Over the last three years, several law enforcement agencies have responded to police violence—including racially-motivated killings—by implementing body cameras. By 2015, 95 percent of police departments in major U.S. cities reported the use of, or a plan to implement, body cameras. And the Department of Justice has awarded $40 million to police departments around the country to equip their officers with the technology.

But a study released today (October 20) by researchers at The Lab @ DC concluded that the use of body cameras had no significant impact on policing tactics.

The study, “Evaluating the Effects of Police Body-Worn Cameras: A Randomized Controlled Trial,” partnered with the Metropolitan Police Department in Washington D.C. to analyze the actions of more than 2,200 law enforcement officers. Half of them were selected at random to wear police body cameras, while the rest of them did not wear them.

For seven months, researchers tracked police behavior in use of force and civilian complaints, checking to see if police body cameras made any difference in their actions. By the end, researchers found no statistically significant effects on any of the measured outcomes. The study was conducted by David Yokum (from the DC Office of the City Administrator), Alexander Coppock (Yale University’s Department of Political Science) and Anita Ravishankar (DC Office of the City Administrator and the Executive Office of the Chief of Police).

“Our experiment suggests that we should recalibrate our expectations,” it concluded.  “Law enforcement agencies … that are considering adopting [body worn cameras, or BWCs] should not expect dramatic reductions in use of force or complaints, or other large-scale shifts in police behavior, solely from the deployment of this technology.”

From a breakdown of the results on the study’s website

We learned that BWCs do not have detectable average effects on documented uses of force or civilian complaints. That said, we cannot rule out that BWCs cause small decreases or small increases on these two outcomes.

For example, we find that a group of 1,000 officers with BWCs is estimated to report 74 more uses of force in a year than officers without BWCs. This is our best estimate. However, the data are also consistent with the real effect of BWCs being anywhere from a decrease of 97 uses of force to an increase of 244 uses of force per 1,000 officers, per year.

Because this range spans negative, zero, and positive values, the result is considered “statistically insignificant,” or “null.” More plainly, we interpret this to mean that BWCs had no detectable, meaningful effect on documented uses of force.

Read more about the study via The Lab @ DC