美国国家公共电台 NPR--Many midterm races focus on rising crime. Here's what the data does and doesn't show

（单词翻译）

Many midterm races focus on rising crime. Here's what the data does and doesn't show

Transcript¹

Crime and public safety are among the issues that have taken center stage in many midterm races nationwide, from Wisconsin to Pennsylvania to New York.

Republicans are repeating a page from their 2020 playbook by ramping² up ads accusing Democrats³ of being soft on crime. The GOP spent nearly $40 million on crime-related messaging in September alone, as NPR has reported — and ad spending tends to be a good indication of what candidates and their backers think will sway voters.

The ads are hard to miss, warning of "defund the police Democrats" and liberal incumbents⁴ who have failed to bring violent crime under control. They point to rising crime rates and paint casting midterm votes as a life-or-death decision (Democrats are arguing a similar case, but about abortion⁵ rights).

The messaging taps into a fear clearly shared by many voters, even if it's not their top or only concern. Some three-quarters of respondents of a POLITICO/Morning Consult poll released earlier this month said that they thought violent crime is increasing nationally, and 88% said it is either increasing or staying the same in their own communities.

For many, those fears are grounded in jarring incidents and unmistakable trends, from successive mass shootings to brazen⁶ attacks during daylight hours. So are crime rates in the U.S. actually skyrocketing, as some claim?

"We don't really have a common definition of what crime means, when you ask that question," says Jeff Asher, a data analyst⁷ who specializes in crime statistics. After working with the Pentagon, CIA and New Orleans Police Department, he co-founded the consulting firm AH Datalytics.

The FBI's annual crime report for 2021, released earlier this month, says violent crime decreased by 1% from the previous year. But the report is also incomplete, as only 63% of the country's police departments submitted data — and New York City and Los Angeles were not among them.

The FBI tracks seven major types of crime, Asher explains on Morning Edition. Property crime — theft, auto⁸ theft and burglary — has been falling regularly for the last 20 years. Violent crime — aggravated⁹ assault, murder, rape¹⁰ and robbery — increased at least in 2020, but remains¹¹ lower than it was in the 1990s.

"But even that, when we talk about crime, is not what people think of," Asher says. "What they're really thinking of is murder and gun violence, and murder makes up 0.2% of all big-picture crimes every year. But it's the crime with the most societal harm. It's the thing that people tend to care about the most."

There's been a dramatic uptick in murder over the last several years. FBI data shows that it rose nearly 30% from 2019 to 2020 — the largest single-year increase ever recorded in the U.S.

The number of homicides increased 4.3% nationally in 2021, according to the FBI report, and Asher says it likely decreased by a similar amount so far this year. Nonetheless, it's higher than it was in 2019. In other words: The murder rate is significantly lower than it was in the 1990s — Asher says 30-40% lower — but much higher than it was just three or four years ago.

Importantly, that's a national rate. Asher thinks the reason crime carries so much weight in these midterm elections is because the increase in murders in 2020 was truly a national phenomenon.

"It happened in big cities, it happened in small cities. It happened in counties that voted for Trump¹², it happened in counties that voted for Biden," he says. "It was really everywhere, and so I think that most places in America are grappling with at least some increase in gun violence over the last two years, which brings it to the forefront of these elections as they're taking place now."

But the data doesn't tell us everything

The data informing these reports is "both reliable and unreliable," according to Asher. And it got even more confusing this year, thanks to a change in the FBI's data collection methods.

Most big cities tend to produce and make crime data available in some format¹³ in "near-real time," so that people can see numbers from the last few weeks, months or quarter, he explains, adding that smaller cities, suburbs and sheriff's offices don't necessarily make it as accessible. And on a national level, the FBI collects data from local law enforcement agencies and reports it out a full ten months later.

That makes it really difficult for researchers and policymakers to get any sort of nationwide estimates about these trends as they're happening, Asher says. And they can't solve a problem that they don't fully¹⁴ understand.

"We've really got this change in murder, which almost happened overnight and has been sustained for two years, and we're not able to measure progress against it or regression against it," he says, "Because at this really inopportune moment, our data collection has suddenly gotten much worse."

This year, the FBI changed the way it collects crime data — and many of the nation's roughly 18,000 law enforcement agencies were slow to get on board.

The data collection system, called the National Incident Based Reporting System (NIBRS), has been around for years — but it requires departments to log crimes in greater detail, and wasn't their only option for doing so before.

In 2015, the FBI announced that it would be switching from its previous system to only using NIBRS beginning in 2021, meaning that as of this year, departments had to submit data via NIBRS or not at all.

As it turns out, many chose the latter. The 63% of law enforcement agencies that submitted data for 2021 marks the lowest level of participation¹⁵ the FBI has reported in decades, and only 52% of them submitted a full 12 months of data.

This new methodology, and the numbers themselves, come at a precarious¹⁶ time, Asher says. It doesn't just make it harder for stakeholders to work on solutions; it also makes it harder to refute false claims about crime, especially this election season.

So how should people be sorting through all this messaging? Asher says that as a data analyst, he wants to focus on the data — while a chart isn't necessarily the most effective means of communication, he says the numbers behind the claims are crucial.

"Absent that, we get a lot of politicians that are saying a lot of things that frequently are based on anecdote¹⁷ or sort of the vibes of the moment," he says. "And we get, then, a lot of misinformation and poor decisions being made in the name of data-less arguments."