babble-intro-img
babble is rabble.ca's discussion board but it's much more than that: it's an online community for folks who just won't shut up. It's a place to tell each other — and the world — what's up with our work and campaigns.

Latest polling thread - May 10, 2012

99 replies [Last post]

Comments

Sean in Ottawa
Online
Joined: Jun 3 2003

I have raised this before.

When a party's support is low it is spread more evenly-- when it increases it spikes creating winnable seats.

All the models presume that a party's support rises and falls in the same shape not getting that a part tends to concentrate as it goes up and spread out as it goes down. There is one exception to this which is some popular incumbents may hang on.

I can explain in detail not only how this happens but why if people need me to go through that again.

The lack of recognition of this is the reason all the models over-estimate parties going down in power and underestimate those going up.

You cannot assume that when a party is at 20% that when it increases by another 10% that this support will go proportionately to the same places where the first 10% is located. That is a nutshell of the fundamental weakness of the models.

Try this: add the next ten percent disproportionately on a curve with more of it going to the already better ridings and when it is going down disproportionately from the more popular ridings. Then add a little bump to any party going down only in the ridings they have incumbents who are popular locally (and I mean little). Avoid giving this to less popular incumbents. then you have the start of a more workable prediction model.

This type of model does not assume all parties have the same concentration at the same level but it does recognize that support levels change the dynamic. If people had applied this approach to the last federal election they would have been able to predict more closely the depth of the loss of the BQ and the gain of the NDP. They would have been able to see the loss of the Cons in 1993 (I actually was very close on that having predicted the PCs at 4 seats only 2 off what they got).

Roughly mathematically the model may work like this: if a party is going up by 10%-- increase them across the board by 5% then take another 5% and concentrate it in the better ridings. When a part is going down by 10% decrease them across the board by 5%, decrease them a further 10% in the better ridings and then put back 5% in the ridings where they have more popular incumbents. Then you have to go in with local intelligence and adjust individual results to account for other factors-- new star candidates, unpopular incumbents who may be punished, local policy impacts etc.

This is an art not a science and there is no simple mathematical formula that can give you a good result -- you have to feel your way. Anyway, that is an indication of what you might need to look at.


David Young
Offline
Joined: Dec 9 2007

So how do these numbers play out if the results in Etobicoke Centre are overturned?

Is that a 416 riding, or a 905 riding?

 


Ippurigakko
Offline
Joined: May 30 2011

Environics today

NDP 36% (+6)
CON 32% (+2)
LIB 19% (-1)
GRN 7% (same)
BQ 6% (-2)

 

Quebec
NDP 45% (+11)
BQ 25% (-5)
CON 15% (+4)
LIB 13% (-2)
GRN 2% (-2)

Ontario
CON 36% (+2)
NDP 32% (+6)
LIB 26% (-1)
GRN 6% (-3)

no info in all provinces yet

Male
CON 36% (+1)
NDP 34% (+7)
LIB 20% (same)
GRN 5% (-3)
BQ 5% (-2)

Female
NDP 38% (+5)
CON 28% (+1)
LIB 17% (-3)
GRN 9% (+3)
BQ 8% (same)

18-29
NDP 35% (-4)
LIB 21% (+8)
CON 21% (-7)
GRN 14% (+6)
BQ 9% (same)

30-44
CON 36% (+10)
NDP 31% (+3)
LIB 17% (-6)
BQ 9% (same)
GRN 7% (same)

45-59
NDP 39% (+11)
CON 34% (+2)
LIB 19% (-2)
GRN 5% (-3)
BQ 4% (-3)

60+
NDP 39% (+11)
CON 37% (+2)
LIB 17% (-4)
BQ 4% (-1)
GRN 3% (-2)

 

http://www.environics.ca/uploads/File/Environics-Research---Federal-Vote...


socialdemocrati...
Offline
Joined: Jan 10 2012

Etobicoke Center is technically 416. But it includes some of the wealthiest areas in the city. It cuts off right at the edge of some subsidized housing, which does drag things a little progressive. I don't think the NDP stand a chance there, but I suspect the Liberals would have a decent shot if there were a new election.


Arthur Cramer
Online
Joined: Nov 30 2010

Sean in Ottawa wrote:

I have raised this before.

When a party's support is low it is spread more evenly-- when it increases it spikes creating winnable seats.

All the models presume that a party's support rises and falls in the same shape not getting that a part tends to concentrate as it goes up and spread out as it goes down. There is one exception to this which is some popular incumbents may hang on.

I can explain in detail not only how this happens but why if people need me to go through that again.

The lack of recognition of this is the reason all the models over-estimate parties going down in power and underestimate those going up.

You cannot assume that when a party is at 20% that when it increases by another 10% that this support will go proportionately to the same places where the first 10% is located. That is a nutshell of the fundamental weakness of the models.

Try this: add the next ten percent disproportionately on a curve with more of it going to the already better ridings and when it is going down disproportionately from the more popular ridings. Then add a little bump to any party going down only in the ridings they have incumbents who are popular locally (and I mean little). Avoid giving this to less popular incumbents. then you have the start of a more workable prediction model.

This type of model does not assume all parties have the same concentration at the same level but it does recognize that support levels change the dynamic. If people had applied this approach to the last federal election they would have been able to predict more closely the depth of the loss of the BQ and the gain of the NDP. They would have been able to see the loss of the Cons in 1993 (I actually was very close on that having predicted the PCs at 4 seats only 2 off what they got).

Roughly mathematically the model may work like this: if a party is going up by 10%-- increase them across the board by 5% then take another 5% and concentrate it in the better ridings. When a part is going down by 10% decrease them across the board by 5%, decrease them a further 10% in the better ridings and then put back 5% in the ridings where they have more popular incumbents. Then you have to go in with local intelligence and adjust individual results to account for other factors-- new star candidates, unpopular incumbents who may be punished, local policy impacts etc.

This is an art not a science and there is no simple mathematical formula that can give you a good result -- you have to feel your way. Anyway, that is an indication of what you might need to look at.


Sean-in-Ottawa steps up to the plate, here's the pitch. SIO Swings and its a hit! Its a deep fly ball to Center Field; that ball is going, going, gone! Another Home Run!

PS: Sean, I've been waiting all day to get home so I could write that. Man, you are "cooking with gas" these days!


NorthReport
Offline
Joined: Jul 6 2008

Those results is Ontario are excellent, and will help to eventually ensure an NDP government.


Mucker
Offline
Joined: Mar 8 2012

Huge numbers from Environics, but what's up with the younger age groups?  Drastic decrease, though the NDP seems to be picking up the older crowd.


Ippurigakko
Offline
Joined: May 30 2011

Yeah some of them discourage to Mulcair because of Cannabis/marijuana or something that what i read on Mulcair's facebook page by anger youth who 420 supporters.

 

Cons lose 7 switch to Lib (I think they try stop the NDP) and NDP lose 4 switch to Green (support 420)

Is it right?


Ottawa Centre-Left
Offline
Joined: Mar 6 2012

The first completely undisputable NDP lead in the recent series of leads, and strangely enough I can't find any media mentions of it despite polls usually getting quick postings...


knownothing
Offline
Joined: Mar 24 2011

Yeah I even searched for Environics and didnt get anything...Canada is totally owned by the wealthy minority


NorthReport
Offline
Joined: Jul 6 2008

Can you imagine what polling would be like if there was only one polling company owned by the rich?

The only reason some of the polls show the NDP doing as well as they are is that they have competition so they are forced to somewhat be truthful.

But there are ways to get these results out nevertheless. Feed the local papers, and publications like the tyee, the georgia straight, even perhaps 24 hours in the Vancouver area.

Also rabble could help - how could we publicize/document these results here so come up on a google search for NDP polling results  or canada polling results 

 


Brachina
Offline
Joined: Feb 15 2012

JeffWells wrote:

nicky wrote:

Forum Research gives almost identical Ontario federal numbers as Ipsos Reid

Con     35

NDP     35

Lib       22

Their seat projection is Con 48, Lib 37, NDP 22.

My first reaction is "Whaa -?" Then I remember how bad our system is and think Yeah, I could see that.

Our system maybe bad, but not that bad, its probably a typo, they reversed the Liberal and NDP seat numbers. I mean that would be laughably ridiculous if they thought that such a huge spread in number could produce the liberals having a huge advantage for the liberals like that. Remember the NDP under Bob Rae •Spits• won a majority with just 36 percent of the vote in Ontario. Demographics may change, but not that much.

NorthReport
Offline
Joined: Jul 6 2008

Ipp

How did you find out about this Environics poll?

Did you just randomly check the Environics website or did someone else refer it to you?

And who do you think Environics did this poll for or is this just a way for them to promote themselves?

 


Ippurigakko
Offline
Joined: May 30 2011

From Twitter by someone post it and i find it. ;)

I believe Environics did it? Environics Research has twitter account

 


adma
Online
Joined: Jan 21 2006

socialdemocraticmiddle wrote:

Etobicoke Center is technically 416. But it includes some of the wealthiest areas in the city. It cuts off right at the edge of some subsidized housing, which does drag things a little progressive. I don't think the NDP stand a chance there, but I suspect the Liberals would have a decent shot if there were a new election.

Actually, because of the seat-specific circumstances which'd trigger a byelection, the Liberals have an excellent shot, particularly if it's Borys Wrbftsplk running again--this is a rare case where left-leaning "strategic voting" still definitely works on the Grits' behalf.  And polling doldrums and all, the Liberals aren't in the laughing-stock basement the PCs were in the last time such a byelection actually occurred (York Centre in 1990, where Maurizio Bevilacqua won by a landslide and his 1988 recount-deadlocked PC opponent was reduced to a catastrophic 10% of the vote--I wouldn't expect Borys to do *that* badly).

I'd agree that the wealthy parts cancel out the NDP's chances of winning; though for the sake of optics, there's nothing wrong with the NDP maximizing its reach...


Policywonk
Offline
Joined: Feb 6 2005

Ottawa Centre-Left wrote:

The first completely undisputable NDP lead in the recent series of leads, and strangely enough I can't find any media mentions of it despite polls usually getting quick postings...

Same lead as Harris-Decima at the top of the page.


Ottawa Centre-Left
Offline
Joined: Mar 6 2012

knownothing wrote:

Yeah I even searched for Environics and didnt get anything...Canada is totally owned by the wealthy minority

 

It's now the next day and I still can't find it posted anywhere in mainstream media!


finois
Offline
Joined: Jun 10 2004

60+
NDP 39% (+11)
CON 37% (+2)
LIB 17% (-4)
BQ 4% (-1)
GRN 3% (-2)

 

This environics Poll of seniors is truly Unbelievable

If this holds, this is truly the NEXT BIG breakthrough FOR US.

Significant inroads in Quebec and atlantic canada over the lst 20 years.

Seniors have always been out of our reach. Their larger percentage actual voting on Election day ususally benefits the tories.

They have the most money to donate to political causes

They are on the most boards / action groups / and have the time to be political active.

 

I've been a member since 1974 and The Seniors have always been out of reach for us other than a small percentage of progressives.

Usually they change their vote only rarely as voting patterns become established.

If they are moving our way as the environics and seniors organizations suggest then This will be monumental

AN ALLIANCE OF WORKERS/SENIORS/SMALL BUSINESS/ACADEMICS/ ENVIROMENTALSITS is what we need to complete

I'm so glad after seeing the MSM go after the Dutch Disease issue that we picked Mulcair. This is going to get extremely dirty and a tough fight as these Cons will stop at nothing i fear to maintain power. Having Brian Topp help elect an NDP government in b.c. will be crucial in ramping up our fight in 2015 and creating some buffer in the west or before if something actually sticks to Harper.


bekayne
Online
Joined: Jan 23 2006

Stockholm wrote:

Forum has a seat projection model that makes no sense whatsoever. I have to think they have a bug in their computer program. Last Ontario election the Liberals took 37%, Tories 35% and NDP 23%. Their latest poll says its now PC 34%, NDP 32% and Liberals 27% - yet their seat projection suggests that the NDP would only gain 5 seats. That is mathematically impossible. If the Liberals have lost 10 points and the NDP has gained 9 points that is a NINETEEN POINT swing. Even if i used the crudest estimate and just added 9% to the NDP and subtracted 10% from the Liberals in every riding in Ontario (in other words the NDP wins every seat that they lost to the Liberals by a margin of less than 19%) - I would see the NDP instantly pick up 12 Liberal seats and that is not even counting some Tory seats that would likely be won over.

If you go by the swing method the Lib-NDP swing would be 19%, PC-NDP 10%, Lib-PC 9%. So NDP picks up 12 Liberal seats & 1 PC seat. PCs pick up 6 Liberal seats. NDP +13, PC + 5, Lib -18

http://en.wikipedia.org/wiki/Ontario_general_election,_2011#Marginal_seats


flight from kamakura
Offline
Joined: Nov 24 2006

big big mountain to overcome, considering - http://www.cbc.ca/news/politics/canadavotes2011/#/130


jerrym
Offline
Joined: May 30 2009

If there is a turnaround in seniors support towards the NDP, its likely related to the realization that the Cons feel they have seniors in the bank and are now finding out that a significant number of seniors care enough about the OAS eligibility age being raised and about medicare being totally abandoned to provincial control where many governments are ready to privatize it. Many seniors are also worried about the alarming unemployment among their grandchildren and in some cases, such as mine, their children (I am 64 and have two children 18 and 11 years old) after seeing what the austerity program has done in Europe and now that they see it implemented in Canada.


Ippurigakko
Offline
Joined: May 30 2011

Forum Research on Quebec provincial polls - May 16


NDP 39% (-3)
BQ 21% (+2)
LIB 17% (same)
CON 16% (same)
GRN 4% (same)

Montreal
NDP 41%
BQ 20%
LIB 19%
CON 13%
GRN 4%

Quebec City
NDP 35%
CON 28%
BQ 16%
LIB 15%
GRN 2%

North Shore
NDP 36%
BQ 22%
LIB 18%
CON 17%
GRN 5%

South Shore
NDP 41%
BQ 23%
LIB 14%
CON 14%
GRN 5%

Francophone
NDP 40%
BQ 23%
LIB 15%
CON 15%
GRN 3%

Non-Francophone
NDP 33%
LIB 32%
CON 19%
GRN 7%
BQ 4%

18-34
NDP 40%
BQ 24%
CON 13%
LIB 12%
GRN 7%

35-44
NDP 42%
BQ 18%
CON 18%
LIB 14%
GRN 4%

45-54
NDP 35%
BQ 26%
LIB 19%
CON 15%
GRN 3%

55-64
NDP 43%
BQ 18%
LIB 18%
CON 15%
GRN 2%

65+
NDP 37%
LIB 23%
CON 19%
BQ 15%
GRN 3%

Male
NDP 39%
BQ 21%
CON 20%
LIB 15%
GRN 3%

Female
NDP 39%
BQ 20%
LIB 19%
CON 12%
GRN 5%

Current Federal Party Preference
NDP -> PQ 30%, PLQ 27%, CAQ 26%, QS 14%, GRN 4%
BQ -> PQ 75%, CAQ 12%, QS 9%, PLQ 3%, GRN 1%
LIB -> PLQ 76%, PQ 11%, CAQ 9%, QS 3%, GRN 2%
CON -> PLQ 60%, CAQ 26%, PQ 9%, QS 3%, GRN 1%
GRN -> PVQ 33%, PQ 24%, CAQ 19%, QS 14%, PLQ 10%

Current Provincial Party Preference
PLQ -> LIB 40%, NDP 29%, CON 26%, BQ 2%, GRN 1%
PQ -> BQ 50%, NDP 35%, LIB 6%, CON 5%, GRN 3%
CAQ -> NDP 52%, CON 21%, BQ 14%, LIB 8%, GRN 3%
QS -> NDP 57%, BQ 21%, LIB 6%, GRN 5%, CON 5%
PVQ -> GRN 37%, NDP 34%, LIB 10%, BQ 9%, CON 8%

 


Brachina
Offline
Joined: Feb 15 2012
Its strange how the most solid chunks of NDP support come from such idealogical different Quebec parties, the right wing CAQ and the left wing socialist QS, plus around 30 percent of each of the other parties.

Ippurigakko
Offline
Joined: May 30 2011

Yeah, it is weird.

Like last year, swing to NDP from 30% ex-BQ (490,203), 27% NDP (441,203 in previous election 2008), 20% ex-LIB (321,247), 11% ex-Independent (174,072), 10% ex-CON (156,911), 3% ex-Green (44,851)

total 1,628,483

BQ 30%, NDP 27%, LIB 20%, IND 11%, CON 10%, GRN 3% = NDP 43% (May 2, 2011)
PQ 30%, PLQ 27%, CAQ 26%, QS 14%, GRN 4%, IND 1% = NDP 39% (May 16, 2012)

similar now eh?

 

Edit:

I doing Montreal and Quebec City last election 2011 and calculated it

May 2, 2011 to May 16, 2012 (Forum Research)

Montreal
NPD 39% to 41% (+2)
LIB 25% to 19% (-6)
BQ 19% to 20% (+1)
CON 13% to 13% (same)
GRN 2% to 4% (+2)

Quebec City
NPD 42% to 35% (-7)
CON 27% to 28% (+1)
BQ 21% to 16% (-5)
LIB 8% to 15% (+7)
GRN 2% to 2% (same)


Wilf Day
Offline
Joined: Oct 31 2002

adma wrote:
the Liberals have an excellent shot, particularly if it's Borys Wrbftsplk running again.

Wrzesnewskyj, pronounced zhez-NEF-ski. You're welcome.


JeffWells
Offline
Joined: Dec 15 2003

An Abacus poll, their first since March, conducted May 15-16:

Cons: 37% (unchanged)

NDP: 35% (+7)

Lib: 17% (-3)

Quote:

Coletto suggests, at least initially, Mulcair's NDP has been able to "consolidate that anti- Harper, anti-Conservative, centre- left vote."

 

Meantime, 37% of respondents say they have a favourable impression of Prime Minister Stephen Harper, while Mulcair's favourability rating has risen to 36%.

 

 

http://www.thesudburystar.com/ArticleDisplay.aspx?e=3567564

 

 


Ippurigakko
Offline
Joined: May 30 2011

Gee Abacus Data seem favour Conservative top.

but it is dead heat, yea?

Abacus Data May 15-16


CON 37% (same since March)
NDP 35% (+7)
LIB 17% (-3)
BQ   6% (-2)
GRN 5% (-2)

Atlantic
NDP 39% (+11)
LIB 31% (+4)
CON 27% (-12)
GRN 4% (-2)

Quebec
NDP 41% (+13)
BQ 23% (-10)
CON 18% (+1)
LIB 13% (-3)
GRN 5% (-1)

Ontario
CON 38% (same)
NDP 35% (+9)
LIB 22% (-6)
GRN 6% (-2)

Prairies
CON 55% (-5)
NDP 28% (+10)
LIB 10% (+2)
GRN 7% (+4)
 
Alberta
CON 68% (+4)
NDP 19% (+1)
LIB 7% (-5)
GRN 5% (-2)

BC
CON 43% (+4)
NDP 34% (-2)
LIB 18% (+1)
GRN 5% (-3)

Men
CON 41%
NDP 30%
LIB 18%
GRN 6%
BQ 5%

Female
NDP 41%
CON 32%
LIB 16%
BQ 7%
GRN 4%

18-29
NDP 37%
CON 27%
LIB 22%
GRN 7%
BQ 7%

30-44
CON 37%
NDP 35%
LIB 16%
BQ 7%
GRN 5%

45-59
NDP 37%
CON 37%
LIB 16%
GRN 5%
BQ 5%

60+
CON 44%
NDP 33%
LIB 16%
BQ 4%
GRN 3%

http://abacusdata.ca/2012/05/22/canadian-politics-conservatives-and-ndp-...

 

Edit: LOL! Jeff you beat me post it same time. but oh well lol


JeffWells
Offline
Joined: Dec 15 2003

But you beat me with the all-important breakdown, thanks!

Also, interesting Abacus titles its story "Conservatives and NDP neck and neck," and the first media report I found called it "NDP closing in on Tories."

 


Stockholm
Offline
Joined: Sep 29 2002

Your BC numbers are wrong 


Arthur Cramer
Online
Joined: Nov 30 2010

I don't believe the BC numbers.


Comment viewing options

Select your preferred way to display the comments and click "Save settings" to activate your changes.
Login or register to post comments