Saturday, April 30, 2011

More NDP gains, closing on Liberals

Four national polls (EKOS, Angus-Reid, Nanos, Ipsos-Reid) were added to the projection this morning. Unfortunately, I didn't get the details of the newest Léger poll before I ran the numbers so it will have to be included in tomorrow's final update, which should also include the latest numbers from at least three other new polls that will be released between now and then.
The Conservative slide in national support has stopped, and they are up 0.1 point to 36.8% in the projection. They are unchanged at 144 seats.

The New Democrats have gained 1.3 points nationally and have moved into second with 25.1% support. They are also up six seats to 59, which still places them in third.

The Liberals are down 0.9 points to 24.1% and five seats to 65, while the Bloc Québécois is down to 7.1% nationally. They are also down one seat to 40 in Quebec. Unless the final polls added to the projection are radically different, we can expect the Bloc to drop below 40 seats tomorrow.

The Greens are down 0.3 points to 5.7% and remain at no seats.

Regionally, the Conservatives made big gains in Alberta and the Prairies, and finally stopped dropping in Ontario. They are down in Quebec and Atlantic Canada, however.

The New Democrats are still making big gains in the projection, with increases between 1.1 and 2.4 points in British Columbia, Ontario, Quebec, and Atlantic Canada. At 31%, they've moved comfortably ahead in Quebec and are even starting to challenge the Tories and Liberals in Atlantic Canada.

The Liberals are down everywhere, dropping about a point in British Columbia, the Prairies, Ontario, and Atlantic Canada, while the Bloc is down slightly in Quebec to 29.8%.

Eight seats have changed hands in the projection.

In British Columbia, Conservative candidate Troy DeSouza is now projected to take Esquimalt - Juan de Fuca from the Liberals, while the NDP's Ronna-Rae Leonard is now the projected winner in Vancouver Island North, a Conservative riding.

In Ontario, Alicia Gordon of the Conservatives is now favoured in Kingston and the Islands.

In Quebec, the New Democrats have captured four more ridings. Raymond Côté is projected to win Beauport - Limoilou (Conservative), Hélène Leblanc is projected to take LaSalle - Émard (Liberal), Isabelle Morin is projected to win Notre-Dame-de-Grâce - Lachine (Liberal), and Alexandre Boulerice is projected to win Rosemont - La Petite-Patrie (Bloc).

And in Atlantic Canada, Ryan Cleary of the New Democrats is now the projected winner in St. John's South - Mount Pearl.

The final week's ceilings will be posted later this afternoon, with the final projection being posted tomorrow late in the day in order to capture every last poll.


  1. Hi Eric,

    Did this latest update include the batch of riding polls done by Project Democracy here:

    Thanks for all your hard work over the campaign. You've done a fantastic job! Looking forward to your post-election analysis!

  2. I wonder how much of the last minute Tory surge will be captured in the last polling update ?

  3. I don't get how the NDP can have 30% in the polls, with the Libs at 21%, and the Liberals are still ahead by six seats in your projections...

  4. Sorry to harp on this point, but it has persisted, and seems to go to the heart of the uncertainty in how the campaign is going. I am referring to the Ontario graph in your 6-graph "regional poll" figure. Early on April 26, the Tory support in Ontario for April 24 was presented as 48%. Later on April 26, the April 24 number was reduced to 43%. Your response to my query indicated that you had added additional polling results for the 24th during the day of the 26th. Fair enough. However, it is now April 30, and the Tory support in Ontario for April 24th is presented as 40%. I don't understand how this number can continue to move around. No wonder it is difficult figure out how the campaign is going, when the support of the leading party in the largest province at a certain key point of the campaign is subject to revision of 8% during the subsequent 6 days.

  5. Lasalle-emard was the place that rebecca blakie first ran as a sacrificial lamb against paul martin. now she is going to make it to ottawa and lasalle-emard is projected to go NDP. What a crazy election.

  6. I still figure that with the volatility we've seen in the past about 10 days this whole thing is a major crap shoot.

    Despite all Eric's hard work I don't think, and I'm not alone in this, that anybody can predict before Tuesday morning how it will turn out. The only constant in the last few days is definitely no CPC majority !!

  7. I know you said your model is designed to be slow in response to changes, but that seems like a serious flaw to me (akin to beauty is in the eye of the holder, I guess). These numbers seem slightly more realistic, but still not what the polling data is showing.

  8. Great work! Your blog is addictive.
    When there are such quick swings as we've seen with NDP, any thoughts of adding an adaptive filter/trending which accelerates the model to weight the polls more for recency?
    You admit there will be changes in the next few days because you see it, but the model does not handle it.

  9. Still showing Guelph as Liberal. I'll place a $1000 bet on that seat going Conservative, and not by a hair on Monday.

  10. Hi,

    Do you think your model might actually predict Duceppe losing in Laurier-Sainte-Marie tomorrow? That would really be something. This riding, along with Rosemont-Petite Patrie now predicted to go to the NDP's Alexandre Boulerice, really is the heart of the Montreal francophone vote.


  11. Thank you Eric! Enjoy your weekend!

  12. No EKOS today, so Graves to do a sort of rolling poll for the last week of the election. What????

    He's like that restaurant that you never know is going to be open or not, and you just stop going.

    Have a rolling poll or don't have a rolling poll, one or the flipping other man.

  13. Anonymous 12:04,

    The Harris-Decima poll added yesterday stretched back to April 20th.

    Matt Grande,

    My projection has the NDP at 25% and the Liberals at 24%.

    Anonymous 12:23,

    I still have one more update tomorrow, and then we'll see how they compare to the results on Monday.

  14. This mornings Nanos Poll shows Conervatives up to 38.0 % The NDP at 29.6% and the Liberals at 23.2%.

  15. I see at least 7 more seats for the Conservatives.

    I don't think Pontiac or Central Nova are going down without a fight, and they had significant advanced poll turnout.

    Van Island North, North Surrey, those aren't going NDP.

    etc. etc.
    The Conservatives have micro targeted this campaign, and a lot of their vote is in the box already.

    I know you can't project riding level factors, but they are quite significant this election, and have been for at least a year. I think the ridings that are close that have been targeted by the Conservatives, are going Conservative.

  16. The Conservatives look to be diving fast and hard in Atlantic Canada, from a high over 40% to a final result that may put them below 30% out east. Which seats do you think they could lose if this trend holds?

  17. Peter you'd be very, very foolish to rule out a slim CPC majority at this point.

    Polls seem to show a last minute surge for Harper going into monday. Same as '08.

    When you watch a program like "The House" on the CBC and they tell you its a crapshoot you need to know that's them thinking the CPC is going to win a majority and not wanting to say so.

  18. So it would seem that in regards to the much feared "vote-splitting" - at this point any NDP gain is equally capable of provoking a LIB -> CON swing as it is a CON -> NDP swing (leaving the Conservatives stable but not increasing). The question is, can the NDP breach the threshold where new gains will consistently unseat Conservatives rather than installing new ones?

  19. Btw people, we have the undecideds at 12%. That's VERY, VERY high at this point, given that at least 11% of people have already voted, (probably more)

    That's closer to 14% of people who haven't voted being undecided. That's huge for two days before e day.

  20. Well that clarifies things re Tory support in Ontario on April 24th. Averaging one polling result at 48 with other polls brought your estimate to 43, and then subsequent averaging with still more polls brought the overall estimate for that day to 40. This suggests that all the others were in the mid to high 30s. I wonder how one gets 48 in that environment. Not unimportant as 48 would point to majority. Having to revise backward like that makes it seem like change was more abrupt than perhaps it really was in the population.

  21. Nanos is showing the CPC surge that comes at the end of the campaign.
    CPC 38 NDP 30 Liberal 23

    This means that the 400 people sampled on Friday had the most likely out come of
    Cpc 41.35
    NDP 26.4
    Lib 25.9

    Looking back in 2008 Nanos poll with 2 days left had the
    CPC UNDER by 4.6
    NDP Over by 3.8 and
    Liberal over by .8

  22. I have a couple questions about your projections for PEI.

    Compared to their 2008 standings, you show the Conservatives up 0.3 points in Charlottetown and 0.1 in Egmont, and down 0.8 points in Cardigan and 2.8 points in Malpeque. If Conservative support is at 33.4 in Atlantic Canada, up 3.8 points from 2008, why isn't this reflected in your projections for PEI? Also, you show the NDP up between 3.2 and 5.5 points on the Island but only 2.4 for Atlantic Canada.

    I completely understand that the poll #s for Atlantic Canada aren't an accurate picture of what's happening in each province, but I haven't seen any riding polls for PEI or even a poll at the province level. Just wondering what you use to determine the numbers PEI?

    Fantastic site by the way!

  23. Ashley,

    Support in Atlantic Canada is not handled by just looking at 2008's results at the regional level. Because of the huge shift in support in Newfoundland, poll results in Atlantic Canada are distributed according to the proportions reported in provincial-level polls that were released by Angus-Reid and CRA in February.

  24. Steve Beaulieu30 April, 2011 14:31

    No way Bloc will keep 40 seats.
    Quebec's specific survey this morning in La Presse by Angus-Reid gives 45% to NDP vs 26% to the Bloc.
    You need to add AT LEAST 10-15 more seats to NDP in Quebec.

  25. We are at a magical point with the vote splitting where even a minor systemic error in the polls means hugeshifts in seats.

    The polls, collectively, are saying (more or less)

    Tories 36
    Dippers 31
    Grits 22

    It the Toroes are actually at 39, we likely have a majority Government for Mr. Harper, it it's actually 34 for the Dippers it's likely PM Layton...

    The Grits are toast is about all we know right now....

    Monday will be exciting!

  26. blaffergassted30 April, 2011 16:58

    The NDP surge will have separate impacts in the different regions.

    They'll probably pick up 8-10 ridings in Quebec, but will not achieve the lofty numbers shown here due to their lack of organizational abilities on the ground. Glad to see Sylvie Boucher and her copyright bill being turfed!

    I'd also guess that the numbers for Saskatchewan, the Atlantic and the North are unreliable ... the pollsters are relying on too small a sample. Nettie Wiebe will win in Rosetown, and Jack Hicks may take Nunavut.

    In BC, watch for Van Isle North and Surrey North to go orange. Saanich may still surprise.

  27. "Jack Hicks may take Nunavut."

    A poll of 400 released yesterday showed him 63.2% behind Aglukkaq

  28. I wouldn't change the St. John's South - Mt. Pearl projection so quickly. The nanos poll puts Liberals in the lead in Atlantic Canada (14% higher than the NDP) and I live in that riding, and I have to say while Cleary is definitely going to be a close second, they still don't have enough support in Newfoundland to put him over the top.

  29. Merci beaucoup pour votre travail et aussi pour vos analyses. J'ai hâte de de comparer avec les résultats du 2 mai.

    Prochaine élection, une version française?

  30. This is indeed a crazy election. I can't predict my own riding from the responses I'm getting at the door in last-weekend canvassing. Today has been no exception. There's a massive Green/Liberal "Undecided" response amongst committed, thoughtful, informed voters. This was supposed to be a Tory stronghold. It's anything but.

    If other ridings are anything like this one, pollsters might as well set up a dart board to give their predictions.

    Tomorrow night will be very interesting. Our Victory Party may run well into the evening...

  31. Truthfully, I'd like to see the NDP get a crack as the official opposition. I'm curious to see how and where they would make concessions on their ideologies, as is necessary when applying ideals to practical situations. I'm not especially hopeful, but I am curious.


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.