Excluded and ashamed: Shame proneness interacts with social exclusion and testosterone reactivity to predict behavioral aggression.

Exclusion from social relationships is a painful experience that may threaten an individual's status and dominance. The steroid hormone testosterone, which fluctuates rapidly in response to such threats, may be implicated in subsequent behavioral action (e.g., aggressive or prosocial responses) that aims to protect or enhance one's status after exclusion. Past research, however, indicates that the link between acute changes in testosterone and behavior depend on context-relevant individual dispositions. In the context of social exclusion, an individual's level of shame proneness-characterized by a tendency to experience shame and to react submissively-is theoretically relevant to the testosterone-induced aggression relationship but has yet to be examined empirically. Here, men (n = 167) were randomly assigned to be socially included or excluded in the virtual ball-tossing game, Cyberball, after which aggressive behavior was examined using the Point Subtraction Aggression Paradigm (PSAP). Testosterone reactivity was measured via salivary hormone samples collected pre- and post-game. Moderated multiple regression analyses were run to examine the extent to which testosterone reactivity and shame proneness moderated the effect of Cyberball condition on aggression. Results revealed a significant two-way interaction between Cyberball condition and testosterone reactivity, as well as a three-way interaction including shame proneness. For individuals low in shame proneness, exclusion was associated with higher post-cyberball aggression among those who experienced a rise in testosterone but was associated with lower post-cyberball aggression among those who experienced a decrease in testosterone. For individuals high in shame proneness, however, exclusion did not meaningfully affect aggressive responses, regardless of whether they experienced an increase or decrease in testosterone. These findings extend our understanding of the moderating roles of context and disposition on the neuroendocrinology of aggression in social interaction.