Friday, March 09, 2007

The Tide Has Crested!

I kind of suspected this would happen. As the wave of attack ads and mud has tapered off, the Tories numbers have started to creep downwards; the Libs upwards.

Note that this poll is from the infamous Angus Reid, but the polling dates (March 6 and March 7) given do not seem to match up with the on-line poll they've been conducting recently, and which I wrote about here. Or at least if this is indeed that poll, Cherniak's responses to it (given yesterday) must have been excluded from the total.

Anyway, if it is an on-line poll, I still question the methodology, and would point to the weird Quebec numbers which seem to be at odds with every other polling source:

In Quebec, the Bloc has the support of 37 per cent of decided voters, with the Conservatives in second place with 25 per cent, and the Liberals third with 24 per cent.

Rationally or not, however, I am encouraged by the direction of the national trend (which is pretty close to Decima's).


Anonymous said...

You've got to be kidding! Your post just demonstrates deluded partisans can become.

A 2% shift is not statistically significant. Even if it were, it would be a miniscule change. You're doing some extremely wishful thinking.

I never thought I'd see the a Liberal celebrating a 10-point lead for the Conservatives! How much more desperate for good news can you get?

Anonymous said...

3% margin of error you dummy. Not statistically significant.

bigcitylib said...

Trend in Decima is similar: Libs up a bit, CPoC down a bit. Taste the pain, Tory boys.

Oxford County Liberals said...

This is indeed the online version BCL. Look at the numbers where the Tories go down and the Liberals go up compared to last week. Those comparison #'s are from their infamous "Tories have a 14 point lead" online poll, which I and others have already said to take with a large grain of salt.

I take this with an equally large grain of salt even if the trends are reversed. That means a few more Liberals registered to participate in the poll to get the #'s turned around. It is methodology I find rather suspect, regardless of who is in the lead.

Anonymous said...

Like I've said in response to you elsewhere, how is a scientifically controlled online poll any less accurate than a phone poll, where you're missing out on people who don't have landlines or those who won't answer calls they see on their call display from numbers they don't know (and hence assume it's a telemarketer).

Anonymous said...

As well, the Quebec numbers do jive with the more regionalized Quebec-only polls done by CROP and one other Quebec polling company, which have a higher sampling rate for the province, and more info on the specific regions (incl. Quebec City where the Tories dominate) than the national polls do.

bigcitylib said...


But when you do get somebody with a land line, you can associate that phone number with a street, you can associate that street with Stats Canada census tract data, and you can start making assumptions about class, ethnicity, age, all sorts of stuff. The phone number allows you to put together a demographic profile of the people who answer your phone survey and weight their responses.

Whereas with the online poll I lied about all that stuff, and so can anyone. Five Alberta Tories could, I think, become fifty Toronto Metrosexuals for the Angus Reid poll and indicate a Tory preference. Headline: Tory breakthrough in Toronto.

So I'm not sure what counts as "scintifically controlled" under these circumstances.

bigcitylib said...

Scott, you're right, and I agree. Interesting (or maybe not) that people were answering the surveys yesterday (cuz Cherniak says he did) but these responses were presumably not counted.

Ti-Guy said...

The Liberals are polling in majority territory.

...oh, c'mon connies. Stop crying.