Preliminary results from a longitudinal study of the effectiveness of cognitive-behavioral treatment with sex offenders are presented. The study's research design included 3 groups: a treatment group, a volunteer control group (those who volunteered for but did not receive treatment), and a nonvolunteer control group (those who refused the opportunity for treatment). Although the treatment group had the lowest reoffense rates for both sex and other violent crimes, main effects analyses did not yield conclusive results regarding the program's effectiveness. The results highlight the importance of including appropriate comparison groups, managing attrition from both treatment and methodological standpoints, examining sex and other violent offenses separately as outcome variables, employing tests with adequate statistical power, and analyzing data while taking into account time at risk for reoffense. (PsycINFO Database Record (c) 2019 APA, all rights reserved)