Author
|
Topic: Federal polling thread
|
Northern54
rabble-rouser
Babbler # 5325
|
posted 25 May 2007 05:05 PM
http://tinyurl.com/2z9pgsCon 35% Lib 28% NDP 18% Green 9% BLOC 9% Also, latest Decima poll (a couple of days or so before) http://tinyurl.com/22ykjm Con 33 Lib 31 NDP 17 Green 10 BLOC 8 [ 25 May 2007: Message edited by: Garth Brasseur ] [ 30 May 2007: Message edited by: Michelle ]
From: Yellowknife | Registered: Mar 2004
| IP: Logged
|
|
John K
rabble-rouser
Babbler # 3407
|
posted 25 May 2007 06:03 PM
Yawn. Wake me up in September when polls may begin to count for something again.Can anyone explain how Angus Reid can conduct an online poll with any degree of scientific accuracy? Unsolicited advice to Angus - stick to reporting other people's polls. I take it you forgot to ask Ipsos for a non-compete payment when you sold the company.
From: Edmonton | Registered: Nov 2002
| IP: Logged
|
|
|
Stockholm
rabble-rouser
Babbler # 3138
|
posted 25 May 2007 07:57 PM
quote: Can anyone explain how Angus Reid can conduct an online poll with any degree of scientific accuracy?
As matter of fact on-line polls can be highly accurate. They are not to be confused with those silly newspaper o-called internet polls where anyone can log in and vote - those are not polls at all - they are parlour games. More and more on-line polls are being done using random sample methodologies. On-line polls in the last elections in the US and UK were at least as accurate as anything done over the phone and Angus Reid's on-line polls in the recent Quebec and Manitba elections were the most accurate of all.
From: Toronto | Registered: Sep 2002
| IP: Logged
|
|
|
|
|
|
West Coast Lefty
rabble-rouser
Babbler # 3697
|
posted 26 May 2007 10:07 AM
There's no mystery IMHO - the NDP had a brutal few months in late 06 with the London by-election disaster, the total lack of strategy to respond to the emergence of Elizabeth May, Jack's petty and hyper-partisan reaction to Stephane Dion winning the leadership, etc. Those events caused NDP support to drop to near the rock-bottom NDP base that we were stuck in during the Alexa years - around 8-12 percent of the vote. In the spring of 07, Dion has faded, May has made a series of gaffes, Harper has gone up and down, and most important, Jack and the NDP have regained their focus and have distinguished themselves from the other parties on key issues like Afghanistan, Kyoto, prosperity gap, etc. Layton doubled our support from 8% in the 2000 election to 15-17% support in 2004 and 2006. Those "Layton New Democrats" briefly switched to the Libs and Greens for a few months and now have mostly come back "home" to the NDP, as the latest polling numbers demonstrate.
From: Victoria, B.C. | Registered: Feb 2003
| IP: Logged
|
|
|
FraserValleyMan
rabble-rouser
Babbler # 13970
|
posted 26 May 2007 12:36 PM
quote: Originally posted by thorin_bane: I don't think we can afford any more Harper,... it takes the libs 5 years to do their neo-lib plans the cons do it in 2 months, plus they are extraordinarily secretive for an 'open and accountable' government.
This is a viewpoint I don't accept. It simply is not in accordance with the facts. The Conservatives have done some stupid things, like killing the child care plan and cancelling literacy programs. But they did some good things too, like the Accountability Act and taxing the income trusts. The latter action shows that on basic fiscal and economic policy, any claim the big L Liberals have to being more centrist than the Conservatives is the exact opposite of the truth, and given who funds the parties no one can claim to be surprised. [ 26 May 2007: Message edited by: FraserValleyMan ] [ 26 May 2007: Message edited by: FraserValleyMan ]
From: Port Coquitlam, BC | Registered: Mar 2007
| IP: Logged
|
|
|
FraserValleyMan
rabble-rouser
Babbler # 13970
|
posted 26 May 2007 10:53 PM
quote: Originally posted by thorin_bane: The accountability act is a farce ... Wow harper is going to tax income trusts!!! Thta is only because there would be no tax base in 2 years to support the military keynsian economics of CONs, ...
The Accountability Act is not a farce. It represents a major improvement in campaign financing by restricting all contributions to individuals, something people have talked about for ages. Taxing the income trusts is a necessary element of tax fairness. What the revenues are used for is a separate policy choice on which reasonable people can have honest disagreements.
From: Port Coquitlam, BC | Registered: Mar 2007
| IP: Logged
|
|
FraserValleyMan
rabble-rouser
Babbler # 13970
|
posted 26 May 2007 10:59 PM
quote: Originally posted by West Coast Lefty: There's no mystery IMHO - the NDP had a brutal few months in late 06 with the London by-election disaster, the total lack of strategy to respond to the emergence of Elizabeth May, Jack's petty and hyper-partisan reaction to Stephane Dion winning the leadership, etc.
The was a upsurge in Liberal support around the time of the Liberal convention, that's true. And that surprised me since I didn't expect an old fashioned delegated leadership convention, complete with wheeling and dealing and endorsements from the Joe Volpe type candidates to be a big seller with the public. But for a few weeks, it was actually flying for them.
For most voters I really find it hard to believe that Elizabeth May is on their radar. Same thing with Jack's reaction to Dion's win, which wasn't so much his doing as that of the party communications "experts", ... but still, I find it hard to believe that even two voters in 100 have an opinion one way or the other on how the NDP reacted to Dion's winning the Liberal leadership. And today, the idea that anyone at all would be worried how someone reacted to Dion doing whatever is a bit of a stretch!
From: Port Coquitlam, BC | Registered: Mar 2007
| IP: Logged
|
|
saga
rabble-rouser
Babbler # 13017
|
posted 26 May 2007 11:15 PM
quote: Originally posted by Stockholm:
As matter of fact on-line polls can be highly accurate. They are not to be confused with those silly newspaper o-called internet polls where anyone can log in and vote - those are not polls at all - they are parlour games. More and more on-line polls are being done using random sample methodologies. On-line polls in the last elections in the US and UK were at least as accurate as anything done over the phone and Angus Reid's on-line polls in the recent Quebec and Manitba elections were the most accurate of all.
Ya but I was part of the random sample and (gulp!) I didn't respond. Response rate is always an issue, and should be reported along with the stats. Selective responding by certain types of voters is likely too, and can't be easily determined. Look at the differences ... 35-28 33-31 I would say neither of them is likely accurate, or only accurate within 3 points either way perhaps ... which makes them useless in detecting anything less than a seven point spread. Essentially these are the same results, but I expect the tories will prefer the Angus Reid numbers. [ 26 May 2007: Message edited by: saga ]
From: Canada | Registered: Aug 2006
| IP: Logged
|
|
|
Stockholm
rabble-rouser
Babbler # 3138
|
posted 27 May 2007 08:03 AM
quote: Ya but I was part of the random sample and (gulp!) I didn't respond. Response rate is always an issue, and should be reported along with the stats. Selective responding by certain types of voters is likely too, and can't be easily determined.
You could just as easily have been part of the random sample for a telephone survey and have refused to take part when they called you. There is no evidence that people who respond to surveys vote any differently from people who do not. - whether its an Internet or a telephone survey. The proof is in the pudding, Angus Reid did on-line surveys just before the recent Manitoba and Quebec elections and in both cases those surveys were even closer to the final results than were the final published telephone polls.
From: Toronto | Registered: Sep 2002
| IP: Logged
|
|
josh
rabble-rouser
Babbler # 2938
|
posted 27 May 2007 08:06 AM
quote: Originally posted by Stockholm:
As matter of fact on-line polls can be highly accurate. They are not to be confused with those silly newspaper o-called internet polls where anyone can log in and vote - those are not polls at all - they are parlour games. More and more on-line polls are being done using random sample methodologies. On-line polls in the last elections in the US and UK were at least as accurate as anything done over the phone and Angus Reid's on-line polls in the recent Quebec and Manitba elections were the most accurate of all.
Not in the US. The Zogby online polls done for last year's midterm elections were way off base. They're not worthy of being considered on a par with phone polls.
From: the twilight zone between the U.S. and Canada | Registered: Aug 2002
| IP: Logged
|
|
|
|
KenS
rabble-rouser
Babbler # 1174
|
posted 27 May 2007 01:59 PM
quote: Look at the differences ... 35-28 33-31I would say neither of them is likely accurate, or only accurate within 3 points either way perhaps ... which makes them useless in detecting anything less than a seven point spread.
Margin of error of 3 points covers most of the polls.
From: Minasville, NS | Registered: Aug 2001
| IP: Logged
|
|
FraserValleyMan
rabble-rouser
Babbler # 13970
|
posted 27 May 2007 08:34 PM
quote: Originally posted by Red Partisan: Back in the days of Jean Chretien, when a man wanted to look good he baught a Hugo Boss or Georgio Armani suit. Now it's Harper, all you need is Moore's.
I think a lot of Red Party disciples and devotees were shopping at Harry Rosen's, ... especially after they cashed in on their Sponsorship loot!
From: Port Coquitlam, BC | Registered: Mar 2007
| IP: Logged
|
|
Malcolm
rabble-rouser
Babbler # 5168
|
posted 27 May 2007 09:47 PM
quote: Originally posted by John K:
Can anyone explain how Angus Reid can conduct an online poll with any degree of scientific accuracy?
Glad to see the Liberals are still able to get their talking points out to the field.
Spare us the oft-repeated Liberal lie that the Angus Reid polls are the same sort of self-selecting crap as the web-poll in the Toronto Sun. They aren't, and Angus Reid should sue your ass for slander. Reid uses online contact in much the same way as everyone else uses telephone contact. The respondents are not self-selecting. I suspect many people with knowledge of polling and sampling methodology see potential pitfalls in online sampling - just as many in an earlier age saw potential problems in switching from in person interviews to telephone contact. But your blithe parroting of Liberal talking points really is slanderous. [ 27 May 2007: Message edited by: Malcolm French, APR ]
From: Regina, SK | Registered: Mar 2004
| IP: Logged
|
|
Malcolm
rabble-rouser
Babbler # 5168
|
posted 27 May 2007 09:55 PM
quote: Originally posted by saga:
Look at the differences ... 35-28 33-31 [ 26 May 2007: Message edited by: saga ]
You Liberals really are getting desperate, trying to show wild fluctuations that aren't there.
Yes indeed, Saga - let us look at the results. In one poll, the Conservatives have 35%. In the other they have 33%. Both polls report a margin of error of about 3%, meaning that in one poll the Conservatives have 32-38% and in the other they have 30-36%. The numbers are 2 points apart. I guess I'd have to call the two polls CONSISTENT within the margin of error. Again, In one poll, the Liberals have 28%. In the other they have 31%. Both polls report a margin of error of about 3%, meaning that in one poll the Liberals have 25-31% and in the other they have 28-34%%. The numbers are 3 points apart. I guess I'd have to call the two polls CONSISTENT within the margin of error. Do I need to go on?
From: Regina, SK | Registered: Mar 2004
| IP: Logged
|
|
|
brookmere
rabble-rouser
Babbler # 9693
|
posted 28 May 2007 11:59 PM
quote: Originally posted by Malcolm French, APR:
I suspect many people with knowledge of polling and sampling methodology see potential pitfalls in online sampling - just as many in an earlier age saw potential problems in switching from in person interviews to telephone contact.
Precisely.Look, if you're doing online polling, isn't it obvious that you're restricting your sample to the computer-literate? And just maybe that group isn't necessarily representative of the voting public. Reminds me of the famous "Dewey defeats Truman" headline in 1948. That was based on opinion polls that showed Dewey ahead. Just one problem - those polls were done by telephone, and in 1948 a lot of Americans still didn't have phones. And those who didn't were much more likely to vote for Truman.
From: BC (sort of) | Registered: Jun 2005
| IP: Logged
|
|
|
|
|
Malcolm
rabble-rouser
Babbler # 5168
|
posted 29 May 2007 07:29 PM
The real point here is that the Liberals, with their usual contempt for honesty, are trying to equate the Angus Reid polls with their online methodology to the usual single wedge-question "polls" we find on an assortment of websites where any and all are free to "freep" the poll.Why are the Liberals doing this? Because former Liberal pollster Angus Reid isn't showing any Liberal growth. This is a common Liberal response. When the facts aren't to your liking, slander somebody.
From: Regina, SK | Registered: Mar 2004
| IP: Logged
|
|
|
Michelle
Moderator
Babbler # 560
|
posted 30 May 2007 03:32 AM
These polling threads make me crazy.There's no way to know from the thread title or even the information in the opening post what the poll is about, whether it's across the country, whether it's for a certain province, etc. Seriously folks, why do we need to start a new thread every time a new poll comes out? I propose putting the federal or provincial jurisdiction in the title so that everyone knows what the thread is about. Don't put a date or a polling company in the thread title, and that way the thread can continue being used for future polls in the same jurisdiction. Do people agree with me on this?
From: I've got a fever, and the only prescription is more cowbell. | Registered: May 2001
| IP: Logged
|
|
Stockholm
rabble-rouser
Babbler # 3138
|
posted 30 May 2007 06:48 AM
quote: Because, at least in the U.S., they've been proven to be highly inaccurate.
Who cares. This is Canada.
From: Toronto | Registered: Sep 2002
| IP: Logged
|
|
|
HeywoodFloyd
token right-wing mascot
Babbler # 4226
|
posted 30 May 2007 06:55 AM
quote: Originally posted by BetterRed:
Can you please decipher this cryptic statement?
It's so cutely elitist.
From: Edmonton: This place sucks | Registered: Jun 2003
| IP: Logged
|
|
josh
rabble-rouser
Babbler # 2938
|
posted 30 May 2007 07:00 AM
quote: Originally posted by Stockholm:
Who cares. This is Canada.
SES is not good enough for you?
From: the twilight zone between the U.S. and Canada | Registered: Aug 2002
| IP: Logged
|
|
|
Stockholm
rabble-rouser
Babbler # 3138
|
posted 30 May 2007 08:50 AM
I'm told that it is getting to be over 80% and that furthermore, the voting intentions of those without an internet connection do not differ significantly from the voting behaviour of those who do and once some demographic weighting is done - it is not a problem at all.Anytime you have a number of polls released, they will cover a range of results - but there is no reason to believe that on-line polls are systematically any less accurate than those done by telephone. This latest survey from Angus Reid strikes me as quite credible - it has all the parties within a point of what they got in the last election and this is the same company that used on-line polls to be almost exactly accurate in the recent Manitoba and Quebec elections. Instead of trying to be dismissive of a poll just because it was done on-line - why not accept that technology is on the move and that this will only become a more and more common polling methodology. Is it perfect? NO. But as we all know, there are also all kinds of problems with telephone polls...and yet if we look at recent elections in France, Ireland, manitoba, Quebec etc....the polls were all pretty damn close to the final results.
From: Toronto | Registered: Sep 2002
| IP: Logged
|
|
|