Thursday, September 24, 2009

New Ekos Poll: 7.1-pt Conservative Lead (Updated)

EKOS released its weekly poll today, taken between September 16 and September 22 and involving 2,706 Canadians.

The polling firm headlines the poll saying that the Tories have the most committed voters. On a scale of 0-6, the Conservatives average a 3.5 level of commitment, while the Liberals are at 3.2, NDP 3.0, and Greens 2.7.

The Bloc's level of commitment is 3.6, more than the Tories, but apparently they aren't real voters.

Anyway, here are the national results:

Conservatives - 37.0%
Liberals - 29.9%
New Democrats - 13.8%
Greens - 10.2%
Bloc Quebecois - 9.1%

A big result for the Conservatives. And with EKOS being normally so consistent and using such large samples, that is a huge result. The Tories have improved 1.9-points from last week's EKOS poll. Significant, though, is that the Liberals are unchanged. So it isn't that the Liberals are dropping in the polls, it's that the Conservatives are rising.

The NDP result is pretty horrible for them, and they posted the biggest loss with 2.7 points. The Greens seem to have benefited, rising 1.2 points.

The regionals are very good for the Conservatives, though they are still struggling in British Columbia. They lead, but with only 34.2%. The NDP and the Liberals are tied at 24.1% and the Greens posted an excellent 17.6% result.

Alberta is what you'd expect, but the Liberals had a good 19.1% result there. In the Prairies, the Tories lead at 41.8%, followed by the NDP at 23.8% and the Liberals at 20.7%. Below average results for the two major parties, and a decent result for the NDP.

Ontario is the biggest news from this poll, however. The Tories lead with 41.8% - a huge result. The Liberals follow at 35.1%, still better than 2008, and the NDP and Greens are in the basement at 11.9% and 11.2%, respectively. Something struck me as odd about the Ontario poll, however. The sample size in "Toronto" represents only 27% of the provincial sample, which puts them somewhere between the "City" and "Urban" population designation on Wikipedia. Fair enough, but Ottawa represents 15% of the provincial sample, or the equivalent of almost 2,000,000 people. Without Gatineau, Ottawa is only about 800,000 people. And the Tories polled at 55% to the Liberal 28% in the federal capital. This leads me to believe that the Conservative result in Ontario isn't as high as 41.8%. (No longer a concern, see below)

Quebec is run of the mill, with the Bloc leading at 36.4%, the Liberals following at 28.6%, and the Conservatives posting a decent result at 19.4%. The NDP, at 9.3%, are floundering.

In Atlantic Canada, the Liberals have moved back into the driver's seat at 41.1%. The Conservatives are at 31.7% while the NDP has taken a big hit at 19.2%.

This poll would result in the following seat totals:

Conservatives - 142
Liberals - 96
Bloc Quebecois - 49
New Democrats - 21

Still no majority. The Conservatives can't win a majority with such weak (relatively speaking) results in Atlantic Canada, Quebec, and especially British Columbia. The Liberals would, in this situation, rebuild their caucus at the expense of the NDP.

The Conservatives lead in all demographics except those under the age of 25 and among university graduates. The Bloc leads in Montreal.

The Conservatives will be very pleased with this poll, but it only gets them exactly where they are right now. The Liberals would be back to 2006 levels, but the NDP would take a huge hit. They've apparently been the ones who were hurt the most by the recent events in Ottawa, which isn't too surprising.

This doesn't make Jack Layton any more likely to support the Liberal non-confidence motion (which, La Presse reports, will be worded very simply - "We have no confidence in this government").

I'll have a projection update later today or tomorrow.

11:14 UPDATE - EKOS has kindly responded to my questions (and multiple follow-ups) concerning the Ottawa sample. While they have over-sampled the capital, it's weight is adjusted within the provincial and national contexts to give it the proper proportion. So all is well, and thanks to the people at EKOS for being so helpful.


  1. The note about the size of the Ottawa sample is interesting.

    Of further note, the reported sample size proportions have been very inconsistent.

    This poll has the Ottawa sample size as 15.3% of the Ontario total.
    The August 27 Ekos poll has the Ottawa sample size as 10.9% of the Ontario total.
    The July 23 Ekos poll has the Ottawa sample size as 19.7% of the Ontario total.

    We do need to keep in mind that these numbers are for decided voters, so it could be that for some reason there are fewer undecided voters in Ottawa than elsewhere.

    Also, while the sample sizes indicate how many people in a given area were surveyed, Ekos is likely weighting the sub-regional results when producing either provincial or national results. This would mean that the Ontario results are not being skewed.

    It would be interesting to have a definitive answer on this though.

  2. I'm not sure the Ottawa result is neccesarily skewing the Ontario totals (and I think Martin's right that EKOS weights their result by region, at least that's what they say they do). I mean, the according to this poll the Tories are leading in Toronto (or at least the broader Toronto CMA)! That's unprecedented. If that number is right (and given the large margin of error it might not be) there's something going on in Ontario that's more than just statistical games. That would make the Tories more popular in Toronto than in their core base in rural Ontario.

  3. I don't doubt that the Tories are ahead in Ontario, just that the breakdown of Ontario voters seems odd. I've emailed EKOS about it, we'll see if they have a response.

  4. Wow, my prediction was really wrong - the NDP did take a big hit for their moves last week.

    It'll be interesting to see if their numbers stay down in the weeks to come. Perhaps as the Liberals start to fracture when they see the chance of power slipping away the NDP can regain some momentum.

    I guess they should just lay low, continue to make parliement work, and hope for better numbers in the spring.

  5. If you adjust the polling numbers with the commitment of the various party members you end up with the following results.

    C - 37%
    L - 29.9%
    N - 13.8%
    G - 10.2%
    B - 9.1%

    Commitment scale
    C - 3.5
    L - 3.2
    N - 3.0
    G - 2.7
    B - 3.6

    Adjusted polling with commitment taken into account

    C - 39.6%
    L - 29.5%
    N - 12.7%
    G - 8.4%
    B - 10.0%

    Very close to majority territory for the CPC; near half-life for the NDP.

    Eric, you want to give me your thoughts on this? Am I wrong to take into account partisan commitment?

  6. How did you come up with those results? EKOS puts it at 38% CPC, 29.3% LPC, 13.7% NDP, 9.7% BQ, 9.3% GPC among committed voters.

    I don't think this is significant, because uncommitted and non-partisan Canadians will still have to make a committed, partisan choice come election day.

    In any case, these calculations seem to improve the Bloc's standing in Quebec, meaning the Tories will hit a wall there. I am of the opinion that the Conservatives can't win a majority unless they win more than their 10 seats in Quebec.

  7. Eric what if people who aren't committed at this point simply don't vote ?

    Despite some blips, turnout seems to be on a downward trend:

    Given that this will likely be a highly negative campaign, the fourth election in five years, and one nobody wants assuming it is held this fall my best guess would be a very low turnout election.

    Could you give us a projected seat count using the committment adjusted numbers provided by Ekos ?

  8. No, I need regional results.

    And you're right, I think turnout will be down. But I hate when people say "no one" wants an election. Polls show that about 1 in 4 Canadians want an election.

  9. In other words, more people want an election than support the NDP, but no one says "no one" votes for the NDP.

  10. I guess it would be better to say an election would be "unpopular" then.

    I think Ekos must have a regional breakdown of commitment adjusted numbers on hand since they say in their information that fully committed voters are most likely to live in Alberta and Ontario and that not committed ones are most likely to be in BC. Maybe they'll release those numbers later.

  11. The best thing about these numbers is it would still take all three opposition parties to form their coalition.

  12. I think it is extremely unlikely that any sort of coalition agreement would be made in the future that included the Bloc in any way. The Conservatives have already poisoned that well.

  13. Eric wrote:

    "The Conservatives have already poisoned that well."

    That's a bit of lazy, partisan claim isn't it?

    Such statements seem to reflect a notion that the Conservatives have some sort of svengali-like power over language and can manufacture concern among the public out of thin air.

    In reality, significant numbers of Canadians were deeply uncomfortable with the proposed coalition. The Conservatives didn't invent that concern, they merely tapped into it.

    You may disagree with the concern those Canadians felt, but it seems almost contemptuous of them as persons to imply that they were simply pawns in some sort of exercise in hypnotic control.

  14. I'd say the Conservatives exploited the misinformation and misperceptions Canadians have about the Bloc Quebecois.

    That's politics, though. The one with the good clip wins, the one with the nuance is boring.

    Every party exploits the fact that Canadians spend an average of seven seconds paying attention to politics every day (or whatever it is).

    I think it is absolutely undeniable that Canadians as a whole have a muddy understanding of the Bloc Quebecois, which is unfortunate. It certainly doesn't help our democracy.

    Politicians on the Hill have often recognised the Bloc as hard working MPs - not separatist ideologues. They're a social democratic party with no ability to split the country up. Referendums don't happen in Ottawa.

    Each party has created caricatures of each other party, which is a problem. No one is innocent of that, but I don't think that saves any of them from criticism. "They did it first" isn't a real argument.

  15. Eric, the well didn't need to be poisoned. The Conservatives played up the fact that the BQ and NDP would be in a coalition sure but that was never the central problem.

    The issue was that Dion had ruled out a coalition during the campaign.

    To be Prime Minister you must have a mandate from the people. It isn't just about seat counts and getting a majority in parliement.

    The GG has broad discretion in these matters and traditions, common sense, and the will of the people are part of her decision making process just as much as what's theoretically legal.

    There is just no way a coalition would be accepted by the Canadian people after a leader expressly ruled it out.

    If the parties want to form a coalition next time around they need to say that they are open to the option up front, before or during an election campaign.

  16. Eric wrote:

    "misinformation and misperceptions Canadians have about the Bloc Quebecois."

    I don't think that's it at all. I think that you have misperceptions about the political concerns of a large swath of the Canadian people.

    The Bloc is a separatist party. They even describe themselves primarily by the phrase:

    "Parti politique souverainiste"

    No, a referendum won't be held in Ottawa. So why is the Bloc there? In their own words it is to help create the "winning conditions" for an independence referendum.

    The fact that Bloc MPs are "hard working" is of little relevance if the end towards which they work is anathema to a particular voter.

    Many Canadians hold a deep, patriotic love for their country and do not look kindly upon anyone whose stated goal is to destroy that country.

    You don't have to share their position but it doesn't make it any less legitimate.

  17. As for the BQ, I think a lot of people recognize that they've grown past being scary separatists and basically vote similiar to the NDP on a lot of matters. And, of course, not all their supporters are actually separatists themselves.

    But make no mistake about it, their express purpose in Ottawa is to further the decentralization of powers.

    It has nothing to do with referendums, it has to do with shrinking the role of the federal government and enlarging the scope and size of provincial responsibilities.

    In this way the BQ is more like the Conservative party with their Albertan tradition of not wanting centralized decision making in Ottawa. Where as the NDP and Liberals prefer top down Ottawa decision making and big national programs.

    Unfortunately the Bloc tends to demonize Harper and vice versa.

    Its too bad because they could actually work together on a lot of issues surrounding federalism.


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.