This research compared independent ratings of criminal psychopathy (Hare's Psychopathy Checklist, Hare, 1991) from National Parole Board case files alone with ratings based upon file information plus a semi-structured interview. Notwithstanding high interrater reliability using National Parole Board (NPB) case files alone (n = 35), Psychopathy Checklist (PCL-R) scores had to be prorated because 30% of the items could not be scored. Comparisons between file only and independent file plus interview ratings of criminal psychopathy for a larger sample (N = 120) resulted in relatively low interrater reliability. Interrater agreement for ratings of psychopathy with and without an interview was statistically significant (p < .006), yet 40% of the cases received different diagnoses when the different procedures were used. Contrary to earlier findings (Wong, 1988), file only PCL-R ratings were not routinely an underestimate of file plus interview PCL-R ratings.