Tuesday, January 18, 2011

Conservatives gain in Harris-Decima poll

The newest poll from Harris-Decima shows that the Conservatives have pulled out of the close race that the polling firm had found in its late-2010 polling. But the party is still not in a position to make any gains.Since Harris-Decima's last poll taken in December, the Conservatives have gained five points and now lead with 36%. The Liberals have dropped only one point to 28%, while the New Democrats are steady at 15%.

The Bloc Québécois and Greens have both dropped two points to 9% nationally in this telephone poll, which has a margin of error of +/- 2.2.

That is a big gain for the Tories, and puts them generally where Angus-Reid and EKOS have recently pegged them. This poll from Harris-Decima is also very close to my own current projections.

In fact, the only outlier result to be found in this poll, if an outlier can be defined as a result with a difference from my own projection greater than the poll's MOE, is in British Columbia. There, the Conservatives lead with 34% (+2 from the last poll), while the Liberals are close behind at 32% (+11). That Liberal result is about three points higher than the poll's MOE should allow. The NDP is steady at 24%, while the Greens are down 12 points to 9%.

In Ontario, the Conservatives have gained three points and lead with 39%, followed by the Liberals at 34% (-2). The NDP is up one to 15%, while the Greens are down two to 10%.

The Bloc has dropped four points in Quebec but still leads with 40%. The Liberals follow with 20% (-3), while the Conservatives are up seven to 18% and the NDP is up three to 13%. That about lines up with what the other pollsters have found this month.

The Liberals are leading in Atlantic Canada with 42%, with the Conservatives not far behind at 35%.

In Alberta, the Tories are up 14 points to 61%, while the Liberals and NDP sink six and eight points to 18% and 6%, respectively. The Greens are at 12% here.

In the Prairies, the Conservatives lead with 49%. The NDP is in second with 20%, while the Liberals have dropped six points to 19%.

With these poll results, the Conservatives would win 18 seats in British Columbia, 27 in Alberta, 22 in the Prairies, 54 in Ontario, seven in Quebec, and nine in Atlantic Canada for a total of 138 seats. That is 13 more than my projection for Harris-Decima's last poll, thanks to a gain of seven seats in Ontario.

The Liberals would win 12 seats in British Columbia, one in Alberta, three in the Prairies, 40 in Ontario, 14 in Quebec, and 22 in Atlantic Canada for a total of 94 seats, seven fewer than last time. The drop came primarily in Ontario, where the Liberals lost eight seats (but they made small gains elsewhere).

The Bloc would win 53 seats in Quebec, three fewer than the last projection.

The New Democrats would win six seats in British Columbia, none in Alberta, three in the Prairies, 12 in Ontario, one in Quebec, and one in Atlantic Canada for a total of 23, two fewer than last time. The largest change is in British Columbia, where the NDP drops two seats.

The Greens win no seats, compared to one that was projected from Harris-Decima's last poll.

Nothing really new in this poll, though it does give the Tories a bit of a bigger lead than either Angus-Reid or EKOS Research found to be the case. However, the difference is within the margin of error. The poll should embolden no one, least of all the New Democrats, who have been stumbling in the polls lately.


  1. The Tories and the Liberals seem to be evening out in Quebec. This is becoming a significant sub-narrative in recent polls. Any thoughts?

  2. It's not so much an evening out, as it is a poor result for the Liberals and slightly-better-than-average result for the Conservatives. This is because the NDP are keeping their above-10% vote share in the province, taking most votes away from the Liberals.

  3. I really think the provincial margins of error skew the seat projection.

    Take the NDP in Nova Scotia, for example. Despite the provincial party taking a severe dip in the polls, I think it's a stretch to think that Megan Leslie will drop 15% or that Peter Stoffer will dip 40%, leaving the province with only one New Democrat MP.

    And while I realize that it is purely a mathematical formula and can't account for a range of factors, I think the projections are misleading. Leslie will likely be safe, seeing as how her 5% dip in the last election was largely due to the fact that she wasn't Alexa McDonough. Now that she's become an MP in her own right, I think the strange numbers gained from the small sample size distorts this picture.

  4. Hey Justin,

    As a follower of this blog, I would care to much about the margins of error for the provinces. As Eric has stated in several previous postings, his model is based on a universal swing of each riding. So if the NDP are polling very poorly in Atlantic Canada then his projection will show a reduced NDP number of seats.

    All in all Eric work and website is entertainment for us polical junkies. Most of us recognize that local campaigns and individuals will matter, then information of which cannot be account in a poll of a few hundred in the small Atlantic region or any region.

    Mots importantly this polls show that there is a distinct distance of support between each party, where in an election scenario, the CPC could win as much as 160 seats or a low as 125, the Liberals as high as 100 seats or as low as 70 seats, and the NDP as high as 40 seats and as low as 15 seats, (or something like that)

  5. Sorry, I don't use a uniform swing model.

  6. Eric: "Sorry, I don't use a uniform swing model."

    You've mentioned that before, how do you translate the regional numbers into riding numbers? Is it a proportionate swing (or whatever other term you want to use)? (I apologize if you've explained this elsewhere. If so, maybe you could just provide the link)

  7. Generally speaking, I've looked at what various levels of support in each province/region meant in terms of seats for each of the parties. So, I take poll results (or my projection), compare them to the historical chart I have, and calculate the seats that way.

    In my view, it's a good system that takes into account the intricacies of each region.

    But I am working on a riding-level projection, which will use a combination of factors including uniform swing and each riding's individual particularities. My look at the influence of star candidates and cabinet ministers on voting behaviour is part of that process.

  8. Don't get me wrong, ThreeHundredEight to me is like sugar to a toddler, I love this stuff.

    But as Eric points out, riding-specific projections would offer a really great picture of what the actual landscape is. It's not Eric's fault here, as he have a very general picture of what the polls indicate. What really gets my goat is when the national media leeches onto these polls/projections as the solid, accurate truth.

  9. The media being my own articles I write for the media?


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.