learning from a racial in-group vs. racial out-group individual. Thus, conditioned skin conductance responses in the same-race condition were significantly higher when fear was transmitted by an in-group (same team) vs. an out-group (rival team) individual, and were related to supporter team identification. However, supporter group membership did not influence learning from a racial out-group demonstrator, suggesting that the presence of an alternative alliance does not necessary reduce the influence of racial biases on social fear learning.

Figure 1

Mean CR (CS+ > CS−) across the experiments displayed for the Racial In-group and Racial Out-group separately as a function of Supporter group membership. (A) During observational acquisition SCRs to the CS+ were overall higher than to the CS− in the absence of significant group differences. (B) During the early stage of the direct test, supporter group membership significantly modulated CRs but only when fear was transmitted from a racial in-group demonstrator (C). During the late stage of the direct test, there were no between group differences left. Errors bars denote standard error of the mean (SEM). *p < 0.05.