Author ORCID Identifier

Date of Award

Spring 5-2-2022

Degree Type


Degree Name

Doctor of Philosophy (PhD)



First Advisor

Erin B. Tone

Second Advisor

Jessica A. Turner


Recent evidence suggests that individuals who exhibit socially anxious (SA) symptoms endorse patterns of maladaptive interpersonal behavior that can be parceled into three subtypes based upon interpersonal circumplex theory: friendly-submissive, hostile-submissive, and hostile-dominant. It remains unclear, however, whether these subtypes translate into observable social behavior in laboratory contexts. I used two economic-exchange tasks, the prisoner’s dilemma game (PDG) and the ultimatum game (UG), as models of domains of social behavior to detect interpersonal differences in a sample of college students (N= 88) who endorsed mild-to-severe levels of SA based upon responses to the Liebowitz Social Anxiety Scale Self-Report (LSAS-SR). Using a two-step automatic clustering procedure, the sample was divided into three groups according to their responses on the Inventory of Interpersonal Problems – 32 (IIP-32). Interpersonal profiles were constructed for these groups and two of the three expected subtypes were identified (friendly-submissive and hostile-submissive); however, instead of hostile-dominance, friendly-dominance emerged as a potential subtype. Hierarchical and quantile regressions were conducted to examine whether SA severity and interpersonal subtype predicted cooperation and acceptance rates in the PDG and UG respectively. The data revealed that in the PDG, SA severity significantly predicted an increase in cooperation rate, while the interpersonal subtypes did not have a significant effect. However, when analyses included only those individuals who met a clinical cutoff for severe SA (N = 66), SA severity no longer predicted cooperation rates. But friendly-submissiveness predicted cooperation rates exceeding 65% during gameplay, while friendly-dominance predicted a ceiling cooperation rate of 65%. Hostile-submissiveness did not predict variance in cooperation rate. In the UG, the interpersonal subtypes and SA severity did not significantly predict acceptance rate. These findings build upon a burgeoning literature substantiating links between self-reported interpersonal problems and unique interindividual psychopathological presentations. However, improvements in sample recruitment, the implementation of economic-exchange tasks, and data-analytic methods need to be put into practice before stronger assertions can be made concerning the therapeutic relevance of these games as social decision-making paradigms.


File Upload Confirmation