Lott, cherry picking

Over the past few years crime rates in Australia, Canada and England have fallen dramatically.
For example, in NSW crime plunged to the
lowest level in 20 years, in Canada, the 2003 homicide rate was the lowest in 36 years, while in England the crime rate was the lowest since the BCS started in 1981. While crime has been plummeting, John Lott has been drafting a steady stream of op-eds blaming gun control for increasing crime in those places. His secret? Cherry-picking.

Lott’s latest column is a little unusual amongst his cherry picking efforts in that he provides links to his sources. He writes:

And with Canada’s murder rate rising 12 percent last year

The trouble with this approach is that readers can click through and read the parts of the report that he chose not to mention. Here are the section headings, leaving nothing out.

Violent crime down but homicide rate up … Robberies with a firearm continue to decline … Property crime resumes downward trend … Drug incidents resume upward trend …Youth crime down

And look at how the increase in homicide was reported:

Canada’s homicide rate rose 12% in 2004 after hitting a 36-year low the year before.

Lott conveniently left out the second part of the sentence. He also says:

With Canada’s reported violent-crime rate of 963 per 100,000 in 2003, a rate about twice the U.S.’s (which is 475), Canada’s politicians are understandably nervous.

Lott does not tell his readers that the ‘violent crime rate’ in the Canadian statistics includes simple assaults but in the US statistics it only includes aggravated assaults. If you compare the same crime categories, violent crime rates are lower in Canada.

Lott also cherry picks some English crime statistics:

The 2000 International Crime Victimization Survey, the last survey completed, shows the violent-crime rate in England and Wales was twice the rate of that in the U.S. When the new survey for 2004 comes out later this year, that gap will undoubtedly have widened even further as crimes reported to British police have since soared by 35 percent, while those in the U.S. have declined 6 percent.

Lott does not mention that the crime victimization rate in England has decreased significantly—the increase in crimes reported has occurred because the police have improved their record keeping, not because there has been any increase in crime.

And he cherry picks Australian ones:

Australia has also seen its violent-crime rates soar immediately after its 1996 Port Arthur gun-control measures. Violent crime rates averaged 32-percent higher in the six years after the law was passed (from 1997 to 2002) than they did in 1995. The same comparisons for armed-robbery rates showed increases of 74 percent.

I’ve put the Australian statistics in a spreadsheet so you can see for yourself that he has selected the crime rate and the basis for comparison to conjure up some crime increases. There was a temporary increase in the armed robbery rate after the 1996 laws, but since then the armed robbery has fallen below what it was when the law was passed. More importantly (and you will never hear this from Lott), the firearms murder rate has halved, falling from 0.32 per 100k in 1995 to 0.16 in 2004. The non-firearms murder rate did not change significantly.

I’ve long been opposed to the 1996 laws because I didn’t think they would have a significant effect on crime, but the latest Australian crime figures are making me waver because it’s likely that laws were responsible for at least some part of the reduction.

Comments

  1. #1 Hiram Hover
    August 23, 2005

    A great analysis. You’ll no doubt see it soon enough, but Lott argues to similar ends in an op-ed in today’s Washington Times, claiming that we’d all be safer in the workplace if we just toted pistols along with us.

    ‘Stop
    Endangering Employees’

    The “proof” for this jaw-dropping claim comes from his wider research about the effect of right-to-carry laws.

  2. #2 Tim Lambert
    August 23, 2005

    That op-ed is pretty much a cut and paste from some of his other op-eds, so it won’t take long to debunk.

  3. #3 z
    August 23, 2005

    Regarding simple assault in Canada, the common wisdom there is that mild bar fights are much more common than in the US, largely because the chances of getting shot in one are less. (This is considered to be a point in favor of Canada.) One may take that as one wishes; perhaps that’s the proof of the “armed society is a polite society” cliche, or maybe it’s just a totally fictitious urban legend.

    Also, following your link to the Canadian crime stats, given their 2/3 of American homicides involve a gun, 1/3 of Canadian, and 5.5 and 1.8 homicides per 100,000 respectively, gives 1.8 and 1.2 nongun homicides per 100,000, obviously much more comparable than the rate of gun homicides in the two countries. This suggests that the general level of murderous intent is perhaps not so much greater in the US as is sometimes postulated, as well as that all those guns are not exactly preventing nongun homicides.

    Re the Australian stats, I don’t suppose there are any figures breaking them down into domestic type homicides, Kellermann style, are there?

  4. #4 z
    August 23, 2005

    Me again… Checking out the Lott oped in the Wash. Times,
    “The real question is why the two firms bringing the case, ConocoPhillips and the Williams Co., are doing so.”
    This being the US, without much doubt the answer would be so that they can demonstrate that they took “all reasonable precautions” to ensure the safety of their employees, as a preemptive measure vs. any future lawsuits.

  5. #5 ben
    August 23, 2005

    >This being the US, without much doubt the answer would be so that they can demonstrate that they took ‘all reasonable precautions’ to ensure the safety of their employees, as a preemptive measure vs. any future lawsuits.

    Yeah, but they can then get sued because an employee was prevented from carrying their pistol at work (prevented from keeping it locked in the car, I think) and then they were hurt during a crime when they could have otherwise defended themselves.

    >Regarding simple assault in Canada, the common wisdom there is that mild bar fights are much more common than in the US, largely because the chances of getting shot in one are less. (This is considered to be a point in favor of Canada.)

    Two things, I just got back yesterday from Canada where my 20 year old brother in law was sucker punched at a bar and had his cheek fractured and got a big gash under his eye. Nice. Secondly, it’s called *hockey*.

  6. #6 Ian Gould
    August 23, 2005

    >This suggests that the general level of murderous intent is perhaps not so much greater in the US as is sometimes postulated

    I haven’t looked at the most recent figures but for the first few years after the 1996 gun laws, the number of homicides in Australia fell – and the number of attempted murders rose by an almost equal number.

  7. #7 Steve McIntyre
    August 24, 2005

    There are a lot of variables affecting bar fights in Canada over a generation. First, drinking ages in bars have been lowered from 21 to 18, so you get a lot more young people out in bars (which might also account for a difference in US/Canadian rates). Secondly, the number of young people in bars and clubs on a nightly basis has increased dramatically. A huge industry of clubs has developed with tens of thousands of young people in the entertainment district virtually every night. Parking lots in downtown Toronto near University and Queen are as busy at Thursday midnight as during business hours. Thirdly, police charging policies for teen-age scuffles have changed over the years. Things that wouldn’t rise to a charge 20 years ago always lead to charges now. This is partly social policy, but police union contracts have changed over the years to increase overtime rates for witness fees, so there are considerable incentives for them to charge people, rather than warn them. I don’t think that guns have anything to do with changes in bar assault rates in Canada. In passing, Canada is proportionally more urbanized than the US; the bar entertainment scene in Toronto is much larger than comparable US cities for a variety of reasons: one of them is that people feel safer downtown.

  8. #8 ben
    August 24, 2005

    IFIRC, in most provinces the drinking age is 19, not 18. Alberta is 18, but BC is 19.

    From living in both cities, I don’t see any difference in how safe people feel in downtown Seattle or downtown Vancouver. In fact, I feel much safer in downtown Seattle because I can go there armed. I remember feeling somewhat helpless in downtown Vancouver.

  9. #9 dave heasman
    August 25, 2005

    In countries where murder is rare, selecting murder rates for a single year is always going to give results that can’t be used to prove a point. When a mass-murder is identified, like Harold Shipman in England, his murders are all attributed to the year in which they were identified, rather than the years in which they were committed. Obviously, when Shipman commits 200 murders and the annual national average is less than 1000 you’re going to get skewed figures.
    Similarly wasn’t it only last year that the missing whores of Vancouver who ended up as pig-feed were classified and counted as murder victims? That’ll skew your figures, too.

    As for Ben “I feel much safer in downtown Seattle because I can go there armed”, well, I’d therefore feel less safe in Seattle because some inadequate might be tooled up.

  10. #10 Mike Maltz
    August 25, 2005

    I don’t believe that Dave is correct in his comment (9, below). In the US a murder is recorded as occurring on the date it is discovered, not on the date an arrest (or “clearance”) occurs — irrespective of whether the offender is known or unknown at the time. I assume that it is the same for the UK and Canada.

    I agree with his wider point, that “selecting murder rates for a single year is always going to give results that can’t be used to prove a point.” That is the essence of Lott’s cherry-picking; moreover, if he can’t find evidence in the rate, then he looks at the change in rate, or in some other statistic that will “prove” his argument.

  11. #11 dave heasman
    August 25, 2005

    If you check the Shipman & the “whores as pigfeed” cases and indeed Fred West, too, you’ll find that murders that took place over a number of years were “discovered” at a single point in time. Thus screwing the stats.

    It’s not as relevant to the US, because I don’t think the US qualifies as a location where murders are sparse.

  12. #12 Meyrick Kirby
    August 25, 2005

    So does anyone disagree with the notion that Lott’s comparisons are overly simplistic?

  13. #13 ben
    August 25, 2005

    >As for Ben ‘I feel much safer in downtown Seattle because I can go there armed’, well, I’d therefore feel less safe in Seattle because some inadequate might be tooled up.

    In Vancouver I guarantee there are *inadequate’s* tooled up, so to speak. They are criminals who are tooled up illegally. This just made *their* parts of town all the more uncomfortable. At least here, the adequate people with guns outnumber the inadequate. And even so, Seattle is not particularly high on the gun crime scale, compared to the rest of the US. But guns and concealed carry permits are perfectly available here for the law abiding.

    In fact, the Seattle police are *required* to carry a concealed pistol whenever they are *off duty* within the city limits.

  14. #14 Ian Gould
    August 25, 2005

    Mike,

    Dave is correct, in the Shipman case he prepared the death certificates for most of his victims himself and recorded the deaths as being of natural causes.

    It was only when the bodies were exhumed and re-examined that they were found to have been murdered.

    Similarly in the Vancouver case, because the bodies were never recovered the dead women were officially recorded as “missing” even though it was strongly suspected they’d been killed.

  15. #15 wcw
    August 26, 2005

    In re: armed in Seattle, why? There will be maybe one violent
    crime there this year for every one thousand residents. The
    chances that some “inadequate” notices your concealed weapon
    and decides to prove his mettle by singling you out would seem
    higher than the one-in-a-thousand you stumble into bad luck.

    If your work necessitates violence, of course, my criticism
    is misspecified. I understand the utility of weaponry to
    law enforcement and lawbreaker alike.

  16. #16 ben
    August 26, 2005

    It’s unlikely, but it does happen. And since it hurts no one, why not?

  17. #17 Meyrick Kirby
    August 28, 2005

    >>And since it hurts no one, why not?

    I wonder how many serious or fatal firearms accidents occur each year in the US?

  18. #18 Ian Gould
    August 28, 2005

    I tend to be cautiously in favor of concealed-carry laws on the basis that the best current evidence is that while they don’t decrease serious crime neither do they decrease it.

    In a liberal state, personal choice should only be constrained when there is clear evidence of an over-riding public interest.

    When one considers concealed-carry laws solely from a crime-prevention perspective there is no apparent over-riding public interest.

    however, I have to wonder if there are other, less obvious, costs resulting from such laws.

    The cost of accidental injury and death is one such cost.

    The burden on bars and night-clubs and other businesses which have to choose between requiring customers to surrender their weapons and risking a negligence law-suit if a client is shot in their premises is another.

    I don’t know of any data which would help to answer these questions.

  19. #19 ben
    August 29, 2005

    >I wonder how many serious or fatal firearms accidents occur each year in the US?

    I’m glad you asked. One of the best sources for this type of info on the web is at [guncite](http://guncite.com/gun_control_gcgvacci.html)

    The chart on that page is for deaths only, but it seems reasonable that serious injuries follow a similar trend. As you can see, the number of accidental firearms deaths is pretty small, compared to other accidental causes. Note also that injury and death rates from firearms accidents have been on the decline for quite some time. Mandatory firearms safety training in schools certainly would help this as well, but the anti gun groups are ardently opposed. Why?

    And as I reported on my blog a while back, it’s a lot less dangerous than [owning a dog](http://carnabyfudge.blogspot.com/2005/01/is-dog-more-effective-for-home-defense.html).

  20. #20 Jonathan Dursi
    October 30, 2005

    For those keeping count, Lott has drafted a lightly edited (and link-free!) version of this same op-ed in Canada: http://www.canada.com/national/nationalpost/news/issuesideas/story.html?id=ef2f258a-b6bb-4f85-b1cc-cdcdf3218c24

The site is undergoing maintenance presently. Commenting has been disabled. Please check back later!