The federal government is pushing hard to make a significant change in the way Statistics Canada conducts the 2011 Census.  If the Minister responsible for Statistics Canada, Tony Clement, has his way, the long form portion of the Census will be eliminated, and replaced with a voluntary data-gathering instrument.

The news of this proposed change has sparked debate across more than just those involved in the Census.  Advocacy groups, academics, some business leaders and many within the labour movement are expressing their deep concerns about the prospect of a shift to voluntary measures within the Census.

From a statistical perspective, the elimination of the long form portion threatens data reliability in a major way.  Statisticians point out that having the current long form ensures far greater reliability when it comes to weighting information collected in other sample surveys.  Without that comprehensive, detailed data gathered through the long form, estimates become less reliable and bias creeps into the results.  Don Drummond, the former chief economist with TD Bank and member of the National Statistics Council that advises Statistics Canada, summarized the problem recently by noting that with a voluntary survey, the results will reflect the bias of those who are most likely to complete the survey.  In fact Drummond stressed that increasing the size of the voluntary sample won’t fix the problem.  It will only “magnify the bias”.

Adding to the controversy was the unexpected resignation of the head of Statistics Canada, Munir Sheikh, who has stated publicly in his letter of resignation that the proposed voluntary survey is no substitute for the long form.  For Mr. Sheikh, the integrity of Census data relied heavily on having the long form remain a mandatory part of the Census.

Mr. Clement has tried to deflect criticism of his proposed changes by suggesting that the long form was intrusive and raised privacy concerns.  However, the privacy issue seems to be more anecdotal given that there have never been any formal complaints to the Federal Privacy Commissioner regarding any previous Census.  Moreover, compliance with the mandatory long form has never been a critical issue, at least not until Mr. Clements decided to say it was.

What is particularly shocking about the federal government’s approach on the Census is their apparent lack of regard for how a diminished standard to data collection will work its way through economic and social policy decision making.  Reliable Census data provides the platform for hundreds of other surveys done by Statistics Canada.  Those surveys, in turn, provide information on critical things like the rate of inflation, household formations, cultural diversity, the urban-rural mix of our population and other critical components needed to develop sound public policy.  Businesses too rely on this data for everything from where to target new investments to what products should be part of next year’s inventory.

Unfortunately, The Harper government seems to have ignored all these factors hoping, instead, to gain a partisan advantage by championing an ill-considered cause.  Hopefully, the growing public backlash will convince the Federal Cabinet to acknowledge its mistake and reinstate the long form in the 2011 Census.