What crime statistics don't tell you

A lot of crime takes place in the dark for obvious reasons. That's no reason to conduct the public debate about it under similar conditions. Yet while informed discussion is the cornerstone of self-government, on this 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.

The problem is not just that we don't have some numbers we ought to have. It's that we have a high-profile, apparently excellent source of data on crime that is unsuitable in important ways. Once a year Statistics Canada releases a comprehensive review of police-reported crime statistics (the "Juristat report"), generally suggesting that crime rates are low, and falling, and generally leading commentators to suggest that anyone who thinks crime is a serious problem in Canada is an ignorant fear monger and probably a hayseed to boot. But in a new study from the Macdonald-Laurier Institute, where I am managing editor, former Crown prosecutor Scott Newark makes clear in detail that there are a number of serious flaws in the way the Juristat numbers are collected, presented, and interpreted. The result is to deprive Canadian policy-makers, opinion-makers, and citizens information on which to make difficult decisions about the complex social phenomenon known as crime.

For instance, the annual cycle of news stories on the Juristat report should give far more attention than it does to the fact that the report covers crimes reported to police rather than overall crime. This distinction matters, especially in public debate, because as StatsCan itself regularly reports in another publication (its General Society Survey of crime victimization undertaken every five years), a growing volume of relatively minor and sometimes not so minor crime is not reported to the police. When citizens increasingly do not bother to report crimes or, worse, are afraid to do so, it gives a falsely reassuring picture of the state of law and order in Canada. It is also troubling evidence of an erosion of citizens' faith in the competence and compassion of their government when it comes to the criminal justice system. When misinterpreted as proof that citizens are loud ignoramuses, this statistical distortion further contributes to that erosion.

By itself this shortcoming might only call for commentators to be a great deal more alert when using statistics. Unfortunately, the Juristat report suffers from a significant number of other genuine flaws in the way it collects and reports data and in how it explains its work. Some of those flaws have been getting worse rather than better over time in ways that are as puzzling as they are regrettable. The Juristat report thus further contributes to the misleading impression that crime is not a major problem in Canada, and conspicuously fails to meet StatsCan's usual high standards.

Thus in addition to glossing over the problem of unreported crime, the Juristat report only counts the single most serious offence in a criminal incident, so that if someone attempts to commit murder driving a stolen car drunk while on bail, it only shows up as one crime, namely attempted murder. This approach doesn't just undercount crime, it undermines accountability. If things like bail violations, parole violations, or breaches of deportation orders are widespread, it tells us the justice system is doing a poor job of dealing effectively with offenders it has already identified and convicted. Yet they are almost never the most egregious aspect of a serious crime and the Juristat report does not collect and address them separately.

Once again Canadians are deprived of essential information about how they are governed on a highly controversial topic. The report should do a better job of collecting, presenting and explaining data, instead of kicking off with a series of tendentious highlights that get largely uncritical attention despite the shaky assumptions and difficulties with data that underpin them. In short, the Juristat report unhelpfully and unnecessarily creates obscurity on this important topic. Canadians deserve better light and Statistics Canada can certainly provide it.

John Robson is managing editor of the Macdonald-Laurier Institute in Ottawa, which has just published Why Canadian Crime Statistics Don't Add Up: Not the Whole Truth ( macdonaldlaurier.ca).

[First published in the Ottawa Citizen]