Saturday, October 1, 2011

McGuinty down to 54 seats, Aylward to 2

Five polls were released over the last 24 hours, four of them for Ontario and one of them for Newfoundland and Labrador.


All four polls in Ontario showed a statistically insignificant gap between the Liberals and the Progressive Conservatives. It's a remarkably close race.

The oldest poll was conducted online by Environics for the Canadian Press between September 26 and 28, interviewing 1,000 people. It found that the Progressive Conservatives stood at 36% to the Liberals' 35%, with the New Democrats trailing at 25% and the Greens at 3%. The poll found the Liberals ahead in Toronto and tied with the PCs in the GTA. The New Democrats were ahead in the north and the Tories in eastern and central Ontario.

The next poll was by Nanos Research for CTV and the Globe and Mail. This telephone poll was conducted on September 28 and 29, and surveyed 600 people. Nanos found the Liberals leading with 37.7% and the Progressive Conservatives trailing with 34.4%. The NDP stood at 25.5% and the Greens at 1.8%. Compared to Nanos's last poll from early September, the Liberals are down 0.4 points and the PCs 0.3 points. The NDP is up 1.2 points, all insignficant.

Then Sun Media released a Léger Marketing poll conducted online on September 28 and 29 and that interviewed 1,012 people. In this poll, the Tories led with 34% to the Liberals' 32%. The New Democrats came in at 29% while the Greens were at 5%. This is a two point drop for the Tories and a one point drop for the Liberals since Léger's mid-September poll, while the NDP picked up three points.

Regionally, Léger found the Tories to be leading in eastern Ontario (38% to the Liberals' 31%) and southwestern Ontario (33% to the NDP's 30%). The Liberals led in Toronto and the GTA (38% to the Tories' 34%) while the New Democrats led in Hamilton (38% to the PCs' 30%) and northern Ontario (39% to the Liberals' 30%).

Finally, Angus-Reid and the Toronto Star released a poll this morning. Conducted online between September 28 and 30 and interviewing 1,002 people, Angus-Reid pegs PC support at 34% with the Liberals at 33% and the NDP at 26%. The Greens trail in fourth with 6%. This is a two point drop for the Tories since Angus-Reid's mid-September poll and a one point gain for the Liberals. The NDP are unchanged.

Regionally, the Tories lead in the 905 (36% to the Liberals' 33%), southwestern Ontario (39% to the Liberals' 32%), eastern Ontario (36% to the Liberals' 30%), and northern Ontario (37% to the NDP's 33%). The Liberals lead in the 416 (46% to the NDP's 27%) while the NDP is ahead in the Hamilton/Niagara region (34% to the Tories' 28%).

It's a lot of new data. Interestingly, all of the online polls show an insignificantly slim lead for the Tories, while the one telephone poll has a less insignificant (but still within the margin of error) lead for the Liberals. Support for the Tories is very consistent, ranging between 34% and 36%. Liberal support is pegged at being between 32% and 38%, a much wider range that means the difference between winning and losing. NDP support is generally solid, at between 25% and 29%.

In terms of trends, this is Environics's first poll of the campaign so they have nothing to tell us on that score. But the other three all show stability for the two main parties, with changes being comfortably within the margin of error. However, they also all show the Tories slipping a little and two of them have the NDP gaining. Taken together, it would appear that Liberal support is solid, PC support is sliding, and NDP support is on the rise.

Regionally, Environics, Léger, and Angus-Reid all put the Liberals ahead in Toronto and running neck-and-neck with the Tories in the GTA. They all show the PCs leading in eastern, central, and southwestern Ontario, while the NDP is ahead in the Hamilton region. Two of the three have the NDP leading in the north, while the other shows a close race. For the smaller sample sizes that regions have, this is pleasantly consistent.
The projection has only moved slightly, however, as these polls have generally fallen on either side of the last set of numbers. The Progressive Conservatives still have a 0.1 point edge (if it can be called that) with 34.2% support. The Liberals follow with 34.1%.

Both parties have dropped 0.4 points since the last projection of September 27.

The New Democrats are up 2.2 points to 25.9%, while the Greens are down 1.5 points to 4.5%.

Only one seat has changed hands, with the Liberals losing it to the New Democrats. This puts the Liberals at a projected 54 seats, the bare minimum for a majority government. With a Liberal Speaker, the legislature would be tied at 53-53 between the government and the opposition.

That opposition is now projected to be made up of 32 Progressive Conservatives and 21 New Democrats.

The one seat that switched hands was in southwestern Ontario. The Liberals are now projected to win 11 in the region, with eight going to the Tories and two to the New Democrats.

At dissolution, the Liberals held 15 seats in the region and the PCs six.

There are enough close races to keep the result up in the air. The Liberals lead in 13 of them and trail in four others, while the PCs lead in four and trail in 10.

This puts the Liberal range at between 41 and 58 seats. The Progressive Conservative range is between 28 and 42 seats.
This means that the Liberals could win anything from a majority to one less seat than the Tories. The odds are still heavily favouring a Liberal government, however.

The New Democrats lead in one close race and trail in four others, meaning they could win between 20 and 25 seats.

Newfoundland and Labrador

One poll from MarketQuest Omnifacts Research was released yesterday for Newfoundland and Labrador, and it was a little controversial.

Taken online and over the telephone between September 28 and 30, the poll of 464 people pegged Progressive Conservative support at 54%, up one point since their last poll earlier in the campaign. The New Democrats were up four points to 33%, while the Liberals fell five points to a woeful 13%.

The controversy erupted when the Liberals questioned the credibility of the poll. But in my experience, when a party starts saying that the numbers in a poll are cooked it is more of a sign that the party itself is cooked. Standard operating procedure is usually to claim that "I don't pay attention to the polls."

Many polling firms employ people with former ties to political parties and governments, and governments are often clients. It's just how it goes, and it seems unlikely that a parent company would sabotage its polling firm's credibility by putting out numbers which can be shown to be false on election day. How easily people are ready to believe that professionals are capable of lying, cheating, and breaking every ethical code in the book for marginal partisan gain is, frankly, depressing.
On that cheery note, let's take a look at how this poll has shifted the projection. The Progressive Conservatives are now projected to take 53.8% of the vote, up 0.8 points since the last projection of September 21.

The New Democrats are up 4.1 points to 32.8%, while the Liberals are down 4.8 points to 13.2%.

With this shift in support, the Liberals have dropped two seats to the Tories. The Progressive Conservatives are now projected to win 42 seats, with the New Democrats taking four and the Liberals two.

The PC gains come in the regions of the Avalon and Burin Peninsulas and the main part of Newfoundland.

The Progressive Conservatives are now projected to win eight seats in the Avalon and Burin Peninsulas (excluding St. John's), with one seat going to the NDP. In the rest of Newfoundland, the Tories are projected to win 19 seats and the Liberals and NDP one apiece.

In these two regions, the Progressive Conservatives held 27 of the seats at dissolution, while the Liberals held three. That score is still 27 for the PCs, but the Liberals are down to one and the NDP stands at two.

There are only a few projected close races at this point. The Tories lead in one of them and trail in another. The Liberals lead in one of their own, while the NDP trails in one.

This puts the Progressive Conservative range at between 41 and 43 seats. The New Democrat range is now between four and five seats, while the Liberals are projected to win between one and two seats.

In other words, it appears that the NDP are well positioned to form the Official Opposition.

It should be an exciting week. Prince Edward Island is heading to the polls on Monday, while Manitoba votes on Tuesday. There may be another poll out of Manitoba before that happens.

Ontario goes to the polls on Thursday, and I suspect we will hear from every firm at least once before that happens. Newfoundland and Labrador then votes next Tuesday, and I have an inkling we could have three or four more polls out of the province before then. That would be welcome, as we have only heard from MQO Research since the campaign began

Final projections will be posted on the day of each province's vote. Post-mortems may have to wait until after the Newfoundland and Labrador vote, but I am very curious to see how the projection model performs. PEI, Newfoundland and Labrador, and Manitoba are all small provinces which makes things more unpredictable. Newfoundland and Labrador has the NDP in uncharted territory and Manitoba is incredibly close. Ontario is even closer, so plenty of surprises could still be in store.


  1. So basically nothing has really changed. I predict that when the votes are actually counted there will be a slight swing towards the Liberals. Even papers on the right, like the Toronto Sun, just can't warm to Hudak.

    My prediction is a very small, 1 or 2 seat, Liberal majority.

  2. Windsor West says the NDP has 41% and Liberals 37% why are the liberals still projected to win (highlighted)?

  3. Oops, forgot to fix the graphic. That is the seat that the NDP gained.

  4. So you're going to ignore all of those Forum Research riding polls published a few days ago? They show different numbers than the big poll they did with the 40,000 people. According to those polls, the NDP now leads in Sudbury and Bramalea-Gore-Malton. This proved that the large poll they did was poor for predicting urban ridings with crossed telephone exchanges, but I reckon it's pretty accurate for rural ridings. That means you're wrong about Glengarry-Prescott-Russell, Prince Edward-Hastings, Northumberland-Quinte West, Ancaster-Dundas-Flamborough, Niagara Falls, Brant and pretty much all of the ridings in SW Ontario you've called Liberal (except Perth-Wellington). The north was probably under polled, but the Tories are up in Kenora-Rainy River as well.

  5. Any word on when we might hear from Harris Decima? I was phone polled by them in the middle of the week and still haven't seen any results posted.

  6. If the Leger poll indicates a trend upward for the NDP could it be possible that they could win more than 25 seats? What is the Threshold for them to be competitive along with the PC and Lib?

  7. Earl,

    We'll see. Riding polls did not do very well in the federal election, and by the time people vote the Forum riding polls will be out of date.

    I'm not sure how two sets of contradictory polling by the same firm taken over a short period of times "proves" anything.


    Harris-Decima polls continuously and their polls generally span one or two weeks. So, I suspect that we'll hear from them this week.

  8. I agree with some of your numbers for Newfoundland and Labrador, particularly the Liberals two seats. Those two districts should definitely go Liberal but at their current numbers I don't think others will, supposedly the NDP may take enough support away from the Liberals in the Straits White Bay North for the PCs to squeak out a win. Grand Falls Windsor Buchans however will not go NDP, the 2007 numbers in that districts got a bit screwed up. First the Liberal candidate died half way through the campaign, then there was a deferred election, and then days before the vote the new Liberal candidate tried to take his name off the ballot but it was too late, so I believe the NDP support was inflated for that reason. The district also belongs to cabinet minister Susan Sullivan and from what I understand the Liberal candidate is a high profile multi millionaire, while the NDP candidate is pretty much a nobody who has no chance of winning. Though you would never know it by looking a the 2007 numbers the NDP supposedly have a very good chance of winning both St. John's Centre and St. John's North. I think the NDP have a chance at winning 6 seats.

    So possible seat projections for me are:
    PC - 40
    NDP - 6
    Liberal - 2

  9. Anonymous 12:59,

    The NDP would have to get to at least 31% or so to start getting on the same level as the other parties. But if they take too much from the Liberals, the PCs will start winning a lot of seats, not the NDP.

  10. Eric,

    Do you have all per riding data from the massive (40,000) Forum research poll? If not, here are the data from the local newspaper (Mountain News, Hamilton), from this poll, for the local 5 ridings (the order is NDP, Lib, PC):

    Hamilton Mountain: 35.4, 30.1, 27.2;
    Hamilton Centre: 50.5, 24.2, 17.4;
    Niagara West-Glanbrook: 16.5, 22, 52.6;
    Hamilton East-Stoney Creek: 47.5, 22.2, 22.9;
    Ancaster-Dundas-...: 21.3, 35.8, 34.9.

    They don't provide numbers of the polled, but I think one can just assume ~400 per riding.

    To compare with your current prediction for these ridings, I averaged the party support across the 5 ridings, both for your prediction and the Forum data. With 2000 polled, POE~2%:

    NDP Lib PC
    Eric 38.6 26.2 29
    Forum 34.2 26.9 31

    The only significant difference is in your model predicting much higher (by 4.4%) NDP support (compared to the Forums' data).

    In terms of riding winner prediction, the only difference between your prediction and Forum data is for Ancaster riding, where you predict PC win by 2%, and Forum suggests a slim Liberal win (by 0.9%). It's insignificant, of course.

  11. Out of curiosity, at what level would the NDP have a shot at winning in Ontario? We all remember 1990 and with the dislike Hudak and McGuinty have it wouldn't be a complete shock for the NDP to climb up to 30%. Heck, given the mood I keep hoping we'll see a Green shift :)

  12. That's interesting, thanks First.

  13. John_Northey

    According to (
    In 1990, the popular vote broke down as follows:
    NDP 37.6%
    Liberals 32.4%
    PC 23.5%

    Barring some kind of seismic shift in support from the PC's to the NDP it seems pretty unlikely that we'll repeat that scenario. Particularly since, as Eric mentioned earlier, if the votes bleed from the Liberals to the NDP it's going to disproportionately benefit the PC's.

  14. New numbers from Nanos: OLP 36.5%, PCO 34.0%, ONDP 26.8%, GPO 1.9%.

    It looks like Nanos may be doing a rolling poll for the last few days of the election.

  15. Eric,

    Have you seen these new data from Forum research - they did bigger size polling in 9 (GTA) ridings (690 people per riding), with the conclusion that NDP will win 8 of them:

  16. First,

    Eric already gave most of those seats to the Ndp, most of them are currently held by the Ndp anyway and the other ones are in the north where the Ndp is really strong.

    The Bramalea-Gore-Malton one can actually go Ndp, the candidate running for the Ndp ran for them in the federal election earlier this year and almost won that seat.

    The poll was commissioned by a group that seems to be Ndp friendly so it looks like they chose strong Ndp ridings. I still believe the Liberals would keep 2 seats in the north, they have 2 ministers running there if anything.


  17. Let's have a comment or two on the Red Tory victory in the Alberta PC leadership race.

  18. As much as I dislike the Liberals, it will be poetic justice if they get a majority in Ontario with fewer votes than the PC's get given how the PC's fought against election reform. Under virtually any method of reform the winner of the popular vote would be the winner overall. Only FPTP gives the loser a majority.

  19. Ekos, with its new poll, plus the analysis of the voter turnout by party affiliation, now also predict a slim Liberal majority:

  20. I'm not sure how Ontario is really going to play out, but I do suspect that the NDP is going to get more votes than most people think they will. It appears that everyone has finally caught on to the fact that Bob Rae is a Liberal.

    How that will affect the PC vote is open to question. I know people who were going to vote PC because they can't stand McGuinty, but don't really like Hudak either. With Horvath looking like a real option, a lot of those Fair Weather Conservatives will be happy to jump ship.

    Also the parties of today aren't the parties of yesterday. Today's Liberals are to the right of the Big Blue Machine. Today's Progressive Conservatives are to the right of the Harris Common Sense Revolution Tories. Which leaves the Horvath NDP as the embarrassed inheritor of the Big Blue Machine.

    Scary, isn't it?


  21. New Forum research poll (20,000) for 27 swing ridings in Ontario is out. Now it has huge number (almost 900 polled) per riding - but the result is again a draw (45-45 seats for PC and Liberals). I wonder why your model consistently over-predicts Liberal seats compared to direct per-riding polls? Could it be that your incumbency factor doesn't work in this election?

  22. It could be, or it could be that Forum's riding polls are off.

    They polled 27 new ridings, not all 107 again. So the 45/45 tie contains a lot of the old information from the other 80 ridings.


COMMENT MODERATION POLICY - Please be respectful when commenting. If choosing to remain anonymous, please sign your comment with some sort of pseudonym to avoid confusion. Please do not use any derogatory terms for fellow commenters, parties, or politicians. Inflammatory and overly partisan comments will not be posted. PLEASE KEEP DISCUSSION ON TOPIC.