Friday, April 29, 2011

Yet more Quebec riding polls

The three national polls added to the projection this morning (Harris-Decima, EKOS, Nanos) have little original to say, though that in and of itself is something. The New Democrats are still in second, the Conservatives are still looking weak in front, and the Liberals are still looking like the old New Democrats. A few days remain, and it seems unlikely the polls will be showing anything other than this state of affairs over the weekend.
Where these three very similar polls differ most is of great importance. EKOS shows a large Conservative lead in Ontario that would likely result in a huge swathe of Tory seats. Nanos shows a closer race that would likely result in a similar landscape as in 2008. And then Harris-Decima puts the Liberals marginally ahead of the Conservatives, meaning the landscape in Ontario would be greatly changed.

How this all shakes out over the weekend and what emerges from it on Monday will decide the campaign. While Quebec will provide us with the most surprises and determine whether Jack Layton moves into Stornoway on Tuesday, Ontario will determine whether we have a minority or majority government. A minority means another election in two or three years and the possibility of either the New Democrats or Liberals forming a government after a Conservative defeat in the House of Commons. A majority means no election until October 2015, and by then all four parties could have different leaders or the polls could be showing something completely unthinkable today.

How will Ontarians vote? Will the Conservatives put up a big lead? Will they benefit from vote splitting? Will the Liberals show surprising strength? We don't know the answers to these questions, and we haven't had a single riding poll from Ontario during this five week campaign to help us answer them. But we have had riding polls for fully 1/3rd of all ridings in Quebec, thanks in part to a new blast from CROP and Cible Recherche.
Ten - count 'em, ten - riding polls. And Cible Recherche put up another one today for Berthier - Maskinongé (which puts the Bloc ahead of the NDP by seven points). The first poll in this graph was conducted in Trois-Rivières, a Bloc riding ready to vote NDP despite only 24% of voters being able to name the NDP candidate (Robert Aubin). At 42%, that's a 33-point bump for the NDP.

Throughout the Quebec City region, CROP found the New Democrats to be extremely competitive. They are leading in Beauport - Limoilou, Charlesbourg - Haute-Saint-Charles, Louis-Hébert, and Portneuf - Jacques-Cartier, while being tied or closely trailing in Louis-Saint-Laurent and Québec. In other words, the NDP could sweep Quebec City, this after supporting ran in the 9% to 13% range in the 2008 election. The New Democrats could, depending on what happens elsewhere, block a Conservative majority in Quebec City of all places.

Moving over to the Saguenay-Lac-Saint-Jean, we see the NDP with a good degree of support but only challenging in Jonquière - Alma, where Claude Patry trails Jean-Pierre Blackburn by two points. While this riding did have a very inaccurate riding poll done for it in 2008, we've now seen two independent polls done for this riding showing that Blackburn is being hotly trailed.

This poll also shows that the Bloc is capable of holding on to some of its ridings. It is running neck-and-neck with the NDP in Louis-Hébert and Québec and has a good lead in Chicoutimi - Le Fjord. In close races like those in Quebec City, the Bloc's existing organization will be key, as will that of the Conservative Party. Polls published in the media like these may certainly demoralize the troops - but it could also galvanize them to get those votes out.

Note that Le Devoir will be publishing my seat projections tomorrow in a slightly different format. As I intend to post updates here on the site on Saturday and Sunday, with Sunday being my final projection, the newspaper will be publishing seat ranges for each of the parties in order not to have two sets of different numbers out in the ether. The Globe and Mail will be doing the same on Sunday, as those projections will be supplied to the newspaper tomorrow and so would have been different from my own projections here on ThreeHundredEight on Sunday. The ranges will give you a good idea of what the possible likely outcomes of Monday night could be, rather than the precise, riding-by-riding breakdown I will have for you on Sunday.


  1. Project Democracy used OraclePoll to do 400 respondent polls in 8 ridings - 4 in PQ, 2 in SK, 1 in BC and NU

  2. Eric, how thoroughly do you plan on discussing the performance of your model and that of the other pollsters in the coming weeks (post-election)?


  3. Ipsos reid poll is out

    CPC drops from 43 to 38, NDP gains to 33 and LPC down to 18

  4. I look forward to that discussion Eric. The poll results today in Ontario are just way too different among the three.

    Isn't stopping the partisans from having their way with things trying to swing momentum one way or another. As a few people have said, anybody that thinks they know exactly what is going to happen on Monday is fooling themselves.

    Hopefully there will be time for a fine nuanced discussion of what polling, and thus seat projections, really means during a fluid election.

  5. Is there any sense that in Ontario ridings where Liberals were in a close race with an incumbent Tory or challenging Tory, that voters are switching their votes to the NDP?
    I would find it hard to imagine that that’s the case. I would imagine that the first concern for those Liberal voters would be to defeat the local Tory, in which case the polls are indicating a complete collapse of Liberal support in ridings where Liberals are not competitive.

  6. How come the two week Harris Decima poll wasnt shown? It shows the surge in context of the bigger picture, and that the NDP and Bloc are tied at 28%. I personally think the QC numbers have been completely overblown, and that poll showed it.

  7. There is a more recent EKOS poll. I don't know what the Ontario numbers are, but there is a slight shift to the NDP nationally to 29.5%. All others about the same.

  8. Oh wow Harper's taken the lead again in the Nanos leadership index!

    Last minute surge in CPC support ?

    While everyone else appears to have dropped.

    I saw commenter Peter saying this was a big deal when Layton went ahead of Harper.

    What do you guys think ?

  9. Anonymous 18:22

    The Harris-Decima numbers are in there. If I had to guess, I would say Eric did not include the two week samples because they are badly outdated and therefore irrelevant beyond putting the surge into historical perspective.

    AverageCanuck, I think it would be important if there was more time before the polls. This election it has taken more than two days for changes in the leadership index to be reflected in the polls, and the CPC is almost out of time for that change to have a major effect.

  10. Of course the NDP numbers have begun to dip. They have had a free ride. The press and the other parties only started to push back a couple of days ago.Their numbers will go down slightly before eday in the polls, maybe 2% and then another 3% from lower turnout. Local candidate is the only issue that went up on Nanos today, because of the press on the NDP candidates.

    Many more people will definitely be voting NDP, don't get me wrong but not as many as the polls indicated from about 2 days ago.

    33% for the NDP is phenomenal on IPSOS today. However, if you look at Ipsos certain to vote and certain to keep your vote intention the same it's 67% of 63% that's 42% of their 33% share, which is 14% certain to vote on eday for the NDP.

    So don't be counting any NDP chickens until they've hatched. It might happen, but I'm thinking it's somewhere between the 14% and the 33%.

  11. I think if there is any significance to be attached to the huge leap it is that (1) the previous drop was probably a mirage or (2) that the leap is matched with drops in EVERY other leader's numbers, split evenly between them. If you don't like PMSH, then you don't like either choice. The first means the Tories are as strong as ever, and the other means that the voters are recoiling from another minority and lining up behind the Conservatives.

    Let's give it another week and see if there is a trend.


  12. Anonymous Prime29 April, 2011 19:09

    "What do you guys think ?"

    I think you should probably re-read that most recent Nanos results.

  13. Even the most CPC friendly poll has the Tories out of majority status. Ipsos has the Conservatives 38% nationwide. They are way ahead in Ontario 40-34 over the NDP no less but losses in Quebec and the West will likely bring them to around 145 and that is assuming a net pick up of 10-12 seats in Ontario with likely a small loss to the NDP. Cant seem to find Prairie/Atlantic numbers but the Conservatives would lose 4-5 seats in Quebec as well as a couple in BC most likely. Remains to be seen but likely theyd lose a couple prairie ridings as well and maybe even some Nova Scotia seats.

  14. @AverageCanuck

    It was a big deal when Layton passed Harper on Nanos' leadership index because Harper had been so dominant on it during all the campaign that it prove that something had changed in a major way. When Harper passed Layton again today, it wasn't really a big deal because it was only a slight return to the usual, even though Layton is still way higher than he was at the start of the campaign.

    Considering that the Conservatives still lead the NDP in popular vote, I'd say that it's normal that Harper gets higher leadership approval than Layton considering how personal the campaign has become. I'd say that Layton being above seems to have been a fluke, his popularity has increased, no doubt about it, but above Harper? I doubt that, considering how the Conservatives have built their campaigns around him (go see the Conservative website, he's in the center of basically all the images on the site) and that they're still in the lead.

  15. Thanks Bernard for the first riding poll I've seen for Esquimalt Juan de Fuca. I had told Eric that - model aside - the contest is between Cons De Souza and NDP Garrison, and the poll confirms that. The model shows the seat staying Liberal, but the poll shows the Liberal candidate far back in 3rd place (15%).

    Patrick from E-JDF riding.

    I'm very disheartened to see De Sleeza ahead in the polls (he has basically said in the past three elections that if we want an overpass at McKenzie avenue we have to elect him).

    Hopefully those who have supported the progressive platform of Liberal Dr. Keith Martin in past elections will see this poll and switch their support to De Souza's only real contender, Garrison.

  16. Harper leads the leadership index because almost all of his supporters automatically think he is the best and there is always a core of 25-30% never changing conservative supporters who love any leader no matter what. When a previously 4th place party leader takes the lead, it is clear that he is supported not only by his own core support but by significant percentages of Canadians of all political stripes.

  17. 38.46% Chretien 1997, 155 seats. Majority.

    Eric's average for the CPC yesterday was 37%.

    It's possible, and every pollster has admitted to the built in increase of about 2% at the polls for the CPC, not reflected in their poll.

  18. I would imagine the leadership numbers are related to the popular vote numbers. Tomorrow morning the NDP is bound to drop in the Nanos poll. This will be spun any number of ways. I am sure that many will attribute it to the attacks working. In actuality it will have more to do with the Tuesday numbers being replaced. Those numbers were clearly an outlier, the NDP jumped more than 4 points in a single day. That means they were polling 13 points higher on Tuesday than the average of Thursday-Saturday-Sunday (Friday and Monday were skipped for Easter). They almost assuredly polled first on Tuesday with something like 36%. That was obviously a fluke outlier that bent the results. It's departure tomorrow should not be taken as proof of anything except the return of more accurate numbers.

  19. anon 20:08. Chretien won that 155 in 301 seats. We now have 308. I don't know how that will affect the equation

  20. Eric everyone is saying this stuff from Jack's past is going to hurt him, especially with women voters.

    Can I get your perspective ? Will the french language media even report it ?

    My understanding is that cultural attitudes are much more open in Quebec.

    There's where the big NDP support is.

    So i'm not so sure this is such a game changer after all.

  21. "38.46% Chretien 1997, 155 seats. Majority."

    There were four competitive parties during that election (PC, Reform, Libs, NDP). Today there are only three and the traditional spoiler (the BQ) is in very poor shape. A higher percentage of the national vote is likely required to secure a majority this time around.

  22. 38.46% goes further in a five-party setup than it does in a four-party setup. When they swept Ontario, the Liberals had about 50% and the PCP and Reform each had about 20%. No polling shows this as a possibility.

  23. Did anyone notice Gilles Duceppes is in grave danger?

  24. AR confirms late Tory surge!

    Ontario lead is HUGE with liberals/NDP tied.

    Vote splitting = perfect.

  25. The CPC are better at targeting ridings than the Liberals were in 1997.

    And um we have a 4 party split of the vote and seats, very much like 1997. Reform was concentrated in the west as the Bloc is concentrated in QC.

    We'll see. it might not happen, but I maintain that it still could. Latest poll of Esquilmault Juan de Fuca taken over the last two days shows Cons ahead by 6, when this model anyways doesn't.

    "This poll was conducted for Project Democracy, a group seeking to defeat the Conservatives. The polled 400 people in Esquimalt Juan de Fuca on April 27-28. That is a margin or error of +-4.9% 95%.

    * Troy DeSouza - Conservative 40.4%
    * Randall Garrison - NDP 34.6%
    * Lillian Szpak - Liberal 15.4%
    * Shauna Salsman - Green 9.6%

  26. Conservatives up 2 points in Angus Reid to 37%, the highest of the election for that firm was 38% on April 4-5, with this caveat added.

    “I’m not exactly sure what it means for the seats,” Mukerji said. “At this point, ground game comes into it, too, in terms of your ability to motivate people to get out to vote.

    “The Conservative base really has maintained a sense of cohesion. They’re a well-organized political machine.”

  27. If Sun TV is going to engage in a last minute smear, they should do so less transparently. I also believe the NDP promises are 70% fantasy, yet if I lived in an NDP-friendly riding I'd definitely be supporting them.

  28. A-R has 37/33/19 now.

    While that closes the gap a little more, I wonder if its possible for the NDP/LPC to be combined larger than the CPC or have likely needed majority between them with only a little more than half the vote.

    Oddly, the TorStar article also indicated that 57% of 2008 Liberal voters weren't voting for the Liberals... which would seem completely impossible given the headline poll numbers. Why do sub questions for polls always give such screwy results?

  29. God help Canada! Get out and vote able bodied individuals. We are gonna go from first to worst in the G8 if this NDP horror continues. It's a sin that 25% of voters fail to do research on where the truth and lies begin. The media should be ashamed for their spin which they are only now beginning to correct.

  30. I really do wish the hyper-partisan commentary would cease on this blog. One of the refreshing things about reading threehundredeight and the comments thread is that it's all about analysis. I've read comments from supporters from all parties and most keep their politicking at the door. It's a nice change from the nonsense you read elsewhere. Sure, everyone makes predictions and when they do, their party affiliation tend to show through but it's usually based on sound reasoning. Comments such as "It's a sin that 25% of voters fail to do research on where the truth and lies begin" are useless and take away from some engaging discussions.

    I'm not suggesting that Éric remove such comments, I just wish we could all discourage it.

  31. just finished listening to CBC's The House. Towards the end there are a couple of really experienced reporters dissecting the campaigns and prognosticating the results.

    If you can get the show definitely worth the time.

    As to results I'm going to have to agree with Tom Clark. I've got NO idea and I don't think anybody else does either.

  32. Kevin the 57% of previous Liberal voters voting Liberal this time could be correct, with the top line number of 18%, because people like Andrew Coyne, who have never voted Liberal say they are going to this time, for some bizarre reason.

  33. " I'm not suggesting that Éric remove such comments, I just wish we could all discourage it."

    Agree 100%. This is a very informative site and those type of comments only make for an annoying read. I might very well agree with those comments; however, I don't think here is the place for them.

    Now, back to the polls. If the latest NDP numbers are to be believed, they could very well have over 110 seats. However, there are several big question-marks:
    - advance vote (before the NDP surge in polls)
    - NDP supporters actually going and vote
    - NDP bias in polls (think 2004)
    - latest attacks on NDP and/or Jack Layton (i.e massage parlor story)

    It seems to make a huge difference in seats projections if NDP is at 28% or 32%.

    An interesting election for sure.

  34. I have no idea what the Kindergarden Kops are talking about, but I haven't seen any hyper-partisan comments on this site. They sound like Audrey McLaughlin. How's that for hyper-partisan?

    Anyways. Chances are that the polls are now underestimating both the Conservatives and the Liberals, unless the right liberal vote goes to the Tories to stop the NDP, which it may have in some of the latest polls.

    There is just no way that this "intention to vote" (which is all a poll measures) holds up for the NDP on eday, with increased scrutiny, second thoughts, less committed voters and no ground game in much of the country.

    A lot of the NDP vote will be there, but not 33%, more like 27% if the wave has a lot of genuine sentiment behind it, not just a fad, more likely 25% by Monday. It's already going down in some polls.

    i.e. There are vote mob sites on Facebook where 400 students sign up to participate, and in reality, 40 show up to the vote mob. We'll see how much of the NDP vote materializes.

  35. Full details from Leger here: CPC 36%, NDP 31%

  36. What is the deadline for polling. Is it 11:59 ET tomorrow or is it earlier?

  37. So basically ALL the vote projection sites and pollsters have the Conservatives at 146 seats, with current poll numbers.

    However, all of them, Nanos, Ekos, Democratic Space, Eric etc, say a Conservative majority is also still possible.

    If one assumes that polling is under representing Tory votes, and I do, by about 2%, that's a majority.

  38. Eric

    Using the information at this site I was able to come within 3 seats of predicting the number of CPC elected.

    I was way off in Quebec and have officially given up on trying to read the intention of the Quebec voter. No reason for them to elect so many unqualified NDP based on Mr. Layton's personality and the NDP policy.

    Basically there is no one who can truly say they speak for Quebec. They might not have any people that really believe in separation... they just like the feel that Bouchard and Parizeau and Duceppe put out.

    Here are my prediction and my analysis of Ontario

    BC Voice of Reason said...
    My predictions (locked them in 27 Apr)

    CPC 164 NDP 55 Lib 49 Bloc 39 Ind 1

    The core separatists are motivated to vote.... the NDP supporters will be finding out on Tuesday or Wednesday or for sure by Friday that they were supposed to go to a polling station on Monday.

    The NDP support will hold strong enough to undercut the Liberal votes.

    30 APRIL, 2011 18:31

    and Ontario went as I said:

    BC Voice of Reason said...

    The CPC were a close second to the Liberals < 15% in 19 ridings in Ontario.

    If the CPC holds at 39% in Ontario and the Liberals drop from 33 to > 25 the CPC will take most of these ridings

    The NDP were within 15% in only 4 Ontario ridings... 3 won by the liberals and one CPC.

    The leftist voter in Ontario has a dilemma--- Continue voting for the Liberals ... which has turned into a pure protest vote or jump aboard the Jack train.

    The Liberals need to have the CPC majority so they have time to rebuild and not get totally wiped out. If the NDP have 20 more seats than the Liberals and the Liberals support a NDP minority government the next election will be the last the liberal party fields candidates.

    Their other option is to support Harper's Minority .... as they have for the last 2 years.... That is another awful choice.

    The best the Liberal party can hope for is to have a strong 3rd place showing and a CPC majority.

    Left the NDP Quebecois implode.

    30 APRIL, 2011 19:29


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.