Let me say I would favor some sort of compromise. I had a coffee with my MP the other day. The story had just broke. I pointed out the problem relates to randomness. I don't know why no reporter gets into this.
In order for the various long forms to have significance for everyone else. That is done by an attempt at random sampling. The holy grail is the pure-raw data. Pure in that it samples people across Canada in such a way so they can do a statistical average... and infer for example the average Canadian has 2.3 bathrooms lets say. The only way you can deduce this is if the sample is random.
Here is the problem, and its one I haven't heard stats-can own yet. The sample isn't pure. Stats Can tries to make it pure, hence the mandatory aspect. How many homeless people do the census? How many migrant people perhaps with extended visas. ( The implication is that they may be on a path to citizenship, have their own house, bank accounts, jobs, etc.
How many lie in protest to the whole procedure? How many people refuse to do the census and get away with it? Jail is not enforced.
I am only taking a stab at some of the problems that impact the purity of the data.
I was surprised to read this article in The Economist: ...Gah it won't let me have it unless I subscribe.
You can see the article at http://paradoxicalx3.blogspot.com/2010/07/from-economist.html
In it it mentions that a number of Scandinavian countries have ceased to use a mandatory census. A very puzzling piece of info. But apparently Big brother over there is so big, that they are tracking their citizens all the time.... gah!
Apparently Germany is next...
It is going to be interesting watching this unfold.
No comments:
Post a Comment