Statistics Canada is in talks with the country's police representatives about collecting and publishing data on "unfounded" cases, after revelations that a disproportionate number of sex-assault complaints are deemed baseless – a phenomenon that erodes the quality of Canadian crime statistics.
The agency stopped releasing unfounded numbers 14 years ago despite warnings from criminologists and policy analysts that it would prevent researchers from scrutinizing statistics that many say reveal how often police believe complainants.
As part of a 20-month investigation, The Globe and Mail obtained unfounded rates by filing individual freedom-of-information requests with each Canadian police service.
"We really want to look into this and very seriously," said Yvan Clermont, the director of the Canadian Centre for Justice Statistics, a division of Statistics Canada.
"We're working in collaboration at this time with the Canadian Association of Chiefs of Police's subcommittee … to see what are the options that we can follow and establish recommendations with respect to that type of data."
Mr. Clermont said his agency will be addressing the issue some time in early spring at its biannual meeting with the association's Police Information and Statistics committee (POLIS committee).
Last week, association president Mario Harel cleared the way for this review by announcing that he had requested the committee to "examine how statistics are recorded and reported to Statistics Canada and make recommendations on how reliable and consistent statistical information may best be collected." The statement was particularly significant given that it was the POLIS committee that originally recommended Statscan stop collecting "unfounded" information over concerns about the quality of data.
On Wednesday, Navdeep Bains, the Minister of Innovation, Science and Economic Development, which is responsible for Statscan – although the agency itself is arm's length from government – told The Globe in a statement that the federal government understands the need for more and better data when it comes to sexual assault.
"Policy makers and police services must have accurate data that allow them to make informed decisions. I'm pleased to see that Statistics Canada is continuing to work with its justice and policing partners to identify ways to improve the accuracy of the data collected on incidences of sexual assault," he said.
Two weeks ago, after The Globe revealed that police forces are dropping 19.39 per cent of sex-assault cases as unfounded, which is nearly double the rate for common physical assault (10.84 per cent). Once an accusation is closed using the unfounded code, it is no longer considered valid. It is not reflected in local or national statistics and it is not reported to Statistics Canada. This means that on average, roughly 5,500 people who report sexual assault to Canadian law enforcement each year are not being counted.
To date, about 40 Canadian police services – including the two largest, the RCMP and Ontario Provincial Police – have announced plans to review unfounded sexual assault cases as a result of The Globe's series. Mr. Clermont said the statistics agency will be releasing a communiqué to police services shortly that will urge them to take case classification into consideration as part of those reviews. Every police force in the country is required to use classification codes outlined in the Uniform Crime Reporting protocols, a system that has been in place since 1962 to ensure crime statistics across the country are comparable.
The unfounded code only applies in the rare event when an investigation has determined that no crime occurred or was attempted. It is not to be used to clear cases where there is not enough evidence to lay a charge or if a complainant does not wish to proceed. But going back to the 1990s, Statistics Canada officials have expressed concern that police services misuse the unfounded category to clear cases.
In 2003, the decision was made to stop releasing the data. (The last year available is 2002, when the national unfounded rate for sexual offences was 16 per cent.) Three years later, the POLIS committee recommended that police services stop sending the data to Statscan altogether.
"Because we were seeing a very low quality of data regarding unfounded incidents … and also a lot of police forces were not reporting or only partially reporting those incidents, [it was determined] that we would no longer require police forces to submit their information to Statistics Canada," Mr. Clermont said. "Now based on … the statement made by the president of CACP, we think it's time to review this and see what can be done."
A CACP spokesperson said that beyond the statement made last week, "we are not doing any further interviews until we have had an opportunity to undertake our recommendations."
University of Ottawa law professor Blair Crew, one of the few researchers to look at unfounded rates in the last decade, said that a thorough review of the issue now is a good thing, because it will limit the ability of police services to blame high unfounded rates on coding confusion.
"It would be very helpful if they made it plain that unfounded is only to be used when no sexual assault occurred. The more transparent they can make police practices, the better it is for everyone," he said, adding that the committee may decide it needs to provide more closure-code options for police.
"Our understanding and even our definition of sexual assault is so different than in 1962, if what we need is more nuanced categories in the [Uniform Crime Reporting protocols], I do think it is important. If it's clear, then it cuts out the argument that high unfounded rates are only statistical errors."