Posted in:
StatsCan’s Crime Severity Index
Statistics Canada has introduced a new measure of police-reported crime, the crime severity index, in which more serious crimes are weighted more heavily, by comparison with the usual crime rate in which all crimes affect the outcome equally. The Daily has a summary of the current index data and the chart reproduced below. For a detailed description of how the index is calculated see “Measuring Crime in Canada: Introducing the Crime Severity Index and Improvements to the Uniform Crime Reporting Survey.”
While I haven’t yet performed a full-scale in-depth review of the dizzying array of statistics that can be mined in this fascinating report, one key fact becomes clear very quickly — crime is going down.
This of course runs contrary to every telecast, newspaper article, and press release that we are bombarded with in our daily lives but the fact remains that whatever way you spin the numbers, we are living in a less crime-ridden society today than we were 10 years ago.
The new weighted format of the index even accounts for fluctuations in minor versus major crimes putting the lie to the claim that, while “crime as a whole” is down, “serious crime” is up. In fact, the survey demonstrates that serious violent crime is modestly down from last decade (though the decline is not as pronounced as the more general crime index).
The survey also busts the generally held Canadian myth that Toronto is a frontier badland where gun-toting criminals walk the malls with impunity. In fact, amongst major Canadian metropolitan centres, Toronto comes in behind Montreal and “champion” Vancouver when it comes to serious crime, and Toronto’s serious crime rate falls markedly below the national average.
All food for thought the next time police demand new sweeping investigative powers and new recruits or politicians fulminate about the need to “get tough on crime”.
I believe the Conservative crackdown on crime is misguided, with more mandatory minimums and longer sentences, but you need to be cautious about what you infer from the crime severity index. It suffers the same weakness as traditional crime stats. It is based only on police-reported crimes, and we know that most crimes are never reported to police (9 out of 10 sex assaults are never reported). Does it really tell us that crime is down, or does it simply tell us that in some communities, reporting of serious crime is low, and in some, much lower than others?
\The new weighted format of the index even accounts for fluctuations in minor versus major crimes putting the lie to the claim that, while “crime as a whole” is down, “serious crime” is up. In fact, the survey demonstrates that serious violent crime is modestly down from last decade…\
The survey shows no such thing. As Table 1.2 of the report indicates (see here), the violent crime rate is down 0.2% and the violent crime severity index is down 1.5% – but that’s just a statistical blip, since if the report had been published a year earlier, it would have shown that both the violent crime rate and the violent crime severity index were higher than at the beginning of the decade. The violent crime severity index is largely unchanged over the decade (see here). But even to point out such obvious facts is to miss the larger point: the violent crime rate in Canada is more than quadruple what it was when StatsCan began keeping comprehensive figures in the early 1960s (see here). Are we \living in a less crime-ridden society today than we were 10 years ago\? Possibly, depending on how you choose to spin the numbers and what definitions of \crime\ you want to use. But we are also, regardless of how you choose to manipulate the numbers, and regardless of what definitions you want to use, living in a society which is vastly more ridden with violent crime than it was 20, 30, 40 or 50 years ago.
Bob is right. There’s a decent graph in a 2005 Juristat release [pdf] that shows the trends:
The Juristat article is good on the data and on the problematic aspects of crime stats.
Thanks, Simon, for providing that link. However, due to the scale of the Juristat graph, it doesn’t properly convey the magnitude of the increase in violent crime rates. I think the second StatsCan graph, headed \Violent crime rate, 1962 to 2007\, at this source is of assistance (source here, for some reason Slaw keeps translating quote marks in my comments as backslashes, which corrupts the hyperlink – anyone interested in viewing the graph should click on the link and then delete the final backslash from their address bar – or just go to the StatsCan Daily for July 17, 2008).