Protesters face off against police officers on the street in front of the Ferguson Police Department building in November 2014, in Ferguson, Mo., three months after the fatal shooting of 18-year-old Michael Brown. (Jahi Chikwendiu/The Washington Post)

John Shjarback is an assistant professor in the department of criminal justice at the University of Texas at El Paso. Scott Decker is a professor in the School of Criminology and Criminal Justice at Arizona State University. Scott Wolfe is an associate professor in the School of Criminal Justice at Michigan State University. David Pyrooz is an assistant professor of sociology at the University of Colorado at Boulder.

American police are under intense scrutiny. A series of highly publicized deadly-force incidents involving minorities has focused media and public attention on police practices. There has been much speculation that this increased criticism has led some law enforcement agencies to pull back, possibly contributing to rising crime in many cities — an alleged relationship that has come to be termed the “Ferguson effect.”

Unfortunately, our understanding of this kind of decline in proactive law enforcement — known as “de-policing” — is long on guesswork and anecdote but short on data and research. We have little hard evidence about de-policing. Does it even occur in the wake of high-profile shootings? If it does, what are the effects?

The answers to these questions may not be as straightforward as they seem.

In late 2015, then-FBI Director James B. Comey made headlines when he suggested in a speech to the University of Chicago Law School that “a chill wind [is] blowing through American law enforcement over the last year. And that wind is surely changing behavior.” He was speaking a little more than a year after Michael Brown, an 18-year-old African American man, was fatally shot by an officer in Ferguson, Mo. The unrest and backlash that followed this shooting — and speculation that it would cause police to take a more passive approach to law enforcement — give the “Ferguson effect” its name.

In his speech, Comey said it was his “strong sense” the recent spikes in violent crime in some cities were connected to such de-policing. But what do the numbers say?

We can shed some light here. For a study published in the Journal of Criminal Justice in May, we examined data on police activity in Missouri before and after Brown’s death. Police departments in Missouri are obligated by state law to submit data on traffic stops, including whether a search was conducted, contraband (such as drugs or weapons) was found or an arrest was made.

Was there a change in police behavior? In Missouri, more than 100,000 fewer vehicle stops were made in 2015 than in 2014, a 6 percent reduction. This was a significant contrast to changes in stops between 2013 and 2014, when just a 2,000-stop swing was recorded. In 2015, something was happening in Missouri that could not be attributed to normal fluctuation.

When we took a more in-depth look into 118 police departments serving municipalities with more than 5,000 people, a similar pattern emerged. These agencies made approximately 67,000 fewer stops in 2015 than in 2014. There were no significant changes in the searches or arrests stemming from those stops; however, contraband “hit” rates did improve slightly — by 11 percent — from 2014 to 2015, which suggests officers were making better searches.

Where was such de-policing most likely to occur? Departments serving jurisdictions with larger African American populations conducted fewer stops, searches and arrests in 2015 than in 2014.

Crucially, however, in both the statewide data and the figures from individual municipalities, we found that — far from the fearful predictions of some — this de-policing had no effect on violent or property crime rates. Which raises a surprising question: In some places, particularly those — like Ferguson — that were using stops to generate additional city revenue, could de-policing represent a correction of practices that served no genuine public safety purpose?

Policing is best employed not as a blunt instrument but rather as a surgical tool. Such tools are most effective when deployed in highly focused ways on crime hot spots and against high-rate offenders. More study, in more places, is needed, but the data from Missouri suggest a provocative possibility: The Ferguson effect may be real. And it may be an improvement.