Projection Update: NDP Majority Government (Surprise!)

I don't think anyone would be terribly surprised at this point to learn that we are producing our first projection for this election, which predicts an NDP majority in BC. The majority of the polling and the consensus among the pundity seems to confirm this is the case. Indeed, as we said in our last entry, this hasn't been the most interesting "horse race" type of campaign.

What's more surprising perhaps for our purposes is that we're predicting a few Conservative seat gains, as well as Green Party gains.

Posh-posh! You say.

Perhaps as we refine our model in the next few days this will change, but consider a few things that we don't think others are taking into account.

1. Seat totals have to be decontextualized from 2009 and then re-contextualized for 2013. Leaders in 2009 aren't around any longer, and new leaders have come to take their place, leaving their marks on the electoral territory. That will have an effect on their seats. The same goes for incumbents, "notable" candidates, and other in-riding controversies.

2. The Conservative party didn't run candidates much of anywhere in 2009. But where it did, it got an average of 7.4% of the vote. More candidates will be running this time, and the current EKOS polling shows the Conservatives at twice the level of support as the 2009 election.

That support has to go somewhere, and the new territory for Conservative candidates have to base their support on some particular level, which must be estimated somehow (we can't multiply their 2009 level of support by the change in polling, because if you multiply by 0 you will get 0).

Averaging the vote totals of all the ridings gives us a rough idea of what the BC-wide vote will be for each party. If a party's added-up BC-wide score is significantly less than what the polling says it should be, there needs to be a correction.

Based on the change from the average level of support in a Conservative-contested riding, we currently project four conservative breakthroughs in Boundary-Similkameen, Chilliwack, Kelowna-Lake Country and Kelowna-Mission. Langely, even with a few generous bonuses thrown at leader John Cummins, is very close, but still comes up a bit short in our numbers.

"Other parties" are a perennial problem for projections because they are generally very small parties scattered about the territory, mixed in with perhaps a few people in serious contention for a riding (or in BC's case, Delta South, where independent MLA Vicki Huntington holds the riding). Pollsters count the support for all these various columns in one big "others" number (usually less than 5% in polls, and much less on election day). Still, we have Ms. Huntington keeping her seat.

The Green Party may very well be in contention in Vancouver-False Creek, based on previous levels of support. Though perhaps with former Vancouver mayor Sam Sullivan running, this might not happen. More tweaks to our model will allow us to see if this bears out.