Title: Canadian Census and Privacy:Historically speaking, the census represents us

By San | September 18, 2007


Special to Globe and Mail Update

September 12, 2007 at 1:53 AM EDT

In Canada, census releases are no dreary catalogue of numbers. They spark debate and fuel conversation about the country. Statistics Canada has released the results of the 2006 census on marital status, families and housing, and we’re at it again. So important is the census that we debate not only its findings, but also access to its content. The latter debate, decades old, has resurfaced again in 2007. Let me explain why.

The census is the only complete inventory of our population, an indispensable historical record of the Canadian people. It’s critical to genealogy, our most popular form of history. Of all visitors to our national archives today, half are doing genealogical research. If you had ancestors in Canada in 1901 or 1911, you can find them in the censuses of those years, online from Library and Archives Canada. Your children will also be able to find their grandparents and great-grandparents in the censuses of the past century — but only after a legally mandated delay of 92 years.

The problem is that we have now denied to our descendants the same right to history that we grant ourselves. For 44 per cent of Canadians, information from the 2006 census will not be available in 2098. Why? Because, in complying with an act of Parliament in 2005, Statistics Canada included in the 2006 census a question asking whether respondents would consent to public release of their personal information after 92 years. Just 56 per cent of Canadians agreed.

The question was presented with virtually no context or explanation. How many read the fine print explaining that consent was being sought for “important historical and genealogical research”? How many read Statistics Canada’s “genealogy corner,” where “yes” responses were clearly encouraged? Likely many respondents transferred to the census a general anxiety about privacy. An instinctive “no” may easily result from inapposite fears about identity theft or suspicion of government.

The effect is that the 2006 census has been seriously impaired as a historical record, and if nothing changes, all future censuses are likely to be similarly damaged. We bequeath to our descendants no complete record of the country’s population, and to future governments, we bequeath a political and legal nightmare. In 2098, our descendants will surely insist that a family’s right to know their ancestors takes precedence over uninformed responses to an ill-framed consent question in the distant past. And why, after 92 years, should Canadians be denied access to their ancestors’ names, when people inside a government agency can see those names? Statistics Canada does an impressive job in making information from recent censuses available, in ways that prevent identification or personal details. But census information, with names of individuals attached, has many important uses after 92 years. Medical scientists use census information to trace ancestry and the genetic inheritance of disease. Researchers in pharmacogenetics use ancestry and nominal information in studying genetic responses to drugs.

Our American neighbours are much more enlightened than we are on this issue. They release census information through their National Archives after a delay of 72 years. They apply the principle of “implied consent” — a principle well known to privacy experts. When completing their census forms, Americans are consenting to the present-day use of their information by the Census Bureau, and to its use by other researchers in the distant future. Americans do not complain about the future use of their information, and there is no evidence that public release after 72 years has made them reluctant to participate.

We thought the issue was solved in 2000, when the Expert Panel on Access to Historical Census Records reported to the industry minister. Drawing from U.S., British and Canadian experience, the panel concluded that most of the information collected by the census is not highly sensitive, and that information that might be sensitive for some respondents, such as income data, loses its sensitivity over time. The panel recommended releasing census information after 92 years; it did not recommend the use of a consent question in future censuses.

The expert panel got it right. Now, government should sort out the mess created by an incomplete application of its recommendations. We do not ask individual citizens to consent to paying taxes; individual interest is served by tax revenues and the common goods obtained from public spending. Censuses are similar. We contribute, and the benefit — a complete historical record of the country’s population — is a common good overriding any claim of individual harm.

I will not be alive in 2098 to see what happens if we fail to strike an appropriate balance between this common good and the confidentiality of personal information. But if we fail, I will hear very clearly the distant voices of my great-grandchildren when they accuse us of hypocrisy and short-sighted stupidity.

Eric Sager is professor of history at the University of Victoria.

Posted on behalf of Bo Wandschneider