The study, conducted by researchers at North Carolina State University and North Carolina Central University, found some troop districts and individual troopers with higher citation rates of African-Americans than would normally be predicted. The study also found troop districts and individual officers with lower-than-predicted citation rates of African-Americans. This disparity, the researchers say, could be attributed to a number of factors, including where and when troopers patrol certain areas, the demographics of the road itself, time of day, day of the week, and the behavior of drivers. The researchers found higher-than-expected rates of searches of African-American vehicles for 1997, but also found that more recently the highway patrol searched fewer vehicles driven by African-Americans. No systematic bias in searches is now evident, the researchers say.
The researchers could not rule out individual officer bias processes that could lead to greater scrutiny of the behavior of vehicles operated by African-American drivers. A fuller examination of such processes, in particular a largely unconscious "cognitive bias" that stimulates more attention to the behaviors of African-Americans drivers, would require information that was not contained in the statistical data available to the researchers, they say. If such bias is occurring, it is not so pervasive as to produce large-scale racial disparity in highway patrol stops.
The high rate of citations of African-Americans in several districts could be due to practices of consistently patrolling areas of a highway where African-American drivers are overrepresented. High levels of patrolling areas where African-Americans are more likely to drive could, in turn, account for some of the high rates for the few areas with relatively high disparity. The reasons for patrolling some areas more than others may be due to the high accident rates of such highways, or other factors such as the availability of safe areas for stopping vehicles.
"There is no pattern to suggest wholesale bias or discrimination across the highway patrol as an organization," said Dr. Matthew Zingraff, professor of sociology at NC State and the primary investigator of the study. "There are outliers at both ends - some troops and some individual troopers stop more African-Americans than expected and some stop fewer. Still, there are also some disparities not accounted for by our study. On balance, the strongest factors appear to be when the troopers are on the road and where they are deployed. A fuller understanding of the relationships among who is on the road, what drivers are doing, and enforcement deployment patterns are essential to the study of racial profiling."
Dr. Donald Tomaskovic-Devey, professor of sociology at NC State and a co-investigator of the study, suggests that if there is bias in police stops by the North Carolina State Highway Patrol, it is most likely to be the work of individual officers, and is best addressed as a managerial problem solved by monitoring the race composition of stops by individual officers.
The North Carolina Highway Traffic Study - which can be accessed on the Web at www.chass.ncsu.edu/justice - was conducted by Zingraff and Tomaskovic-Devey and fellow NC State sociologist Dr. William R. Smith, as well as Dr. Harvey McMurray and C. Robert Fenlon from N.C. Central University. A $472,231 grant from the National Institute of Justice supported the study.
While the researchers examined the utility of several baselines constructed to represent the demographic compositions of drivers, they concentrated on drivers in an area who were involved in reported accidents as their primary baseline against which to compare the percent of drivers who were cited. "We find a very high correlation between the percent of accidents involving African-American drivers and the percent of those issued citations who are African-American," said Smith. "Accident data, where we can identify the race of each driver involved, are probably the best source of information about who is on the highway in specific locations and at specific times." This is because these are also areas where highway patrol officers are patrolling. African-Americans have a slightly lower rate of accidents than do whites, but are more likely to be driving in areas of the state patrolled by the highway patrol.
The study had two goals:
To determine if N.C. State Highway Patrol officers engaged in racial profiling in making their traffic stops; and
To determine North Carolina citizens' thoughts about racial profiling
The report found that, across the state as a whole, African-Americans were stopped and cited at higher rates than were whites, relative to their representation as drivers. African-Americans, for example, comprised 21.2 percent of all licensed North Carolina drivers, yet they received 24.9 percent of all citations from troopers in 2000, the data showed. The researchers show that trooper deployment is an important factor in accounting for racial disparity. They also suggest that driving behavior may account for some of the disparity, although they cannot rule out the influence of some cognitive bias in police selection of whom to stop. Zingraff says the data provide no direct measures of individual trooper decision-making processes, including potential bias processes.
"There is an amount of statistical disparity that cannot be explained by our analytic models," Zingraff says. "Cognitive bias could be the source but we can't substantiate that currently." Some districts and some officers with particularly high citation rates of African-Americans might warrant further study and possibly ongoing or intermittent monitoring, the study says.
The study also examined automobile searches of troopers, including the Highway Patrol's Criminal Interdiction Team (CIT). Researchers found that the CIT troopers are more apt to search a vehicle driven by an African-American than by a white, although the proportion declined dramatically over a four-year period from 1997 to 2000. Requests for consent to search sometimes result from a conversation with the driver and/or occupants of the vehicle. This question-and-answer method - where a person's perceived nervousness or lack of consistency in what is said could lead to a search - is another area where cognitive bias could be operating, the researchers say. Data also revealed that the CIT was less likely to find contraband in a vehicle driven by an African-American in 1997, but by the year 2000 the race differences in hit rates was very small.
To obtain a more thorough story than one told solely by the numbers, the study also included a telephone survey of North Carolina drivers and focus groups with citizens and troopers.
In surveying citizens, the researchers discovered a greater level of disparity in reported police stops among local police than among the North Carolina State Highway Patrol. African-American drivers are slightly more likely than white drivers to report a traffic stop involving the N.C. State Highway Patrol (10 percent vs. 8.2 percent). This closely parallels the official data results for highway patrol stops. Local police were much more likely to have stopped African-Americans relative to whites among those surveyed (18.8 percent vs. 11.2 percent). Stops by the highway patrol were strongly associated with driving behaviors. Stops by local police were strongly tied to race, gender, age, and the age of the car being driven. African-Americans were slightly more likely to report having been told by the officer that the stop was for a discretionary reason - rather than a substantial moving violation like speeding, for instance. They also report more instances of disrespect by the officer after the stop than do whites.
The citizen survey also found that experiencing disrespectful treatment during a stop, learning of disrespectful stop experiences of friends or family, and the perception that police racially profile all dramatically reduce trust in the police. These results are identical for African-American and white citizens, but African-American citizens are more likely to report these experiences and beliefs.
Focus groups provided another avenue for researchers to learn about citizen views on racial profiling. These discussions made it clear to the researchers that African-American citizens' perceptions of biased policing are influenced by not only their own experiences, but also the experiences of family and friends, a more general expectation of racial bias in all walks of life, and a specific expectation of police bias.
Troopers, in their focus group sessions, said that the nature of their work is one in which they are largely reacting to the behavior of vehicles. In such instances, they stress, it is not easy to determine the race of a driver until after the decision has been made to stop the vehicle, and often not until the trooper approaches the vehicle. Regular road troopers expressed a general reluctance to search vehicles, and the rarity of searches in part validates this claim. The troopers believe that the quality of their work can be assessed through record reviews by senior officers, the use of cameras in patrol cars, court visits to assess the quality of evidence and charges, and occasional ride-alongs by sergeants and supervisors.
Dr. McMurray said that it's difficult to draw strong conclusions from trooper focus groups. "Overall, troopers who participated in focus groups believed that it is likely that there are some incidents of racial profiling; however, they believe that such incidents are infrequent. Two additional issues emerged from our analysis of the trooper focus group discussions. First is the myth or belief that a specific lower-income person or person of color is more likely to commit an offense than his or her counterpart. One need not be a racist or ill-intended to make race-based law enforcement decisions and, whether intentional or not, the impact might constitute racial profiling. Second, there is the extent to which other extralegal factors, such as social status and community contexts, can consciously or unconsciously impact law enforcement decision-making and lead to disparities in outcomes. For example, is one as likely to set up a DUI checkpoint near a country club as compared to the neighborhood bar on a given Saturday night?"
Zingraff cautions that the data collected and conclusions drawn are particular to the North Carolina State Highway Patrol and to the reports of North Carolina citizens, and are not applicable necessarily to other law enforcement agencies within North Carolina or other states.