Methodological shortfalls and differences across the studies make statistical appraisal of the results difficult. Recidivism rates were dependent on variables such as the selected sample population, the definition of recidivism, and the length of follow-up period after release for detention. Nevertheless, results revealed a consensus that incest offenders are less likely to reoffend than are nonfamily child molesters. Rapists and exhibitionists were thought to be at a higher risk for recidivism than are child molesters. Figures varied across the studies, but the overall sex-offense recidivism rate was probably well under 10 percent. Findings suggested that a combined actuarial predictive approach in conjunction with empirically guided clinical assessment is probably the best method to predict sex offender recidivism. Psychometric and physiological instruments such as the Psychopathy Check List-Revised Scale, the Michigan Alcohol Scoring Test, and penile phallometry should be part of the empirically guided clinical risk assessment of sex offenders. 57 references (Author abstract modified)