Thursday, October 29, 2009

NDP in British Columbia Change

As promised, I've taken another look at the projection model in British Columbia. From 21 Conservative, 10 Liberal, and 5 New Democratic seats, the projection is now giving the NDP two more seats, one each coming from the Conservatives and the Liberals.

So, that brings the Tories down to 138 seats, the Liberals down to 94 seats, and the New Democrats up to 27.

And that's final!

12 comments:

  1. Your credibility is really at risk here. You simply cannot name 10 seats the Liberals will win in BC. And with the numbers we've seen out of that province recently, there is no way the NDP gets as few as 5. Clearly you don't read the news from out that way, or your assumptions are wa-a-a-ay off.

    Your numbers for the NDP's seats are well below most other predictors, and your Liberal numbers are well above.

    Could you please clarify your methodology, because on the face of it, the results it's predicting make no sense.

    ReplyDelete
  2. Liberal vote in 2008 in BC - 19.3%, 5 seats.

    Liberal vote in 2006 in BC - 27.6%, 9 seats.

    ReplyDelete
  3. "A reader" doesn't seem to understand that what DEric is trying to do here is to purposely NOT read the news etc... this is about purely using polling data to project election results. There are plenty of other sites where people can exchange anecdotal tips about this riding or that riding.

    I'm glad that the BC projection is a bit more sensible - and i have to assume that there are a bunch of seats that are right on the borderline in BC and that if the NDP projection goes up from 25% to - say - 26% or 27%, that seven will quickly become 9 or 10. But I'll be patient.

    ReplyDelete
  4. Eric,

    When you say that you have "taken another look ", what do you mean? What was the nature of the adjustment?

    Was there some sort of data entry or similar error that was corrected?

    Have you adjusted the formula for how your model maps popular vote somehow (for example, by changing the weightings of how the three elections are used)?

    Or was this adjustment simply an ad-hoc adjustment?

    ReplyDelete
  5. I re-evaluated how I determined the NDP's vote based on the last three elections. The popular vote to seat ratio for the last three elections for the NDP in BC have been inconsistent. Using the 2004 and 2006 results made it more consistent, but in the end I agreed that it was not more accurate. So I put more weight on using the 2008 result and comparing that to the other two results to give me a better ratio of votes to seats for the NDP in BC.

    ReplyDelete
  6. I think there is a simple explanation for why the NDP got 26% of the vote in BC in 2004 and got 5 seats, but had 26% in 2008 and got 9 seats. In 2004, the NDP was coming out of nowhere having taken only 11% of the vote in BC in 2000 and had only 1 incumbent running for re-election - so the NDP vote tended to rise evenly across the province and wasn't all that efficient in yielding seats. In 2008, the NDP had 9 MPs running for re-election - many of whom had become very well-established if not impregnable. As a result the NDP vote became much more concentrated in seats they already held or were targeting and dropped in other seats.

    So while the NDP vote in BC was 26% in both 2004 and 2008 - it was distributed very differently. I think that the projection would take this into consideration if you gave more weight to incumbency and to the results of the last election.

    ReplyDelete
  7. BC is notoriously hard to predict by any method because so many of the ridings feature three-way races. Literally any one of the LPC, CPC, and NDP could win any or all of a dozen close ridings.

    ReplyDelete
  8. There actually aren't that many three-way races in BC. Outside of Vancouver, the Liberals basically don't exist. In the last election, the only riding in the whole province that could be described as a three-way race was Vancouver-Kingsway and maybe to a lesser extent Newton-North Delta and Esquimalt-Juan de Fuca - and that's about it.

    ReplyDelete
  9. I'm thinking of ridings like West Vancouver-Sunshine Coast - but good point about the Liberals being nonexistent outside Vancouver.

    I overstated how many ridings have three way races. I do wonder if BC is the place where vote-splitting between the NDP and Greens might hand more seats to the CPC.

    ReplyDelete
  10. West Vancouver-Sunshine Coast is not a good example of a three way race - the NDP can never win there since its mostly multimillionaire dollar monster homes in West van, plus a few ski chalets in Whistler.

    I don't think the Green party is much of a factor for the NDP in BC. If you look at the two ridings they lost to the Tories last time (Surrey North and North vancouver Island) and NWC which they almost lost - they were all ridings where the Green vote was extremely low. Some of the best Green showings in BC were in supersafe Tory seats in the Okanagan.

    ReplyDelete
  11. I agree on the Green votes. Note how well the Greens often do in Alberta ridings where the outcome hasn't been in doubt for 30 years.

    ReplyDelete

COMMENT MODERATION POLICY - Please be respectful when commenting. If choosing to remain anonymous, please sign your comment with some sort of pseudonym to avoid confusion. Please do not use any derogatory terms for fellow commenters, parties, or politicians. Inflammatory and overly partisan comments will not be posted. PLEASE KEEP DISCUSSION ON TOPIC.