Thursday, October 29, 2009

Projection Update: 139 CPC, 95 LPC, 49 BQ, 25 NDP

The Conservative rise and Liberal decline continues.The Liberals have dropped three seats, two of them going to the Conservatives and one of them going to the New Democrats.

Nationally, the Conservatives have picked up 0.5 points while the Liberals have lost that amount. The gap is now 6.7 points, 36.1% to 29.4%. The NDP and Greens have each also picked up 0.1 points.

In British Columbia, the NDP has gained 0.4 points. The Greens have lost 0.3 points and the Liberals have lost 0.2 points. The Conservatives remain steady at 38.3% and 21 seats. The New Democrats have moved into second with 25.0% and five seats, while the Liberals are at 24.8% and ten seats. The Greens, at 11.2%, are far from winning a seat.

In Ontario, the Conservatives have opened up their lead by another 0.8 points, and stand at 39.1% and 52 seats. This is where one of their seat gains has come. The Liberals have lost 0.5 points and a seat and sit at 35.3% and 43 seats. The NDP is steady at 15.1% and 11 seats, and the Greens bring up the rear with 10.0%, a gain of 0.2 points.

In Quebec, the Bloc Quebecois has dropped 0.1 points and stands at 36.9% and 49 seats. The Liberals have lost a massive 0.8 points and one seat and are not at 27.3% and 17 seats. The Conservatives have gained 0.5 points and one seat, and are at 18.3% and eight seats. The NDP has picked up 0.3 points and is at 10.8% and one seat. The Greens are steady at 6.2%.

As for large movements in the other regions, the Conservatives have gained 0.9 points in the Prairies and 0.4 points in Atlantic Canada. They've lose 0.3 points in Alberta, however.

The Liberals have lost 0.6 points in Atlantic Canada and a seat, which has gone to the NDP. The NDP has also lost 0.6 points in the Prairies.

Things are slowly returning to October 2008 levels. Peter Donolo has his work cut out for him.


  1. As Nate Silver so deftly demonstrated at FiveThirtyEight, convention bumps are real, but they're also predicable and temporary.

    The Liberals weren't popular with Dion as leader, and selection Ignatieff gave them a signiicant bump in the polls. But, that bump had to subside eventually, and in the intervening time nothing has really changed. The Liberals haven't released a signficant policy document; the Conservatives haven't have a major gaffe - so we see the numbers return to what they were before the Liberals selected a new leader.

    I would argue that this was predictable. The Liberals need to DO SOMETHING to change their fortunes. Hope is not a strategy.

  2. OK, so now the popular vote projection in BC has the NDP just ahead of the Liberals, but the seat projection still has Liberals at 10 and NDP at 5. Still doesn't make sense. Otherwise, most of the projection makes sense.

  3. Apparently, Liberal vote spreads more efficiently in BC. E.g the Liberals may have just enough to win in their urban strongholds with little or no support anywhere else, while the NDP vote spreads more equally across the province.

    Compare that to national prediction, where the same level of support - 9.2% gives 49 seats to the Bloc and none - to the Greens.

  4. I return to the fact that in the last election the Tories had 44% of the vote in BC and the NDP 26% - and that yielded 23 Tories and 9 New Democrats. According to the projection the CPC-NDP gap has narrowed from 18% to 13% - and with one exception the Liberals are not a factor in any of the 9 current NDP seats.

    I challenge anyone to look at the results riding by riding in the last election and give me a list of 4 NDP seats that would be lost if the popular vote in BC went from C44/N26/L19 to the projected C38/N25/L25. You can't do it.

  5. Hey Leonard,

    Liberal support IS very concentrated in BC. In ridings where they are the third party and its a close NDP-Cons race their vote tends to collapse.

  6. The fact that Liberal support in BC is so concentrated is what makes it hard for them to gain very many seats. They are only competitive in a handful of seats - everywhere else if their vote goes up - it will just take votes away from the Tories and deliver seats to the NDP. See North Vancouver Island, Surrey North and even Kamloops - in each case if the Liberals go up six points and the Tories down 6 points and the NDP vote is stable - guess who gains the seat.

    In 10 days time we will have a good test of what is really happening in BC with the NWC by-election. Of the 9 NDP seats in BC, it was the most marginal in the last election. I prediuct it will be an NDP hold and by a significantly wider margin that the 3% margin last year.

  7. The problem with the NDP in BC is that in 2004 they won only five seats with 26.6%. Then last year, with 25.0%, they win nine seats, and win 10 with 28.6% in 2006. It doesn't compute easily.

    But, since the 2004 is the older result, I'll give more focus on the 2008 result.

    I have taken another look at the numbers, and I've tweaked it. I'll update the projection with the change before the next weekly update.

  8. Eric,

    will you remove 2004's results from your model after the next election?

    It was so long ago now and I get the feeling that it does some weird stuff with the numbers, mostly to the benefit of the Liberals over the NDP/Conservatives.

  9. Not entirely. I will keep the 2004 results to help project seats, but will probably remove it, or severely lower the weight of it, from the popular vote aggregator.

  10. Eric,

    I was wondering if you could display the polling trends on a time scale instead of the current scale, which appears to be 1 tick per poll released. I don't know about others, but I think a time scale would be a bit more intuitive and may show trends a bit more accurately (with respect to time).

    For example, if I look at the charts right now, it would appear that the Liberals had a slow rise and slow fall, where as the gap that has appeared recently only started to show itself in the middle of September.

    Anyway, just a suggestion.

  11. Yes, I'm aware of that issue. I'm not sure how to manage it easily but will take a look at it.

  12. while we are on the subject of improvement. One thing I would like would be if the tables showing all the polling data for the country as a whole and for each region could be in chronological order. Right now they are all over the place.

  13. It's not unthinkable for the Liberals to get 10 seats in BC, but they'd have to be much stronger there than the polls are currently showing them to be.

    Lots of people have already picked on the BC results, so I'll leave them alone except to say that the Liberal rise in BC has actually helped the NDP to an extent, as their gains seem to be coming mainly at the expense of the Conservatives (Cons have fallen 6.2% in BC since 2008, Liberals have risen 5.5%).

    I'm a little confused by the projection that the Liberals will pick up one NDP and one Conservative seat in Sask/Man. I'm guessing those would be Desnethe-Missinippi-Churchill River (SK) and Churchill (MB), but they had about 30% in those ridings in 2008 to their opponents' >40%. That seems like a pretty big swing, but I guess it's due to wins there in 2006.

  14. Thanks again for the great work on the website Eric. Your projection looks fine.


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.

Note: Only a member of this blog may post a comment.