Wednesday, April 13, 2011

Seven-point gap in Ontario

I was waiting on a new poll from EKOS to be reported on, as well as hoping to get my hands on the details of the newest (and oddest) COMPAS poll, but in the end I only have today's Nanos poll for CTV and The Globe and Mail to add to the projection. Accordingly, there is little change to report.
But I suspect this will not be the case for much longer. While last night's debate was interesting, I'm not sure it will move the numbers in any significant way. However, it can take time for this sort of thing to register, and often the effect of the debate is only known well after the leaders have finished arguing.

In terms of poll results, we shouldn't begin to see any of the effects until tomorrow at the earliest, and by the weekend we might still only be seeing a piece of the puzzle. The French debate tonight will, of course, have a large influence on the polls in Quebec but if anything especially remarkable happens it could resonate in English Canada as well.

My initial impression, however, is that all of the leaders generally shored up their own support but did little to capture votes from the other parties. 

There are no changes in the seat projection today. The Conservatives are still at 152, the Liberals are at 73, the Bloc Québécois is at 50, and the New Democrats are at 33. The Tories are back up to 39% while the Liberals and NDP are unchanged at 28.1% and 16.7%, respectively.

There has been little change at the regional level as well. Today's Nanos poll replaced yesterday's in the model, so there were only tiny variations.

The most important are those that took place in Ontario and Quebec. The Conservatives are unchanged at 42% in Ontario, but the Liberals are now at 35%. It is a gain of only 0.1 points, but it nevertheless narrows the gap between them and the Tories. The seats I highlighted yesterday as liable to fall back into the Liberal fold are all the more likely to go Grit now.

In Quebec, the Liberals are down and the Bloc is up - the first time the Bloc has seen their numbers improve in the projection in a few days.

Tonight's debate is very important for Gilles Duceppe, as he has to galvanize what has been, so far, a relatively uninspired campaign. Expect him to go hard at all three of the federalist leaders, as he is being pressed by all of them: the Conservatives in Quebec City, the Liberals around Montreal, and the New Democrats everywhere.

The Bloc isn't likely to lose any seats at the hands of the NDP (except maybe Gatineau), but the real threat to the Bloc that the NDP poses is as a spoiler. Many ridings in Quebec were close BQ/Liberal or BQ/Conservative races, and if the NDP siphons off enough support from the Bloc it could benefit the two other parties.

Undoubtedly, the NDP is also taking some votes from the Liberals and perhaps even the Conservatives. But the Bloc and NDP are both social democratic parties, and many polls show francophones and Bloc supporters have a generally high regard for the NDP and their leader.

As Quebecers already have a pretty clear picture of how the Conservatives and Liberals see their province, expect Duceppe to go after Jack Layton on issues that revolve around Quebec. We saw that a little last night when he talked about Bill 101. The NDP is more of a centralizing party than their two federalist counterparts, and that is a particularly problematic position for many former Bloc supporters now considering to vote for the NDP.

It will be up to Layton to ward off these attacks. I think he might be able to do it, as unlike the other two anglophone leaders Layton speaks decent working-class French (he did grow up in Montreal, after all). The more educated French of Michael Ignatieff and the heavily accented (and at times difficult to understand) French of Stephen Harper will not likely resonate with francophones once they are pressed by Duceppe and knocked off message. 

In any event, the French debate will be an interesting one. It's unfortunate we only get one in each language.

My apologies for the late update today, my alarm clock is to blame.


  1. I think it is interesting to note that the above poll shows the Conservatives ahead in EVERY area of the country today except Quebec, and are ahead of the Liberals in Quebec. I only found this page several days ago, but this is a first in that time. It will be interesting to see what effect the debates have on the polls. I expect that The Bloc will get a bump in Quebec after tonight's French Debate.

  2. That COMPAS poll has some crazy numbers... Drastically higher Conservative support and drastically lower Liberal support west of Quebec. The numbers are just too far off to trust it on it's own. It will only be believable if other polls start showing similar trends.

  3. Hmm. Elizabeth May's numbers in Saanich—Gulf Islands are still rising, it seems, despite the lowering in BC overall. I'm don't think the Green leader's plane has enough runway left, so to speak, but it's getting.

    Portneuf—Jacques-Cartier is against André Arthur today, though Justin Trudeau seems to be shoring up support in Papineau.

    *sigh* My riding is still boring. Yvon Godin is going to win in a walk in Acadie—Bathurst. Though the Green candidate isn't on the official list. Wonder what's up?

  4. In a Leger poll on the debate 1,037

    Harper 37 Ignatieff 21

    "In fact, among those who said the Liberals were their first choice before the debate, just 54% said they thought Ignatieff won the debate. One in five Liberal supporters said Layton was the winner."

  5. A 21 point gap isn't crazy. We've seen a poll with a 19 point gap this election, just after the writ was dropped. It's on the far end of Conservative support for sure, but it isn't crazy and had a substantial number of those polled.

  6. Éric,

    Your alarm clock is to blame for your late update? Are you sure you aren't hung over from playing the "Let me be clear" drinking game whilst watching the debate?

    - PJ

  7. Here is the COMPAS poll:

  8. Thank you Eric for best site on the internet.

    Only poll that matters is on May 2. I think we will see another Conservative minority government, no majority for Harper.

    I know it is not very likely, but I would really like to see this happening:

    Harper is defeated in the house (on budget or whatever). Coalition talk starts again with usual rhetoric about socialists, separatists, etc .
    Liberals go public and say: We will support conservative government for X number of years if they:

    cancel jet purchase
    keep gun registry and corporate taxes unchanged

    Harper will be in very tight spot and I have feeling that knives will come out.

    It will definitely make news at 6 more interesting to watch

  9. Eric, I wonder if COMPAS asked their leadership questions (best judge, best military commander, best money manager) before they asked the federal vote intention question. That could actually cause the 5% swing we're seeing. Mind you, all the other trends in the poll regionally seem to stack up with industry concensus. Only Ontario is really off from the other polls.
    (...Just noticed this, 35% DK/NR response on vote intention. That's really high, but also dovetails with actual voter turnout levels in recent elections.)

  10. Just out of sheer curiosity I was wondering if you would do a seat projection on the COMPAS poll alone?

  11. I was very annoyed that I was still sober after the debate.. Coalition didn't get mentioned by Harper enough lol. Should have known he'd just sit there and not debate.

    Debate impact: minimal. A couple of sound bytes, but that's about it. All four of them seemed worse than the prior election debates. Everyone seemed like they were unsure of why they were there and had minimal interest in what the actual questions were. (I think Iggy and Layton each actually answered all 6 of them. Can't say the same for the other two)

    - Harper: Could have left a a button to play his sound byte. He didn't even need to be there, since he didn't even seem to answer any of the actual 6 questions.

    - Layton: He was trying to distinguish himself from both Cons/Libs, but even he seemed oddly boring.

    - Iggy: Well, not a drooling idiot, so that's a plus. Just not as good as he should have been.

    - Giles: Was a drooling idiot. Very strange given his last election performance in english was actually credible.

    Funniest moment of the debate goes Iggy though. Iggy vs. Giles - Iggy opens with praise for Quebec. Giles counters with a look of wth? Iggy kept Giles from saying anything bad about the Liberals (important for Quebec) and effectively made him useless there.


  12. Not surprising from COMPAS - Dr. Conrad Winn is as right wing as they come, and has been under the direct pay have the Harper Government in the past (at least according to here: I trust that poll about as much as I trust Fox News (or QMI) to be fair and balanced.

    BVoR - I agree that Ignatieff didn't get the knock out punch he needed, but according to this site's projections the Conservatives are polling ~45% outside of Quebec. 37% isn't a great result for Harper either.

  13. The COMPAS (is all caps right on that?) poll had 35% undecideds. That's a lot. I'm going to go out on crystal ball gazing limb here and say undecideds are more likely to be Liberal supporters.

    Also, we saw another IPSOS (all caps?) poll last week that state the Conservative support would be 44% if they only counted those sure to vote. This would also be in the 21% spread range.

    Eric do you know, but I don't believe Canadian pollsters screen for likely voters, like they do in the states? I believe, and I could be mistaken, that we poll support, not if someone is going to vote, which is part of why we see third party vote collapse at the polls.

    We take out undecideds, not unlikely voters.

    So, if you have high undecideds for whatever reason, and screen for likely voters, 21% as a gap is not that unreasonable. High for sure, but possible depending on who votes.

  14. Yes in the states, pollsters will ask leading questions before question number 14 or 21 is the money question, "Who are you going to vote for?" That doesn't happen in Canadian polling as far as I'm aware that question is always asked first.

    The only difference in polling companies is whether or not they give the party names and leaders aka prompt or not.

    However the COMPAS poll asked some hillarious follow up questions, this being the last one.

    "If a young friend of yours were serving with the Armed Forces in a battlefield in Afghanistan, which one of these leaders would you have preferred as unit commander in charge?

    Harper emerges as an exceptionally strong battlefield commander in the eyes of the public,"

    Harper won that question by a 42 point spread over Ignatieff, 56-14.

    This reinforces my notion that Canadians are ok with tyrants as PM, and the opposition constantly going on about his tight leash style of governing doesn't hurt him.

    We don't want to "let the flowers grow", which was Ignatieff's lamest line of the debate.

  15. @Anonymous (11:37)

    The COMPAS press release indicates the leadership questions were the last set asked.

    Ontario isn't the only one that's really off - MB/SK, Alberta and BC are all showing some significant differences from most other polls for the Conservatives, the Liberals or both. They've got drastically higher Conservative numbers for all 3 Prarie provinces (not that it actually matters much for the seat count in Alberta). In BC, the 22 point lead on the NDP and 26 point gap on the Liberals is way off of the ~14 point gap we've seen on the 2nd place party from other polls.

    The 35% undecided makes me wonder if it's either a higher bar for decided (cutting out more "soft" support) or a lower bar for intending to vote.

  16. Just as a comment towards the NDP's centralizing tendencies. It is true the NDP, being a social-democratic party, tends to support a more active and interventionist federal government, which clashes with the Bloc supporters' desires for more Québec autonomy. However, the NDP is also perhaps the party that is the most friendly towards Québec nationalism. This contradiction has resulted in the NDP adopting a position called "asymmetrical federalism" towards Québec.

    Meaning, the Federal would enact policies, but grant Québec the right to remove itself from these policies, getting the money that would otherwise have gone to it in that program so that they may enact their own program or use the money elsewhere. So the NDP can use this as a way to go after the soft nationalist vote in Québec.

    However, this poses a dilemma for Layton, because this policy risks not being too popular in the rest of Canada, especially in Western Canada, due to resentment against Québec. That's a main reason why the NDP hasn't made its policy of asymmetrical federalism known too well in their campaigns. It's there, but they prefer not to talk about it too much, at least not in the national campaign, the Québec candidates use it as a main selling point on the other hand.

    If the NDP were to win with a strong Québec delegation, I have a feeling that sparks would fly in the caucus on that policy. I don't think Ontarian and Western NDP candidates are keen on the whole thing. So the policy is there and Layton can use it, but he also has reasons not to expose it too much.

  17. Here is the EKOS poll:

    5% gap. Looks like somebody's COMPAS is pointing in the wrong direction.

  18. If you think that the 21 point spread is unrealistic go read the Toronto Star

    Photo caption: "Prime Minister Harper and three other guys"

    Dave Olive opens:

    "I soon had the feeling the PM was on his own, delivering a briefing to Canadians on present and future conditions. And the others were the proverbial dog that finally has caught up with the car it's been chasing and now doesn't know what to do with it. Having defeated the government, they were now unable to make a convincing case for having done so."

    This is the most Liberal friendly media. The Star is incorporated and mandated to support progressive points of view.

    I wonder when the CBC will go into damage control.

    A Harper majority getting a mandate to balance the budget earlier would be obligated to look at the 1.1B / year of tax dollars going to the CBC.

    The CBC argument for being the fair balanced purveyor of the news as a reason for continued support will be absurd.

    If I were a CBC executive I would:

    1) have the Liberal Biased voter COMPASS removed from the web site and issue an apology
    2) Drop Greg Weston while he is still on probation.
    3) Assign Terry Milewski to cover Libya
    4) Immediately get the freedom of information requests completed. How ever bad they are for the CBC hiding them will be worse.

  19. @Anonymous (12:00)

    Some pollsters definitely do screen for likely voters. Maybe all - not sure.

    I don't think any of them list the specific question wording or the cutoff for doing so though.

  20. BTW, I was trying to see how COMPAS had performed in the past, and I couldn't find any comparable polls. Is this their first poll of national voting intentions ever?

  21. Ryan,

    I think they have polled in the past, but they are not exactly very active on federal voting intentions. I have no poll from them taken within the last four weeks of any of the last three federal campaigns.

  22. According to,_2011

    That has supporting pdfs attached

    COMPAS did a few national polls during the coalition coup attempt Dec 3,2008

    They had the CPC ahead 51 to the Liberals 20

    Most other polls had the cpc ahead by 20-22 % points.

    To me this COMPAS poll translates to a 14 pt CPC lead. removing a 33% CPC Bias.

  23. @BCVoR

    You can't claim the debate performance as a reason why the 21 point spread is reasonable. There is zero connection between the two - the 21 point spread did not influence Harper's performance in the debate, and the debate was after the polling period that produced the 21 point spread.

    As for the CBC Voter Compass - I have yet to hear a single reason why it's biased that is based in actual fact. Every single one has been on the basis of incorrect assumptions (including a handful of completely ignorant ones) on how it functions.

  24. Do the Liberals pull back on the national campaign and run a targeted campaign to save money for a leadership campaign?

    Will Ignatieff bother with a visit to Alberta / Saskatchewan?

    Funding becomes a serious issue with the $2 / vote being phased out.

  25. No BT, if you go to the pdf of the poll from the COMPAS website, the questions about money management, and combat leadership were asked last.

  26. From the COMPAS poll pdf

    "As my last questions, I’d like to ask you about the personalities and character of the 5 party leaders....'

    Not first questions, last.

  27. @ BC Voice of Reason: The Toronto Star is no more a Liberal(federal) friendly paper. The current publisher John Cruickshank and editor in chief Michael cooke ran the most conservative papers in Canada i.e The Vancouver Sun and the Province under the leadership of none other than Conrad Black during the Chretien years. They then moved to Black's US papers when he sold his Canadian operations and then back to Canada after he got into trouble there. These guys are as right wing as they come. Many progressive journalists at the Toronto Star openly despise these two.

  28. "Measures of enthusiasm and commitment still show the Conservatives with a major advantage but the gap seems to be eroding somewhat."

    This is the most interesting line from the EKOS press release. Ekos still shows the Greens at 9% and the NDP at 20%.

    I think this happens in the EKOS parallel universe where younger voters all come out and vote, as much as older voters, and non-voters come out and vote.

    I'm not sure if they don't screen for likely voters, or if they include leaners and COMPAS or IPSOS calls those people undecided, but I place bets that how you get most of the differences between polls, and it isn't terribly transparent.

    If young people come out and vote, which is historically against the trends seen since the 1980's, and if the 800,000 Liberal supporters come out who didn't vote last time, the EKOS poll is probably fairly accurate.

    Do I think both those things will happen? No I don't. I think we will have record low turnout of around 55-57%, and the 35% undecideds in the COMPAS poll are mostly liberals who aren't coming out again.

  29. Eric I assume you already have this past COMPAS poll:

    That was released on June 24, 2004. The Canadian federal election in 2004 was held on June 28, 2004. COMPAS had 34% Liberals and 33% Cons at that point with a 3.5% MOE.

    Seats ended up being 135 to 99. 36% to 29% Nationally. So I guess that was within the MOE although this new one seems rather strange as how can so many other polls be so off?

  30. The last time COMPAS did national polling during a Federal election was 2004. Their last national poll was December 2008. They had the Conservatives with a 31% lead (51>20)

  31. The 35% undecided is significantly higher than the number for other pollsters. On the other hand, it's awfully close to the percentage of Canadians who don't bother to vote (i.e., 40%+). My theory is that the people who say undecied are, generally, people who are disinterested in politics and who are unlikely to vote (although there is, no doubt, a small minority who are genuinely undecided voters). If Compass' methodology is more effective at identifying actual voters than those of the other polling firms, that might explain the significant differences.

    Moreover, while the Tory leads in some provinces seem unusually high, it's not so implausible if you think that Liberal voters are more likely to be disengaged by this campaign (perhaps as a result of the pre-writ Tory add campaign against Iggy), while Tory voters are more likely to be energized. Moreover, some of their other party numbers (particularly for the NDP, Bloc and greens) aren't inconsistent with what other pollsters are getting.

  32. There was also a poll yesterday from Innovative Research. Con 39% Lib 28% NDP 17% BQ 9% Green 7%

  33. @Anonymous (12:52)

    Not sure why you seem to be disagreeing with me... I said they were asked last.

  34. Leger was polling for QMI/Sun. And now COMPAS is...what's the deal with that?

  35. BC Voice of Reason as a name13 April, 2011 13:51

    There are many anonymous posters making contributions.

    You can maintain your anonymity by just selecting the name/url out of the drop down box. All you have to do is fill in a fake name... don't even have to provide a fake e-mail address.

    It would not even have to be consistent (although that would be better) and makes it easier to have a conversation or make reference to a particularly wise post.

  36. BCVoR,

    You're right, there was a poll in late 2008:

    It'd be nice if there was one right before an election so that we could get a bit of a better estimation of what their house effect was though. I think your guess of +7 for the Conservatives is probably close to the truth though - a 14 point lead would still make them one of the more Conservative-favourable pollsters, but still within the MOE of what seems to be the industry average of around 11% right now.

    EKOS' numbers were pretty odd too, but at least we know that we should probably take about 3 points from the Greens and give them to the Conservatives (or at least that's what I do).

  37. Hi BC Voice of Reason, I did not realize you could do that. Thank you. I was Anon @ 13:14.

  38. Esquimalt - Juan de Fuca is not accurate13 April, 2011 14:58

    At this point I'm willing to put down money that your projection for the Liberals is at least 6 percentage points too high for the Liberals. Most other outlets are putting this complicated race in the "too close to call" column.

    The Liberal candidate is no longer the popular incumbant and maverick MP Keith Martin. They got their candidate literally just before the election started--Troy Desouza from the Conservatives has been campaigning in this riding for many elections and came within 80 votes of beating Keith last time.

    I heard somewhere that federal parties are actually allowed to spend their spending total every year rather than every election in each riding. If this is true, you may want to look at this for Esquimalt - Juan de Fuca, where I believe the Conservatives have spent boatloads of money every year pre-writ since last election to get Troy Desouza in.

    I know you need uniform polling data... but maybe there's a way for you to get a closer poll by poll analysis of Esquimalt - Juan de Fuca that could actually adjust your numbers back down to reality for this important BC riding.

  39. Do you have any data on which pollsters were most accurate last election


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.