The publication of a review of the treatment outcome literature by Furby, Weinrott, and Blackshaw in 1989 raised the issue of the failure of research to support adequately the efficacy of treatment for sexual offenders. Now, many years after the publication of Furby et al., there are still major flaws in the outcome research being conducted; this has led certain researchers to conclude that there is still no empirical evidence of the effectiveness of sexual offender treatment. Specifically, the current literature has been criticized for reliance on single-group, follow-up-only designs and designs that include nonequivalent treatment and comparison groups. These types of designs fail to provide adequate information about treatment outcome due to threats to internal validity and construct validity of putative causes and effects. Internal validity can best be maximized in offender outcome studies by taking care in posing hypotheses. When there is no control or comparison group, or when the comparison group differs from the treatment group in systematic ways, there are always threats to internal validity. These threats must be identified, measured, and, when possible, controlled. Differences in posttreatment behavior cannot be confidently attributed to treatment effects, since they may be due to extraneous factors or may result from an interaction of treatment with extraneous factors. The best way to test the effectiveness of a given treatment intervention is to identify the elements of that intervention and the desired changes expected as a result of participating in it. Outcome studies should then measure the implementation of an intervention, the realization of its goals, and finally, recidivism. Such a strategy can provide data from which more confident conclusions about treatment can be drawn and can allow for computation of effect size in a meta-analysis. 33 references