Tag Archives: long-form census

All the data that’s not fit to print

DSC_0153

Can the irony be any more succulent?

Mere days before Statistics Canada took the extraordinary decision to sit on the final batch of data stemming from its 2011 voluntary survey because, it said, the numbers don’t accurately reflect current conditions, the 1921 Census of Canada went public, showing what life was like, in authentic detail, for citizens nearly a century ago.

That’s great for fans of family trees and downright awful for anyone else who seeks to obtain a faithful picture of our present times – though some might yet extrapolate from the form and fit of great-grandma’s bloomers the spending habits of the modern, smartphone-addicted tween demographic.

Still, StatsCan insists the problem with the 2011 data had little, or nothing, to do with the controversial shift to an optional household survey, from a mandatory nose count of the population, three years ago.

Indeed, Marc Hamel, a census manager, told The Globe and Mail  this week, “We were in the final stages and some of the results seemed odd, a bit. When we went back to the data-processing steps, we discovered that one of the steps was not applied correctly. . .It is unfortunate that it was in the late stages. But it’s lucky we found it before it was released.”

The actual statement on the numbers-crunching agency’s website is a marvel of circumspection: “The release of the third and final set of data from the 2011 National Household Survey is postponed to September 11, 2013. The release focuses on income, earnings, housing and shelter costs. Statistics Canada found issues in data processing that need to be addressed prior to release. All the data previously released from the National Household Survey are not affected.”

I guess we’ll just have to take its word on that. It’s not as if the agency has any real context for assessing the verisimilitude of the results from the voluntary questionnaire. Apart from the fact that the household survey boasts a much lower response rate than the census (according to the Globe piece, it’s 68.6 per cent versus 93.5 per cent), the new system is still in its infancy.

But why is any of this necessary?

For decades, Canada led the developed world in the quality, comprehensiveness and accuracy of its census data. The numbers served a useful, and often crucial, purpose when legislators sought to craft and implement social and economic policies. The findings materially contributed to health, education and infrastructure programming.

In a 2010 letter to Tony Clement, who was the Minister responsible for Statistics Canada, the Canadian Sociological Association (CSA) argued that “Long‐form data are used by businesses, provinces and municipalities, economists, urban and community researchers, policy analysts, sociologists, and other scholars in the humanities and social sciences (including geographers and historians).

“Religious and ethnic groups are also users. They all rely on the mandatory long form census for solidly representative and accurate data – especially when data are disaggregated to community or minority‐group levels. Whatever the unit of analysis, an accurate statistical portrait of the population – one that allows for cross‐tabulation – is required.This cannot be provided by the voluntary NHS because bias – due to the under‐representation of specific groups – is likely. Aboriginal people, recent immigrants, low‐income families, and perhaps even busy professionals may fail to respond.”

A subsequent CSA blog post rather archly observed, “If the minister responsible for Statistics Canada is to be believed, the long-form census was eliminated so that upright citizens would no longer be threatened with jail time for failure to complete and return a census form that asked intrusive personal questions. A more convincing reason is that we have a government that not only says ‘Don’t bother me with the facts!’ but also wants to ensure that no one else has access to the facts.”

Without facts, of course, we are left with assumptions, suppositions and, in the words of American commentator George F. Will, “factoids” plucked “from the ether.” As he wrote in a piece that appeared recently in The National Post, “implausible and utterly unsubstantiated claims flourish when there is indifference to information.”

How odd that such sentiments should belong to one of the continent’s more notable conservatives.

You might even say, it’s ironic.

Tagged ,

Survey is no replacement for a proper census

DSC_0133

Greater Moncton’s labour force is alive and well and kicking, which may be one reason why fully a-fifth of New Brunswick’s workers will never achieve “Freedom 55”.

Here are a few other “facts” about Canada, now celebrating its 146 birthday, we may not have appreciated, courtesy of Statistics Canada’s National Household Survey 2011, as faithfully reported in The Vancouver Sun:

“We love driving our cars to work. In 2011, almost 93 per cent of people in the workforce drove to work and most drove by themselves. . .English dominates Canadian workplaces, with 85 per cent of the population using the language at work, compared to just 25 per cent who say they use French. . .Chinese languages are the most-commonly used languages in the workplace after English and French. . .Possessing a post-secondary education increases Canadians’ chances of being employed. But there’s no obvious employment benefit to a graduate degree. . .Canadians ages 25 to 34 are far more likely to have trained to be a cook than an auto mechanic or construction worker, when compared to workers closer to retirement.”

Now we know all we need ever comprehend about our families, friends, neighbours; about ourselves in this great and peaceable land. Or do we?

In 2010, back when the federal government announced it was scrapping the mandatory long-form census in favour of a “voluntary” household survey, editorials in just about every major newspaper in Canada screamed their disapproval. The nation’s two top numbers-crunchers Munir Sheikh and Phil Cross actually resigned their posts at Stats Can in evident, if dignified, protest.

In a news advisory at the time, Mr. Sheikh wrote that while he could not “reveal and comment on (the) advice” he gave the government “because this information is protected under the law,” he wanted to “take this opportunity to comment on a technical statistical issue which has become the subject of media discussion. This relates to the question of whether a voluntary survey can become a substitute for a mandatory census. . .It can not.”

Only a month ago, Robert Gerst, a partner in charge of operational excellence and research and statistical methods at Calgary-based Converge Consulting Group Inc., declared in an opinion piece for the Waterloo Region Record, “Take the first data releases from the national household survey of Statistics Canada. . . .The quality of the results has come under criticism because the voluntary survey replaced the compulsory long-form census questionnaire. In effect, this replaced a random sample with a non-random sample. Non-random samples have their place, but making conclusions about the population isn’t one of them.

Naturally, then, “no conclusions about the Canadian population can be drawn from the national household survey. Since making these types of conclusions is the whole point of a census, the survey data is worthless. (This is also true for any survey where participation is voluntary, including citizen, customer and employee satisfaction surveys).”

What, then, justifies the heavy media coverage of the survey? (The Globe and Mail devoted its entire Folio section to the data). Aren’t we, in the Fourth Estate, supposed to suspect this sort of information. In fact, isn’t that our job description? On the other hand, when faced with a bunch of lemons, there’s really only one thing to do.

Intuitively, the survey results seem valid. They track neatly with the more rigorous findings of the 2006 census: Canada’s population was getting older; most commuters in major cities did tend to drive themselves to work; possessing an undergraduate degree did improve an individual’s likelihood of landing a good job. That much appears not to have changed.

But the farther out we go from a proper census (broad, random sampling), the less verisimilitude the findings display. A decade from now, we may no longer be able to trust that the survey results are grounded in enough truth to say anything accurate or useful about our families, friends, neighbours; about ourselves.

In an age when information, increasingly, unlocks the door to economic growth, this seems oddly and regrettably retrograde.

Tagged ,
Follow

Get every new post delivered to your Inbox.

Join 580 other followers

%d bloggers like this: