The authors examined gender bias in the diagnostic criteria for (4th ed. 32 diagnostic criteria). Results were not significant for pathology ratings of the four criteria sets (and were significant for only three of the 32 criteria at = .05). The second study was almost identical, except that criteria from the antisocial, histrionic, narcissistic, and dependent disorders were examined. Results were largely consistent with those of the first study. Infrequency was statistically different for the antisocial, narcissistic, and dependent criteria, but there were no differences by gender for pathology ratings on any of the four criteria sets. The authors concluded that the diagnostic criteria from these five disorders have differential sex prevalence rates. but professional clinicians who apply these diagnostic criteria to men and women do not perceive the diagnostic criteria as having different implications for maladaptivity or impairment (Anderson et al., 2001, p. 667). Only one study has examined differential prevalence ratings on all criteria 677297-51-7 IC50 as reported by participants using questionnaires (Morey, Warner, & Boggs, 2002). Gender differences reached significance for 9 of the 79 criteria. As in previous studies (Anderson et al., 2001; Sprock, Crosby, & Nielsen, 2001), participants were also asked to provide pathology ratings (e.g., a 677297-51-7 IC50 man with this characteristic would have much more trouble functioning than a woman with this characteristic and vice versa). Results were largely nonsignificant, but when a criterion was viewed as more problematic for one gender, it 677297-51-7 IC50 also tended to be more prevalent in that gender. Morey et al. (2002) concluded that extremes of sex-typed behaviors are viewed by others as most problematic, and personality problems simply tend to manifest differently in men and women (p. 62). Gender bias in the diagnostic criteria for borderline, schizotypal, avoidant, and obsessive-compulsive PDs has also been studied by Boggs et al. (2005). Using data from the Collaborative Longitudinal Personality Disorders Study, these investigators examined relations among diagnostic criteria (measured using semistructured diagnostic interviews) and levels of 677297-51-7 IC50 functional impairment in male and female patients. The data indicated relatively little evidence of gender bias. In other words, specific diagnostic criteria were associated with equivalent levels of impairment in men and women. In the current study, we investigated whether gender bias is associated with diagnostic criteria for PDs using differential item functioning (DIF), a psychometric method for evaluating whether a construct is expressed equivalently across different groups. Because it is assessed within an item response theory (IRT) framework, DIF can determine Thbs4 whether a person’s response on an item (in this case, a PD criterion) depends on both his or her trait 677297-51-7 IC50 level (level of personality pathology) and group membership (gender). In other words, given the same level of personality pathology, are men more likely than women to endorse a PD criterion (or vice versa)? DIF Using IRT DIF occurs when two individuals with the same trait level but different group membership do not have the same probability of endorsing a test item. In the current study, the detection of DIF in the PD criteria was calculated using IRT.1 In IRT, an individual’s trait level, , is estimated from responses to items, and the model specifies how both the trait level and item properties are related to an individual’s item responses (Embretson & Reise, 2000). IRT relates the characteristics of the items (in this case, PD diagnostic criteria) and the characteristics of individuals (here, gender) to the probability of endorsing the individual items (Zickar, 1998). Although a full conversation of IRT and DIF is definitely beyond the scope of this article, excellent evaluations and articles pertaining to this methodology are available elsewhere (observe Embretson &.