Wednesday, September 16, 2009

New AR Poll: 7-point Conservative Lead

The Toronto Star is reporting on a new Angus-Reid poll. The paper is giving very few details, but what this article mentions is that the poll involved 1,002 people and completed polling on September 13.

The result:

Conservatives - 36%
Liberals - 29%
New Democrats - 17%
Bloc Quebecois - 10%
Greens - 7%

Angus-Reid has been pretty bang-on in the polls the last few elections, so their findings have to be given due respect. With a 3.1 margin of error, this puts the Ipsos-Reid poll from yesterday in a better light, though it still seems they under-polled the NDP at 12%.

The article also mentions that the Tories have a 12-point lead in Ontario. Massive.

I am curious to see what the EKOS poll will say tomorrow as well as polls next week. The polls we've been getting this week all took place in the context of a probable election. Polls that will be released at the end of this week and next week will be taken in the context of an election avoided. What we will have to watch is whether the NDP will benefit for trying to 'make Parliament work' or whether they will be punished for folding to the Conservatives. We will also have to watch whether the Liberals will be punished for their adamant opposition to the government, or whether they will be rewarded for growing a back-bone.


  1. Let's not get carried away with the brilliance of Angus Reid. AGR's track record is, I'm pretty sure, exactly one election long. Angus Reid Strategies was created in 2006 and I'm fairly sure it was started after the election was held. From 2001 to 2006, Reid was barred from working in polling because of his non-compete agreement with Ipsos after he sold Angus Reid Group.

  2. I'm not here to promote or defend Angus Reid, but they do have a pretty good record - not just the last federal election, but also provincial elections in Ontario, Manitoba, Nova Scotia, Saskatchewan and BC. In fact in BC they were the only ones who correctly predicted that the BC Liberals would end up with a low single digit lead over the NDP while everyone else was off by a mile.

    Its true that in political polling you're only as good as your last election, but so far Reid seems to have quite a bit to crow about.

  3. I'm not disputing DL's statement, but ARS also has had a few of dogs as well.

    For example, in the last federal election, ARS had an 800 sample size poll of Saskatchewan near the end of the federal election campaign and here were their findings with the actual election results in brackets:

    CPC - 40% (54%)
    NDP - 35% (26%)

    ARS 5% spread turned into a 28% spread on election night. That's a huge differential.

  4. AR was also closest in the 2008 Quebec election.

    Nobody's perfect, of course, but some are less imperfect than others.

  5. I agree that that federal poll by ARS in Saskatchewan was a "dog". I can only assume that maybe in a province as small as Saskatchewan, they may have had an online panel that was heavily skewed somehow - but they seem to have redeemed themselves with good results in the Nova Scotia election. Every company has its flubs: remember Ipsos and Strategic Counsel just before the 2006 election forecasting 15 point Tory leads when it ended up being a 6 point spread?

    For the most part ARS seems to be as accurate as anyone else - if not more so.


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.