Counterfeiting Down!

The coverage devoted over the past couple of days to the recent StatsCan release about Canadian crime rates can best be described as "hallucinatory".  Where news reports do not rely on obfuscation and misdirection, they are forced to fall back on outright lies.  As an example of the latter, check out this passage from this Toronto Star story:

"It's not likely that it's easy to draw cause," said criminologist Neil Boyd, an author and professor at Simon Fraser University. "Any change on a year-to-year basis isn't likely to be significant unless it's extremely dramatic."

Although the effect of the Baby Boom generation on the crime rate is lessening, the aging of the population remains the single most important reason why violent crime rates have dropped over the past 30 years, said Boyd.

Um, no.  Would it have killed Neil Boyd, or, say, Toronto Star writer Francine Kopun to actually read the StatsCan report on which they're commenting?  Because if they would have deigned to do so, they would have discovered that "violent crime rates" have not "dropped over the past 30 years", they have skyrocketed, as helpfully demonstrated by this handy little graph contained in the middle of the StatsCan report:

What's that, nearly a doubling of violent crime rates over the past 30 years?  Nearly a quintupling over the past 40 years?  That the country's largest newspaper feels comfortable in printing flat-out lies should be cause for concern.

Have either Boyd or the reporter bothered themselves to look, maybe they would have noticed the following little nuggets in the StatsCan report:

The youth crime rate, which has remained relatively stable over the past decade, declined 1.5% in 2007 following a 3.3% increase in 2006. The 2007 drop was due to a decrease in non-violent crimes.

The youth violent crime rate remained stable in 2007 after increasing steadily over the past two decades. The 2007 rate was more than double the rate reported in the mid-1980s.

Or that, say, the rates for the most violent of "violent crimes", like, f'rinstance, the most serious assaults have increased over the past ten years by rates of, oh, 32.3% (assault level 2 - involving use of a weapon/bodily harm) or 18.6% (assault level 3 - aggravated).

Or, maybe, that the drop in the total crime rate (i.e., violent and non-violent) over the last ten years is driven largely by drops in property crime?  Scroll to the bottom of the StatsCan report, and you'll be pleased to learn that while, say, theft over $5,000 rates droped by 32.7% over the last decade, things like "forcible confinement/kidnapping" have increased by more than 120%.

Even with all the attempts by media outlets to spin the story by focusing on the year-over-year results (which even someone like Neil Boyd thinks is a bad idea) and efforts to commingle drops in crimes few people actually spending much time worrying about with massive increases in crimes people actually fear, the Canadian crime rate is, what, three times higher than it was in 1962?

Were headlines writers across this country more honest, rather than "CRIME RATES DROP!" being plastered across newsstands in 248-point font, we would be seeing "MINISCULE DROPS IN CRIMES YOU'RE NOT WORRIED ABOUT STILL DON'T PUT MUCH OF A DENT IN OUR ABSURDLY HIGH VIOLENT CRIME RATES!" - but I suppose that wouldn't "fit", now would it?

Comments

[Dr.Dawg]

Bob:

The graph you helpfully provide, though, shows a steady decline after a peak in 1992, for both violent crime and all crime.

That seems to me cause for cautious celebration--especially because you know, and I know, that the reporting rate for a lot of crime (e.g., sexual)has increased in the past forty years as well.

Looking a a three-year drop might be problematic, but so, I think is your analysis. Too many confounding variables.

[DCardno]

Must be a pony in there somewhere eh, Dawg?

Yes, here is a minor drop in violent crime from ~1992 until now. Just by eyeball, I would put it at slightly less than 10%. Big whoop - we are still in a dramatically more violent society than we were at any time prior to ~1989. In any event, I think the point is that the (greatly quoted) crimnology expert couldn't even be bothered to ascertain the facts before he started blathering.

[Dr.Dawg]

"we are still in a dramatically more violent society than we were at any time prior to ~1989."

Or, put another way, we have returned to the lower levels of violent crime that we had nearly two decades ago--an over-all decline over a twenty-year period.

I'm more interested in the slope: steep going up, steep coming down. Perhaps reporting rates reached their peak? Maybe there are a whole lot more variables in the mix? In any case, it seems a little late (almost 20 years too late in fact) for the Conservatives to start screaming about law and order.

[Bob Tarantino]

In any case, it seems a little late (almost 20 years too late in fact) for the Conservatives to start screaming about law and order.

*grumps* I've been whining about the crime rate for at least 20 years. */grumps*

[DCardno]

I'm more interested in the slope: steep going up, steep coming down

Only when looking at total crime. When looking at violent crime we see something more along the lines of "steep going up / barely coming down." Given the greater level of concern about violent, rather than non-violent crime, and the much greater likelihood of under-reporting of the latter (or of changes in the reporting frequency), I would be very careful not to oversell the idea that crime is coming down at all, let alone violent crime over a thirty year period, as did the good Professor. Of course, facts don't matter when you've got a story to sell, do they?

[Dr.Dawg]

Damn, you guys'll find a cloud for every silver lining, won't you? The graphs here do indicate a lower rate of decrease for violent crime. But it's a decrease. Part of what look like a more or less steady decrease over two decades.

What's not to celebrate here? Why poke around looking for flawed reasoning (Bob, I agree with you about the "30-year" stuff) when we have been observing a progressive decline in crime of all kinds, including violent, for many years? Let's just keep doing what we're doing, seems to me--it's working, whatever it is, across successive Conservative and Liberal federal regimes and several NDP, Conservative, PQ and Liberal regimes provincially.

All this, by the way, not to exclude the possibility of many variables at work--like the reporting rates that I mentioned earlier.

[dcardno]

What's not to celebrate here?

Um, a deliberate lie fed to the general population under the colour of authority? Willful ignorance of (or disregard for) known facts by a (supposedly) competent expert in order to support a convenient fiction? Lack of investigative accumen bordering on professional incompetence by the reporter? Naw, that couldn't be it.

Given the correlation between crime and demographics, it is not clear that the decline either in crime or in violent crime has anything to do with public policy - it could well be that crime (in either category) declined by less than we would expect based solely on an aging population. Unfortunately, it looks like we can't ask the "expert" whose salary comes from public coffers, because he is too busy making up shit to actually examine what is happening around him, or comment on it honestly to his fellow citizens. Sure, Dawg, let's all raise a beer to that one.

[Dr.Dawg]

And another one to dismissing math entirely and going by gut-feel:

Prime Minister Stephen Harper has dismissed empirical evidence that crime rates are actually falling, suggesting that emotion is a more telling barometer.

[Mike H]

"In any case, it seems a little late (almost 20 years too late in fact) for the Conservatives to start screaming about law and order."

Dawg, I can't decide whether revisionism or contradiction best applies with regard to this extremely perplexing statement of yours. We conservatives have been accused (quite rightly, I might add) of screaming about law and order for more than the past 20 years. Our accusers have been people of your ideological mindset, Dawg. You know, claims that we are vengeful sorts who want to lock 'em up and throw away the key, etc, etc. Is there some other meaning you intended to convey with this statement? I can only see the obvious, straightforward interpretation, and it makes absolutely no sense, given the historical position on crime that has both been adopted by and attributed to those of us with right of center views.

[Mike H]

"Or, put another way, we have returned to the lower levels of violent crime that we had nearly two decades ago--an over-all decline over a twenty-year period."

Or, put another way, the number of "excess" victims of serious violent crime (defined as those Canadians who would not have become victims of serious violent crime had violent crime rates remained relatively consistent instead of rising by nearly 600% from 1962 to 1992) is in the hundreds of thousands.

Hardly cause for celebration, that. Especially if you happen to be one among the unlucky hundreds of thousands.

[Peter]

"It's not likely that it's easy to draw cause,"

Actually, it's quite easy to draw cause. The vast majority of violent crime is committed by young males, especially unemployed young males. You bring the crime rate down by sublimating their aggression and changing their behaviour, which means controlling and sanctioning it legally or socially. The most promising palliatives are: A) employment and/or vocation; B) limited alcohol consumption; and C) marriage and family. Religion probably plays a role too. Tough jail sentences are actually quite far down the list. So the way to go is to tighten up EI and welfare for single males and maybe introduce civil conscription, restrict alcohol access and public drinking, repeal no-fault divorce laws and encourage church membership.

Any takers from the left? OK, how about you folks on the right? Nah, I didn't think so. Well then, I guess it's back to your stats and your discussion on whether we had a 1% increase or decrease last year.

[dcardno]

OK, how about you folks on the right? Nah, I didn't think so.

Gee, a program of encouraging employment instead of welfare, limiting access to alcohol [and, might I add, other drugs], encouraging family, and possibly promoting religion. I can't see why you would conclude that no one on "the right" would be interested Peter, although in Bizarro universe, you'd be right on the money.

[dcardno]

Dawg - you seem so upset by Prime Minister Stephen Harper has dismissed empirical evidence...

Why is it so much a non-event when the talking head leftist does exactly the same thing? In fact, the blabbering professor was worse; he didn't claim that perceptions were more important than statistics, he simply misrepresented the statistics he was talking about.

[Dr.Dawg]

Damn. Mike H. has a point. I have no idea why I made such a silly comment. He's right: conservatives have never stopped beating the law and order drum. What the heck was I thinking? Sorry, all, for that odd lapse.

But still: only conservatives would greet a decrease in crime since 1989 with gloom and excessive explanations.

[Dr.Dawg]

dcardno:

I was simply making the case for two beers. What's the matter with you?

[dcardno]

I was simply making the case for two beers....

I suppose that made sense when you typed it, but it made no sense when I read it.

...only conservatives would greet a decrease in crime since 1989 with gloom and excessive explanations.

Dawg, if the editorializing by the likes of Professor Boyd had recognized a small but encouraging improvement, I think we all would've seen that as a reasonable contribution to the discussion. Instead, every account I have read has trumpeted in one way or another a 'major reduction in crime' - and incidentally crowed about how that may undercut a plank in the Tory platform. Neither the spin nor the partisan cheerleading is particularly subtle.

[Bob Tarantino]

only conservatives would greet a decrease in crime since 1989 with gloom and excessive explanations

I guess, but then I suppose only progressives would look at the 450% increase in violent crime in the forty-odd years during which comprehensive statistics have been maintained, and then look at the 14% decrease in violent crime from its statistical peak and view that as, well, "progress". (I had to use the statistical peak in 1992 to obtain that 14% drop, by the way - violent crime has increased since 1989.) In any event, all you're doing is exactly what the Star is indulging in: arbitrarily picking the peak of the crime curve to make the facts fit a pre-determined narrative. If you look at violent crime rates using any year after 1989, then, yes, violent crime rates have declined - but why pick that frame of reference? Use a 10-year frame and violent crime has declined; use a 15-year frame and violent crime has declined; use a 20-year frame and it has increased; use anything more than a 20-year fram and it has increased; use anything more than a 25-year frame and it has increased by (hyperbolically) orders of magnitude.

[Bob Tarantino]

Just as an addendum to my 2:49pm comment, if anyone is interested in seeing the full set of Canadian crime rates, see page 10 of this StatsCan report [pdf document].

[Mike H]

Dawg:

Any comments in relation to my second post, concerning "excess" violent crime victims, and the actual human face of this discussion?

[Dr.Dawg]

dcardno, old buddy, you're being thick.

Sure, Dawg, let's all raise a beer to that one.

Posted by:dcardno | July 18, 2008 at 07:20 PM

And another one to dismissing math entirely and going by gut-feel:

Prime Minister Stephen Harper has dismissed empirical evidence that crime rates are actually falling, suggesting that emotion is a more telling barometer.

Posted by: Dr.Dawg

Frankly, I'm in the middle--no manipulated stats, but please, no Harper gut-feel as a substitute.

Bob:

A reduction of 14% is progress.

[Mike H]

"Any takers from the left? OK, how about you folks on the right? Nah, I didn't think so. Well then, I guess it's back to your stats and your discussion on whether we had a 1% increase or decrease last year."

Peter, I'd be interested in your response to the same question I put to Dawg just now.

[Dr.Dawg]

Mike:

Crime, particularly violent crime, obviously exacts a human toll. What do you want me to say? I'm just pleased it appears to be dropping.

Why has nobody commented about the confounding variable of the reporting rate? Or did I miss something?

[dcardno]

dcardno, old buddy, you're being thick.
Yes - more than two beer last night, and fewer than two coffees this morn.

I'm in the middle--no manipulated stats, but please, no Harper gut-feel as a substitute.
The two issues are orthogonal, Dawg: a politician perforce responds to public concerns - that's his job. That response can either be to try and convince the public that their concerns aren't valid, or to introduce policies that respond to the concern. An unacceptable response is to simply lie about the state of nature - which is what raised the whole discussion. In the quoted material Harper didn't misrepresent the Stats Can results, he asserted that they were not going to replace public concern as the basis (or at least, the only basis) for government policy.

Why has nobody commented about the confounding variable of the reporting rate? Or did I miss something?
Okay, I'll comment: variations in reporting rate are much more likely in non-violent crimes than violent crimes - which suggests that the noted reduction in non-violent crime overstates the improvement. I don't see that lending much support to the assertion that overall crime has declined, and I don't see that it bears very much on violent crime at all. So what?

[Mike H]

"Crime, particularly violent crime, obviously exacts a human toll. What do you want me to say? I'm just pleased it appears to be dropping."

I wanted you to address the concept of "excess violent crime." We know violent crime exacts a human toll, and we also know it is unrealistic to expect that we can eliminate it. However, we have a baseline rate of violent crime in the 60s and 70s that is dramatically lower then what we have experienced for the past 30 years, and that period of greatly increased violence has resulted in hundreds of thousands of victims of violent crime who can be categorized as "excess" victims of violent crime; that is, people who would not have been murdered, robbed, seriously assaulted, or sexually assaulted, had crime rates remained near the 60s and 70s baseline rate. I, along with like-minded people, are objecting to the creation of a new baseline standard, that being the current more "stable" but much more violent year to year rate of violent crime incidence.

I'll give you a parallel example, to illustrate how you would be taking the exact opposite position (in other words, the same position as myself, Bob and D Cardno), if the context is changed, but the statistical argument remains the same.

As you know, Johns Hopkins University sponsored two highly controversial mortality studies in Iraq, commonly known as the Lancet surveys. I'm not going to risk sidetracking this discussion by making any comment about the accuracy of the results, so I'm going to omit their country wide death extrapolations. The survey methodology involved population surveys to determine the mortality rate prior to regime change. The researchers established a baseline mortality rate for the 14 month period immediately preceding invasion, then compared it to the mortality rate for fixed periods following invasion (in the case of the 2004 study, for the first 17 months following the date of invasion. In the case of the 2006 study, for the first 40 months following the date of invasion). Both studies extrapolated a pre-war baseline extrapolated violent death toll of roughly 3,000 people for the entire nation (essentially, victims of violent, non-war related crime). Obviously, their extrapolations of the death toll after regime commenced is much higher, and the number of deaths over and above 3,000 are described as "excess deaths (in other words, the number of violent deaths that would not have occurred had there been no invasion).

Now, let's say, for the sake of argument, that approximately 20,000 Iraqis died violently in Iraq last year, 2007. And let's also assume that 20 years down the road, the U.S. is still in Iraq, and still fighting the combination of a stubborn insurgency and serious sectarian violence, and violent death has hovered in and around that 16,000 - 17,000 mark year in and year out. Sometimes a couple thousand higher, sometimes a couple thousand lower, but on average, around 16,000 - 17,000. In 2027, the yearly death toll is one of those "average" years, around 16,000 - 17,000. The U.S. government points to its critics, and says "yes, we're still seeing violence in Iraq, but it's down about 14% this year from what it was in 2007."

Would you buy their argument, Dawg, or would you point to the obvious huge yearly increase above the 2002 pre-war baseline mortality rate and remind them that some 13,000 - 14,000 Iraqis have, on average, died violently every year over and above what would have been expected, had their been no invasion?

I'm making the exact same point with regard to Canadian violent crime stats.

[Mike H]

"Why has nobody commented about the confounding variable of the reporting rate? Or did I miss something?"

Sorry Dawg, I forgot to speak to this. Can you be more specific about the variables? I haven't had the time to read over the entire report.

[Dr.Dawg]

dcardno:

"he asserted that they were not going to replace public concern as the basis (or at least, the only basis) for government policy."

Even if public concern is ill-founded? (Reference Dan Gardner's recent book on fear.) Wouldn't a responsible politician confront misconceptions, not opportunistically and cynically make use of them?

Mike:

You raise, of course, an interesting and pertinent question. Being familiar with the Iraq studies--including the child death controversy--I under stand the benchmark issue. The problem is, how do we establish that benchmark here?

For example, does that comfortably low rate 40 years ago define that benchmark, or were the fifties lower? The forties? The thirties? The stats aren't available for comparison purposes, but it's interesting to speculate why the wild sixties apparently had such a low crime rate, even with "trace busts" and that sort of thing.

I didn't pluck the notion of reporting rate from the report, Mike. I just recollected reading a long time time ago that sex offences (for example) were grossly underreported until recently. That would obviously affect the graph. Forensic science has also improved markedly. I just mention these (without quantification" to raise the possibility that there are numerous factors involved in these rates.

[Mike H]

"For example, does that comfortably low rate 40 years ago define that benchmark, or were the fifties lower? The forties? The thirties? The stats aren't available for comparison purposes, but it's interesting to speculate why the wild sixties apparently had such a low crime rate, even with "trace busts" and that sort of thing."

As you say, we don't have data to establish whether crime rates were even lower prior to the 60s. However, I see that as a red herring. Post 1960, setting aside which time frame's statistics are more skewed, and in which direction, we still know the 60s experienced a much lower incidence of crime than subsequent decades. "Trace busts" relate to non-violent crime, and therefore are not a factor in attempting to explain away why, moving forward from the 60s, violent crime rates skyrocketed in tandem with the overall crime rate.

"I didn't pluck the notion of reporting rate from the report, Mike. I just recollected reading a long time time ago that sex offences (for example) were grossly underreported until recently. That would obviously affect the graph. Forensic science has also improved markedly. I just mention these (without quantification" to raise the possibility that there are numerous factors involved in these rates. "

From the latest Stats Can crime report linked to by Bob, all sex offences combined constituted only 8% of all violent criminal offences in 2007. Accordingly, even allowing for a doubling or tripling of sex offences due to increased reporting will have a negligible effect if one is attempting to account for the 450% increase in violent crime from 1962 to today. I'm not sure I follow your reasoning with regard to the effect improvements in forensic science techniques have on the incidence of crime. They certainly have an effect on clearance rates.

I couldn't agree more with your observation that there are "numerous factors" exerting influence on crime rates. These factors interact, pushing and pulling on one another, offsetting and aggravating each other. Societal factors like demographics and poverty levels play a role, as does the increased availability of drugs like cocaine and crystal meth. Separately, we have numerous factors working within the justice system itself. The introduction of The Charter of Rights and Freedoms, a dramatic erosion in sentencing severity, increased leniency with regard to parole, stricter rules governing crown disclosure, the additional work load for police forces on a variety of fronts, all these, and many more I've failed to touch on, have contributed to the huge increase in crime rates.

It's my view (and I think I speak for Bob and many other of us law and order types) that the single most meaningful measure we can implement to bring down violent crime rates significantly and quickly would be the introduction of severe mandatory minimum sentences for the most serious violent offence categories, accompanied by a radical reduction in parole eligibility for these offences. Arguments that such measures won't work are completely counterintuitive.

[Dr.Dawg]

I'm not sure I follow your reasoning with regard to the effect improvements in forensic science techniques have on the incidence of crime. They certainly have an effect on clearance rates.

I had thought, foolishly, that the crime rates we were discussing were based upon convictions, not filed reports of crimes being committed. I was, for the second time in this thread, wrong. I need to pay more attention. This discussion is not going well. : )

Now, I'm indeed worried about a narrow law-and-order approach, because it treats symptoms rather than causes. But nonetheless, when you state:

that the single most meaningful measure we can implement to bring down violent crime rates significantly and quickly would be the introduction of severe mandatory minimum sentences for the most serious violent offence categories, accompanied by a radical reduction in parole eligibility for these offences.

...it might surprise you (although Bob knows this) that I agree with you.

[Mike H]

"...it might surprise you (although Bob knows this) that I agree with you."

Dawg, "surprised" doesn't begin to cover it. Just when I think I have you figured out...

[Dr.Dawg]

I normally don't like mandatory minimums, for a host of reasons, but you did say "the most serious and violent offence categories." I've never been soft on violent crime, I'm highly sceptical of "rehabilitation" for the perps, and I've posted more than once about shocking parole decisions. You should visit my place more often. : )

[dcardno]

Sorry to let this sit - I didn't notice the little "next page" marker...
Even if public concern is ill-founded? (Reference Dan Gardner's recent book on fear.) Wouldn't a responsible politician confront misconceptions, not opportunistically and cynically make use of them?

Are (violent, since that's what we are talking about) crimes down so far that 'correcting misperceptions' is the appropriate response, Dawg? My guess is that most voters look at today's rate of violent crime as too high, based on their own experience from 20-30 years ago of a time when there was less violent crime. Professor Boyd's attempt to soft-peddle the problem is unconvincing: Crime is more-or-less stable over the last decade+, after a dramatic increase over the four previous decades.

I agree that Harper is (was) pandering and playing to his base; but so was Neil Boyd, and so was the reporter who was credulous enough to take his comments without fact-checking them: they were just pandering to a different audience, and playing to a different base. The difference is that the media was universally skeptical of Harper's comments, and universally accepting of the politically-acceptable narrative that violent crime is coming down, and in particular that it has been doing so for a long time.

I just recollected reading a long time time ago that sex offences (for example) were grossly underreported until recently...
Yes - and some activities (ie "date rape") that are now considered criminal (rightly so, in many cases) used to be "just one of those things" that women had to live with. I agree that both of those factors will influence reported crime rates - but they have no effect on ordinary (non-sexual) assault, homicide, kidnapping, and so on. Again, if Dr Boyd had wanted to make those sorts of distinctions (although they wouldn've have fit in his alloted 25 words) then it would have been a valuable contribution - but he couldn't be bothered to derail the narrative.

[Mike H]

"I normally don't like mandatory minimums, for a host of reasons, but you did say "the most serious and violent offence categories."

Yes, I did, and I was being deliberately precise. There's no shortage of sentencing travesties in cases involving less serious crimes, but I believe these should be corrected without resorting to mandatory minimums.

"You should visit my place more often. : )"

I "visit" quite a bit, Dawg. I just don't comment much. Now, if you were to ban Ti-Guy....

[James Goneaux]

I'll have to check, but does the murder rate, for instance, count deaths that have somebody actually charged with murder, or something like this:

http://preview.tinyurl.com/65sopz

wherein a man is involved in 43 murders, but isn't actually convicted of a murder? Are they still counted as murders?

I would imagine that if one were to look coldly at the numbers, something isn't going to add up. Particularly when you look at cases that get pleaded down to manslaughter.

As well, StatsCan has some bizarre tables:

http://www.statcan.ca/english/freepub/84F0211XIE/2005000/t012_en.htm

is one. Just check out BC's data here for instance...

And as someone who is opposed to the death penalty, I still like to point out that when executions began to be commuted to "life" in "prison" (somehow translated into 25 years, tops), the life expectancy of a Canadian was a bit lower than today. Any thought of actually RAISING "life" to mean, say, 30? Would be logical, for all those who believe this is solely an emotional debate.

[stephen.reeves]

We only had six shootings in Toronto last week, so it must be going down, I suppose we can thank Miller and his buddies.