Wednesday, July 21, 2010

On the Canadian Census

The ideal census involves a purely random sample. I believe this is why the director resigned. A voluntary statistics gathering from a statistical point of view violates randomness. The problem with the Canadian census is that purity. As was pointed out by
Anthony in Ontario, no matter how "mandatory" was every Canadians requirement to provide census data, you can never get pure compliance.

Particularly when it comes to certain people-groups like the homeless, with almost 100% non-compliance the whole purity of the process is statistically corrupted. The homeless remain invisible to Canadian society as a result.
Then there is the people-group of the stubborn, who will go to jail rather than comply with such a violation of free speech. How many of those long forms are swept under the carpet? Then there are people who live in such remote places, they are never touched by the census taker. There are people in Canada like that. Then there are the ex-pats. Some live outside Canada for 6 months less a day. They may be missed. I could go on.
While statisticians say that they can calculate out the anomalies of the impurity, they really can't without becoming a problem--tinkering with what is supposed to be pure data. That's the thing about samples, you touch them to allow for this or that, and they are not the same. People and companies who push for the census don't want to hear about that.

If this whole census phenomena is driven by big business, how did that happen? Why does the government need to know how many toilets I have? Big business might want to calculate toilet paper... but why should that be pried from Canadians under threat of censure?

Ironically then, wouldn't the government quite rightly be standing up to big business? Shhh....big business doesn't want to hear that stats-can's data is "massaged".

For me, I wish we could dump the whole thing. Its an invasion of my privacy that I and many other Canadians deeply resent.

No comments:

Post a Comment