Ads are being blocked

For us to continue writing great stories, we need to display ads.

Un-block Learn more


Please select the extension that is blocking ads.

Ad Block Plus Ghostery uBlock Other Blockers

Please follow the steps below

How to Stop Misreading Homicide Statistics

Yes, some U.S. cities have experienced an uptick in murders this summer. Here’s what that doesn’t mean.

AP Photo/Patrick Semansky (This July 30, 2015 picture shows a blighted home in west Baltimore. Murders are spiking in Baltimore—but that doesn't necessarily mean it's a trend. )

On Tuesday, The New York Times joined other outlets in addressing a recent spike in homicides in a number of U.S. cities, and in so doing entertained suggestions that Black Lives Matter might be to blame.

"Law enforcement experts say disparate factors are at play in different cities, though no one is claiming to know for sure why murder rates are climbing," theTimes reported. "Some officials say intense national scrutiny of the use of force by the police has made officers less aggressive and emboldened criminals, though many experts dispute that theory."

Of course, it's also possible that an uptick in violence in some U.S. cities does not represent a long-term or widespread phenomenon, and it is highly, highly unlikely that a movement like Black Lives Matter is responsible either way. Crime reporting, like the human perception of crime more generally, often misinterprets spectacularly horrible anecdotes as trends. The Times does mention that the murder rate is still well below stratospheric 1980s and ‘90s highs, but that should be the lead rather than a caveat.

Still, the fact remains that there are a number of U.S. cities that have experienced an uptick in homicides so far this year. Here’s how to make sense of it all in a more responsible, data-based manner.

Ask yourself, is this random or comprehensive nationwide data?

The 30 cities cited by the Times aren't randomly selected. Rather, they suggested their own selection because they are the cities—Milwaukee, Baltimore, New Orleans, and so on—where there has been a recent uptick in killings. The article does mention, as an aside, that cities such as Cincinnati, Los Angeles, and Newark are not experiencing such an uptick.

"The New York Times story ... focused on the cities that have gotten a lot of national attention for being more violent this year than last year," says John Roman, senior fellow at the Urban Institute's Justice Policy Center. The problem with this approach is that focusing on whatever cities happen to be struggling in any given year "really doesn't tell you anything about crime in America as a whole," as Roman puts it. It's a mistake to "start with the cities that are having the worst years and create a story around them, rather than trying to figure out whether the average city is having a bad year."

And it appears that they aren't. In response to the Times' story, Bruce Frederick, a senior research fellow at the Vera Institute of Justice, analyzed homicide data from 16 of the 20 most populous cities. Of those cities, just three showed a statistically significant increase. And looking back farther at recent years, Frederick found that murder rates have seesawed back and forth—evincing no stable trend at all.

Cherry-picking the cities that prove a doom-and-gloom hypothesis quickly leads to misinterpretation and obscures the big picture: violent crime is still way down.

"The story with violence is, for the last generation violence has gone down," says Roman. "The implication that violence is now somehow erupting or becoming a new epidemic is completely unsupported by any available data. It's the equivalent of telling me that a city set a record low temperature as evidence that there's no climate change."

We also need to take into account the number of shootings and other aggravated assaults. Shootings can be a better gauge of violent crime trends than murders because whether a shooting victim ends up dying owes a lot to the chance of aim, the type of gun being used, and the success of emergency room doctors.

Is the data short-term or long-term?

More than 100 homicides so far this year in Milwaukee is a lot of killing in terms of the human cost and pain inflicted. But statistically speaking, these are very small numbers. A set of discrete factors in one city—say, a series of neighborhood beefs spilling out of control—can result in a seemingly substantial increase in murders over a short period of period time. But a nasty few months doesn't necessarily represent a trend in one particular city, let alone one taking place on a  nationwide level.

"With a year or two of data it is premature to try to figure out whether we are seeing the beginnings of a long-term trend or a random fluctuation that will quickly reverse itself," emails David Greenberg, a professor of sociology at New York University. "The question about causes is being posed prematurely."

Implications of causality require an even higher burden of proof

The causes of violent crime are hard to pin down and the right answers might not be the obvious ones. Roman points to 2005-2006, when violent crime increased for the first time in 14 years. The headlines were dire, and police chiefs reportedly contended that music and movies were fostering a "thug mentality."

Roman and his colleague Aaron Chalfin, however, marshaled a lot of data to argue that the sudden ubiquity of pricey iPods may have driven the spike. Roman says that today, the economic recovery is causing people to spend more time outside, shopping and in transit. An increase in targets for crime could once again be a factor, he suggests.

By contrast, there is really no evidence to support the notion of a so-called "Ferguson effect"—the idea that criminality is thriving because protesters over the past year have bullied cops into not doing their jobs. The Times article strikes a tone of he-said-she-said false balance, as The Atlantic’s Ta-Nehisi Coates wryly notes: "Some think brain cancer can be cured with roots and berries, but others say proof has yet to emerge."

The idea of a "Ferguson effect" was popularized in The Wall Street Journal by Manhattan Institute fellow Heather Mac Donald this May, and it keeps perniciously reemerging. There are several reasons it is probably wrong.

Just to start, take the case of the murder in St. Louis—which began its upward move before Michael Brown was killed.

Deterrence has been a key rationale put forward by proponents of aggressive policing. In New York, the NYPD's argument in favor of its massive stop-and-frisk program was that it deterred people from carrying guns. This is hypothetically possible, civil rights concerns aside: widespread knowledge of a strong possibility that any young black man will be stopped and searched for no constitutionally valid reason could no doubt deter a young black man from carrying a gun.

But the flip side of this argument—that a year or a few months' worth of decreased enforcement since Ferguson or Freddie Gray could radically reshape potential criminals' conception of their chances of committing a crime with impunity all over the country—that is hard to comprehend. Law enforcement officials readily acknowledge that many of this summer’s urban homicides stemmed from personal disputes—not the sort of crimes for which broad-based efforts at deterrence would seem to have much success.

For that matter, it’s not at all clear that harsh criminal justice policies have had much of an impact on violent crime. For years, defenders of mass incarceration have argued that locking up 2.2 million people at a time works because it has caused violent crime to plummet. But the research suggests otherwise: in 2014, a major study from the National Research Council of the National Academies found that "the increase in incarceration may have caused a decrease in crime, but the magnitude of the reduction is highly uncertain and the results of most studies suggest it was unlikely to have been large."

Violent crime trends are just very hard to understand, as Jesse Walker points out at Reason. "We aren't even sure why crime across America was just FALLING, and that went on for more than two decades, so it's a little early to draw sweeping conclusions about the last eight months."

It is possible that there is a Ferguson effect—just not the one its proponents are suggesting. My own reporting on gun violence in Philadelphia has confirmed what should be obvious: young black men shooting each other often grow up shut out of good schools and jobs and stuck in marginalized neighborhoods where respect is the only real coin on offer. Outrage over recent police killings has likewise, understandably, stoked anger in black communities.

But Black Lives Matter is a politically engaged and pro-social movement. Are we really supposed to believe that those carrying a placard protesting the state of systemic injustice are somehow more likely to shoot their neighbor over a turf slight?

Gun violence is a nightmarish and persistent feature of poor, often segregated black neighborhoods, and it deserves close attention from reporters (see The Baltimore Sun's heartbreaking profiles of 45 murder victims in 31 days). But rushing to conclusions about recent violent crime risks urging Americans to accept the false premise that we must choose between criminal justice reform and public safety. This sort of framing only makes it harder to accomplish much on either front.

About the Author

  • Daniel Denvir
    Daniel Denvir is a Rhode Island-based contributing writer to CityLab and a former staff reporter at Philadelphia City Paper.