PA state police data shows no racial profiling in traffic stops: Study

The question of racial bias in police traffic stops is a highly debated issue. Some analyses have shown that certain racial or ethnic groups are disproportionately represented in traffic stops, leading many people to allege racial profiling and discrimination. However, it is not accurate or fair to claim that all police traffic stops are inherently racist. Other factors can contribute to racial disparities in police stops, including differences in driving behavior, geographic location, and crime rates in specific areas. As more research is conducted on this topic, it becomes more apparent how much context and other factors can play a role in traffic stop decisions. For example,  research that adequately accounts for the impact of contextual and situational factors has found that racial disparities may not be as pronounced as previously thought, and that they are often explained by other factors unrelated to race.

This was the case in a recent study examining traffic stops in Pennsylvania, which found no evidence of racial profiling. According to the full-length report, the Pennsylvania State Police (PSP) stopped more than 440,000 drivers in 2022, 78.5% of whom white. In comparison, 14.4% were Black, and 8.2% were Hispanic. To conduct the study, the PSP partnered with Dr. Robin Engel and the National Police Foundation to ensure that the evaluation was independent and external to the department. Overall, these data should inspire public confidence in the police. It also suggests that PSP’s approach could serve as a promising model for other agencies.

The document released earlier this week details the findings of statistical analyses conducted on PSP’s 2022 traffic stop data. The PSP initiated voluntary data collection efforts in 2002 in an initital attempt to promote transparency and accountability. This was discontinued in 2011, but was renewed in 2021 in light of recent debates regarding discriminatory policing. The study was conducted in partnership with the National Policing Institute (NPI) and researcher Dr. Robin Engel, which allowed for an independent and external evaluation of PSP’s practices and alleged racial profiling. This lends more credence to the results because the evaluation was conducted by people who are not internal to PSP.

Methods

Researchers examined the data for all traffic stops conducted by PSP from January 1 to December 31, 2022. NPI performed an initital audit of the data to confirm that it was valid and reliable, where they determined that PSP’s traffic stop data collection system is likely one of the most comprehensive and accurate in the entire country. The PSP’s data collection protocol goes beyond the minimum reporting standards that are used by other states and jursidctions, and includes some additional data fields that provide important context regarding the legal reasons for the stop, characteristics of the vehicle, and characteristics of the driver, passenger, and trooper involved.  Post-stop outcomes that were examined included verbal warnings, written warnings, citations, arrests, searches, and seizures.

It is important to note that this report does NOT use the traditional “benchmarking” approach that is frequently employed across other studies of racial bias. Benchmarking studies compare the racial/ethnic percentages of stopped drivers to an external benchmark representing the “expected” population, which is typically based on the demographic makeup of the residential population. However, there is a lot of contention regarding this approach, as residential population data has been shown to be flawed in determining a benchmark for drivers stopped for traffic offenses. Numerous factors such as driving location, time, frequency, quality, vehicle conditions, traffic conditions, and police priorities can influence the risk of being stopped for a traffic offense. No benchmark can adequately account for all these factors. Therefore, the report focuses on examining patterns and trends associated with trooper decision-making and post-stop outcomes, rather than using benchmark analyses. For a more detailed description of “benchmarking,” as well as the other methods for studying racial bias in policing, refer to my previous post.

Stop Characteristics

Pennsylvania State Police (PSP) stopped more than 440,000 drivers in 2022. The majority of people stopped by the police were male (67%), and almost everyone (98%) displayed civil behavior toward the officer during the traffic stop. Regarding race, the majority of people stopped were White (78.5%, 8.2% of whom were also Hispanic), while 14.4% were Black. The most frequent reason for the stop was speeding (40%), with an average speed over the posted limit of 21.4 mph. Moving violations accounted for 26.8% of stops, and equipment problems accounted for 18.8%. Most traffic stops occurred on weekdays (69%), during the daytime (66%), and on state highways (53%) or interstates (34%). The majority of stops lasted between 1-15 minutes (88%), involved vehicles registered in Pennsylvania (80%), and had no passengers (80%). There was some variation in stop characteristics, reasons for the stop, and driver characteristics across PSP organizational units. However, these differences are expected due to variations in geography, roadways, jurisdiction, traffic flow, and the demographic makeup of residents and travelers across the state.

Warnings, Citations, and Arrests

Researchers used different statistical techniques to examine verbal warnings, written warnings, citations, and arrests. First, bivariate analysis were used to examine the impact of race/ethnicity. Then, multivariate analysis was used to identify and understand the influence of various factors, such as drivers’ characteristics, stop reasons, and other relevant variables, on the probability of specific outcomes. Searches and seizures were examined separately because they comprised a small percentage of overall stops. There was some variation in outcomes across PSP organizational units, though, these differences are expected due to variations in geography, roadways, and jurisdictions managed by different units.

Overall, 56.8% of stops resulted in a warning being issued to the driver (38.3% written and 18.5% verbal), 57% resulted in a citation being issued to the driver, and 4.6% resulted in an arrest of the driver. They found statistically significant differences in outcomes across racial/ethnic groups. Whites were more likely to receive a citation or written warning (57.3% and 39.4%) than Hispanics (55.1% and 36.1%) and Blacks (54.3% and 36.7%). Blacks were more likely to receive a verbal warning or be arrested (21.2% and 6.6%) than Hispanics (19.7% and 5.8%) and Whites (17.7% and 4.3%).

However, these initial percentages do not account for other factors that can impact stop outcomes, such as legal reason for the stop. To further understand these complexities, the researchers used multivariate analysis. Binary logistic regression models were used to predict verbal warnings, written warnings, citations, and arrests, while controlling for stop characteristics, demographics, and other factors that could influence stop outcomes. When accounting for other driver, vehicle, and situational characteristics, there were no substantial racial/ethnic differences in the likelihood of receiving warnings, citations, or arrests. Rather, the strongest predictors of all post-stop outcomes included legal variables related to the reason for the stop, whether the person had multiple violations, and whether evidence was seized. The characteristics of PSP members, except for their assignment to patrol, did not strongly predict stop outcomes.

Discretionary Searches and Contraband

Search and seizure activities were examined separately, with a focus on discretionary searches conducted by PSP troopers.Of the 440,000+ traffic stops that were examined, 2.8% (n=12,236 stops) resulted in a discretionary search. Of the 12,236 searches conducted, 53.6% resulted in the seizure of contraband. Binary logistic regression analyses showed some racial/ethnic disparities in terms of being searched, though the overall likelihood was low (2.7% chance for Blacks, 2.1% for Hispanics, and 1.4% for Whites). However, legal factors related to the stops (i.e., reason for stop, multiple violations) remained the strongest predictors of discretionary searches.

The majority (72.7%) of searches were performed after the driver gave consent to the officer to search; in 45.9% of these cases, contraband was found and seized. 27% of searches were conducted on the basis of reasonable suspicion or probable cause. These types of searches had a much higher contraband seizure rate of 74%. The most commonly seized contraband included drugs (46.1% of seizures) and drug paraphernalia (38.6%), followed by weapons (5.1%).

Seizure rates for both types of discretionary searches varied significantly across drivers’ race and ethnicity, particularly for Hispanic drivers. Hispanic drivers searched based on probable cause or reasonable suspicion were less likely to have contraband seized (65.1%) compared to searches of White and Black drivers, whose seizure rates were similar (75.8% and 73.5%, respectively). Similar racial/ethnic differences were observed for consent-only searches, with Black (41.5%) and Hispanic drivers (32.9%) being less likely than White drivers (52.4%) to have contraband seized. However, data limitations hindered further examination of the relationship between drivers’ race/ethnicity and contraband seizure. Specifically, it’s important to note that traffic stop data alone cannot determine the legality of individual searches. To gain further context, the research team was invited to observe PSP criminal interdiction training classes, which emphasized professionalism, civil rights protection, totality of circumstances, and focusing on behavioral indicators of criminal activity rather than individual characteristics.

Conclusion

After rigorous analysis, this study found no evidence of racial and ethnic bias in traffic stops and post-stop outcomes within the PSP.  While some racial disparities existed initially, these differences disappeared when accounting for legal factors related to the stop (e.g., reason for the stop, multiple violations). There was some unexplained racial/ethnic disparities in consent searches, though these patterns mirror those reported in many jurisdictions across the country, which suggests it is not a result of individual police officer or trooper bias. PSP’s commitment to ongoing data collection and analysis should be noted. This continual monitoring ensures that the department practices unbiased policing, providing valuable information to the organization while simultaneously institutionalizing a culture that inspires fair and impartial policing.