The construct of psychopathy is essential in explaining criminal behavior, but unfortunately the empirical research on psychopathy in women has been inconsistent. In this study the underlying structure of psychopathy in women was examined by testing the two-factor model by Hare (2003) and the three-factor solution by Cooke and Michie (2001) using confirmatory factor analysis. We replicated the study by Warren et al. (2003) using a nationwide sample of 97 female homicide offenders in order to facilitate the comparison of results. The prevalence of psychopathy in the present study was 9.3% with a cut-off of >or=30 and 21.6% with a cut-off of >or=25. The best fit for the data out of the tested models was the three-factor model with six testlets. The two-factor model proved to be too simple a model for the female homicide data. The findings regarding comorbidity of psychopathy with personality disorders show that the concept of psychopathy includes diagnostic criteria of several personality disorders, but further research is needed to establish a possible superordinate dimension. Further research on the PCL-R and putative gender differences in the expression of psychopathy in women and men as well as on the putative impact of cultural differences on the instrument is clearly needed.