Tuesday, November 23, 2010

Kamouraska-Témiscouata a close race in new CROP poll

A rare by-election poll was conducted by CROP for the riding of Kamouraska-Témiscouata, whose voters will be heading to the polls in less than a week. It shows a very close race, with the provincial Liberals, the Parti Québécois, and the ADQ within seven points of one another.With 34%, France Dionne of the Liberals (PLQ) has a narrow two-point lead over André Simard of the PQ. With the 4.9% margin of error in this telephone poll, that puts them at a statistical tie.

With 25%, Gérald Beaulieu of the ADQ is not very far behind.

Serge Proulx of Québec Solidaire and Frédéric Brophy-Nolan of the provincial Greens bring up the rear with 6% and 3%, respectively. About 13% were undecided.

CROP also broke down the numbers by likelihood of voting, which more pollsters should do. Of those who will "probably" vote, the PQ leads with 31% to the Liberals' 27%. Of those who are "certain" to vote, the Liberals lead with 37% to the PQ's 31%.

In other words, the PQ needs to up their "get out the vote" campaign if they want to win this riding, which they appear to be capable of doing.

Of course, it is a by-election and anything can happen. What's more, the margin of error is large in this poll and riding polls are notoriously unreliable, especially when you consider that 59% of people said they would vote. By-election turnout is usually about half of that.

This is a riding the PLQ should not lose, even in this political climate. The by-election was called after the death of Claude Béchard, who succumbed to the cancer he had been fighting for several years. He won the riding with 46% of the vote in 2003, dropped to 40% in 2007, and took 54% in 2008. That is a solid base of support for the Liberals.

The wild card in this race will be whether the voting angst against the Charest government will coalesce around one party. Over the last three elections the ADQ has been the alternative, with as much as 37% support in 2007. But the party still had 26% and 22% in 2003 and 2008, when the party did less well provincially as a whole. The PQ's share of the vote has been relatively steady, going from 26% in 2003 to 19% in 2007 and 21% in 2008.

Neither the Greens nor Québec Solidaire have ever had much traction here.Using Léger Marketing's last provincial poll, and applying uniform swing to the 2008 electoral results (2007 for the Greens) according to how the vote has shifted, we still get a large Liberal lead: 48% to 24% for the PQ and 18% for the ADQ.

This seems unlikely, based on Quebecers' dissatisfaction with the Charest government at the moment, and how more unknown Dionne is as compared with Béchard.

We will have to wait and see. The election will be taking place on November 29, the same day as the three federal by-elections in Manitoba and Ontario.


  1. Interesting results.

    While K-T is certainly a nominally Liberal seat, 2007 demonstrated that it isn't out of reach once opposition solidifies behind one party - the perfect "bellwhether" riding if there ever was one.

    If the Adequistes and the Pequistes manage to split the vote enough to allow the Liberals to pass through, just imagine the grins on the Liberal strategists faces on the new way they'll manage to somehow keep power after all the bruisings they've taken.

  2. I realize that on paper this seems like a safe Liberal riding, but don't you think that this was largely a personal vote for Bechard?? Kamouraska-Temiscouata voted PQ in '76 and '81 and maybe when Parizeau won as well and the area has voted solidly BQ in every federal election until the byelection loss to the Tories last year - so you would think that a government as ridiculously unpopular as the Quebec Liberals would have ZERO chance of holding on to a seat like this which I believe is also 99.9% francophone - so no knee-jerk anglo- and allo-phone Liberal voters like in St. Laurent.

  3. what a great polling question and added information:

    1) Are you going to vote?

    a) For certain
    b) probably
    c) not likely
    d) no

    How hard of a question is that?

    It would reduce the polling sample of voters and require a larger sample to make the statistical claims.

    I would think that the internal polls done for the parties include this question in every poll.

  4. The trouble is that in poll i have seen about 90% of people will CLAIM that they are "certain to vote" - then the turnout ends up being 60%. Keep in mind though that many (if not most) people who don't vote also do not take part in surveys - so its all a wash.

  5. Keep in mind though that many (if not most) people who don't vote also do not take part in surveys

    Hmm. Interesting.

  6. BTW: I checked and PQ only lost Kamouraska by 100 votes in 1994 and it was also very close in 1998. The riding also votes Yes in the 1995 referendum.

  7. DL

    If 90% of people say they will vote for certain and only 60% actually do vote does that mean the MOE due to inability to get a representative sample in any given poll would be 33% rather than the 2.5% published??

    The pollsters understand that they are not going to get a representative sample with only 1000-2000 people polled so make adjustments to reflect a representative sample that reflects the actual population. If they end up with 60% female in the survey the female input will be adjusted down ward to reflect the 51% reality. If a higher portion of female say they wont vote that would have to be worked into the equation to get to the published results.

    If even 10% say that they are not going to vote that could dramatically impact the poll. If half the people that indicate they would vote Green but will not vote drops the Green polling support to 5% from 10% and leaves Ms. May out of the debates.

    Is this information the pollster understand is easily available but choose to ignore?

    That moves the bias from inadvertent based on polling methods that Eric tracks to biased data manipulation favored by the Global warming supporters.

  8. suggestion to Eric....

    Your seat predictions might be more accurate if you provide more weight to pollsters that identify the "will not vote" people and remove them from the polling results.

    Now that you have gone mainstream with the G&M you might have some influence in making the pollsters do a better job

  9. Its not that simple BC VOR. There seems to be little correlation between people CLAIMING that they will vote and people actually voting - since almost everyone will CLAIM that they will definitely vote - because it is socially desirable. One thing we do know, at least anecdotally, is that if you are the kind of person who has no interest in politics and can't be bothered to vote - you are probably also not going to pick up the phone when your call-display says "Ipsos-Reid" or "Nanos Research"


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.