I won’t spend as much time going through the NDP or Bloc vote as I did for the Liberals and Conservatives. Unlike for the other two parties, the NDP vote isn’t quite as predictable – the best model I can come up with only accounts for a little over half of the variation. Personally, I tend to think this is proof that voting NDP defies any sort of rational explanation, but there may be another explanation.

However, if we do look at the model, it does suggest the following ridings as possible long-term gains for the NDP, based on the types of people living there: Nunavut, Edmonton East, Edmonton Centre, Saskatoon-Rosetown-Biggar, Winnipeg South Centre, Sydney-Victoria, Yukon, Desnethé–Missinippi–ChurchillRiver, Regina-Qu’Appelle, and Palliser.

And it suggests their seats in Outremont, Western Arctic, Nickel Belt, Sackville-Eastern Shore, and Algoma-Manitoulin-Kapuskasing are being won for reasons other than the demographic composition of the riding. That doesn’t mean the NDP can’t keep taking them, but it certainly means that, long term, it will take some work to hold them.

The model works well for the Bloc, with french as mother tongue being the strongest predictor. What it shows are that the 48 “best” Bloc ridings are all held by either the Bloc or the Conservatives. In short, Conservative wins in Quebec have come mainly in ridings we’d expect to go Bloc. This is hardly earth shattering news, which is why I’m not going to talk about it for more than a paragraph.

As for the Greens, well, once again, their vote is difficult to predict, so there’s not a lot to say about it. I will say that Elizabeth May appears to have made a wiser choice when it comes to picking a riding this time than last – Saanich-Gulf Islands is the 8th best Green riding in the country based on demos, whereas Central Nova is the 245th best.