Although the findings suggest emerging evidence that DWI courts may have beneficial effects on participants' behavior, the findings are questionable due to flawed research methodology. Only 5 of 14 evaluations that met eligibility criteria were determined by independent raters to have used good to marginally acceptable research methodology, and several of those evaluations still had serious flaws. These included evaluating immature programs, failing to conduct intent-to-treat analyses, and the failure to maintain intervention and nonintervention conditions across treatment and control groups. Most of the evaluations reviewed provided insufficient information for determining how DWI courts work and for what types of offenders. Seventy-nine percent of the evaluations failed to report any information on the level of services received by participants, such as the number of counseling sessions or status hearings attended. The evaluations also generally limited outcome analyses to recidivism rates and graduation rates. There were no measures of indirect or short-term outcomes, such as counseling attendance or abstinence rates and whether these effects impacted long-term outcomes. What is needed are scientifically defensible studies that can objectively establish the effects of DWI courts compared to adjudications as usual and as compared to alternative intervention approaches. This article presents a table that shows best practices for the field of evaluation research, which should be used to guide future designs of DWI court program evaluations. Each evaluation report was scored for methodological rigor by at least two trained independent raters, using established scientific criteria for evaluation research. 4 tables, 1 figure, 63 references, and addendum