Gender differences have been observed in the frequency and severity of psychopathy. Research suggests that levels of psychopathy, as measured by the PCL-R, are substantially lower among women than men. A study sample was composed of 119 female inmates from a large metropolitan jail in Texas. The following confirmatory factor-analytic approach was used to test the fit of three models: an established two-factor model derived from male samples; a proposed two-factor model for female psychopathy; and a proposed three-factor model, which was tested almost exclusively on male offenders. Results show that the prevalence rate of psychopathy is consistent with previous research. Females are much less likely than their male counterparts to be classified as psychopaths. The two-factor models of psychopathy represented a poor fit for the current data. In contrast, the three-factor model appears to have considerable promise in capturing the underlying dimensions of psychopathy among female offenders. These data question whether current evaluations accurately represent the underlying dimensions of psychopathy for all individuals. The lack of programmatic research on the predictive validity of the PCL-R with females should persuade clinicians to be especially circumspect in utilizing psychopathy for risk assessment in female offender and clinical populations. Future research must continue to refine the underlying dimensions of psychopathy and also test their criterion-related validity. Before a three-factor model of psychopathy is adopted for clinical and forensic practice, its clinical usefulness must be demonstrated, especially regarding predictions of recidivism, interpersonal and adjustment, and treatability. 2 tables, 4 notes, 27 references