Wednesday, June 30, 2010


What what?

Statistics Canada has quietly made major changes to the country's census in time for the forthcoming round of national sampling in 2011. The long census questionnaire that provided information on a broad range of topics such as ethnicity, education, employment, income, housing and disability has been eliminated. Instead, those questions will be asked on a new, voluntary National Household Survey (NHS) and the results will never be released, in contrast with the treasure trove of census data that become public after 92 years.
First, some clarification on the 92 year figure. The Vancouver Sun article implies that the data collected in previous Census isn't released for nearly a century. Not true, or this wouldn't exist. The 92 year rule is about identification of respondents. This means that in 92 years, your name, birthday etc with your answers is made publicly available, but ONLY if you check the consent box when you complete your Census form. The other data, from which you cannot be personally identified, is made available in the present.

I use Census data a lot. I've also worked on the Census in the past. And while I have some serious questions about the ethics around methods of data collection and enumerator training for certain populations, I am left in no doubt as to the necessity of collecting the sort of detailed data one enters on the long form. In fact, it would be very hard for me, and many others, to do research without it.

It would make it difficult for public interest groups, citizens, business, and the government itself to collect the sort of information that enables them to track things like changes in employment, marriages, housing, ethnicity, and migration. Governments of all levels rely on this stuff for figuring out funding and service delivery. Businesses use it find profitable locations. Researchers use it to track social trends and generate the sort of papers and reports that further society and reveal problems. Smaller population centres and rural areas, where for sampling significance, the long form is used much more frequently would become black holes of statistical data.

BCL is right when he says,
So expect to see the response rate to StatsCan's new voluntary survey fall towards nil, which in effect means that the nation has taken a scalpel to its own brain and removed that part devoted to memory.
If the changes are as advertised, the result would be a hodgepodge of returned surveys that might show interesting trends regarding the sorts people who actually return the form, but little else. The data would be useless. It would be akin to trading in your GPS for a Mappa Mundi.

As to why the government would intentionally lobotomise the nation, I can only speculate. Is it just brainless cost cutting (Census is expensive and these are reality challeged Conservatives were dealing with)? Or some misguided notion of libertarianism (again, the reality challenged)?

No comments: