Data availability and quality have improved over the 3 years of the evaluation. Remaining data issues are largely due to equipment limits or relatively infrequent human errors. The evaluation did not find evidence of departmentwide racial bias in officer decisions to stop certain vehicles in 2006. Similarly, data from the entire evaluation period (the first-year report included data from 2003) did not show departmentwide bias in decisions to stop vehicles. The report recommends identifying and giving attention to officers whose vehicle-stopping patterns in relation to a driver's race depart from the departmentwide norm. A comparison of all stops of Black and non-Black drivers shows that the stops of Black drivers take longer on average, and Black drivers are subject to searches at a higher rate; however, much of the differences are apparently due to the location and time of the stop, the type of stop, whether the driver was a Cincinnati resident, and whether the driver had a valid driver's license. Videotaped police-motorist interactions showed that White officers were more likely than Black officers to use proactive police tactics during vehicle stops, and communication with White drivers was more apologetic and less argumentative than with Black drivers. These results from year 3 of the evaluation are largely consistent with the year-one and year-two evaluation findings. The evaluation has used a variety of methods, including a survey of citizen satisfaction with the police, a survey of citizens who have interacted with the police, and a survey of police officers about their perceptions of community support and working conditions. 5 figures and 30 tables