We assessed the association of personality pathology with romantic couples' observed interpersonal behaviors. Couples engaged in four discussion tasks, after which observers used the Continuous Assessment of Interpersonal Dynamics method to continuously rate each participant's dominance and warmth over the course of each discussion. Using these ratings, we derived indices of average behaviors and changes in behaviors over the course of discussions. Generally, results indicated that the more personality pathology either spouse reported, the colder husbands were on average, and the colder they became toward their wives over time. However, personality disorder symptoms and overall interpersonal problems were largely unassociated with wives' behaviors. Results also indicated that the more dominancerelated problems husbands and wives reported, the more dominantly and coldly they behaved, the more submissive or withdrawn their partners were, and the colder wives became over time; and the more warmth problems wives reported, the more dominantly, they behaved.

Results suggest that social anhedonia may undermine the functioning of romantic relationships by reducing positive sentiments toward partners and security in the partner's sentiments toward the self.

In addition to replicating examinations of gender bias in the diagnosis of all cluster B personality disorders (PDs), this is the first study to examine the extent to which patient sexual orientation biases the diagnosis of antisocial, histrionic, and narcissistic PDs as well as whether or not such sexual orientation bias differs by patient gender. Furthermore, this study is the first to examine how such gender and sexual orientation biases are moderated by (1) the model of personality pathology used (i.e., traditional DSM vs. dimensional Alternate Model of Personality Disorders [AMPD]) and (2) measurement specificity (i.e., global PD measurement vs. symptom-level measurement). To undertake these examinations, it utilized a vignette describing a patient whose gender identification (man or woman) and sexual orientation (heterosexual or gay/lesbian) were experimentally manipulated. Clinicians (N=435) were randomly assigned to examine one of the resultant four vignettes, after which they each completed three measures of personality pathology. Though there was evidence of gender bias, such bias was twice-to-four times as weak as gender bias found in past similar studies. There was no evidence of significant diagnostic bias based on patient sexual orientation and sexual orientation bias did not differ by patient gender. Broadly, neither gender nor sexual orientation bias was moderated by the model of personality pathology underlying the measures used, by the specificity with which the pathology was measured, or by clinician characteristics (i.e., age, gender, sexual orientation, licensure status, race). Results suggest a decrease in gender and sexual orientation bias within experimental contexts relative to that which was found by prior studies. Further examinations should elucidate the mechanisms moderating diagnostic bias.

Sunk costs have been known to elicit violations of expected utility theory, in particular, the independence or cancellation axiom. Separately, violations of the stochastic dominance principle have been demonstrated in various settings despite the fact that descriptive models of choice favored in economics deem such violations irrational. However, it is currently unknown whether sunk costs also yield stochastic dominance violations. In two studies using a tri-colored roulette wheel choice task with non-equiprobable events yet equal payoffs, we observed that those who had sunk costs selected a stochastically dominated option significantly more than did those who had no costs. Moreover, a second study revealed that people chose a stochastically dominated option significantly more when the expected value was low compared to high. A model comparison of psychological explanations demonstrated that theories that incorporate a reference shift of the status quo could predict these sunk cost-based violations of stochastic dominance whereas other models could not.

No abstract