Are Americans’ perceptions of the economy and crime broken?

This election cycle showed that our evaluations of external reality are increasingly partisan. Can the media bridge the gap?

By Joshua Benton @jbenton Nov. 26, 2024, 2:23 p.m.

A responsible news media has a lot of jobs, but here’s one of the most important: giving audiences an accurate image of the state of the world around them. How’s the country doing, overall? Is the economy booming or busting? Is crime climbing or dropping?

Anyone can, of course, reach their own conclusions on those questions, independent of the news they consume. But their views will necessarily be influenced by their own individual circumstances. Did they just get a promotion — or laid off? Do they feel safe sleeping with their front door unlocked — or did they just get mugged? Their own personal data points might align with a larger trend — or they might not. And news stories have traditionally been a big part of how people figured out which was which.

But we’ve just concluded an election cycle that suggests something important has broken in that feedback loop. How people perceive the economy and crime are major factors in whether they reward or punish incumbents with their vote. And decades-old patterns in that process seem to have gone a little haywire.

Take the economy. This spring, everyone was talking about the “vibecession.” Americans believed the economy was in the tank, despite loads of evidence showing the opposite. Just before Election Day, 54% of Americans surveyed said the U.S. was currently in a recession, versus only 30% who said it wasn’t. (It wasn’t and isn’t.) Americans feel quite good about their own financial situation, but they’re convinced the American economy writ large is in dire straits.

A huge part of that is partisanship. Between 1999 and 2020, the gap between Republicans’ and Democrats’ views of the economy roughly doubled. (“The size of the partisan divide in expectations has completely dominated rational assessments of ongoing economic trends,” one researcher wrote. “This situation is likely to encourage poor decisions by consumers and policymakers alike.”) Republicans’ views have been especially responsive to political events. When pollsters ask people whether their own financial situation is better or worse than it was a year ago, Republicans are much more likely to change their view when control of the White House changes.

But there also seems to be something more fundamental happening. Before the covid pandemic, consumer sentiment was relatively predictable based on economic fundamentals. The hard data and the survey responses tended to move up and down in something like unison. But since 2020, they’ve become disconnected, with a wide and pessimistic gap opening up between them. It’s hard to look at that phenomenon and see the impact of a changed media environment.

Or look at crime. Crime in the United States has been on a decades-long decline — a steep drop starting in the early 1990s, then on a shallower grade starting in the 2000s. There was a spike during the dislocation of pandemic lockdowns, but that has faded and the general trend is down again.

But throughout this decline, most Americans have said they believed crime was increasing. As with the economy, they typically had a rosier view of crime in their own communities — but the nation as a whole was getting more dangerous year after year.

And now we’re seeing the same gap open up between Democrats and Republicans that we see with the economy. Here’s Gallup data on how people respond to the question: “Is there more crime in the U.S. than there was a year ago, or less?” You can see that both Democrats and Republicans tended to overestimate crime increases across this span — but check out those 2024 numbers. 90% of Republicans say crime is up over the past year, versus only 29% of Democrats.

These new perceptual gaps are why I wanted to talk with Jeff Asher, a crime data analyst based in New Orleans and cofounder of AH Datalytics. He’s often cited for his expertise in both what crime data can tell us and its many limitations. In September, he announced the debut of the Real-Time Crime Index, an attempt to provide an accurate, big-picture view into the state of crime in America. Rather than wait for the FBI’s annual release of year-old (or older) data, the index gathers close-to-real-time numbers — the lag’s about a month and a half — from hundreds of individual police agencies across the country and pulls them into a set of common trend lines. (They also post all the data on GitHub.)

If you check out the Real-Time Crime Index right now, you’ll find that murders between January 1 and September 30 were down 16.6% from a year earlier. Violent crimes more broadly (including rapes, robberies, and assaults) were down 4.1%. Property crimes dropped 8.6% over the same span. Here’s the data on murder:

I don’t pretend that a single chart will change many minds about crime levels — or the economy’s strength, for that matter. But I wanted to ask Asher how he thought the news media and the public were doing in navigating these seas of numbers. Here’s a lightly edited version of our conversation.

Jeff Asher: A lot of my training has been in how to communicate complex ideas that are filled with uncertainty. You start with “Here are the trends,” and then it moves to “Here are the trends, but this is the uncertainty.” You have to live with some of the uncertainty. We’ll never know the exact number, but we have an estimate.

In crime data, we’ve got the reported data — “This agency reported X number of murders last year.” You’ve got the unreported data, which comes from the Bureau of Justice Statistics’ National Crime Victimization Survey, which is surveying people. We know from the survey data that not every crime is getting reported. So we think that when it comes to murders, we have a pretty good handle on them. Sometimes an agency will make a clerical error that pushes things up or down, but usually we get a good share of the murders that occur.

But we know that with some crimes, like rape and sexual assault, you’re looking at sometimes a quarter of those getting reported. Property crimes, it’s usually about a third or so of those getting reported, and about half of violent crimes getting reported. Obviously, you’re not surveying murder victims, but everything else we know gets underreported. Sometimes an agency will say, “Oh, we had a 6% increase in rape, but we think people trust us more and are reporting them more often.” There’s all sorts of uncertainty.

Joshua Benton: It’s seemed to me that the gap between perception and reality in terms of crime has been particularly large in this election cycle, and that the gap has gotten more partisan. How healthy do you think that feedback loop is, between reality and perception, in terms of how people understand the level of crime and make political judgments about it?

Asher: People’s ability to evaluate crime relative to the trends has been completely broken. And it’s a new thing, in terms of the level of brokenness. Gallup just came out with its poll of people’s perceptions of crime. They found overall that 64% of people in 2024 said there was more crime than a year ago — which was down from 2023, when it was 77%. But the split is 29% of Democrats and 90% of Republicans say there was more crime now than there was a year ago, nationally.

It’s shocking the degree to which that persists. That 29% is really low — lower than I would expect. But the 90% didn’t change, even though we’re dealing with what is undoubtedly the largest decline in murder ever recorded. Generally, the survey data suggests in 2023 that things were about even in terms of violent crime, about even in terms of property crime, maybe down a little bit — but the reported crime showed declines. So if people were accurate, you probably think it’d be about 50-ish percent of both. So that you’ve got this insane partisan split.

And what’s interesting is the degree to which it’s new. I mean, historically, in the ’90s, everybody knew it was going down. Then during the Bush administration, there was a little bit of a differentiation, where Republicans thought it was going down and Democrats up, but the gap was small. The gap grew a bit under Obama, then got a little bit smaller under Trump, but there’s still a gap there. In 2014 — which is the year with the lowest murder rate recorded since the 1960s, or maybe ever — 51% of Republicans said crime was going up compared to 39% of Democrats.

I think part of the challenge is that, when we talk about crime, what do you mean by crime? Do you mean murder? Do you mean violent crime? Do you mean auto theft, which surged for several years? Do you mean disorder, or homelessness, or things that maybe aren’t criminal, or at least not major crimes, but that are what people think about then they think of safety?

Benton: It’s interesting, because there’s long been research showing that media consumption has a big impact on people’s perceptions of crime. Most famously, people who watch a lot of local TV news — “if it bleeds, it leads” — tend to be more afraid of crime and think it’s a bigger problem than people who don’t. But at the same time, watching local TV news was never an especially partisan behavior — Republicans did it and Democrats did it. There’s a sense that, as bad as the pre-internet mass media environment could be in many ways, it provided more of a shared universe of information for people.

Asher: I think the challenge is that people’s attitudes reflect leadership. And because it has been seen as politically expedient to show crime going up, you’ve got an environment where, regardless of what happens, it’s either “crime is going up” or “the data that shows crime going down has flaws.” Which it does! I’m the first person to talk about the flaws of crime data. We could talk for an hour about all the flaws in the FBI crime data, but that doesn’t invalidate the trends that we’re seeing.

And so I think that’s what’s frustrating, and part of the reason that we did the Real-Time Crime Index is to create another independent source that is accurately and transparently communicating crime trends without any partisan bias, being as upfront and clear as to the methodology as possible. Because I don’t think the FBI data has a partisan bias, but they’re clearly struggling with the communication element of the trends.

Benton: So let’s talk about the index a bit. For how long had you been thinking about this as an idea for a project? And, for someone who hasn’t seen it, what’s its main selling point? What does it add that other sources don’t provide?

Asher: So it adds speed and it adds accuracy, at the expense of a little bit of precision. What we do is we basically take as many cities as possible — the goal is to get somewhere in the 500- to 1,000-city range — and we get their data with a month and a half delay, and we publish it to show the trends. So it’s one thing to say, here’s one city’s data, or here’s another city’s data. It’s another to say this is a big-enough sample that it’s capturing half or 60% of the murders this year. Yes, it might change a little bit up or down, based on how whether or not agencies adjust their numbers, but by gathering this large of a sample, we’re able to show what the trends are.

This is something that I really started doing this in 2016, looking at just murder data, and it started with a couple of dozen cities. And I was able to show that, yeah, you know, if you’re showing murders up 8% in your 30-city sample, it might be up 4% or it might be up 10% — but you’re able to show what the trend is nationally. We eventually expanded that murder dashboard to about 250, 280 cities.

The Real-Time Crime Index is larger, it’s more standardized as far as what’s being collected, and it’s collecting all the major crime types. So you can show here’s the auto theft trend, here’s the murder trend, here’s the robbery trend. And you can see it nationally. You can see it statewide for all the cities in your state that we have data for, or you can see it at an individual city level. So there are all sorts of ways you can analyze it and hopefully help to answer the question of what is actually happening and allow people to focus on the why, rather than focusing on the what.

Benton: Who’s your target audience for the index?

Asher: It’s policy makers who want to use it to show whether or not what they’re doing is working. It’s members of the media, to be able to use it in the same way you say, “BLS [the Bureau of Labor Statistics] reported X” or “Here’s what the stock market market is doing.” The stock market is not the economy, but the stock market can be a good way of evaluating the health of the economy and understanding the degree to which people have confidence in it. And it’s audiences that want to know crime trends.

Benton: It strikes me that, in contrast to political polling, the trend lines seem to be just about all people think about. People don’t spend a lot of time asking, “Is this the right level of murders to have?” It’s “Is the number going up or going down?” It’s as if, in political polling, the emphasis was only on which candidate had the momentum, not which one was actually leading. We sort of accept as background noise whatever the current level of murder is, the current level of violent crime, and it’s the increase or decrease that gets noticed more.

Asher: Yeah — that’s sort of the challenge in talking about this. Like, “Murders are down 18%” — that’s great, fabulous. You’re still talking about 16,000 people dying from murders in a year, and you’re not talking about things like overdoses and other tragedies. So I think it’s important to sort of celebrate the trend, or acknowledge the trend, without thinking, “Okay, we’ve achieved our goals, murder is back down to where it used to be, everything’s achieved.”

Benton: So what are the next steps for the index — what are the biggest problems with it that you can conceive of solving? Additional data sources? More cities?

Asher: More cities. We’re at 300 agencies with complete data. Our goal is in the 500 to 1,000 range. Part of the challenge there is finding which agencies are good and which are not. And so we’ll have an audit procedure. One of the challenges is that agencies do revise these numbers — they’re usually small changes, but they happen. So if we say that murder is down 16%, and after revisions, it’s really down 15.5%, I think that’s okay — you’re showing the trend accurately, and you understand that that this sample is never going to be 100% accurate.

Benton: Let me ask the sort of pedantic question that someone who works for a news organization that covers news organizations would ask: Do you consider the index or your work around it more broadly as journalism? Or do you consider yourself a data analyst who feeds into journalism?

Asher: I don’t consider myself a journalist. I consider this a data project that feeds into journalism. This is a crime data collection for the media, not a piece of media — that’s how I look at it. We’re trying to be open and transparent about the data, and we want to leave to others the explanation-and-understanding part.

Benton: We’re at the end of a political cycle in which crime has been a big topic of focus. I’m curious how you’d evaluate the job that news organizations have done in covering the issue of crime in this election. If you were suddenly appointed czar of all media coverage, what would want us to be doing better?

Asher: I would say the record is mixed. You see journalists who fail to address the uncertainty in the data — or going the other way and saying that because there’s uncertainty, we’ll never know for sure. You know, “One side says one thing, the other side says another thing, and the data is imperfect, so we’ll never know.” Some of the trends are knowable. Most of the trends are knowable. Violent crime is about where it was, down a little bit. Property crime is falling because motor vehicle thefts have started to drop. We can say things, definitively, murder is falling fast. The question is “How far will it come down?” not “Is murder falling?”

Photo of broken eyeglasses by Angus Stewart used under a Creative Commons license.

Joshua Benton is the senior writer and former director of Nieman Lab. You can reach him via email (joshua_benton@harvard.edu) or Twitter DM (@jbenton).

POSTED Nov. 26, 2024, 2:23 p.m.

Show tags

TWITTER FACEBOOK EMAIL