Sunday, August 23, 2009

Rumour: Ipsos Mystery Poll To Appear Monday

Reported in my comments here. As has been pointed out by several readers, Ipsos Reid--which will apparently claim that the Tories are "flirting with 40%"(so what really--about 38% or 39%?) and therefore flirting with a majority --1) skews Tory, very Tory, 2) polls for CanWest, publisher of Canada's most crazy-assed Tory newspapers, and 3) is following upon a Tory plea to its supporters Canada-wide to talk up its need for a "stable majority".

Given the explanation for the figure noted above here, we can probably rewrite tomorrows Natty Post headline as "No Movement Noted!"


Steve V said...

I don't want to question Eric's methodology, but the graph doesn't seem to capture the under-representation for the Liberals. The only real basis you have is election results to compare. In 2004 they had the Libs 5% below their election total on the eve of the vote, 2006 3% below on the eve of the vote. In 2008, they didn't poll for the last days of the campaign, so it's hard to get a true measure of potential bias in their numbers.

I agree in over stating Con support, but there is also a past behavior of not accurately capturing the Liberal vote.

What I find interesting, and this isn't the pollster necessarily, but the client. No polls commissioned in months, then all of a sudden you ask them to go out in the field the week Harper is in photo op heaven, a week anybody with a half a brain would pencil in as probable optimal support potential. Then, you don't release late Friday, per usual, you wait until Monday to maximize the impact. I must say, this strikes me as an attempt by Canwest to get the Liberals to think again about an election, JUST as their are scheduled for their summer caucus meeting. The timing is just so cute, in every way.

Oh well, it will give the Con moonbats something to cuddle with.

√Čric said...

The chart above is a comparison to other polling results, rather than election results. Since the beginning of the year, in any given month Ipsos-Reid polls best for the Conservatives compared to the other pollsters.

Could they be under-polling the Liberals? Maybe. What the chart says is that since the beginning of the year they've been about average with the other pollsters. So it could just be that, on average, all pollsters are under-polling the Liberals, or that other pollsters are over-polling the Liberals, or that Ipsos-Reid has changed its methodology.

Robert McClelland said...

I must say, this strikes me as an attempt by Canwest to get the Liberals to think again about an election

That's been going on for quite some time. I've noticed that nearly every time election speculation heats up to a boil a poll comes out showing the Conservatives approaching majority territory. The most obvious example was when the coalition formed. Conservative support suddenly spiked to 50%. Ipsos really had to cook their poll to generate those results. Once speculation dies down however, the polls resume showing a dead heat between the Cons and Libs.

Gayle said...

Surely the LPC does not use polls commissioned by the media to determine whether or not to call an election. That would be foolish.

I like polls that show the CPC close to majority. Those kind of numbers make people think twice about voting for them.

I suspect that if the CPC really were flirting with majority numbers at the moment they would be anxious to go to the polls.

Tomm said...

Nice graph.