i want to clarify a point in case it's unclear.

if singh hangs around for a few elections, if he can hold off the inevitable leadership challenge, he could end up with a bigger and more durable coalition than layton could put together. this isn't really a question of total appeal - there's lots of indo-canadians here, and the religious left has a history here, too.

but, any kind of longterm coalition or tent that singh is going to build is going to be dramatically different than layton's - and the outcome of this election is probably going to be more about breaking up layton's lingering coalition than building singh's new one.

the older parties, the liberals and conservatives, have roots in their communities that go back centuries. that's not true of the ndp, which shifts in allegiances through different leadership phases. tommy douglas' base was drastically different than broadbent's, which was different than layton's, and singh's will be unrecognizable from any as well, should he be able to build it. so, if you're building a model, it would not be a good idea to rely on past ndp results the same way you'd rely on past liberal results. the same thing is true of the bloc, even if it's only going through it's second shift. remember - it started off as a splinter of the conservatives.

douglas was a religious reformer, a prairie populist, that preached the "social gospel" of wealth redistribution, as well as fire and brimstone for the gays. he supported eugenics, and opposed abortion. he was wildly popular amongst farmers, but viewed rather derisively by people in the cities that saw him as a bit of a nut. in the end, he was pushed out by the first lewis, a socialist jew, who tried to swing the party to the hard left. it wasn't even until broadbent in the late 70s that they became the union party, and only after unionism had started to decline. layton took over after nafta ripped them apart and tried to convert them into a party of latte liberal urbanites that supported small businesses, family values and social programs. in the end, he built a coalition of quebecois leftists, rural libertarians, young people and advocates for the urban poor. mulcair, a right-leaning former quebec cabinet minister who ran on eliminating the deficit, lost quebec and lost young people, but kept some of the party's legacy in tact. this is what singh inherited: a dying union movement, skeptical anti-poverty advocates and a mish-mash of rural parties.

singh is not going to do well with much of any of these groups. his coalition may include students, if he can hold them, but he's not going to win back quebec. he has a lot of work to do in winning over antipoverty activists, as well, and in the end may abandon them. his coalition is more likely to include ethnic voters in the 905 and in vancouver that have historically leaned a little right, but see something of themselves in him. and, where his religious beliefs are a liability in winning secular leftist voters, he may find it an asset in winning votes in more upscale, ethnic-dominated areas of vancouver and toronto.

so, how do you fix the model?

you need the data first.