Think tank questions Canadian crime data

Statscan’s falling crime rate claim is wrong: report


The Macdonald-Laurier Institute, an Ottawa-based think tank, has come out swinging against Statistics Canada’s evidence of falling crime rates in a new report that supports the Harper government’s goal of being tough on crime and expanding prisons. Scott Newark, a former Alberta prosecutor and a co-author of the report, concludes that “on the central question of the state’s duty to protect citizens from crime and public disorder, Canadians are not as well served as they should be” by Statscan’s data on crime. Chief among the Institute’s criticisms is that Statscan’s Juristat report on crime statistics over-emphasizes decreases in crime rates, revises crime categories from year to year and does not factor in unreported crimes. Statscan has responded to Newark’s criticisms, saying that while crime categories do change, data is available to anyone who wanted to compare crime statistics before and after categories were changed. Anthony Doob, a criminologist at the University of Toronto, says that Newark has “cherry-picked” his data, and that “there are no perfect measures of crime.” The Macdonald-Laurier Institute’s work has generally been supportive of the Harper government’s policies, and Scott Newark has previously worked with Treasury Board President Stockwell Day.

The Globe and Mail

Filed under:

Think tank questions Canadian crime data

  1. It's hard to believe it's a coincidence that both Day and his former advisor have come out with this notion that StatsCan is under-reporting crime and the real rates are going up.

    • You're assuming that the real rates are going up.

      Using the Newark method, I can also postulate that the government has been remiss in its duties to provide people with extermination services for faeries because, although people don't report them for fear of being labelled crazy, everybody just *knows* that there's more of them around these days.

    • I'm not sure if they are going up or not, but the General Social Studies has measured unreported crime (by asking people if they were victims of crimes and did not report it) for a long time. They have indeed found that the level of unreported crimes are high, and have been rising consistently.

  2. "…and does not factor in unreported crimes."

    Just how does one gather stats on that? I'd really love to know… esp. as it seems to be a major component of the CPC's "tough on crime" agenda.

    • From your gut, silly. You just watch the news and see how horrible things are and you "know" that there must be more crime out there than is being reported.. because the numbers that are being reported are falling.

      Just like how we "know" that travelling by plane is more dangerous than by car because, hey, how often do you see them reporting on car crashes?

    • Well, if you can stop the "Evil Harper" bashing I will enlighten you. What Day was referring to were University studies that were done with people going into actually neighbourhoods and interviewing people rather than pencil pushers at Stats Can waiting for police reports to come in. Crime has gotten so bad in many underprivileged parts that people know the game and that calling the police is a joke. If they show up at all, they file a report and that is the end of it. All they get for their trouble is being labelled a snitch and left to fend for themselves when the cops leave. Also victims of sexual assault often fail to report it. Also in today's press release it also mentioned that it's up to police to send in all the stats and often they don't as they are up to their necks dealing with…crime. Also Stats Can for years, for reasons beyond me, will report repeat offenders, who have been arrested yet again, having committed multiple crimes, as one crime. Now go back to your cosy neighbourhood, suck on a Starbucks and feel smug.

      • Thank you for the explanation. Just one question, what is an "actually neighbourhood"?

        I personally have evidence to the contrary. I have noticed that in all the areas where I have lived over the last thirty-odd years, actual crime on my streets and neighbourhoods has gone down. I'm sure I used to notice more crime back then than I do now.

        That's evidence, isn't it? I'm not a University Study, but then I thought intellectual elites were out of favour these days. Don't University Types drink Starbucks? I sure don't. Hate the stuff.

        • Really? And what "evidence", other than your own "personal" impressions ("I have noticed…") would that be, exactly?

          Or would it be more accurate to say that, in your opinion, the only crime that matters is the crime that YOU see.

          • Hey! That's not fair! You're talking about evidence again. Of course the evidence of my own eyes, or gut rather, is the only true measure, it's accurate 19 times out of 20.

      • So you trust anecdotal reporting of crimes to rather than the use of information provided by police departments and the courts? And do expand on this repeat offenders assertion – I am more than a little curious as to what you could be referring to?

    • The General Social Survey has measured this for a long time. They measure it by asking a large sample of people whether they were victims of crimes, and whether or not they reported it. They have consistently found that reporting rates have fallen since the early 90's. Many of the crimes people tend not to report are less serious true, but some are indeed serious (the reporting rate for sexual assault is very low). And while a top reason for not reporting crimes is the belief that the crime was not a big deal, another factor is the belief that the police are unlikely to solve the crime. That is a legitimate policy question .

  3. I don't know much about The Macdonald-Laurier Institute. Are they a parking lot of patronage Conservative positions like the Canadian Taxpayer Federation, or do they have aspirations of objectivity?

    • Yes?

  4. Keith in Brampton – if, in statistics, "randomness" and "uncertainty" can be accounted for, I'm sure there's a way that Statistics Canada could factor in unreported crime.

    • Perhaps – but an explanation as to HOW they come to their conclusions would give some weight to their argument. "Because I said so" or "my Magic 8 Ball says" doesn't cut it with most rational people.

    • Err not quite. The reason many of the statistics you read have a margin of error is because they are from samples. So if you randomly select a given number of people you can make abstractions about the general populace with some certainty.

      The Uniform Crime Report is a different kettle of fish. It is not a random sample of crimes. It is in fact a very un-random sample – those crimes that are reported differ systematically from those that are not. For instance reported crimes tend to be more severe, for instance (though some severe crimes, like sexual assault are also often unreported).

      Secondly, when you have margins of error, the idea is that there is some random noise that is equally likely to push something up or down. For instance, in a poll, there is a chance of errors, like accidentally classifying an NDP voter as a Tory. However, these things aren't likely to systematically bias the results one way or the other – it is just as easy to accidentally write NDP voter for a Tory. The effect of unreported crime isn't random noise – it is a systematic factor that always pushes estimates of crime lower than they actually are.

      In fact there is a way around this problem that involves a lot less magic. Statistics Canada does actually measure unreported crimes in their General Social Survey. They interview 22,000 people and ask them (among other things) if they have been victims of crimes and whether they reported the crime. The problem with this is:
      -the survey only comes out every 5-6 years
      -22,000 is a good sample size, but it would be nice to have a population measure
      -people may have subjective notions of what crime is
      -they don't ask about all categories of crime

  5. OK, let's assume for a second that this guy is on the right track with the 'unreported crime' business. How does the current law and order agenda of the government assist in this? Minimum mandatory sentences can only be applied when crimes are reported and successfully prosecuted. And, I'm assuming, people aren't reporting crime not because they think the criminal is going to get a slap on the wrist by a softy leftist judge, but because they think the police can't solve said crime.

    • Or out of fear of retribution. Many victims and witnesses don't come forward out of fear of what might happen if they speak up.

      I agree with your point that tougher sentencing won't help the reporting or conviction rate. Nor will it help with the rehabilitation of those who are capable of being rehabilitated. But for repeat offenders it will at least keep them off the streets longer.

      So there are aspects of their policy I agree with, but not all of it. And they need to explain how they are measuring unreported crime – now and in the past, as we need a comparator – if they want us to believe that crime really is on the rise.

  6. I would pay no attention to the stat. You can make the stats on any topic that show any result you want. It's just a matter of what data you use and how. Example: As is argued in the article: if you leave out unreported crime the stats show a completely different picture. So I don't pay much attention to the "stats".

    Also, even if crime rates are falling they still has reason to expand prisons. The fact that crime rates are falling doesn't change the fact that prisons are already over capacity. Either way, more prison space is required.
    The real question is: will this tough on crime approach negate the expansion and flood the prisons just the same or more.

  7. It should also be noted that there has been an exponential increase in what is now considered a "crime", particularly in the area of domestic violence and school-yard violence.

    While we may or may not support the expansion of relatively minor acts of violence into the criminal realm, the fact is that assuming the same conduct has remained constant, one would EXPECT a serious increase in crime – not because bad conduct has in fact increased, but because the police are paying attention to things today that they would have ignored 10 or 15 years ago.

    I'm a red-neck, Alberta conservative, and the reality is that the current Federal Conservative suggestions of the necessity to "get tough" on crime are over-stated and are little more than the conservative compliment to the liberal gun control effort.

    Emotional responses that energize your base without demonstrated utility to the tax payer.

  8. This whole "unreported crimes" nonsense in order to justify more prisons & longer sentences is beside the bloody point anyhow! If all this crime stays unreported, these new prisons and harsher penalties won't be applied/used after all. Circular/oxymoronic/call it what you will, this think-tank has blown it's credibility right there.

  9. It might have been better if Day had come out with his evidence for doubting Stats Canada prior to just stating his unsupported views. This looks too much like retro-fitting the evidence to support his pre-conceived ideas.
    The MacDonald Laurier Institute has lost what credibility they had by producing this report. Shilling at it's worst by a poor man's Heritage foundation.

    • He did. Its called the General Social Survey – he even referenced it. Perhaps you are the one who is too set in your preconceived notions to consider contrary evidence.

      • No he didn't, so enough with the hand waving and post event justification.
        The report by Newark is the Day report. That is the one which attempts to support Day's assertions..
        Picking holes post spouting without support and then looking for loopholes later is never good practice.

        • Fine, here is proof that you are wrong (from August 2010):

          "A spokesperson for the justice minister said Day was referring to the General Social Survey conducted by Statistics Canada that asks individuals if they have been victims of crimes and if they reported the incidents to police."

          • very good and the report by Newark justifying it was released today.
            He may have been referring to the GSS for figures but it took until this week that the rationale was invented.
            Throwing figures around with out methodology or justification is still not an argument, at best it is opportunism.

  10. Actually, this is a really good idea. There should be a new category of "Unreported Offences." These UOs should carry mandatory minimum sentences that are at least double – oh,hell – triple prison terms. The new prisons to be built should house all of the unreported criminals in maximum security conditions. In fact – YES! – there should be a major government push on ENCOURAGING people to NOT report crimes in order for more of the unreported criminals to fill more of the unreported prisons.

    There, it's solved.

    • Oh, such witty sarcasm!

      One might logically presume that an increased "law and order" agenda which refers to "unreported crime" might, just might, have another plank in its platform which deals with research into how and why crimes are unreported or unprosecuted and, dare I say it, plans to remedy that situation.

      But I presume from your post that it's more socially responsible to make sarcastic comments about tripling sentences for unreported crime, just because crime hasn't increased in YOUR neighbourhood.

      • Don't be silly. I thought we had decided NOT to use such elitist, high-brow, leftist stuff such as research?
        No, crime, panhandling etc HAS gone down in my area, using the currently accepted Canadian measure – my own gut feeling. Or, as it used to be called, "Because I say so."

      • I think it's late in the day for the Conservatives to start championing research. Please.

  11. Simple. Unbuild prisons for unreported criminals.

  12. *the UCR only includes reported crimes, not only prosecuted crimes.

  13. Statscan has little credibility, they have the same hidden agenda as the liberals, and that narrow-minded agenda is the real problem here, it pays them back in spades if they obfuscate, which is exactly what they do.

    • Nice clear cut execution of a talking points. I appreciate the brevity, chet would take 200 words to same the same thing.

  14. The NP has better coverage, more details and less politics. Statscan is looking rather useless.

    “It would seem to be the case that a gun-toting, drug-possessing, impaired car thief, on probation, bail, and subject to a firearms and driving prohibition, who is in possession of stolen credit cards and who crashes into another vehicle, thereafter flees, assaults a police officer, and gives a false name upon arrest would show as the single offence of whatever was deemed to be the most serious,” Mr Newark writes.

    The agency allows police department to report up to four charges for the same offence, but not every police department does, said John Turner, head of policing services at Statistics Canada. So the annual reports include only the most serious crimes to be able to compare rates between police departments. “You can't compare apples and oranges,” Mr. Turner said.

  15. I kind of wish the article had covered the substance of the MLI criticisms. Some of them make sense, while others do not.

    1. They criticize Statscan for using crime rates (per 100,000 people) instead of volumes of crime. This is a rather stupid criticism. Detroit has fewer murders per year than Canada (probably). Does that mean Detroit is safer than anywhere in Canada? However, it is a useful point for one issue – prisons. It is true that volumes of crime, not rates of crime determine the need for prisons.

    2. Statscan counts incidents, classifying each incident based on the most serious violation. This is a sensible critique – being robbed is bad. Being robbed and punched in the face is worse. On the other hand, while this may tell us that the UCR understates crime, it doesn't have any bearing on trends (unless there is a trend towards/away from multiple offences).

    3. The Crime Severity Index captures severity by sentences put forth by judges, who may be liberal pansies. Well sure, but the alternative is to count all crimes of the same type as being the same. I'm going to go out on a whim and suggest that – chardonnay-sipping judge or not – the judge's call does convey more information than a blanket assumption of equality between crimes. Now, they have a stronger point about the use of the CSI for youth crime. Given that the young offender's act limits sentencing of youths, raw incident counts may be more instructive.

    4. Stats Canada revises crime upwards after the fact. Well yeah – because data corrections are much more likely to involve crimes that were missed than things reported as crimes which were not crimes. Well that or Stats Canada is conspiring to make the police services look good.

    5. Stats Canada keeps changing its definitions. This is a problem, but surely at least part of the issue is that laws change as well. It is impossible to have a perfectly consistent measure over time.

    Data collection involves a number of difficult calls. How one makes those calls can have a significant impact at the margins, which makes it essential that people read the caveats and the methodology section in a new report. Most journalists fail to do so, and I think that, less than StatsCan's methods, is what drives MLI's critique.

  16. "The minister didn't provide evidence of an increase in unreported crimes, or a description of what types of crime go unreported, but said his office would follow up."

    His office followed up by mentioning the GSS to cover his ass, that is not an argument that is a holding statement. His office actually provided an explanation via the WLI on Wednesday when a case was made. He still hasn't answered the important question even if he is right about the vast underground occurrences of crime,

    "But even if the minister is correct, Holland wondered how that would result in the need for more prison spaces."

    He couldn't answer that then and can't now, so he attacks the source of information. Right out of the tobacco industry playbook.

    Wow you're calling me an ideologue who bases his ideas on gut instincts. Given the fundamentalist beliefs of Day and the recent right wing gut decision making in the US, being called that by an apologist for gut reacting ideologues is kind of a compliment.

Sign in to comment.