Thursday, April 28, 2011

Conservatives struggling in Quebec

The three national polls added to the model this morning (EKOS, NanosForum) are all telling the same story. And one of those things is that the Conservatives are struggling in Quebec. A few riding polls indicate that this weakness could even put some of their fortress seats at the mercy of the - yep - NDP.
The topline numbers of these three polls are almost identical: the Conservatives around 35%, the NDP around 29%, and the Liberals around 22%. The New Democrats are strongly in second in the two westernmost provinces, and the situation in Ontario is looking dangerous for the Liberals. In Atlantic Canada we have a three-way race, while in Quebec the New Democrats are enjoying a 10-17 point lead over the Bloc Québécois.

But look at the Conservative numbers in Quebec: between 14% and 16%. That is well below their 2008 performance and means the loss of about a quarter of their support. Perhaps they are bleeding that support away to other parties in regions in which they are not a factor. Maybe - but two polls conducted for the Journal de Lévis just when the NDP were beginning to take off indicate that even in the Quebec City region the Tories are suffering.

Let me re-iterate that point. These polls were conducted between April 12-16. The NDP really began to take off starting on April 17. Nevertheless, in the polls conducted for Lévis - Bellechasse and Lotbinière - Chutes-de-la-Chaudière the NDP is polling at 23%.

In Lévis - Bellechasse that represents a growth of 12 points since the 2008 election, while in Lotbinière - Chutes-de-la-Chaudière that is a 10-point gain. Worse for the Tories is that Steven Blaney in the former and Jacques Gourde in the latter have seen their vote drop by eight (Blaney) and 12 (Gourde) points, and this before the Conservatives dropped to the mid-teens in the provincial polls.

In neither of these ridings, however, has the Bloc appeared to lose support. That may not be the case almost two weeks later, but it is remarkable that these two polls were showing New Democratic gains at the hands of the Conservatives before the party took off in the province, and not the Bloc. It is just speculation, but with the large margins of error in these Axiome polls and with the NDP gains in the province as a whole, for all we know the NDP could be leading in these two very Conservative ridings by now.
They are demonstrably leading in Hull-Aylmer, the riding just across the river from Parliament Hill in Gatineau. The riding has famously been Liberal for about a century, but the latest poll from Segma Recherche (conducted April 20-23, so recently) shows the New Democrats opening up a 13-point lead over Liberal incumbent Marcel Proulx.

At 29%, that is a drop of eight points for the Liberals. But at 42%, the NDP has gained 22 points since the 2008 election. The Bloc is down to only 13%, meaning the NDP has taken about the same amount of support from the Liberals as from the Bloc. The party really does cross the divide. The Conservatives are also down. But with this poll Hull-Aylmer looks like a very safe bet for the NDP, and adds fuel to the fire for a potential NDP victory in neighbouring Pontiac.

That brings us to two polls done in the Gaspésie by Segma Recherche for the Journal Graffici. Note that these polls are old, taken between April 6-13. At that time the NDP was roughly at 18% support in Quebec, so still well below their current levels but nevertheless doing respectably.

Nevertheless, the NDP was still at only 5% to 11% support in these two ridings. Even if we double this support, as the NDP has apparently done in the province as a whole since these polls were taken, we still have the NDP out of the race. This is an indication that the NDP will not be able to win just any riding in Quebec.

But what is more surprising is that in the case of Haute-Gaspésie - La Mitis - Matane - Matapédia the Liberals were only polling at 24%. They've sunk in the province since then, so they are unlikely to be doing better now. That's quite a drop from 36% in 2008. The Conservatives are still at about the same level they were in 2008 in this riding, but that has to be a disappointment as the Tories thought they had a good shot with a decent candidate.

Though the Bloc was only at about 36% or so in the provincial polls when this riding poll was taken, the Bloc had grown its support by about 11 points in the riding since the last election. This might not be a surprise, as the previous Bloc MP was a bit of a no-show. Jean-François Fortin, a local mayor, was leading the riding in this poll with 49%.

In Gaspésie - Îles-de-la-Madeleine, the Liberal and Conservative vote had dropped by quite a bit to the benefit of the Bloc's new candidate, Daniel Côté. At 48%, it is unlikely that the recent drop in support the Bloc has seen in the province has put this particular candidate in danger. The NDP was only at 11% in this riding when the poll was taken.

Taken all together, these five riding polls tell us something about what is going on in Quebec. Firstly, the Liberals are completely out of it, even in ridings where they were supposed to be competitive.

The Bloc is looking in rough shape and can lose a lot of its support in NDP-winnable ridings, but they do still have a base of support to fall back upon (the two Quebec City-area ridings and those in the Gaspésie demonstrate that). For the NDP, they have the potential to be competitive in surprise locations (Lévis, Lotbinière) and win in others (Hull-Aylmer), but still have a long way to go in other parts of the province. It is unlikely the Bloc will be reduced to single-digits in seats.

And for the Conservatives, even their fortresses are at risk. The seats held by Blaney and Gourde are among the safest for the Tories in Quebec, yet they look threatened by the NDP's surge. And the polls in Hull-Aylmer and the Gaspésie show that the Conservatives can disappear in ridings in which they never had more than an outside shot at winning.

What this means is that the New Democrats are in a terrific position to win a great number of seats in Quebec on Monday. All of the other parties are looking weak, even in their strongholds. Now the NDP just has to hope (because that's all they can do with the means they have in the province) that Quebecers go out and vote.


  1. At what point do the numbers out west cause the Conservatives to start losing significant numbers of seats to the NDP in 2-way races between them?

    In your last update before Monday, will you be going with only the last polling number, or extrapolating any trends that you see to continue through the weekend?

  2. On Sunday I'll have the final projection and a projection based solely on the last set of polls. I won't be extrapolating anything into the future.

  3. How do you get to your 4% figure for the daily poll average chart for the Greens when all three polls are higher? Equal weighting gives mean 6%.

  4. New Harris Decima numbers:
    CPC 35%
    NDP 30%
    LPC 22%
    BQ 6%
    GPC 7%

  5. Oh now I see that the only poll for Apr 27 is Nanos.

  6. New Harris Decima with precisely the same results as Angus Reid from the other day, for all three major parties.

    However HD's poll was taken over a full week, for what it's worth.

  7. At what point does the NDP support in Ontario cause several ridings to split the left of centre vote and elect Conservatives in ridings where it usually goes Liberal.
    Seems to me Ontario is the key to a Harper majority, would you agree?


    Le NPD 5 point derrière les Conservateurs, et 20 points en avance sur le Bloc au Québec. Imaginez.

  9. Eric,

    Not sure if you got my tweet, but here is Trois-Rivieres

  10. New Harris-Decima -

    CON 35
    NDP 30
    LIB 22
    GRN 7
    BQ 5

    In QC - NDP 42, BQ 22 (!), CON 18, LIB 15
    In ON - LIB 34 (!), CON 33, NDP 25, GRN 6

  11. "Anonymous said...
    At what point does the NDP support in Ontario cause several ridings to split the left of centre vote and elect Conservatives in ridings where it usually goes Liberal. Seems to me Ontario is the key to a Harper majority, would you agree?"

    I have the same question. It seems to me that close ridings like Guelph and Kingston, that already have substantial NDP support, are at risk of turning CON due to the LIB/NDP split.

  12. Huge discrepancy between Harris Decima and Nanos in Ontario numbers. HD has the Liberals now LEADING in Ontario. If that holds true, the Conservatives would essentially be trounced through the entire country.

  13. Liberals at 34 in Ontario?

    But with national numbers at 22, that would put them at the 12% range in the whole rest of the country. Sure, that's a good Ontario number for them, but everywhere else is just a disaster.

  14. That Harris Decima Poll is the worst result for the Conservatives yet. In addition to the bad Ontario numbers (+-0 LIB +7 NDP -6 CPC compared to 08) Quebec is obviously hugely NDP with the CPC down 4 from 08. Not to mention BC is basically CPC -5 NDP +5 from last election.

  15. Wow, I'm getting nervous because Harper looks really nervous and is threatening everyone with doom (sounding like desperate Iggy). I worry that its a sign its over for him too. Plus it looks like he's visiting Ministers ridings and his Quebec seats in an attempt to stop the bleeding. It seems to be getting worse for him.

  16. Gotta figure this is the end of the CPC, eh Eric ?

  17. Poll in Trois Rivieres has the NDP's Robert Aubin running 14 points ahead of the BQ incumbent!

  18. hey
    great site, lots of funn for political and statistical junkies. :-)
    2 minor suggestions for the seat by seat predictions table which I realize is already a very busy picture.. 1) would be useful if in the Held column there was an additional notation such as a star within the box to denote that there is a predicted change from incumbant party- we have to eye ball those now. 2) on a similar vein.. if the seat is arguably too close too call (within 2 points or something like that) .. some sort of notation in the Held column (a question mark perhaps?) or other way of seeing those more readily...
    Cheers... GR.

  19. It will be very interesting to see if there's a vote-splitting 'sweet spot' for the Tories. In a riding like Esquimalt, if you start with Eric's current numbers and suppose that the NDP steals 2.9% from the Liberals, then this becomes a CPC pickup. But if the NDP picks up as much as 4.4% from the Liberals, it becomes an NDP riding.

    Similarly, this might happen on a regional level. Vote splitting could allow the CPC to pick up Vancouver South from the Liberals, but the same trends would likely cause them to lose Surrey North to the NDP.

    Based on the attack ads that the Conservatives have unleashed in the last few days, I'd guess that they're more concerned with losing seats to the NDP than potentially benefiting from vote-splitting.

  20. "the Conservatives would essentially be trounced through the entire country. "

    And this would be a bad thing ???? Get real, they are a severe menace !!

  21. Interesting.

    If HD thing showing Lib rising and Dippers falling in Ontario is more than just a blip, could this be a bit of a reflexive RoC response to the voters of Quebec seizing control of the National agenda, not to mention the proMedia, over the past week or so?


  22. @Ross K
    What polling have you been looking at? The NDP have mostly been in the mid-teens or lower in Ontario the whole campaign. From the last HD poll they are up 3 in ONT with the Liberals down 2, Conservatives down 3. Halfway through
    it looked like the NDP wouldve been lucky to hold much more than 10 seats in Ontario with some numbers around 10 percent. At 25% and with the Liberals and Conservatives less than 10 points up they can keep their 17 and probably add 3-5 more, mostly from Liberals in the GTA. There are also a couple Conservative seats, Essex, Kenora and Oshawa that would possibly fall under this scenario.

  23. Amateur Psephologist28 April, 2011 21:02

    RossK; note that the Harris-Decima poll took place over the last 7 days, thus including both the most recent days when the NDP was flying high and the beginning of the week, when other polls had them still just starting to pass Liberal levels.

    As others have noted, the Nanos numbers for the most recent day must have had Layton and Harper tied.

    Will be interesting

  24. RossK, how to you see Liberals rising and NDP falling in Harris-Decima poll for Ontario? I see the numbers virtually unchanged for current week versus 2-weeks, and significant increase for NDP compared to earlier. Liberals are ahead of Conservatives because the Cons are falling faster than Libs as NDP rises (ruining the vote-splitting theory).

  25. Five pollsters in one week converge around Tory 35, NDP 30, Liberal 22. Remarkably consistent. On the other hand, the fluid situation and inconsistent polling of Ontario really needs sorting out.

  26. At this point I would take the HD poll in Ontario with a grain of salt as there is no confirmation quite yet. I think Eric has rubbed off on me quite a bit and I like to see consistency first. I looked at the five poll average in Ontario (keep in mind MOE is relatively high) and the Tories scored as follows:

    HD: 33%
    Ekos: 40%
    Nanos: 41%
    Forum: 38%
    Angus Reid: 37%
    Average: 37.8

    As all five polls were done April 25-27 HD is far off the mark and outside most other polls MOE. So at this time it is far too early to equate this to the demise of the Conservatives.

    However, when looking at the national five poll average (35.2%), it seems that a confirmation has now been set that a majority is not within reach at this time.

    In my opinion, it would require either 38%+ nationally or 40%+ in Ontario to obtain a majority. At 35% and 38% respectively, even with vote-splitting between NDP/Lib it is probably not feasible for the Conservatives to obtain their majority.

    In saying all this, it's been such a surprising election and we still have 5 days to go, so who knows!!!

    By the way, this is my first post despite being on this site from day 1 so to Eric I cannot express my gratitude for all the work and effort you have done for this site and the fact thatyou are sticking to your guns by not deviating from your formula.

    While I know you want to be as accurate as possible I think this that whether you get it bang on or way off is almost beside the point because you created a blog that allows daily vieweres like me to get first hand information plus the whole logic of how polls and predictions work.

    I also want to point out that in the last couple days I have noticed the comments have become far more partisan. Why here of all places? For those who simply are partisan, put your posts on the various partisan blogs and editorials out there. This site should be informative, not an extension of the marketing department of the party of your choice.

    So no more "if Conservatives get a majority", "if Liberals create a coalition", "if NDP socialists become the PM" than , if elected will eat your children.


  27. There have been 4 defining moments so far: the exchange between Layton and Ignatieff during both debates, Iggy's "Rise up" sketch, Duceppe's speech at PQ convention and Layton's "future Prime Minister" speech in Winnipeg yesterday. The first 2 moments sinked Ignatieff, the 3rd sinked Duceppe and the last one scares the heck out of English Canada. Maybe there is another one around the corner however the timing doesn't allow really big moments (one never knows though).

  28. Hi Eric,

    I am a long time reader and first time commenter. Thank you for the good work you do.

    I just wanted to comment because I wanted your opinion of this new Harris Decima poll.

    As far as I can tell, the Liberal numbers do not mathematically make sense. HD has the Liberals at 22% nationally but 34% in Ontario. Given that Ontario is about 35-40% of the population, the Liberal vote combined in the rest of the country should be somewhere between 14 and 15%. However, HD has the Liberals at 14% or higher in every region of the country, including at 32% in Atlantic Canada and 18% in BC. I do not see how these Liberal numbers could possibly make mathematical sense.

    Keep up the good work!


  29. @RossK: I will give the benefit of the doubt and assume you're proposing something that you don't actually believe yourself. That's because there seem to be a few suggestions to your post: (1) that Canadians outside Quebec will vote-in a bloc--purely to spite Quebecers, (2) that Canadians outside Quebec resent Quebecers voting strongly for the NDP and somehow "seizing the National agenda" in the process, (3) that the media (what is "proMedia"?) is wrong to report heavily on what could well be one of the largest and most unexpected federal voting shifts in a generation.

    This is all utter drivel, and fairly offensive to boot. It projects a rather bizarre vision of Canadian voting behaviour that may reflect the worldview of, say, Jacques Parizeau, but which bears no relation to the thinking of ordinary people.

    And this is coming from a voter in the West.


  30. Jeebus, folks.

    I was only referring to the HD poll discussed by a number of commenters immediately above my previous comment.

    'Twas just a thought (and please note what I mentioned about the possibility of it being a 'blip')

    I am not a provocateur with an axe to grind.


  31. A bunch of riding polls just came out for Québec City and Saguenay Lac St Jean. Here they are :

    Riding ( Incumbent)

    Louis St Laurent (Josée Verner, CPC) : CPC 37% NDP 37% BQ 17% Lib 8%
    Charlesbourg (Daniel Petit, CPC) : NDP 35% CPC 31% BQ 22%
    Beauport (Sylvie Boucher, CPC) : NDP 36% BQ 30% CPC 27%
    Louis Hébert (Pascal-Pierre Paillé, BQ) : NDP 32% BQ 30%
    Québec (Christian Gagnon, BQ): BQ 34% NDP 34%
    Portneuf (André Arthur, IND) : NDP 31% IND 29% BQ 28%
    Jonquière (Jean-Pierre Blackburn, CPC) : CPC 34% NDP 32% BQ 25%
    Roberval (Denis Lebel, CPC) : CPC 47% NDP 27%
    Chicoutimi (Robert Bouchard, BQ) : BQ 43% NDP 27% CPC 18%

  32. Couldn't agree more with Non-Partisan:

    Whether the model is close or not, I've learned much about this process.
    Eric has done a tremendous amount of work and his intellectual rigor is beyond reproach.
    Keep the partisan bickering somewhere else. Question the model, question the poll results, even bet on who will form the next government, but take the political debate elsewhere.

    Fiercely Partisan

  33. 2 million people voted in the advance polls.

    Immediately after that the Tories lost 6% and the Liberals lost 5%.

    Could it be that an older, more engaged base turned out to vote and then stopped answering the "Who would you vote for today" phone calls?

    The advance poll I was at was awash in senior citizens. No #votemob in site.

    The NDP are all hype and there will be a lot of head scratching come Monday night.


  34. New Quebec riding polls, must read

  35. Great work Eric, just curious as to how the bloc is at 31% of the popular vote in QC when I haven't seen any poll with them that high in the province as of late.

    Cheers- Justin

  36. Justin - I'll do Eric's job for him for a moment - I'm sure he's busy preparing today's post.

    The model that Eric uses is a weighted combination of many many polls, previous results, and local factors. As time goes by, the older polls are weighted less and less, but are still there. This is to avoid any one poll (remembering of course, that a Margin of Error means that one out of every 20 polls, on average, is actually outside that margin!) or daily "burp" (The Royal wedding, say, or a big hockey game skewing the demographics of people who answered the telephone polls) affecting things too much.

    Thus, when a trend continues, and thus more and more polls are added as "weighty" (because they're recent), then Eric's model responds. It may feel like a glacial pace to we who hit "refresh" every 5 minutes, but it seems to me that it's a very responsible (and likely accurate) system.

    Hope that helped!

    - Ashley

  37. Kenneth (Ontario Dipper)
    New Nanos numbers have the NDP making VERY significant gains in BC and Ontario, while they may have peaked in Quebec and Atlantic Canada. These Ont and BC gains will seriously eat into CPC hopes for new seats, and endanger many that they hold now.

  38. Majority
    out of reach, Tories say

    NIAGARA FALLS—Stephen Harper's Conservatives must win 23 more
    seats in Ontario to achieve their coveted majority, a task that senior
    party insiders now admit is almost impossible, the Star
    has learned.
    High-ranking sources confide that even with the collapse of
    Michael Ignatieff's Liberals — and NDP Leader Jack Layton's surge,
    which helps split the vote in many Ontario ridings — it will be very
    difficult to make such immense gains in Canada's most populous province.
    At the dissolution of Parliament, the minority Tories held 51
    of Ontario's 106 federal seats.
    Party sources say the possible loss of several British
    Columbia ridings to the New Democrats — and others in Quebec, where
    Layton is surfing an orange wave — has forced them to revise their
    As of Thursday, they said they needed to win at least 74 seats
    in Ontario to achieve a majority.
    “It all comes down to Ontario and we're just not there,” a
    source said, speaking on condition of anonymity because the party's


    He set the bar himself.

    So, plain and simple, if Stephen Harper does not win a
    majority Monday, he has lost.

    He will have to deal with the fact that after five years in
    power, and after four elections as leader, Canadians do not trust him
    with a majority.
    It would seem inconceivable the Conservative leader would get
    a fifth kick at the can.

    Harper made this a referendum on his leadership by moving the
    call for a majority from its one-time status as the radioactive third
    rail of his campaign to his daily mantra.


  40. Ashley, your post is accurate except for one thing. I'm not criticizing just educating. The margin of error takes into account sampling error, but still assumes that the population being sampled is representative of voters. So your examples of hockey/wedding changing who is likely to answer the phone are good examples of what will throw off the polling results, but those influences aren't captured by the margin of error. So while the poll will be within the margin of error 95% of the time, it's still not certain that what the poll is measuring reflects the electorate. That's also why we see such different numbers across polls, beyond what we would expect from simple statistical variance. If we are lucky, the bias in different polls will not be correlated and by looking at the average across polls as Eric does, some of the bias will cancel out. On the other hand the sampling error is guaranteed to go down with the larger samples you get by coalescing poll data.

  41. More offbeat sites like 308 and election prediction seem more conservative this time around and less believing in the NDP wave producing results. This election is hard to predict and that's why. Polls can't measure feet on the ground heading for the ballot. If you look back at 1988 Mulroney did not perform in Atlantic Canada but once he swept Quebec it was over and he had his majority. I say watch the Quebec numbers and if the NDP surge is there it's over for Harper. His minority will be reduced.


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.