On sample sizes and the absence of a mandatory census

Posted by & filed under Canada, Learning is fun!, News, Politics.

With the 2011 Canadian federal election a week away, voter turnout looking to thankfully exceed 2008’s embarrassingly dismal 58.8%,  the polls are starting to look extremely interesting. While earlier polls indicated we could expect very nearly an identical Parliament to the one that existed before the election, an apparent surge in support for the NDP has put the party within theoretical reach of becoming Canada’s Official Opposition.

So if the New Democrats took an unprecedentedly huge share of the votes–for once in their history, actually benefiting from first-past-the-post, rather than consistently taking a far smaller share of seats relative to the popular vote cast…

And if the Conservatives managed not to benefit from this at all either, winning another minority government…

And if the Liberals, NDP and Bloc Quebecois voted for a motion of non-confidence…

And if Governor General David Johnston doesn’t dissolve Parliament, forcing another election–which, as viceroy, (stupid monarchy) he’s perfectly entitled to do…

…then Johnston would have to turn to the newly anointed leader of the Official Opposition and ask him if he was up for giving it a go. And if the answer was yes, as it almost certainly would be, then with a third of the seats, a quarter of the popular vote, and the leadership of the Liberal and Conservative parties more appalled than ever imagined possible…

Then this guy would become Canada’s next prime minister:

Credit: Scott Rogers

As entertaining as that would be, and while it’d be great for galvanizing Liberal and Conservative support for abolishing first-past-the-post, an important question remains: can we actually trust those polls?


The CBC posted an article on the issues facing pollsters today: chiefly, that the average Canadian has neither time nor desire to talk to pollsters on the phone. This, of course, is countered somewhat by the typically Canadian instinct to not actually hang up when asked if we can spare a half hour of our time, but still, the fact is:

But for the purposes of their polling, researchers are obliged to assume that the 15 per cent of callers who agree to spend 20 minutes talking to them are representative of the 85 per cent who are too busy or whatever to participate or who never pick up at all because they can identify a pollster through Caller ID.

(Emphasis added for my own amusement.)

Still, editorial snafu that may be, it’s pretty accurate given what I presume about these folks who do not under any circumstances want to respond. A Freudian slip of the word processor?

One thing is certain: the typical respondent is not the typical Canadian. How different they may be is another matter.

PS: Vote. Seriously.

Comments are closed.