Angus Reid has posted the results of a new poll, showing the Tories 4 points ahead of the Liberals. This was a strong poll for the NDP everywhere except on the Prairies, and indeed Angus pegs the Dippers at 19% nationally, a whopping 4.4% more than this week's EKOS. For the Grits, this was a somewhat weak poll, except in Québec, where they posted a respectable 28%. The Tories did OK, if you discount the dismal 11% in Québec: Angus still has them marginally ahead in Ontario. The 42% for the Bloc was a strong showing.
The aggregate projection is little changed, with the small parties gaining 1 seat each, and the big ones losing 1 each:
CON - 128
LIB - 94
BQ - 49
NDP - 37
Interesting note: the BQ and the NDP are exactly back to their 2008 seat count.
You might ask why the Conservative projection dropped even though this poll was relatively favourable to them. The answer is that the recent midpoint date of this poll (Jan. 25.5) caused a decrease in weight of polls from early January, which also had the Tories higher than recent surveys.
Thursday, January 28, 2010
Wednesday, January 27, 2010
EKOS: Liberals Lead!
This week's EKOS results are similar to last week's, but it is now the Liberals that have a small, statistically insignificant lead (0.5%). The poll was strong for the Tories in Atlantic Canada (3.3% lead; small sample size though), and for the Grits in Québec (29.1%) and Ontario (7.6% lead). As a result, both the Conservatives and the Liberals gain a seat in the projection, while the Bloc loses two:
CON - 129
LIB - 95
BQ - 48
NDP - 36
Several news outlets have also reported the national numbers of a recent Harris-Decima poll, which gave a 1-point edge to the Tories (32-31). However, the poll has not been added to the projection due to a lack of regional numbers.
The national poll average corresponding to the above projection shows a 2.5% Conservative lead. Why does it translate into a large 34-seat advantage? (Remember that the Tories won the 2006 election by 6%, but only 21 seats.) There are several reasons, the first two of which explain the difference with 2006:
1. The Ontario gap (the difference between the Lib-Con gap in Ontario and nationally) is quite small: just over 5%. In 2006, it was just under 11%, while in 2004 and 2008, it was about 6%. A large Ontario gap favours the Liberals, while a small one is good for the Conservatives; by my reckoning, 7-8% would be about neutral.
2. Quirks in the 2008 Ontario results: the Tories won just 3 seats by between 1% and 8%, but 10 seats by 8% to 12%. The projected swing is currently 8%. As you can see, the Liberals are approaching the Red Zone (apt expression here, isn't it?). In fact, if my projection accounted for swing uncertainty, the Liberals would take a few seats from the Tories.
3. Québec: Were the Liberals ahead of the Tories by 8% in any other region with 75 seats, it would get an advantage much greater than 8 seats (for example, it would be about 24 seats out of 106 in Ontario, which is 17 seats out of 75 if pro-rated). That doesn't hold in Québec due to the presence of the Bloc. Of course, it's hard to say if the Liberals would actually lead the Conservatives by 8% in Québec if the Bloc were absent...
4. Liberal/NDP split in MB, SK, BC: The Conservative vote is extremely efficient in these regions due to the split. The Tories get about half the seats in BC with just a third of the vote, and almost all the seats in Saskitoba outside Winnipeg with just 45% of the vote. The Liberals cannot do that in any region.
Wasted Tory votes in Alberta act as a counterweight to the 4 points above, but not enough to balance things out. As a result, to pull even in terms of seats, the Liberals probably need a 1-2 point lead in vote intentions.
CON - 129
LIB - 95
BQ - 48
NDP - 36
Several news outlets have also reported the national numbers of a recent Harris-Decima poll, which gave a 1-point edge to the Tories (32-31). However, the poll has not been added to the projection due to a lack of regional numbers.
The national poll average corresponding to the above projection shows a 2.5% Conservative lead. Why does it translate into a large 34-seat advantage? (Remember that the Tories won the 2006 election by 6%, but only 21 seats.) There are several reasons, the first two of which explain the difference with 2006:
1. The Ontario gap (the difference between the Lib-Con gap in Ontario and nationally) is quite small: just over 5%. In 2006, it was just under 11%, while in 2004 and 2008, it was about 6%. A large Ontario gap favours the Liberals, while a small one is good for the Conservatives; by my reckoning, 7-8% would be about neutral.
2. Quirks in the 2008 Ontario results: the Tories won just 3 seats by between 1% and 8%, but 10 seats by 8% to 12%. The projected swing is currently 8%. As you can see, the Liberals are approaching the Red Zone (apt expression here, isn't it?). In fact, if my projection accounted for swing uncertainty, the Liberals would take a few seats from the Tories.
3. Québec: Were the Liberals ahead of the Tories by 8% in any other region with 75 seats, it would get an advantage much greater than 8 seats (for example, it would be about 24 seats out of 106 in Ontario, which is 17 seats out of 75 if pro-rated). That doesn't hold in Québec due to the presence of the Bloc. Of course, it's hard to say if the Liberals would actually lead the Conservatives by 8% in Québec if the Bloc were absent...
4. Liberal/NDP split in MB, SK, BC: The Conservative vote is extremely efficient in these regions due to the split. The Tories get about half the seats in BC with just a third of the vote, and almost all the seats in Saskitoba outside Winnipeg with just 45% of the vote. The Liberals cannot do that in any region.
Wasted Tory votes in Alberta act as a counterweight to the 4 points above, but not enough to balance things out. As a result, to pull even in terms of seats, the Liberals probably need a 1-2 point lead in vote intentions.
CROP Québec Poll
Cyberpresse reports the results of a CROP poll giving strong results for the Conservatives and NDP, and a weak one for the Bloc. As a result, the Tories gain a seat from the Bloc in the projection:
CON - 128
LIB - 94
BQ - 50
NDP - 36
CON - 128
LIB - 94
BQ - 50
NDP - 36
Monday, January 25, 2010
New Ipsos Poll
ThreeHundredEight.com has reported on a new Ipsos poll showing a 3-point Conservative lead nationally. Relative to other recent polls, this one was strong for the Liberals in Atlantic Canada and Québec, for the NDP in Saskitoba, and for the Tories in Alberta; Ontario and BC numbers are in line with previous polls.
The upshot is that the Liberals are once again up in the seat projection, though by less than in the 4 previous updates (+4, +5, +6, +5; this time +2):
CON - 127
LIB - 94
BQ - 51
NDP - 36
We are approaching the zone where the outcome of an election held today would be in doubt: with EKOS' Ontario numbers (Liberals leading by 4-7 points over Conservatives) and Ipsos' Québec numbers (Grits at 30%), an election would likely come down to a handful of ridings decided by margins of 3% or less. Of course, it's unlikely that actual public mood reflects the best poll for the Liberals in each region, but the fact is that a month ago, even cherry-picking polls like this would still have landed you in rock solid Conservative minority territory.
The upshot is that the Liberals are once again up in the seat projection, though by less than in the 4 previous updates (+4, +5, +6, +5; this time +2):
CON - 127
LIB - 94
BQ - 51
NDP - 36
We are approaching the zone where the outcome of an election held today would be in doubt: with EKOS' Ontario numbers (Liberals leading by 4-7 points over Conservatives) and Ipsos' Québec numbers (Grits at 30%), an election would likely come down to a handful of ridings decided by margins of 3% or less. Of course, it's unlikely that actual public mood reflects the best poll for the Liberals in each region, but the fact is that a month ago, even cherry-picking polls like this would still have landed you in rock solid Conservative minority territory.
Thursday, January 21, 2010
EKOS: Liberals Crack 30
This week's EKOS poll reports a virtual tie between the Liberals and the Conservatives. It's the first time since early September that the Grits are above 30% in a non-Nanos poll. The Liberals now appear to be back in front in Atlantic Canada, and their strength in Ontario (where all polls since the New Year, except for Angus Reid, have them above 36%) is starting to hurt the NDP.
There was also a Léger poll published yesterday showing strong Bloc numbers in Québec (kudos to ThreeHundredEight.com).
The new projection, more than 99% based on polls since the New Year, is:
CON - 128
LIB - 92
BQ - 52
NDP - 36
Suddenly, we're closer to a Liberal minority than a Conservative majority, and LIB+NDP=CON!
It's interesting to note that this comes just as the Left in the U.S. has been slapped by usually friendly Massachusetts voters. We all remember that in Fall 2008, the reverse was true: the Liberals had a miserable election, while the Democrats trounced the Republicans. For all the strong economic and cultural ties between Canada and the U.S., it appears virtually impossible to look at public mood in one country, and say anything useful about it in the other.
There was also a Léger poll published yesterday showing strong Bloc numbers in Québec (kudos to ThreeHundredEight.com).
The new projection, more than 99% based on polls since the New Year, is:
CON - 128
LIB - 92
BQ - 52
NDP - 36
Suddenly, we're closer to a Liberal minority than a Conservative majority, and LIB+NDP=CON!
It's interesting to note that this comes just as the Left in the U.S. has been slapped by usually friendly Massachusetts voters. We all remember that in Fall 2008, the reverse was true: the Liberals had a miserable election, while the Democrats trounced the Republicans. For all the strong economic and cultural ties between Canada and the U.S., it appears virtually impossible to look at public mood in one country, and say anything useful about it in the other.
Monday, January 18, 2010
What a Difference a Week Makes!
I was on vacation over the past week, and upon my return online (though not yet in North America), I realized that I've missed one of the most exciting weeks of polling since last Fall! This is in fact true both in Canada and in the U.S., where the Democrats are in danger of losing a Senate seat in solid blue Massachusetts in a special election tomorrow.
But coming back to the Great White North, four polls were published over the last week: Strategic Counsel, Harris-Decima, EKOS and Angus Reid. Nationally, the Tories are way down, while all opposition parties appear to be up. The numbers seem to be heading back to where they were last summer, and in a big hurry (except for the Bloc, which is doing quite a bit better than in August). This is pretty much what I had predicted for 2010, but I had no idea it would happen in the first two weeks of the year! Prorogation seems to have struck a nerve, much like Ignatieff's election-mongering in September.
Regionally, the Liberals, in a reversal that isn't quite stunning as it is stunningly quick, are now slightly ahead in Atlantic Canada and Ontario. The Bloc got very good numbers in Québec this week, which bumps their seat count up at the expense of the Liberals (recall that the last projection was heavily based on one EKOS poll that had the Grits at 27.5% in Québec - that now looks like an aberration). The Liberals have also reclaimed second place (by a hair over the NDP) in MB/SK and BC, despite the Dippers not doing poorly at all. The Tories are big losers in every single region polled, and they are suddenly no higher than 34% nationally in any of the polls.
The updated aggregate projection is:
CON - 129
LIB - 87
BQ - 53
NDP - 39
This is the worst result for the Tories since mid-September, and the best for the Grits since late September. It is also the best for the Bloc and the NDP since the creation of this blog. The latter suggests that while the Liberals have recovered somewhat, they are not getting as many voters dissatisfied with the government as they should. So the Liberals still have lots of work to do before they can be ready for an election.
Also of note, the Tories' steep drop in BC means that, just like at the end of last summer, the projection model is close to spitting out a seat for Elizabeth May in Saanich--Gulf Islands.
But coming back to the Great White North, four polls were published over the last week: Strategic Counsel, Harris-Decima, EKOS and Angus Reid. Nationally, the Tories are way down, while all opposition parties appear to be up. The numbers seem to be heading back to where they were last summer, and in a big hurry (except for the Bloc, which is doing quite a bit better than in August). This is pretty much what I had predicted for 2010, but I had no idea it would happen in the first two weeks of the year! Prorogation seems to have struck a nerve, much like Ignatieff's election-mongering in September.
Regionally, the Liberals, in a reversal that isn't quite stunning as it is stunningly quick, are now slightly ahead in Atlantic Canada and Ontario. The Bloc got very good numbers in Québec this week, which bumps their seat count up at the expense of the Liberals (recall that the last projection was heavily based on one EKOS poll that had the Grits at 27.5% in Québec - that now looks like an aberration). The Liberals have also reclaimed second place (by a hair over the NDP) in MB/SK and BC, despite the Dippers not doing poorly at all. The Tories are big losers in every single region polled, and they are suddenly no higher than 34% nationally in any of the polls.
The updated aggregate projection is:
CON - 129
LIB - 87
BQ - 53
NDP - 39
This is the worst result for the Tories since mid-September, and the best for the Grits since late September. It is also the best for the Bloc and the NDP since the creation of this blog. The latter suggests that while the Liberals have recovered somewhat, they are not getting as many voters dissatisfied with the government as they should. So the Liberals still have lots of work to do before they can be ready for an election.
Also of note, the Tories' steep drop in BC means that, just like at the end of last summer, the projection model is close to spitting out a seat for Elizabeth May in Saanich--Gulf Islands.
Friday, January 8, 2010
EKOS Confirms Liberal Bounce
Our first update of 2010 shows the extension of the movement observed in the last one of 2009. The first poll of the year, an EKOS, has the Tories down at 33.1%, just 5.3 ahead of the Liberals, who aren’t doing that well either at 27.8% nationally. The NDP and the Bloc are pretty much par for the course, so the big winners of this poll are the Greens, at 13.4%. This contrasts sharply from the 4% reported in the Nanos December poll – a Copenhagen effect?
Despite an uninspiring national tally, the Liberals are probably actually very happy about this poll: they made significant gains in Ontario, where they are marginally ahead of the Tories (0.6 points, which is of course not at all statistically significant), and in Québec, where they stand to pick up some seats with 27.5%. These confirm the good Nanos results for the Grits in the two most important provinces. However, as we’ve seen in most polls since November, the Liberals are second in Atlantic Canada – but they’re probably not too concerned about that, since the Atlantic provinces have followed Ontario’s lead with a lag in the past couple of years. This poll is particularly weak for the Grits in the Prairies: bad for the national numbers, but insignificant seat-wise.
The Tories, however, should be worried: their lead in Ontario seems to have slipped away, and the 14.6% reading in Québec would imply huge losses. The latter, however, is unique to this latest poll, so it could just be an outlier. 34.2% in B.C. isn’t that great either.
For the NDP, this poll suggests continued strength in Atlantic Canada, a very strong result in Saskitoba (as Kady O’Malley would say), but some weakness in Ontario.
As a result, the aggregate projection shows another significant rise for the Grits. This time, the victims are the Tories:
CON – 140
LIB – 81
BQ – 48
NDP – 39
This is the worst projection for the Tories since late September, the worst for the Bloc since early October, and the best for the Grits since early October. The NDP is still doing well, but a few of those 39 seats, mainly in Ontario, have just gotten very vulnerable due to the Liberal rise.
Note: This post was written somewhere over the home state of Sarah Palin. I accessed the EKOS poll from Dulles Airport, and computed the seat projection somewhere over the Yukon. I am posting from Narita Airport, which is not even my final air destination today. Quite exciting for a relatively infrequent traveler and new blogger like me!
Despite an uninspiring national tally, the Liberals are probably actually very happy about this poll: they made significant gains in Ontario, where they are marginally ahead of the Tories (0.6 points, which is of course not at all statistically significant), and in Québec, where they stand to pick up some seats with 27.5%. These confirm the good Nanos results for the Grits in the two most important provinces. However, as we’ve seen in most polls since November, the Liberals are second in Atlantic Canada – but they’re probably not too concerned about that, since the Atlantic provinces have followed Ontario’s lead with a lag in the past couple of years. This poll is particularly weak for the Grits in the Prairies: bad for the national numbers, but insignificant seat-wise.
The Tories, however, should be worried: their lead in Ontario seems to have slipped away, and the 14.6% reading in Québec would imply huge losses. The latter, however, is unique to this latest poll, so it could just be an outlier. 34.2% in B.C. isn’t that great either.
For the NDP, this poll suggests continued strength in Atlantic Canada, a very strong result in Saskitoba (as Kady O’Malley would say), but some weakness in Ontario.
As a result, the aggregate projection shows another significant rise for the Grits. This time, the victims are the Tories:
CON – 140
LIB – 81
BQ – 48
NDP – 39
This is the worst projection for the Tories since late September, the worst for the Bloc since early October, and the best for the Grits since early October. The NDP is still doing well, but a few of those 39 seats, mainly in Ontario, have just gotten very vulnerable due to the Liberal rise.
Note: This post was written somewhere over the home state of Sarah Palin. I accessed the EKOS poll from Dulles Airport, and computed the seat projection somewhere over the Yukon. I am posting from Narita Airport, which is not even my final air destination today. Quite exciting for a relatively infrequent traveler and new blogger like me!
Subscribe to:
Posts (Atom)