Bryan Caplan  

Pessimistic Bias: Crime Edition

PRINT
A Proposal for Masonomics Fiel... Super Freakonomics...
In chapter 4 of Selfish Reasons to Have More Kids, I try to help parents overcome their pessimistic bias.  Kids over the age of 1 have long been the safest people in our society, we're all much safer than we used to be, and kids' safety rose especially quickly.   This is easy to demonstrate for mortality.  But other measures of safety also show sharp improvement.  In 2008, for example, violent crime hit its lowest measured level in the 37-year history of the National Crime Victimization Survey.

That's reality.  What about perceptions?  Gallup has been asking people "Is there more crime in the U.S. than there was a year ago, or less?" for twenty years.  People are almost always more likely to say "more" than "less," but in recent years the paranoia is off the charts.  More than 2 out of 3 Americans have been falsely saying "more" since late 2005; the latest rate of more-saying is 74%.

Also striking: There is a huge difference between the perceived severity of crime in "the area where you live" versus "the United States."  50-60% say the local problem is "not too serious" or "not serious at all"; only 2-5% say the same for the country as a whole.  Since the country as a whole is just of bunch of localities, it's safe to say that there's a contradiction.  And since they observe their localities first hand, but only experience the nation on t.v., it's safe to say that local perceptions are more reliable.  (Still, I should point out even at the local level, more people see crime getting worse).

All this is broadly consistent with a story I've been telling for years.  But I have to admit that I feel like I'm missing something.  Suppose you grant my general thesis about pessimistic bias.  As long as parents correctly perceive that their local areas are safe, why would national news make them fear for their children?

HT: Justin Longo via Radley Balko


Comments and Sharing





COMMENTS (18 to date)
Kevin Dick writes:

This seems to be simple availability bias. Most of the available examples people have for the nation as a whole are of crime incidents covered on the news.

However, for their local area, the available narratives they encounter among their neighbors and friends are mostly of non-crime.

ajb writes:

They're not thinking about averages. They're thinking about the likelihood that their kids wander into some pocket with really serious crime. And there's greater uncertainty about which areas -- which are not obviously bad -- might have worse than average crime.
Furthermore, there may be areas they feel they *have* to visit from time to time (lots of downtown areas) where they're certain crime is bad. Big cities are focal points and if those downtown areas are bad, it doesn't matter if the *average* is bad, then the average of the dominant centers of the US is bad. (Think downtown DC, NY, Chi, LA, SF, Detroit, St L, Philly, etc.)

Indeed, many people would agree that if the 50 biggest cities in the US had horrible crime rates and the rest of the US were crime free, they might correctly perceive that crime is high in the USA.

jb writes:

Just because it is safe where I live *right now* doesn't mean it will be safe tomorrow.

Also, I suspect that at least some of this has to do with our hardwired tribal ancestry - the people on TV who speak my language and share a cultural background are part of my clan. If something happens to them, even on TV, it *feels* local and immediate.

Joseph writes:

This is the kind of thing that I was generally informed of in a sociology class recently. Because the media gives coverage of the more "devious" actions of society, we often think that crime is much more prevalent than it really is. The most morally twisted and despicable stories of sick crimes are often more publicized than political debates and charitable acts. These events receive more public light, and thus, give society and malformed view of how society operates as a whole.
It's obvious the amount of crime is increasing in this country, that may be a fact, but a fact that is laced with more fear than need be. Of course there is more crime when you are only looking at numbers. The amount of people in this country is enjoying steady growth, and thus, there will inevitably be more crime. When one looks at the amount of crime as a percentage rather than as raw numbers, the amount doesn't seem so explosive.

mark writes:

I agree that there is a pessimism bias. Guilt might be a factor as well. Are you really going to tell yourself and some interviewer that your raising your own kids in a dangerous environment.
Also when you see a kid get pummeled on a bus while other kids cheer or at least be indifferent to the kids plight, is that one act of criminality. I suspect that if we could radically reduce crime(like 95%) in this country then all the media in the world couldn't scare us.

Prakhar Goel writes:

Dear Bryan,

You are taking far too short term a view. Crime in the last two to three decades may have slightly decreased but compared to the early 1900s or the 1950s, it is still unacceptably high. Just not as unacceptably high as say in 1995.

This paper might shed some light on this matter: http://bjc.oxfordjournals.org/cgi/content/abstract/41/4/618

What we are seeing here may just be accurate perception from more than 30 years ago.

Jim writes:

I think there's an issue involving the units of measurement that are being used in criminological statistics versus in the survey.

Crime rates are measured in crimes/100K population, but the survey respondents probably gauge the definition of "your area" in terms of distance, land area, or arbitrary politcal boundaries and the definition of "more crime" in terms of the absolute number of crimes committed.

We know population density has been increasing over this time period; the geographic area that survey respondents define as "your area" has an increasing population year to year. From this its easy to see that a constant or even declining crime rate can be consistent with (accurate) perceptions by respondents that there is "more crime" in "your area" than last year.

EX: County X with population 100,000 in 2005 had an annual crime rate of 25/100,000.

(25 crimes in County X/year/100,000 people in County X)*100,000 people in County X = 25 crimes in County X/year

In 2008, County X population is 150,000 and a crime rate of 22/100,000.

(22 crimes in County X/year/100,000 people in County X)*150,000 people in County X= 33 crimes in County X/year

County X's crime rate fell by 12%, but because of the increase in population density, more crimes were commited in County X per year in 2008 compared to 2005. Residents of County X are indeed correct to claim that there is "more crime" in "your area" in 2008. We don't need to assume any kind of bias to account for this survey result.

The same wording in the question about crime in the US, means that this units issue applies here too. This does not explain the disparity between the perceptions of how "serious" the "problem of crime" in the US versus local areas.

dullgeek writes:

How much of the pessimistic bias is just selection bias in disguise?

I wonder if there's a correlation between our pessimism and broadband penetration. If it tracks well, then our pessimism may be a result of just having better access to communication than we did before. In other words, our previous (relative) optimism was due to the fact that we simply didn't have enough of the facts of the world.

8 writes:

People also think their children aren't doing drugs. It could be optimism bias.

Why does increased TV coverage mean people are pessimistic? Maybe crime is falling, but the public didn't know the real level. In that case, the public's tolerance for crime is much lower than previously thought.

Jeff writes:
As long as parents correctly perceive that their local areas are safe, why would national news make them fear for their children?

Maybe most people are kind of, well, you know, stupid?

Steve writes:

My neighborhood/suburb has an acceptable crime rate, but the downtown area I trek to regularly doesn't. Ergo, crime in the US is a problem, by not in my area.

Or maybe I just feel bad for the people in those areas of Detroit, DC, etc. for whom crime is a constant worry. Therefore, it's a serious problem (in the US), but not in my area.

TDK writes:

There is definitely a pessimism bias when it comes to crime. And I believe a major cause has to be media, both local and national. I personally cannot remember the last time I turned on the news and the majority of the stories feature weren't about crime. I just think the reason people feel more secure about their general area is that they feel they have more control over it and therefore feel safer about it.

agnostic writes:

Here's an excellent review of the decline in all forms of child abuse since the early '90s:

http://www.unh.edu/ccrc/pdf/CV137J.pdf

The only category that hasn't plummeted is parental neglect. Petty criminals have been behaving better and better, but those damn lazy parents aren't neglecting their children any less!

Here's some more long-term data on violent crime trends:

http://www.gnxp.com/blog/2008/06/your-generation-was-more-violent.php

It may be higher now than in the '50s, but that's a Nirvana comparison -- people, driven by their pessimistic bias, will always choose the best time as the comparison point, sneaking in the "it's never been worse" view through the back door.

Few will look any farther back than the '50s because they will see that things were pretty bad in the '20s and '30s. Remember the St. Valentine's Day Massacre, that whole string of movies about The Godfather, seedy jazz clubs, reefer madness, flappers pretending they're men, etc.?

Here's what the secular trend over the past several centuries looks like:

http://www.gnxp.com/blog/2009/08/what-does-decline-in-homicide-rates.php

Again, things just keep getting better.

Eric Johnson writes:

I have a very different explanation. I would suspect that people's evaluation of their own locale is more empirical, while their impression of national conditions is more likely influenced by what they intuit is going on in the social order overall.

I mean, that's what I would do. A couple years ago I didn't know basic facts such as empirical crime rates over recent decades, and obviously neither do most people. The question about my locale would have prompted me to say what I actually know - and the national-level question, since I lacked the ability to answer it factually, would have invited me to think more generally and "philosophically" about whether or not things are "going to heck."

From a learned perspective, the pessimistic view has two pillars. First: what Prakhar said. The decline is recent and follows the increase that started in the 60s. Second, it is highly plausible that a major cause of the recent decline is the astonishing spike in imprisonment under Reagan, Bush, Clinton, and Bush (note that this is gross imprisonment, not per capita):

http://en.wikipedia.org/wiki/File:US_incarceration_timeline-clean.svg

I'll bet most non-scholarly people don't know about these secular trends. (For one thing, the media seems likelier to highlight the current imprisonment rate rather than the shocking secular increase - but of course most people probably can't state the current rate either.) However, I'll bet people have other, non-scholarly ways of sensing the ongoing social decay, which is real. They probably note things like the incidence of obnoxious public behavior and the average degree of trust and pro-social behavior.

If I'm right, perception of national crime conditions should correlate strongly with overall view of progress or decline in civilization. For people's view of local conditions, such a correlation should be far weaker. I should probably learn how to use the GSS.

Krinken writes:

As a parent and a behavioral scientist my own anxiety, I think, comes from 2 places. Evolution (instinct) and the media. Although I know I live in one of the safest places in the US, the pervasive stories and images of horrific crimes against children override rational thought and keep me up most nights.

Mr. Econotarian writes:

My pessimism is that the kid will never learn to live on their own, never move out, get pregnant at a young/unmarried age, get hooked on opiates, be a hopeless coach-potato basket case, blaming me for their problems and fighting with me for the rest of their life.

It is scary when you see it happen to someone else...

(OTOH, even the most badly behaved dog is typically only a 15-year problem, you can legally get them "fixed", they can't inject drugs, and you don't need to pay for them to go to college that they might fail out of and waste your money.)


Doc Merlin writes:

1) POSSIBILITY 1
There is a good reason for this type of pessimism bias. In a society with a strong tradition of self defense and mutual protection, this allows people to stamp out initiators of violence more effectively.

2) POSSIBILITY 2
What matters to people isn't actual initiation of violence rate but the marginal rate. So, as crime rate falls, people become more wary to it and more worried about it.

AleG writes:

[Comment removed for supplying false email address. Email the webmaster@econlib.org to request restoring this comment. A valid email address is required to post comments on EconLog.--Econlib Ed.]

Comments for this entry have been closed
Return to top