Wayne Smith, the chief statistician of Canada, is going to have to do a lot better if he wants to convince the public that Statistics Canada's new surveys are a suitable replacement for the mandatory long-form census that was cancelled in 2010. Mr. Smith's statement – "True, it's not a census, true, there's been some loss of small-area data and true there's more volatility in the estimates for small populations and small areas, particularly small populations in small areas. But the data turns out to be remarkably strong" – is not the sort of categorical endorsement that will win many people over.
It was precisely the loss of small-area data involving small populations that raised serious doubts about the first wave of results from the voluntary 2011 National Household Survey, as it is now called. That first wave, in May, related to aboriginal peoples and immigrants, and its lack of depth and breadth left serious gaps in a critical area of public policy. "The country needs more rather than less significant and reliable information about [aboriginal peoples'] lives and circumstances," we wrote at the time.
Mr. Smith bemoaned that sort of criticism this week in comments that come just before the release on Wednesday of the next batch of results from the NHS. "It's irresponsible to try and dissuade Canadians from using what is an extraordinary rich and powerful database," he told The Globe and Mail. "To make them nervous about that is I think irresponsible."
It is not irresponsible to criticize or to promote the NHS. But there continue to be serious doubts about the value of the NHS when compared to the mandatory long-form census – doubts that Canadians should be aware of. There is a general consensus that the NHS results are of some use at the national and provincial levels, but of no use at the local level. The harshest critics of the new voluntary system call its data "worthless," because it does not produce a random, non-biased result, and because it is no longer mandatory – an obligation that had successfully gotten more and more aboriginal communities to take part.
The data from the 2011 NHS is certainly not worthless, but even our chief statistician admits it has not produced as valuable results as the mandatory long-form census would have. He also admits that it cost $22-million more than expected because more copies had to be printed and field workers had to chase down non-respondents. But Mr. Smith also says, rather dismissively, that those who continue to compare the new survey's results to those of the defunct mandatory census should "get over" it, because "that ship's sailed."
Let's hope not. Canada needs a proper, randomized, long-form census and the data it can produce. There is nothing irresponsible about saying so.