The Canadian Census is a major source of data for any researcher in urban planning, sociology, economics, geography, linguistics, and many other fields. While many scholars argue that the Census is prone to error and non-representation (for example, people without a regular address or students living away at college may be underrepresented), it is simply, to quote The Globe and Mail, “Canada’s only complete national database on education, income, employment, ethnicity and language”. It’s also a very costly endeavour undertaken every five years, with the next one scheduled for 2011. Which is probably why Tony Clement, Minister of Industry and Minister Responsible for Statistics Canada, very quietly arranged to scrap the long form next year, although he’s hiding behind alleged privacy concerns. The decision has prompted a quick response from the Canadian Institute of Planners, Metropolis, the Federation of Canadian Municipalities, market research companies, and many other organizations who rely upon the data for research and policy work. Provincial governments, non-profit groups and many other bodies dealing with target populations, such as immigrant settlement services or at-risk youth, depend upon the data to develop and deliver their programs effectively.
The Census is a statistically viable data source because it is a mandatory survey administered by government officials, with every fifth household receiving a more in-depth questionnaire, known as the Census long form. Eight basic questions, such as age, sex, marital status, and the relationship of people in a household, are recorded on the short form, and many of these questions date back to 1871. Fifty other questions (that’s right, 50) , such as mode of transportation used to commute to work, commute distance, detailed questions about income and occupation, and detailed questions about ethnicity and immigration are on the long form. Although many of these questions have been on the long form for 35 years, some are relatively new: the two transportation questions, dealing with transportation mode and commute distance, date back to 1996. In the absence of a national transportation survey, this data can tell us which groups travel by transit the most or which cities have the highest cycling rates, just to give a couple of examples. I published a paper in Plan Canada just six months ago that compared youth and young adults’ transportation modes in the ten largest cities in Canada. I’m currently using Census data from 1986 to 2006 to investigate how immigrants’ housing and transportation choices have changed over time.
I fail to see how any of these questions could be considered an invasion of privacy, especially considering the fact that names or any identifying characteristics are never linked to the data. This on top of the fact that Census data in Canada, unlike in the US where data is free and public, is incredibly restricted. Only researchers in government or academia have access to the Census microdata, that 20% sample that contains the long form data. Plenty of other government agencies collect private information: you need to report your height, weight, hair colour and eye colour to get a driver’s license.
The federal government is planning to replace the long form with a voluntary “national household survey” that will be mailed out to approximately 30% of Canadian households, which the Tories argue will reach more households than the long form did. Anyone done a mail-out survey lately? The response rate is usually around 20-40%…what is 30% of 30%? And critics have already noted that the most vulnerable groups, such as immigrants, Aboriginal communities and low-income populations, will be the least likely to respond.
While the opposition parties are marshalling their efforts to reverse the decision, petitions are circled and we all write madly to our MPs, the media has given the issue a fair shake: the issue was covered in all the major papers and online venues, and not just by journalists (see “Canadians must be able to count on Statistics Canada” by academic Richard Shearmur in The Montréal Gazette). In the past week, the Canadian Medical Association, faith groups like the Canadian Jewish Congress, and economists like former TD Bank chief economist Don Drummond have all voiced their objections to the decision to jettison the long form.
Beyond the appalling lack of respect for the vast amount of data generated by the long form and its necessity to researchers, policy makers and community groups, the troubling issue here is that Harper’s “new world order” even extends to the collection of statistics about the people he is supposed to serve.
An update on this story: the head of Statistics Canada, Munir Sheikh, tendered his resignation July 21st over this issue, saying the voluntary “new Census” cannot be considered comparable to the long form.
