Those polls

I've been following the polling data in Canada with some interest. I'm always curious to see just how the various parties are doing, and trying to correlate that with the behaviour of the various members in the recent past. Call it backwards induction for the purposes of public choice analysis.

At any rate, recent polls have been absurd and unreadable. Consider just two, released today (but you can probably pick any random batch). The Ekos poll (which polled people before Belinda crossed the floor, but still audaciously dubbed their poll "Belinda Bounce?" Words escape me...)--where was I? ah--showed the Grits at 34.7% and the Tories at 28.3%. Compare that to the Compas poll (taken after Belinda's departure) which puts the Tories ahead at 38% (that's heading into majority territory) and the Grits sagging at 29%. (Compas, though, only polled 500 Canadians. That's useless.)

Just in the last election, I covered the pollsters following the minority win of the Grits. They were all over the place too. You can read that piece, entitled "Public opinion says the pollsters blew it" in PDF format.