hi.
the results you got back from edmonton yesterday are going to confuse people, but i saw this coming. i just want to explain what i think is happening and why these polls are useful, and could lead to huge upsets in calgary and edmonton.
if you go to the aggregate sites, like cbc, the apply a uniform swing model across the provinces. the reality is that this is not a smart way to build a model. there's not a lot of difference between rural alberta and rural saskatchewan, so that would make sense; but there are large differences between rural and urban alberta (and rural and urban ontario).
so, for example, the conservatives got 66% in alberta in 2011. they appear to be down about 20%, into the low 50s. the way that the models work is they tale 20% off everybody, under the assumption that a 20% decrease in conservative support is uniform across the province. it assumes the conservatives are down by the same amount in edmonton as they are in yellowhead.
but, i think a little bit of clear thinking concludes that this is a ridiculous assumption. so, let's start from a more reasonable assumption. given that rural conservative support has been roughly steady for roughly a century (there was actually a socialist movement in alberta a hundred years ago, but it's been stable since), let's assume that that holds and that, despite the conservatives being down 20% province wide, they have not changed in the rural ridings. that would mean all the changes are in the urban ridings.
we can make the math easy because, by a fluke, the population of calgary and edmonton is, together, roughly half the population of all of alberta.
that would mean that a 20%
and, that would mean that calgary is probably in play.
you've actually demonstrated this little bit of math i've been throwing around; thank you. conservatives are down far more in edmonton than the models are suggesting, and what i've done here is explain why.
but, consider the possibilities.
.55 - .55*.4 = .33
.65 - .65*.4 = .39
.75 - .75*.4 = .45
they're all in play. even harper's riding.
but, this mathematical argument is useless in determining where liberal and ndp support are or how they're distributed. i would have argued the liberals were ahead in the ridings you've shown the ndp are ahead in. so, getting riding polls is key if you want to, say, knock jason kenney off.
and, hey - best way to get right of stephen harper is to have him lose his seat, right?
j