Flawed Census Isn't Much Of A Census At All

Opinion

Flawed Census Isn't Much Of A Census At All

Politically-motivated tinkering has blocked the collection of reliable, valid data that is so important in properly delivering social programs in Canada.

By Michael Borrelli
Published May 06, 2011

This past week, households across the country began to receive their 2011 Census packages. Enclosed in this mass of paper was an upbeat reminder from the Statistics Canada that, "Communities depend on census information when planning for new schools, roads, waterworks, public transit, and police and fire services. Town planners use census information on households and families to plan current and future housing needs, health care, and day-care centres."

With such pro-social intentions, how could anyone argue against completing the Census, as some Canadians are now suggesting? It would seem uncivil to deny the government this valuable information, and it could also get you thrown in jail.

Since the objective of a census is to systematically enumerate and collect information on a whole population, not a sub-sample of it, participation is mandatory and methodologically necessary. This has been the primary method of conducting censuses for thousands of years: Jesus was born in Bethlehem while his parents were in town to be counted by the Romans; and there are records of censuses going back three or four thousand years before Christ in Egypt and China.

But starting this year, only a small part of the Canadian census is mandatory. Form 2A, the eight questions commonly referred to as the 'short-form', is sent to every household in Canada. In previous years the 'long-form', which includes 53 additional questions about ethnicity, education, and other important demographic indicators, was sent to 20% of households.

Like Form 2A, completion of the long-form census was mandatory, but last year the Conservative government informed Statistics Canada that it would be replaced by a voluntary National Household Survey sent to 30% of households instead. The government insisted that the decision was made in response to "public opposition," though this was greatly overstated.

In the face of critics who argued a voluntary form would result in unreliable data, and despite the fact that the US Census Bureau concluded in 2003 that a voluntary census was not a workable option there, Industry Minister Tony Clement twisted himself into knots to justify the decision. The head of Statistics Canada at the time, Munir Sheikh, even resigned in opposition to the plan.

Why Bother Completing The Census At All?

So what is the point of filling out either Form 2A or the NHS now that the integrity of the data is compromised?

The 2011 iteration of the Canadian Census addresses neither the privacy issues cited by the government, or the broader data validity questions raised by a wide spectrum of economic and social interests.

On privacy, the accurate enumeration of individuals, their ages and their primary language spoken is important for planning where to build new schools, but there's precious little reason for Statistics Canada to know the nature of my relationship with my domestic partner or others in the household.

Confidentiality and security assurances from Statistics Canada aside, what's even scarier from a privacy standpoint is that data processing will be carried out by Lockheed Martin. This US defense contractor works intimately with the CIA, NSA and the Pentagon, and their motto is: "We never forget who we're working for."

But of greater concern than privacy is the loss of valuable socio-cultural information provided by a mandatory long-form. Questions on immigration, Aboriginal identity, religion, education and employment are now optional, and for the delivery of targeted programs for those most in need, it is this detailed demographic information that is most essential.

Because the choice to make the NHS optional has introduced a significant amount of error into the statistical process, there will be less certainty that the data collected from the survey will be accurately generalized to the remaining population. Why? Because Statistics Canada will be unable to determine if certain sub-demographics are systematically declining to fill out the NHS, and will therefore end up unrepresented in the final weighting of the data.

You might think this is statistical pedantry, but I see it much like the Federation of Francophone and Acadian Communities of Canada did when they sued the federal government. They argued that without reliable data about the francophone presence in Canada, the quality of government services in French could suffer.

Because it is increasingly clear that the 2011 Census is broken, some Canadians are vowing to complicate the collection of data. As I mentioned, it's illegal to refuse to answer Form 2A, but the Count Me Out campaign is instead encouraging Canadians to adopt a strategy of minimum cooperation. They've even provided a chuckle-worthy list of alternative methods of compliance that includes ignoring the post-paid envelope Statistics Canada provides and instead mailing your completed 2A to Tony Clement's cabinet office, postage-free!

As for me, I'll comply with the Statistics Act and fill out my short-form eventually, but not before I demonstrate my opposition to the politically-motivated tinkering that has blocked the collection of reliable, valid data that is so important in properly delivering social programs.

In my view, a census that ignores basic methodological standards isn't much of a census at all, and what you do with the form is up to you.

Michael Borrelli is a social researcher living with his family in Hamilton's North End. He tweets @BaysideBadger.

17 Comments

View Comments: Nested | Flat

Read Comments

[ - ]

By mrgrande (registered) | Posted May 06, 2011 at 08:43:40

My household was selected for the NHS. I had a hard time deciding whether or not to do it because it's not statistically sound, dammit! I'm a computer guy, I love stats!

I ended up filling it out...

Raise the Hammer

Opinion