Federal Election Polls

[Big Guy]

I guess one has the right to believe what they want to believe:

http://news.nationalpost.com/news/canada/canadian-politics/skewed-ndp-poll-suggesting-trudeau-trailed-in-his-own-riding-could-be-another-dagger-for-polling-industry

Some of us will wait for the election results.

[cybercoma]

Since 1 out of 5 polls were skewed, all polling is inaccurate? That's a pretty radical stance to take.

[Topaz]

66-70% of Canadians want Harper* gone so, IF that is true, then the idea of voters to vote for the party that is in the lead in the polls, especially those undecided, would put either NDP or the Liberals the winner and so, we just have to wait to see which of the two is leading in the polls the day of the election. Don't count out the Green Party either, she may be the one to put control to one of them.

Edited September 20, 2015 by Charles Anthony
*spelling

[Canada_First]

The Greens who have 1 seat are will have 1 seat Oct. 20 are insignificant.

[Vancouver King]

Nanos newest after debate;

CPC - 30.8

Liberal - 30.3

NDP - 29.1

l

Obviously, the debate did not propel any leader from the pack.

[Big Guy]

Since 1 out of 5 polls were skewed, all polling is inaccurate? That's a pretty radical stance to take.

I Think that when manipulation of polls become a tool of campaigning then one should question their validity.

[ToadBrother]

Nanos newest after debate;

CPC - 30.8

Liberal - 30.3

NDP - 29.1

l

Obviously, the debate did not propel any leader from the pack.

Still orbiting 30%.

[cybercoma]

I Think that when manipulation of polls become a tool of campaigning then one should question their validity.

You have no proof that the poll was "manipulated." They could have just got a bad sample. You're claiming they intentionally tried to be misleading, but there's no evidence of that. And again, that's a single poll with several others that show it's wrong. So in the grand scheme of things it's not even important.

You know, a funny thing happened here Fredericton during the provincial election. The Green Party claimed their candidate was ahead in Fredericton. All the other parties said that was nonsense and their polling didn't show that. It seemed inconceivable too with a Tory incumbent, a former Liberal cabinet minister running for the NDP, and a well-known local personality up for the Liberals.

On election day, the Green Party's internal polling was correct and David Coon won by roughly as much as they said he would.

That's a different situation from this one, where other pollsters went in to try to determine if the NDP lead was as strong as they claimed. No external polling was done in Fredericton before the election, as this was only leaked by the Greens about a week from E-Day.

Anyway, it would be foolish to discount polling because a single one was off. It's also foolish to discount seemingly crazy polls, when there is no better information available. Polls are for the most part accurate when it comes to the popular vote. It's in the prediction of seats that most of the problems come.

[ToadBrother]

You have no proof that the poll was "manipulated." They could have just got a bad sample. You're claiming they intentionally tried to be misleading, but there's no evidence of that. And again, that's a single poll with several others that show it's wrong. So in the grand scheme of things it's not even important.

As was pointed out by Eric Grenier, it was one single poll. We know very little about the full details of the poll, because it was privately commissioned. We don't know how many other similar polls were taken. There are explanations for the poll's seemingly errant results that don't involve CROP being involved in evil electoral skulduggery.

Edited September 21, 2015 by ToadBrother

[cybercoma]

As was pointed out by Eric Grenier, it was one single poll. We know very little about the full details of the poll, because it was privately commissioned. We don't know how many other similar polls were taken. There are explanations for the poll's seemingly errant results that don't involve CROP being involved in evil electoral skulduggery.

I don't think they were either. Those smaller local polls generally have only a few hundred participants, many of whom may refuse to answer. So they could possibly end up with only a couple hundred valid responses. This gives quite a bit of room for non-biased error.

[Big Guy]

You have no proof that the poll was "manipulated." ...

Anyway, it would be foolish to discount polling because a single one was off. It's also foolish to discount seemingly crazy polls, when there is no better information available. Polls are for the most part accurate when it comes to the popular vote. It's in the prediction of seats that most of the problems come.

The only way to find out if a poll was accurate or "crazy" is to compare it with the eventual result - by then it is too late to use it as an indictor of what will happen - because it happened already.

I guess I will continue to be "foolish" and discount different polls until after the election. At that point I will know which ones were correct and which ones were not - but by then - who cares?

[cybercoma]

The only way to find out if a poll was accurate or "crazy" is to compare it with the eventual result - by then it is too late to use it as an indictor of what will happen - because it happened already.

Or, you know, compare it to the 4 other polls taken in that riding.

I guess I will continue to be "foolish" and discount different polls until after the election. At that point I will know which ones were correct and which ones were not - but by then - who cares?

Actually, no you won't. Because the polls don't tell you who people are going to vote for on election day. I guess that's where your misunderstanding is. Each of these polls is only a snapshot of when they are taken. They're not election day results. They give you an idea of who people would vote for if the election were held today, as the questions state. To say polling is useless because one done 5 weeks before the election didn't accurately reflect election day is silly. Edited September 21, 2015 by cybercoma

[ToadBrother]

The only way to find out if a poll was accurate or "crazy" is to compare it with the eventual result - by then it is too late to use it as an indictor of what will happen - because it happened already.

I guess I will continue to be "foolish" and discount different polls until after the election. At that point I will know which ones were correct and which ones were not - but by then - who cares?

Since the poll was commissioned near mid-way through a nearly two-and-a-half month election, how would one even check its veracity in the long term?

[cybercoma]

This is important.

Harper has lost the GTA.

Forum Research indicates voting intention for 905 suburbs is LPC 43% (+16), CPC 33% (-18), NDP 20% (+2). Toronto City LPC 38% (+3), NDP 31% (+1), CPC 27% (-4).

If Harper loses all of those ridings around Toronto that flipped blue last election he's done. This election is going to come down to flipping seats, rather than the perpetual tie we're seeing in the popular vote. In 2011 the Toronto suburbs delivered Harper the win. If he loses those, he's done.

Edit: Link to poll.

Edited September 21, 2015 by cybercoma

[Big Guy]

Since the poll was commissioned near mid-way through a nearly two-and-a-half month election, how would one even check its veracity in the long term?

Or, you know, compare it to the 4 other polls taken in that riding.

Actually, no you won't. Because the polls don't tell you who people are going to vote for on election day. I guess that's where your misunderstanding is. Each of these polls is only a snapshot of when they are taken. They're not election day results. They give you an idea of who people would vote for if the election were held today, as the questions state. To say polling is useless because one done 5 weeks before the election didn't accurately reflect election day is silly.

Putting it that way makes sense. Thanks for the correction.

[ToadBrother]

This is important.

Harper has lost the GTA.

Forum Research indicates voting intention for 905 suburbs is LPC 43% (+16), CPC 33% (-18), NDP 20% (+2). Toronto City LPC 38% (+3), NDP 31% (+1), CPC 27% (-4).

If Harper loses all of those ridings around Toronto that flipped blue last election he's done. This election is going to come down to flipping seats, rather than the perpetual tie we're seeing in the popular vote. In 2011 the Toronto suburbs delivered Harper the win. If he loses those, he's done.

Edit: Link to poll.

We'll see. With a month to go, I'm not willing to call it yet. Some people still believe the Tories have a solid 30% base. I disagree with that. I doubt any party has anywhere near that base. That being said, there is some justice in the claims of some analysts that the NDP may simply have nowhere to grow, which may lead to more battles, particularly in Ontario, between the Tories and Liberals.

All i can see, for all of it, is all three parties orbiting 30%, with a base assumption that the Tories get a marginal "win" (a narrow plurality of seats) if they are at the high end of the error margin. But I question what kind of a win it is to be left with maybe ten or fifteen seats above your competitors, and with a clear negative view of your party among large swathes of the electorate, particularly in areas you were claiming were as blue as the sea even a few weeks ago.

[Derek 2.0]

And from Forum, the firm that historically underpins the Tories by 2-4%:

The Conservatives have one third of the vote (33%), compared to just fewer than 3-in-10 votes for the Liberals and the NDP (29% each). Few will vote Green or for the Bloc Quebecois (4% each) or for other parties (1%). More than one quarter of those who voted NDP in 2011 will vote Liberal this time (26%) and the reverse is true for those who voted Liberal last time, and will vote NDP this election (28%). Just more than one tenth of past Conservatives will vote Liberal this time (13%) but few New Democrats or Liberals from 2011 will be voting Conservative.

A day before the 2011 election, Forum polled the Tories at ~36%......

[Topaz]

The news is saying that BC and Ontario are the provinces to watch and there were many Tory MPs that quit and so the new ones may not have a chance of winning. I think there's going to be more voters this time around and IF the First Nation has a good turn out, they probably won't be voting Tory.

[Triple M]

The Tories might get wiped out in the 416 but will remain competitive in the 905 (except Mississauga). However I honestly didn't understand how they won 9 seats in Toronto in the first place.

[Vancouver King]

Conservatives have pulled even with the NDP in BC according to Nanos, Forum concludes Conservative national support translates into 145 seats. I hope I'm wrong, however, there are less than subtle indications of movement back to Harper's gov't.

[Triple M]

Ipsos Reid

LPC 33

NDP 30

Con 27

+/- 3.4

Seems like a large margin of era but if the libs show signs of momentum than maybe they will get a lot of the time for change voters to swing over.

With that said these polls aren't making to much sense to me because it seems like they come to radically different conclusion on the same questions

[SpankyMcFarland]

Predictions by riding:

http://www.electionprediction.org/2015_fed/index.php

More important than usual in determining whom to vote for.

[ToadBrother]

Ipsos Reid LPC 33 NDP 30 Con 27 +/- 3.4 Seems like a large margin of era but if the libs show signs of momentum than maybe they will get a lot of the time for change voters to swing over. With that said these polls aren't making to much sense to me because it seems like they come to radically different conclusion on the same questions

Every time someone touts the poll that shows their party will ascend the great halls of power, another poll immediately comes out showing that all three parties are mired in the 30% range. Even predictions that someone was going to break out after the last debate have fizzled.

Actually, that's a good word for the leader debates thus far; total fizzle.

I honestly doubt we'll have a good sense if anyone is making great gains until a couple of weeks before the election, when voters start taking it seriously. The first month of this election could best be described as a "phoney war".

[cybercoma]

And from Forum, the firm that historically underpins the Tories by 2-4%:

A day before the 2011 election, Forum polled the Tories at ~36%......

So within the margin of error.

[cybercoma]

Predictions by riding:

http://www.electionprediction.org/2015_fed/index.php

More important than usual in determining whom to vote for.

The predictions are not going to be very good with the parties hovering in a statistical tie. Just keep that in mind. Many predictions had the NDP leading in seats, but I'm not confident in that. It sounds like a lot of seats are going to flip blue to red in the Atlantic Provinces and the 905 from recent polling. This suggests that the LPC will win a minority government despite the tie.