The field interrater reliability of three assessment tools frequently used by mental health professionals when evaluating sex offenders' risk for reoffending--the Psychopathy Checklist-Revised (PCL-R), the Minnesota Sex Offender Screening Tool-Revised (MnSOST-R) and the Static-99--was examined within the context of sexually violent predator program proceedings. Rater agreement was highest for the Static-99 (intraclass correlation coefficient [ICC[subscript 1] = 0.78) and lowest for the PCL-R (ICC[subscript 1] = 0.60; MnSOST-R ICC[subscript 1] = 0.74), although all instruments demonstrated lower field reliability than that reported in their test manuals. Findings raise concerns about the reliability of risk assessment tools that are used to inform judgments of risk in high-stake sexually violent predator proceedings. Implications for future research and suggestions for improving evaluator training to increase accuracy when informing legal decision making are discussed. (Contains 6 tables and 3 footnotes.)