Republican presidential nominee Donald Trump says he favors "stop-and-frisk." Here's what that means and why the policy is so controversial. (Victoria Walker/The Washington Post)

In an interview with Fox News that aired Wednesday night, Donald Trump said the policing tactic called "stop-and-frisk" helped reduce crime in New York City and should be replicated in high-crime areas of the country.

"I would do stop-and-frisk," the Republican presidential nominee said during a town hall with Sean Hannity in response to a question about what Trump would do to control crime. "We did it in New York. It worked incredibly well."

The comment has drawn criticism from people who note that a federal judge in 2013 ruled New York's stop-and-frisk practice unconstitutional. However, law enforcement experts say the phrase "stop-and-frisk" is vague enough to be essentially meaningless, and the specifics of Trump's plan could mean the difference between whether the policy is legal or unconstitutional, effective or worthless.

Trump did not elaborate on his plan, even when asked about his remarks again Thursday morning in a second interview on Fox.

At the height of New York's stop-and-frisk policy, officers were stopping civilians hundreds of thousands of times each year, often on the slightest of pretexts and disproportionately in minority communities. A federal judge ruled that the way New York police practiced the tactic equated to a “policy of indirect racial profiling.”

However, the general practice of stopping and frisking civilians is not unconstitutional, and research shows that doing so can help reduce crime — but only as part of a broader, clearly articulated strategy that targets real criminal activity.

In New York, the practice of stop-and-frisk is usually associated with Raymond Kelly's second stint as New York's police commissioner, which began in 2002. The number of stops recorded by New York police increased from 97,000 to 161,000 that year, and then nearly doubled the following year, according to police data obtained by the New York Civil Liberties Union.

At that time, New York City's crime rate of violent crime was already in decline. In 1990, there were nearly 31 homicides in the city for every 100,000 people — more than the average for other major American cities even in a year of frequent violence across the country. A decade later, that figure had declined by nearly 75 percent, to 8.4 homicides per 100,000 people.

As New York police abruptly moved away from the practice of stop-and-frisk toward the end of Kelly's tenure in 2013, the rate of homicide continued to decline as it had previously.

"To say that stop-and-frisk is the reason crime went down is wrong," said Peter Moskos, a former Baltimore police officer who is now a sociologist at the John Jay College of Criminal Justice in New York.


Philip Bump / The Washington Post

From these broad trends, it is difficult to tell whether the stop-and-frisk policy had an effect on crime that was obscured in the data by other factors.

To answer this question, legal scholar Jeffrey Fagan of Columbia University and other researchers studied data from specific neighborhoods in which New York police focused their manpower in periodic efforts to control crime, comparing those areas with others nearby that were not saturated with police officers.

The researchers found that a high rate of stop-and-frisks was not associated with reduced crime when the practice was indiscriminate.

Trump's claim that the stop-and-frisk policy was effective in New York City was "not true, simply not accurate," Fagan said.

But again, the details of implementation were crucial. Fagan's group did find that police reduced crime by stopping and frisking civilians after directly observing someone acting violently, selling drugs or "casing a joint" -- that is, checking out a store or another potential target for a burglary.

The stop-and-frisk policy was ineffective because civilians were regularly stopped on inconsequential pretexts and vague justifications, such as that a person was moving furtively. The result was officers wasting their time with civilians who were not criminals, Fagan said.

"You can achieve really very positive crime control, reductions in crime, if you do stops using those probable-cause standards," Fagan said. "If you just leave it up to the officers, based on their hunches, then they have almost no effect on crime."

In the long term, Fagan argued, stop-and-frisks could prove counterproductive by making young people less likely to share information with law enforcement. "You traumatize kids, young adults who are the focus of the stops, and you completely alienate them so they don’t cooperate with the police in future investigations," Fagan said. "That puts everybody at risk."

"Stopping and frisking people can be part of a legitimate and legal strategy of policing," said Moskos, the former officer, but he added that the stops must serve some real purpose to be effective.

Moskos and other critics say that demands from superior officers to record a certain number of stops each year discouraged cops from using good judgment.

"It was a stat-driven thing," Moskos said. "It was a bad form of policing because it took away discretion."