Traffic Stop Data Can Shape Policy. It’s Often Missing.

Last Monday, we published a data-driven look at the decline of traffic stops in America. The cities and states we highlighted weren’t necessarily those with the steepest declines in traffic enforcement nationally — they were places with steep declines that data enabled us to see.

There is no national standard or source for recording traffic stops, so we tried to gather data from as many different cities and state agencies as possible. Many cities we would have liked to have shown had incomplete data.

And other cities publish no data on traffic stops at all.

The lack of reliable, comparable data doesn’t just make our job as reporters tricky. Communities with good data often have different political and policy discussions than places where nonexistent data makes it hard for the public to know what’s going on. Data collection is also a police reform in itself. Data mandates tell the police that their actions are being tracked. As several researchers told us, the existence of data alone can change how the police behave.

For all these reasons, we came to think of the data as a character in our story — one shaping the other actors around it (including us, and the story we were able to tell).

Below are some of the data quirks and difficulties we encountered.

Lack of consistent historical data

Concerns over racial disparities in enforcement have led to laws mandating data collection on traffic stops. In 2003, an Illinois state senator named Barack Obama sponsored the Illinois Traffic and Pedestrian Stop Statistical Study Act. It required every police department in the state to record data on traffic stops. As a result, Illinois has some of the best available data today.

Other states, like North Carolina and Missouri, passed similar laws around the same time. Most states still don’t have comprehensive data laws, though large states like California and Virginia have adopted them recently.

Another way to track stops is to ask the public — instead of the police — about encounters with traffic enforcement. The Bureau of Justice Statistics conducts a representative survey every few years that yields an estimate of the number of people nationally who have been pulled over. Unfortunately, the most recent publicly released survey was completed during the first six months of 2020. That means federal surveys have not yet captured much of the decline in enforcement starting with the pandemic that we see in local data.

City-to-city comparisons

Many cities have detailed data about traffic citations, but not necessarily all traffic stops (many stops end in a warning, not a citation). This complicates comparisons between cities.

Seattle, for instance, releases data on stop-and-frisk encounters with pedestrians. It also releases data on traffic infractions. But the city has no public count of all the times the police have stopped a vehicle.

And even among vehicle stops, definitions vary by department. Some cities, such as Denver, filter out incidents where a driver is pulled over after being seen fleeing a crime scene. Other cities may include stops like this in a “vehicle stop” yearly total.

Data accuracy

In many instances, officers have been accused of intentionally or mistakenly misreporting the race of drivers. This could have the effect of making racial disparities appear smaller in data than they are in reality.

In California, officers must record “perceived” demographic information about people they stop, but they are explicitly not allowed to ask for it. Police officers must check a box if someone is perceived to be Hispanic, has a mental health condition, or is “L.G.B.T.” — characteristics the police may be guessing about.

In March 2022 in Los Angeles, a new policy took effect requiring officers to record their reasons for making pretextual traffic stops (these stops occur when officers use a minor violation, like a broken taillight, to look for evidence of more serious crimes). After the policy change, the city’s data showed that stops for equipment and nonmoving violations decreased, while those for moving violations increased.

Often drivers are pulled over for multiple violations (say, speeding and a broken taillight). But California’s standardized data collection form allows only one violation to be entered as the “primary reason for stop.” That means that what may look on paper as a shift in police tactics could also mean that the police are changing which violations they record as “primary.”

Road fatality data is messy, too

We became interested in traffic enforcement because of our previous reporting on road deaths. More than 40,000 people in America are killed on roads each year. But it can be hard to find annual road fatality counts for individual cities. In some cities, that data is collected by the police department, in others by the transportation department. Some cities also exclude from their numbers deaths that occur on federal or state highways that pass through town.

To get around these problems, the fatality data we cited in our article came from a federal database in which the location of each death nationwide was recorded. Each road fatality should include a recorded race. But we found that in many states the share of all deaths with race “unknown” has changed meaningfully over the years. Pennsylvania, for instance, listed over one-third of all fatalities as “unknown” race in recent years. These inconsistencies, both by geography and time, make any takeaways difficult.