Are police-reported driving while Black data a valid indicator of the race and ethnicity of the traffic law violators police stop? A negative answer with minor qualifications

Are police-reported driving while Black data a valid indicator of the race and ethnicity of the drivers police stop? This research answered that question by advancing the first multivariate analysis of race and ethnicity missingness in the traffic stop data reported by Boston police during April and May of 2001. The most important multivariate story the data tell was that race and ethnicity missingness was significantly nonrandom on multiple dimensions, including the second month of data collection, for drivers living in zip codes with above average and average people of color, for drivers living in zip codes with above average and average poor people, and for drivers whose stop ended in a ticket. The results therefore supported a clear answer to a fundamentally important question about the validity of the driving while Black data reported by police. Based upon the present research and with minor qualifications, police-reported driving while Black data were not valid because they underestimated the frequency with which police stop drivers of color.