« Hey, me too! | Main | Ask questions later. Or not. »

January 06, 2006

Crime and embellishment

This little tidbit from David Frum has been getting a fair bit of attention:

Canada's overall crime rate is now 50% higher than the crime rate in the United States. Read that again slowly—it seems incredible, but it's true. It's true too that you are now more likely to be mugged in Toronto than in New York City.

This is just little old contrarian me talking, but I think the bold statement above helps National Post readers much less than a source for this statistic would have. Because that stat is out there, baby, and it's loving every minute of it, but until this week I had no idea where it came from. Now that I do know where it comes from, sort of, I'm even more skeptical.

In 2001, the Canadian Centre for Justice Statistics published a "Feasibility Study on Crime Comparisons Between Canada and the United States." It determined, first and foremost, that due to major differences in the way certain crimes are defined in the two countries, "the direct comparison of total crime rates is inappropriate."

(The idea that a woman is twice as likely to be raped in Canada as in the United States is an extreme manifestation of this disconnect. As the feasibility study says: "Due to significant definitional differences, it is impossible to compare sexual assault in Canada to forcible rape in the United States.")

After some statistical adjustment, the study determined that homicide, robbery, aggravated assault, break and enter, theft, motor vehicle theft and arson were comparable between the two countries, and the ensuing analysis, "Crime Comparisons Between Canada and the United States," provided the following figures for 2000:

By combining the above, one arrives at basically identical (though obviously incomplete) crime rates: 3977 per 100,000 in Canada vs. 4124 in the US. This makes sense to me: the Canadian and American rates are roughly comparable except in homicide, aggravated assault and robbery, which are the sorts of crimes associated with highly criminal metropolises like Detroit and Washington that we simply don't have in Canada, but these make up a relatively small portion of the crimes committed in both counties — that is, though more gruesome than "other theft" or "break and enter," they have little effect on the overall crime rate.

So what's with the "50% greater"? First of all, Frum isn't actually talking about the "crime rate" but about the "crime victimization rate," which relies on citizens self-reporting whether they were victims of crime in the preceding year. (That's just what it sounds like — the crime, if indeed there was one, needn't even have been reported to police to qualify.) I'll let his e-mail reply to me take it from there (links and illustrations added, obviously):

You'll have to do a little arithmetic, but ....

If you scroll through this StatsCan report, you'll see that in 2004 Canadian households were victimized by crime at a rate of 248 out of 1000 households

and if you look at this one, from the US Bureau of Justice Statistics, you will see two numbers: the one on the right measures property victimizations per 1000 households - and that figure is 150 [it's actually 161.1 –ed.] - the one on the left measures violent victimizations per 1000 individuals. That figure is 20 is per 1000 [actually 21.4 –ed.].

Here are the charts from which he took his rounded-down numbers:

And here are the numbers themselves:

Now making the assumptions that ALL of those 20 per 1000 victims of violence are in different households from the 150 per 1000 households tht suffered a property crime, we reach the total of 170 per 1000 victims of all forms of crime.

248 is 45.88% greater than 170.

If say half the US victims of violence lived in a household that also suffered a property crime, then we'd have a total US victimization rate of 160.

248 is 55% greater than 160.

So 50% seemed a good approximation.

Confusion #1: Frum added the US property victimizations to the violent victimizations. I don't understand why he didn't do likewise for the Canadian figures, i.e., add the 248 "household victims" to the 106 "violent victims" you see here:

That would have gotten him a Canadian crime rate of 354, which is more than double the American one and roughly twice as unbelievable as the 50% number he arrived at. But I'll go with 248 as the Canadian crime victimization rate to start, for the sake of argument and because it's so easy to disprove. Here goes.

First of all, because Frum got the numbers wrong, the actual number for the US isn't 170 (150 + 20) but 182.5 (161.1 + 21.4), which brings us down to 36% higher, as opposed to 45.88.

Second of all, fully 31% of the household crimes involved in that 248 number were vandalism (see chart at the bottom of this page), and the semi-comparable American number, you'll note above, does not include vandalism at all. Subtract vandalism from the Canadian numbers and you're down to 171, which is pretty much exactly equal to the American rate.

I'll stop there for now, until I understand why Frum seems to have underestimated the Canadian crime victimization rate (according to his methodology, that is, which I obviously distrust). There's certainly plenty more to fret about if one is to consider the two countries' victimization surveys comparable, including the idea that there were ostensibly more than twice the number of sexual assaults in Canada in 2004 than there were in the United States — that's raw numbers, not a per capita rate, and obviously a result of wildly inconsistent definitions.

I note too that Frum suggested to a relapsed Catholic that "Canada presents its crime data in a weirdly propagandistic way," which might explain why he resorted to this agonizing and imprecise manipulation of statistics detailing whether people believe themselves to be crime victims instead of just sticking with easily available statistics detailing, er, the actual crime rates. He hasn't responded to my e-mail asking for clarification of this propaganda angle, so again, I'll hold off on further analysis until I better understand what he's driving at.

I will ask, though, why he couldn't have just gone ahead and called the 50% figure what it was — his own artistic interpretation of crime victimization, and not the "overall crime rate" as he stated.

[UPDATE January 16.]

Posted by Chris Selley at January 6, 2006 11:53 PM

Trackback Pings

TrackBack URL for this entry:
http://www.tartcider.com/cgi-bin/mt-tb.cgi/218

Comments

I presented a similar analysis yesterday at SDA, and reached much the same conclusions (or at least the same level of scepticism). If interested, please see the comments section at:

http://www.smalldeadanimals.com/archives/003272.html

Posted by: 2Sheds at January 7, 2006 01:58 AM

Thanks, Chris, for the work on this one. I was baffled when I first read Frum's assertion in his Post column.

Posted by: Adrian at January 7, 2006 09:11 AM

My take on the Frum statement was that he was referring to specific areas of Canada compared to the overall rate in the US...not a fair comparison with one study being local and the other rationalized and broad based. However, I think your analysis may be flawed in 2 aspects...first we need to work with 2005 figures ( this is where Canada's rate spiked in one or two specific areas) and second we need to compare apples to apples.

If we care to do that, we would find that the US indeed has areas of their nation with lower crime rates than comparable areas of our nation and vice versa...but in overall agrigate numbers the US will always have a higher true crime rate because there are more metropolitan areas, more wealth so more property crime and more people so more opportunity for crime and more enforcement with higher capture and conviction rates to stamp on the numbers....percapta rationalization does not compensate for these obscured imbalances.

Thre best sourc4es I have seen for accurate balanced crime rate studies are David Kopel and Gary Kleck...there an no reliable sources in Canada as the issue is too political and you get vastly disparate numbers dependant on your source.

If you are interested in rsearching US Crime statists there are no better sources for true figures than the ones gleened in the research of David Kopel or Prof Gary Kleck. In Canada we are really devoid of a reliable source but the best study I ever read was commissioned by the justice departement and looked at the arrest profiles, case summaries and sworn complaints from the provincial court systems accross Canada. The conclusion on this report was that the data sets of violent crime were too small to be conclusive of any theory on causation and that most major crimes are perpetrated by individuals who engaged inprevios similar crimes where complaints were not filed...much of Canada's crime rate goes unreported.

Posted by: wlyonmackenzie at January 7, 2006 09:11 AM

Nice piece of deconstruction. Sometimes the counterintuitive is not the truth. :)

Posted by: Dr.Dawg at January 7, 2006 12:08 PM

wlm,

Just to clarify, Frum was looking at 2004 numbers just like I was. As for comparing apples to apples, my whole point was that Frum wasn't doing so.

Posted by: Chris Selley at January 7, 2006 12:36 PM

One thing I really "liked" about his article is that he pointed out the falling crime rates in the US as proof that their approach to prision is working.

In the book "Freakonomics" it is very interesting to read that the lower crime rate doesn't seem to relate at all to whatever the police is doing but (hold your breath and start screaming) the legalization of abortion in the US in the 1970s, which lead to a "domino effect" 20 years later.

Not a very nice notion if you think about it, but the book is a really interesting read when it comes to typical misconceptions about statistics.

Posted by: Michael [TypeKey Profile Page] at January 7, 2006 03:15 PM

Chris - speaking of apple and oranges:

You can't add the numbers for violent crime and property crime because the first is measured per 1000 population (teen and older) and the latter per 1000 HOUSEHOLDS.

Assuming the statscan numbers and the U.S. figures are roughly comparable if vandalism is removed from the Canadian numbers, property crime comes out at (per 1000 HOUSEHOLDS) 161 for the US and 204 for Canada. 25% higher in Canada.

The violent crime comparison (per 1000 pop) is 21 for the US and 80 for Canada (excluding sexual assault/rape). By these numbers violent crime is at least 4 times higher in Canada than the US.

Posted by: JR at January 7, 2006 10:13 PM

The "abortion domino effect", and "Freakanomics" in general, have been challenged and don't stand up well to the criticism (I agree with the counterarguments). Look to other reasons for decreases in US crime rates.

Posted by: lrC at January 8, 2006 01:27 PM

Hope you seen and heard this.....

Bruce Bowie, a former RCMP Inspector, and now working for Investigative Solutions, has released a damning report on the crime rate in Canada on Peter Warren today. To listen to the report you would have to sign up at the CKNW website and go to the audio vault for the Peter Warren show dated today. This report basically slaps all of the government for it's "Huge a Thug" mentality. To verify the information review the audio. The accusation is that we have been lied to for all these years, and his facts and figures show that this is indeed the case.

Stats from United Nations survey of crime trends and other sources.
Canada
population
Current population thru criminal justice system 400,000
Current yearly incarcerating rate both federal and provincial - 32,000
12,000 federal
20,000 provincial

Employees at correction Canada in 2004
15,000 for 12,000 inmates

Police personal
Canada - 186 per 100,000
USA - 326 per 100,000

incarceration Rate
2002(latest year available) does not include young offenders
Canada - 102 per 100,000
USA - 700 per 100,000

Crime rate in Canada have tripled since 1962

Crime rates(2002)
Canada - 8025 per 100,000
USA - 4100 per 100,000
Germany - 7800 per 100,000
Italy - 3800 per 100,000
Mexico - 1500 per 100,000

Fact:

Canada has one of the highest crime rates in the western world, the lowest incarceration rate in the western world, and lowest rates of police personal in the world.

As if we didn't all see that, but it's much worse then anybody thought.

Posted by: Anonymous at January 8, 2006 08:44 PM

Anonymous, the stats you reference don't take very much research to debunk. Please see:

http://www.unodc.org/unodc/en/crime_cicp_survey_eighth.html

and pull down any of the PDFs. (I used “sorted by country”.)

If you look at response line item 2.1 for each of Canada and the U.S., you’ll see the crime rates exactly as you cited.

But you’ll also note that the rates for individual offences in the subsequent line items show that the U.S. typically has higher rates than Canada. (As noted in the U.N.’s survey instructions, the individual crime rates below line 2.1 don’t have to add up to the total shown in that line.)

The following examples illustrate what I mean, using the rates from 2002 (sorry I can’t do a nice table in the comment section…references shown as line item, and US rate vs. Canadian rate):

2.2 Completed homicide: 5.62 vs. 1.67

2.4 Completed homicide by firearm: 3.25 vs. 0.46

2.6 Major assaults: 310.14 vs. 8.56

2.10 Major thefts: 969.14 vs. 63.40

2.11 Thefts: 2445.80 vs. 2195.25

It’s true that some stats are higher for Canada (e.g. line item 2.8 rapes, 2.12 auto theft, and 2.13 burglaries), but hardly enough to justify claiming the overall crime rate is twice as high as the U.S. (If so, what exactly are the crimes that aren’t shown in the line items on the chart? Maybe there needs to be a line for federal government malfeasance.)

Also note that the U.S doesn’t even provide data on items like less severe assaults, frauds, drug offences, and kidnapping.

To my mind, based on the above, it’s obvious that the U.S. simply doesn’t include as many offences as Canada in its line 2.1 “Grand total” number, which is the basis for a key assertion in your comment.

So, once again, claiming that Canada has a substantially higher crime rate than the U.S. is misleading -- it's certainly not "worse than anybody thought".

Posted by: 2Sheds at January 8, 2006 10:57 PM

Why would anyone pay attention to David Frum? It's not like he's honest. Like any good political operative he twists facts to suit his agenda. Frum's agenda has long been that Canada sucks. How dare we offer a viable alternative to Republican fantasy.

Frum is offering propaganda to counter what he sees as Canadian-centric propaganda.

He should stick to writing bullet points on how to end Evil. And Anonymous should look up the meaning of fact.

Car theft and homocide are the only reliable crime statistics. Property crime is totally unreliable as it is badly and inconsistently reported.

Rape is particularly problematic, until recently many juristications didn't recognize 'spousal rapes'. Rape is not consitently reported, there is not even consistency within the inconsistency of its reporting, making an honest comparision nearly impossible.

Anyone with even a slight familiarity with statistics and social science knows how problematic property crime statistics are.

Or you can believe David Frum.

Posted by: wsam at January 9, 2006 11:45 AM

Pay Day Loans http://www.idealpaydayloans.com/pay-day-loans.htm

Posted by: Pay Day Loans at March 28, 2006 01:35 AM

Did you know that everytime you surf the internet you could be at risk for identity theft? Do you know how your identity is stolen? Are you concerned about identity theft happening to you? Do you know what identity theft really is? Do you want to know what identity theft has to do with you? Do you know how to prevent becoming a victim of identity theft? Are you protected from identity theft? Are you looking for help with identity theft or information about identity theft? Do you want a place with a large selection of identity theft information sources and help with identity theft prevention? Get your identity theft information and identity theft help at identitytheft-info.org and learn how to protect yourself from identity theft.

Posted by: IDENTITY THEFT INFO at May 17, 2006 04:37 AM

With the constant rise in violent crime and drug crimes that are committed on homes, how much home security do you have to keep your family safe? How much home security do you need? The police are not always able to protect your home's security and your family as the number of police officers are becoming fewer with budget cuts every year. Have you stopped to consider how effective your home security really is? Could the steps you are taking for home security really protect your home and family against an intruder? Do you want more information about home security and what kind of home security is right for you? Do you need help with home security options? Does your home even need home security, and if so, how much home security do you need? You can find help with home security questio
katherine rice (5/17/2006 11:08:45 PM): questions and home security information on this site and many different home security options. Find the home security option that best fits your home security needs here.

http://home-security-help.biz/

Posted by: home security at May 18, 2006 12:56 PM


According to the Federal Reserve, Americans carry on average, $5,800 in credit card debt from month to month. Making the minimum monthly payment on that debt would take 30 years to pay off and include an additional $15,000 in interest. According to the Administrative Office of the Courts, 2,078,415 bankruptcies were filed in 2005, the largest number of bankruptcy petitions ever filed in any 12-month period in the history of the federal courts. With mounting credit card debt and the new tougher bankruptcy laws, people are looking for alternative ways of managing their debts. Debt consolidation loans have become a popular way to free up money each month by consolidating several monthly credit card payments into a single lower interest loan. But, the question is whether it's best to consolidate your debts into a home equity loan or an unsecured debt consolidation loan.

http://www.debt-consolidation-help.org/

Posted by: Debt at May 19, 2006 06:26 AM

WANT TO EARN $10,000 A MONTH ?
Top earners make $50,000 + a month some well
into six figures. It is not unheard of for
beginners to make $100-200 a day when they
first start out. Don't make the mistake of doing
nothing or spending time, money and effort on scams.
Learn from the masters and start earning today.
Google Cash, Beating Adwords, The Rich Jerk, Adsence

http://www.twofour.biz

Posted by: TwoFour at May 20, 2006 02:28 PM

weightloss or a diet? do you need diet information? Are you looking for a place you can go to find different avenues of weightloss and diet information with different options for weightloss and diet? We have the answer to your weightloss and diet questions and alot of sources for weightloss and diet information and weightloss and diet help. Try looking for your weightloss and diet needs at weightloss-diet-help.com

http://weightloss-diet-help.com/

Posted by: weight loss at May 21, 2006 03:33 PM

WANT TO EARN $10,000 A MONTH ?
Top earners make $50,000 + a month some well
into six figures. It is not unheard of for
beginners to make $100-200 a day when they
first start out. Don't make the mistake of doing
nothing or spending time, money and effort on scams.
Learn from the masters and start earning today.
Google Cash, Beating Adwords, The Rich Jerk, Adsence

http://www.twofour.biz

Posted by: TwoFour at May 25, 2006 11:23 AM

Online Money Making Guide to Success? View great articles that will help you successfully earn money online. Modern life articles to deal with every day situations.

Posted by: Articles Direct at June 1, 2006 10:46 AM

Online Money Making Guide to Success? View great articles that will help you successfully earn money online. Modern life articles to deal with every day situations.

Posted by: Articles Direct at June 2, 2006 10:14 AM

visit http://www.structured-settlements-info.org . Need cash now ? Use you fure cash settlements to get it. From structured settlements, annuities, lottery payments, and on-going litigation settlements.

Posted by: Structured Settlements at June 4, 2006 04:11 AM

Here you will find different ways to make money online. Money can be made several ways. You can make money doing surveys, you can make money reading emails, you can make money selling a product, you can make money doing what you are doing right now searching or money can even be made when others search on your sites. Making money on line can be fun as well as profitable. How you will do depends on what you choose to do and how much time you have to put in it.

Posted by: money at June 7, 2006 12:40 AM