Federal Election Polls

[punked]

The Quebec numbers are a trend though. You also have to realize that they use averaged numbers from three nights, not just one.

But Angus Reid has really really different numbers again I find these numbers weird.

In Quebec B:34 N:24 C:19 L:18

http://www.thestar.com/news/canada/politics/article/970557--ignatieff-s-appeal-improving-but-harper-still-leads-poll-says?bn=1

Again these tracking poll numbers aren't being shown anywhere else. I still think the small sample size is killing its accuracy. It doesn't matter if you take the number over three nights if your accuracy on one night is off, it is also off over three nights. That is how polling works.

Threehundredeight is calling out Nanos methodology of asking first two choices and thinks that is why the Liberals are showing super high compared to every other poll.

Edited April 10, 2011 by punked

[Smallc]

But Angus Reid has really really different numbers again I find these numbers weird.

In Quebec B:34 N:24 C:19 L:18

Those numbers seem highly implausible.

[punked]

Those numbers seem highly implausible.

I am not saying they are right I am saying having the Liberals at 27 after that poll and this one which all came out over the same period of time tells me something is wrong with Nanos polls. Even Ekos says Nanos is way off. Their poll is so far outside the norm it doesn't make any sense.

In Quebec:

NDP: 18 Tories: 22 Libs: 21

http://www.montrealgazette.com/news/Mulcair+walks+tall+Outremont/4589441/story.html

Edited April 10, 2011 by punked

[ToadBrother]

Those numbers seem highly implausible.

Maybe the pollsters get together for drinks and decide that one of them will throw out whacky numbers. "Yeah, tell you what, you guys play it straight and I'll have NANOS predict the Greens will win by a landslide!"

[Smallc]

Maybe the pollsters get together for drinks and decide that one of them will throw out whacky numbers. "Yeah, tell you what, you guys play it straight and I'll have NANOS predict the Greens will win by a landslide!"

With the erratic numbers this campaign, I wouldn't doubt it.

[Harry]

Nanos has had a history of wonky polling.

Here is a similiar example, they used to be called SES, to what they they are presently projecting, which occurred during the 2006 election campaign.

http://www.sesresearch.com/election/SES%20CPAC%20December%2012%202005E.pdf

Edited April 10, 2011 by Harry

[wyly]

Those numbers seem highly implausible.

parties do there own internal polling so it would explain the NDP throwing their lone alberta mp under the bus to court quebec voters...

[wyly]

Nanos has had a history of wonky polling.

Here is a similiar example, they used to be called SES, to what they they are presently projecting, which occurred during the 2006 election campaign.

yup,the least reliable polling source IMO...only the one on election day is accurate...

[Smallc]

yup,the least reliable polling source IMO...only the one on election day is accurate...

Nanos isn't the least reliable:

The firm received great praise when it correctly predicted the outcome of the 2006 Federal Election to within one tenth of one percentage point for the four major parties - a record in Canadian polling history.

http://en.wikipedia.org/wiki/Nanos_Research

[punked]

Nanos isn't the least reliable:

http://en.wikipedia.org/wiki/Nanos_Research

But isn't Angus Reid the most reliable and their numbers put the NDP at 21 and the Libs at 26. So why are they so off?

[Smallc]

But isn't Angus Reid the most reliable and their numbers put the NDP at 21 and the Libs at 26. So why are they so off?

Angus Reid was closest last time, but NANOS wasn't much further off.

[punked]

Angus Reid was closest last time, but NANOS wasn't much further off.

I am just saying when you have two pollsters which are this far off you have to look at the other polls and the other polls say the Liberal numbers in the Nanos polls are inflated by 4-5%. There is something wrong with these Nanos polls.

Nanos showed the same thing last election 2 weeks in they had a 2-3 point gap between the Libs and Cons when every other poll had it much much bigger. Why is that SmallC? AR btw had the same numbers they have today as well last election.

Edited April 10, 2011 by punked

[Smallc]

I am just saying when you have two pollsters which are this far off you have to look at the other polls and the other polls say the Liberal numbers in the Nanos polls are inflated by 4-5%. There is something wrong with these Nanos polls.

Some of them also say that the Conservatives are inflated. I guess we'll find out on May 2.

[punked]

Some of them also say that the Conservatives are inflated. I guess we'll find out on May 2.

Let's be honest the Conservative numbers are almost always underinflated because they are more likely to vote so they almost always get a 3 point bump on election day.

[Smallc]

Let's be honest the Conservative numbers are almost always underinflated because they are more likely to vote so they almost always get a 3 point bump on election day.

I hope you're right.

[Jerry J. Fortin]

Let's be honest the Conservative numbers are almost always underinflated because they are more likely to vote so they almost always get a 3 point bump on election day.

Ya think maybe Iggy taking the position he has has anything to do with a platform designed to acquire a broader base than merely a partisan slant. Iggy seems to have designed an approach to garner a piece of the apathetic voter market. These are the folks who can turn an election around. His appeal to ALL Canadians not simply an ideological segment, is to promote a government for the people instead of the party. A government for the little guy. That is a Liberal approach not used in many years, and it appears to have gained some traction.

[punked]

Let's be honest the Conservative numbers are almost always underinflated because they are more likely to vote so they almost always get a 3 point bump on election day.

Ya think maybe Iggy taking the position he has has anything to do with a platform designed to acquire a broader base than merely a partisan slant. Iggy seems to have designed an approach to garner a piece of the apathetic voter market. These are the folks who can turn an election around. His appeal to ALL Canadians not simply an ideological segment, is to promote a government for the people instead of the party. A government for the little guy. That is a Liberal approach not used in many years, and it appears to have gained some traction.

No I think the Liberals are looking at a higher popular vote but only a gain of 10 seats at the most. Most likely it will be a bunch of seats shifting with no real gain and for the Libs this is will be called a win.

I don't think the polls are moving as much as people think. The debates are the Liberals, and NDP only chance of actually gaining anything and they are going to need a good one.

[Harry Hayfield]

I always believe that when it comes to polls, you need a moving average "poll of polls" in order to get some sense of what is happening. Using the polling data that has been listed on Wikipedia and taking the polls since the election was announced, I get the following (working on the assumption that the election was being held on April 10th 2011)

Conservatives 39.22% (+1.56%)

Liberals 29.84% (+3.21%)

New Democrats 15.36% (-2.81%)

Bloc Quebecois 9.09% (-0.89%)

Greens 5.20% (-1.60%)

Others 1.60% (+0.43%)

Conservative lead of 9.38% on a swing from Con to Lib of 0.83%

According to the calculator used on FairVote Canda that would see the following House:

Conservatives 147 seats (+4 seats)

Liberals 93 seats (+16 seats)

New Democrats 22 seats (-15 seats)

Bloc Quebecois 44 seats (-5 seats)

Greens 0 seats (unchanged)

Others 2 seats (unchanged)

leaving the Conservatives short of an overall majority by 8 seats, but a Lib / NDP / Bloc coalition with 159 seats (and a majority of 12 seats)

[punked]

I always believe that when it comes to polls, you need a moving average "poll of polls" in order to get some sense of what is happening. Using the polling data that has been listed on Wikipedia and taking the polls since the election was announced, I get the following (working on the assumption that the election was being held on April 10th 2011)

Conservatives 39.22% (+1.56%)

Liberals 29.84% (+3.21%)

New Democrats 15.36% (-2.81%)

Bloc Quebecois 9.09% (-0.89%)

Greens 5.20% (-1.60%)

Others 1.60% (+0.43%)

Conservative lead of 9.38% on a swing from Con to Lib of 0.83%

According to the calculator used on FairVote Canda that would see the following House:

Conservatives 147 seats (+4 seats)

Liberals 93 seats (+16 seats)

New Democrats 22 seats (-15 seats)

Bloc Quebecois 44 seats (-5 seats)

Greens 0 seats (unchanged)

Others 2 seats (unchanged)

leaving the Conservatives short of an overall majority by 8 seats, but a Lib / NDP / Bloc coalition with 159 seats (and a majority of 12 seats)

Maybe however too close to call points out if you take Nanos out of the picture (and remember you numbers would weight them more because they have a "new" poll everyday) your picture is more like the last election with an average that looks like this.

Cons 38, Libs 27, NDP 18, Bloc 9

Which by their prediction gives you a seat count which is this.

Cons 146 Libs 78 NDP 37 Bloc 47

That is what it looks like if you take Nanos out and weigh every other poll so far evenly. I don't know why Nanos numbers are so skewed it is really bugging me. Maybe it is Momentum however if Nanos numbers looked like every other pollsters right now this election would have a different narrative.

Edited April 10, 2011 by punked

[ToadBrother]

Maybe however too close to call points out if you take Nanos out of the picture (and remember you numbers would weight them more because they have a "new" poll everyday) your picture is more like the last election with an average that looks like this.

Cons 38, Libs 27, NDP 18, Bloc 9

Which by their prediction gives you a seat count which is this.

Cons 146 Libs 78 NDP 37 Bloc 47

That is what it looks like if you take Nanos out and weigh every other poll so far evenly. I don't know why Nanos numbers are so skewed it is really bugging me. Maybe it is Momentum however if Nanos numbers looked like every other pollsters right now this election would have a different narrative.

So we get to have a 300 hundred million dollar election to elect the same Parliament.

[Smallc]

So we get to have a 300 hundred million dollar election to elect the same Parliament.

It looks that way, doesn't it?

[ToadBrother]

It looks that way, doesn't it?

Which means the really interesting part will happen during the three weeks after the election and before the House reconvenes. I'm betting on Layton and Harper trying ignite that all-too-brief pre-election romance to keep those jealous beaus, Duceppe and Iggy, from toppling the show.

I propose we call the upcoming Tory-NDP marriage the "Taxcut-and-Spend" Coalition.

[Smallc]

I propose we call the upcoming Tory-NDP marriage the "Taxcut-and-Spend" Coalition.

In all serious though, a Conservative - NDP marriage just may work.

[Harry]

In all serious though, a Conservative - NDP marriage just may work.

I think there may be some truth to that.

[punked]

Anything has to be better than the Con-Lib Coalition we have had for the last 2 years.