Examining Friendship Dynamics in Social Anxiety with Iterated Games of the Prisoner’s Dilemma

Abstract

Individuals with social anxiety have been shown to have higher levels of friendship impairment and greater difficulty establishing close relationships than persons without social anxiety. However, the mechanisms associated with such impairment have not been widely examined. Previous research suggests that deficiencies in prosocial behaviors (e.g., low warmth, limited self-disclosure, and constrained cooperation) during interpersonal exchanges may partially explain their difficulties developing close relationships. The present study aimed to examine the effect that rejection may have on prosocial behaviors, as well as other factors associated with developing and maintaining friendships, including trust, perceived likeability, closeness/connectedness, using an iterated, computerized version of the Prisoner’s Dilemma (PD) game. Participants with high (n = 56) and low (n = 35) social anxiety were asked to play the PD game with another participant (actually an experimental confederate) whom they met at the start of the study. Participants were led to believe that they were playing the PD game against the other participant but were actually playing against a computer that was programmed with a strategy meant to initially facilitate cooperation. Cooperation, or giving, in the PD game was measured primarily by the number of tokens that the participant shared with his/her partner. Additionally, participants were randomized to either rejection or non-rejection conditions. Partway through the study, those in the rejection condition were exposed to a programmed decrease in giving by their partner as well as an ambiguous in-person rejection (administered by the experimenter). Participants in the non-rejection condition were not exposed to either the programmed or in-person manipulations. Outcomes of interest included total giving during the PD game, constriction of giving, and the use of particular strategies following rejection in the PD game, as well as various measures of relationship quality collected at the end of each round of play. It was hypothesized that individuals with high social anxiety in the rejection condition would exhibit less total giving and a constricted response to low partner giving and also be less likely to use prosocial strategies to encourage cooperation (i.e., a coaxing strategy) following rejection by the partner compared to those with low social anxiety in the rejection condition. Additionally, we anticipated that those with high social anxiety in the rejection condition would provide lower ratings of the following relationship quality domains following the rejection condition: trust, closeness/connectedness, and perceived likeability than those with low social anxiety in the rejection condition. Overall, results provided little support for these hypotheses. However, there were several significant main effects that highlighted differences among those with high and low social anxiety. For example, individuals with high social anxiety had greater ratings of the expectancy of future rejection and of the importance of their next turn for maintaining the quality of their relationship with their partner. Additionally, there was a trend level (p = .08) social anxiety group by rejection condition interaction on the participant’s trust of his/her partner, such that those with high social anxiety exhibited reductions in relationship trust following rejection whereas those with low social anxiety did not. Implications of these findings as well as limitations and future directions of study are also explored.