Friday, April 27, 2007

New Ipsos Reid Poll

This is the first poll that has results from this week (Tuesday to Thursday). National numbers:
The survey says the Conservatives have 38 per cent support, up two points from the election; the Liberals are up one point to 31 per cent; the NDP is down three points to 14 per cent; the Bloc Quebecois dropped three points to eight per cent, and the Green party is up three points to eight per cent.

For context, last week's Ipsos numbers:
The survey said support for the Tories was up three points to 39%, while support for the Liberals dropped three points to 29%. The NDP climbed two points to 16%, the Bloc Québécois edged up to 9% from eight and the Green party slipped to 7% from eight.

Not much movement, although the Liberals tick upward, while the NDP falls slightly.

The regional numbers are suspicious, as I've brought up before, particularly British Columbia:
Across Canada, Conservative support declined in every region except British Columbia, where it rose 17 points to 47 per cent. It dropped eight points to 62 per cent in Alberta, six points to 22 per cent in Atlantic Canada, three points in Quebec, two points in Ontario and two points to 46 per cent in Manitoba and Saskatchewan.

The Liberal numbers were up five points in Ontario and four points Quebec.

Grit support declined six points in B.C. to 28 per cent, four points in Alberta to 10 per cent and two points in Atlantic Canada to 44 per cent.

The Liberals remained flat in Saskatchewan and Manitoba at 22 per cent

Support for the NDP fluctuated mildly in most regions, while the Green party got its largest bump in Alberta, up seven points to 15 per cent.

The margin of error in regional breakouts ranges from a high of 12 percentage points in Manitoba and Saskatchewan to a low of five percentage points in Ontario.

Someone please explain to me how Tory fortunes rose 17% in British Columbia in a week? Did I miss the announcement that all British Columbians would receive a $1000 tax rebate effectively immediately? Last week Ipsos had the Conservatives down 9 in British Columbia, maybe they should just stop releasing the regionals with high MOE, because it's just useless information.

Ipso concludes the Afghanistan mess hasn't cost the Conservatives:
Bricker said a lack of significant movement in this week’s national numbers from last week’s results suggests the political uproar in Ottawa over what and when the Conservative government knew about the alleged torture of prisoners of war in Afghanistan is not resonating with most Canadians.

I disagree, this issue really boiled over on Thursday(poll ended) and most of the punditry agrees there will be a cost. Coupled with the pale green plan and this is one of the worst weeks I can remember for the Conservatives. It will be interesting to see if subsequent polls show further erosion.

12 comments:

candy said...

hi!!!you blog very special!!!!

Karen said...

Well, if there ever was proof to explain how inaccurate these polls are, specifically Ipsos, this is it.

They are telling us in a week that the con's put out their Environment Plan, that their numbers soared in BC? I'm laughing as I'm typing.

I don't know how or why, but it seems that we don't digest things until a couple of days later. Afghanistan and the Environment are going to hit a bit over the weekend and not go away next week.

Ipsos is premature with it's findings, as usual, and imo, that will be shown next week.

Always the defenders of the con's is Ipsos. They may just be putting themselves out of business.

Steve V said...

knb

"They may just be putting themselves out of business."

How can you just throw out that B.C. number and expect anyone to take you seriously? You will also notice that Green support SURGED in Alberta.

On the national numbers, it might be noteworthy that the gap between the Cons and Libs went from 10% in the last poll to 7% in this one, which is in line with other erosion polling. Having said that, this poll really doesn't digest the fallout from this week, given the dates.

Monkey Loves to Fight said...

The British Columbia numbers are simply due to a large margin of error. With the regional numbers I only look at the Ontario and Quebec numbers as the samples are too small in all other regions. With only 100 interviews, that gives a margin of error of +/-10% so the Tories probably fall somewhere between the two numbers.

As for the fallout from this week, I certainly cannot see it being good for the Tories, but it really depends on how much people are paying attention.

Anonymous said...

It is pollus interuptus - thinking that a poll on the treatment of detainees as an issue is settled on Thursday of th eweek it is discovered is ignorance or hubris. Hey - it may well be both!

This is a serious complex issue that takes time to sink in - but sink in it will.

Anonymous said...

Again, I guess we'll have to wait for SES to compare.

If the Conservs still remain in 36 to 38 territory after this week - Canada is sleeping.

Anonymous said...

I agree with Candy....The ipsos poll is always a sham...They sometimes put out a low tory number but you can bet the next week they will come out with a whopper how the cons have soared..It is a laugh.

Anonymous said...

This poll is obviously a sham. Now where were we? Right... "Time to match the rhetoric with action fellas. BRING THEM DOWN."

Steve V said...

"The ipsos poll is always a sham."

Even the way they framed this poll is suspicious, using the election results as backdrop, instead of the last poll for trends. It makes it seem like Harper is up, when he clearly isn't. I put Ipsos at the bottom, in terms of objective accuracy.

ottlib said...

Darryl Bricker is the poster child I hold up when people ask me why I dislike polls and pollsters.

Half of the time the methodology of the poll is suspect, to say the least, and the other half of the time the analysis by the pollster is superficial. (and I am being generous by characterizing it like that.)

What Mr. Bricker seems to have forgotten (willfully) is up until Wednesday afternoon the only media talking about the Afghan detainee story was Bell Globe Media. It was well toward the end of interviewing for this poll that the story became front-page news in all media.

As well, the polling ended well before Canadians saw David Suzuki call the Conservative Green Plan a "sham" on all of the television networks and it all of the papers.

Mr. Suzuki is well respected amongst Canadians who do not live and breath politics and is perceived as a non-partisan defender of the environment. When he is seen using such stark language about a government initiative it is going to have an impact, especially when he has not been known to do such things in the past.

So Mr. Bricker had better begin thinking about what he is going to say for his next poll, which will be the one that shows the impact of this last week.

Steve V said...

ottlib

I'm surprised Bricker tried to tie the poll into the Afghanistan debate:

Bricker said a lack of significant movement in this week’s national numbers from last week’s results suggests the political uproar in Ottawa over what and when the Conservative government knew about the alleged torture of prisoners of war in Afghanistan is not resonating with most Canadians.

As I stated earlier, this story only went nuclear on Thursday, which is when the polling ended, so I don't see how he can draw the above conclusion.

Monkey Loves to Fight said...

I don't think you can really compare polls of different companies in between elections as each one has a different methodology. What you have to look at is what direction the movement is in. Also the numbers from all the polling companies will begin to converge around the second week of an election since when there is a high undecided vote and people are less firm in their intentions any change in the question can change the answers. As people become more firm in their preferences, then there is less variance. I've actually found Ipsos to generally be pretty accurate on the final results although they do in between elections tend to put the Tories a bit on the high side.