This study compared changes in psychometric scores over treatment for a sample of convicted male sexual offenders who had attended a traditional relapse prevention (RP) program in the community, to those who had attended a revised version of the program based on the good lives (GL) model of offender rehabilitation. The sample consisted of a total of 601 offenders either serving sentences in the community or on probation in the community following release from custody. The sample attended one of two programs operating in the community, both of which delivered an RP and GL version of the program. The groups were compared on their level of psychometric change over treatment on individual measures, on measures grouped by dynamic risk domain, and on overall psychometric change, using a variety of analyses, including examination of clinically significant change. There were no differences in amount of change over treatment or, for those deemed as requiring change, clinically significant change, by program approach, for the majority of the measures examined. However, chi-square analyses indicated that there was an association between attending the GL version of the programs and sustaining a functional score across a number of measures of pro-offending attitudes and across measures of socioaffective functioning, posttreatment. In addition, achieving an overall "treated profile" on the psychometrics posttreatment was associated with attendance on the GL versions of the community programs, although the GL starters were less dysfunctional than the RP starters. Attrition rates were low for both the RP and GL programs, and there was no difference in rates between the two.