When the Conservative Party made the formerly mandatory, once-every-five-year census, optional in 2011, there was an outcry. It sounded bad to me at the time, but I didn’t understand the implications. Four years later, the critics have unfortunately been proven correct. Macleans’ Anne Kingston reports:
When told that his small Prairie town had, in profound ways, fallen off the statistical map of Canada, Walter Streelasky, mayor of Melville, Sask., is incredulous. Streelasky had no idea Melville had been rendered a “statistical ghost town” after the mandatory long-form census was cut in 2010, and fewer than 50 per cent of Melville’s 4,500 residents completed the voluntary National Household Survey that replaced it in the 2011 census. Melville still exists—but as a shadow. We know how many people live there, but nothing about them—where they work, their education levels, whether they’re married, single or divorced, how many are immigrants, how many are unemployed, how many live in poverty. Melville’s numbers, then, aren’t factored into Canadian employment numbers or divorce rates or poverty rates. According to Sask Trends Monitor, the high non-response rate in the province resulted in “no socioeconomic statistics about the populations in about one-half of Saskatchewan communities.” Nationally, we’re missing similar data on 20 per cent of StatsCan’s 4,556 “census subdivisions,” making a fifth of Canada’s recognized communities statistical dead zones.
The justification was of course cost-savings, as usual, but the optional survey both cost more and gave us useless data:
[T]he voluntary National Household Survey added $22 million to the cost of the 2011 census; the response rate dropped from 94 per cent in 2006 to 69 per cent, which makes the data totally unreliable. “A response rate of 75 per cent is the minimum required for sample accuracy,” says StatsCan’s former chief statistician, Munir Sheikh, who famously resigned in 2010 after Tony Clement, then industry minister, stated publicly that the decision to cut the long-form census came from within StatsCan. “The federal government misrepresented my advice,” Sheikh told Maclean’s, adding that ongoing cuts to the agency have undermined its credibility.
StatsCan has become a cautionary tale in the rest of the world:
When it comes to data, Canada has become “a cautionary tale,” says Phil Sparks, a former associate director of the U.S. Census Bureau who is co-director of the Washington-based Census Project, a group fighting to maintain the U.S. equivalent of a mandatory long-term census in the face of Republican calls for its elimination. He calls the Canadian experience “an unmitigated disaster.” The census’s elimination damaged Canada’s international reputation, says John Henstridge, president of the Statistical Society of Australia. “Prior to that, Statistics Canada was regarded as possibly the best government statistical body in the world.”
But it’s the poor and the marginalized who are really hurt by this policy. Making the data useless matches the Conservatives’ (usually) unstated policy of “the poor should just get a damn job” perfectly. You can’t prove that their policies have increased poverty; you can’t say that we have a widening income gap under their rule:
Voluntary surveys also create biased data, says Sheikh: Response rates from the the very poor, rural areas, immigrants and Aboriginal communities tend to be far lower—so these groups are not well-represented. “People who do not respond well to a voluntary survey are the very people social policy tries to help,” he says. “So if you were to base policy on data received, you’d say, ‘Gee, we don’t have a poverty problem in this country.’ ”
And there’s so much more than just the census. Stephen Harper’s Convervatives have truly been closing the Canadian mind.