Manitoba: Projection vs. Results

If you'll allow me a little immodesty, the projection for the 2011 Manitoba provincial election was ThreeHundredEight's Mona Lisa. Tongue firmly in cheek, of course, but of the 57 ridings in the province, 56 were called correctly and the 57th was so close that it is currently under judicial review.

The Manitoba election might seem trivial to people who aren't one of the 1.3 million Manitobans in the province, but this was an election that was thought to be very close. It was an election that demonstrated perfectly the usefulness of seat projections, as though the polls pointed to a tight race my seat projection model indicated that the New Democrats could win a comfortable majority with an uncomfortable margin in the popular vote. And, as I will show, this projection was no mere accident or stroke of luck. Even at the individual riding level the projection model performed very well.

In the end, Greg Selinger's New Democrats won an even larger majority than was projected, defeating Hugh McFadyen's Progressive Conservatives. Jon Gerrard, leader of the Manitoba Liberals, was the only candidate elected for that party.

ThreeHundredEight projected 36 New Democrats would be elected, 20 Progressive Conservatives, and one Liberal.

The result was 37 New Democrats, 19 Progressive Conservatives, and one Liberal. Using the actual provincial vote tally did not change the projection.

In all, the model called 56 out of 57 ridings correctly for an accuracy rating of 98.2%.

The projected seat ranges anticipated this result, as the New Democrats were expected to win between 35 and 38 seats, while the Tories were expected to win between 18 and 21 seats. The result was within the expected seat range. The Liberals were expected to win one seat and one seat only, and so they did.

As the polls by Probe Research and Angus-Reid at the end of the campaign were excellent, the vote projection was excellent as well. It under-estimated Progressive Conservative support by one point, but was within 0.6 points for the New Democrats, Liberals, and Greens.

The chart below shows a few quick facts about the projection at the riding level.

The margin of victory in St. Norbert, the one riding incorrectly called, was only one per cent.

Generally speaking, the Progressive Conservatives were under-estimated at the riding level. In 16 ridings they were over-estimated, in four ridings the projection was accurate, and in the remaining 37 they were under-estimated.

For the other parties the projection was generally solid, with the New Democrats over-estimated in 35 ridings, correctly called in five, and under-estimated in the remaining 17. The Liberals were over-estimated in 31 ridings, correctly projected in eight, and under-estimated in 18.

In all, 79% of individual riding projections for the New Democrats, or 45 out of 57, were called to within 5% of the result. That increased to 81% for the Tories and in almost every case the riding projections were within 5% for the Liberals and Greens. In 40 ridings, or 70%, every party was pegged to within 5% of their actual result. That is no mere coincidence or accident.

On average, the riding projections were 1.9 points higher for the NDP than the result and 2.2 points lower for the Tories. The Green and Liberal estimations were even closer, putting the margin of error for each party's riding projections at +/- 2.7%. That is, I believe, quite extraordinary.

To compare to the Prince Edward Island election, less than 30% of those ridings had every party called to within 5% and the average margin of error was +/- 4.6%.

Many ridings were called very well, but here are the top three:

In Dawson Trail and Selkirk, the projection was only off by one point for the three parties running. This is a good time to remind readers that the boundaries in Manitoba changed since the 2007 election, compounding the difficulty in making accurate projections.

Rossmere was off by a total of three points across four parties.

Now, what of St. Norbert, the one riding that got away? As mentioned, a judicial recount will take place and the results revealed on October 24. So, it could be that the result will be overturned and the projection will end up being 100% accurate. Note also that the vote in Kirkfield Park, which was correctly called, will also be reviewed. Hopefully that one won't be overturned.

If the result is not overturned in St. Norbert, why did the projection get it wrong? St. Norbert only bucked provincial trends by the tiniest amount, and it appears that the lack of an NDP incumbent did not hurt the party at all. But removing that factor still would have given the riding to the Tories.

So, it would appear that local factors were at play. Perhaps a good performance by Dave Gaudreau, the NDP candidate, or a less than good performance by Karen Velthuys, the Tory candidate, is to blame. Handing Gaudreau the star candidate bonus would have given him the riding, but he doesn't fit the profile of a star candidate and Velthuys does not seem to have been especially problematic.

The wrong call in this one riding needs to be attributed to the fact that the model cannot precisely predict how thousands of individuals will behave. St. Norbert being the one wrong call is, I think you'll agree, perfectly acceptable.

It should come as no surprise that the 2011 Manitoba provincial election projection is the best of the seven that ThreeHundredEight has called. And that is by a wide margin - the per party error on the vote projection was almost half the error of the next best called (2008 Quebec election) and the per party seat error was almost one-third of that in the 2008 Quebec election. I do not consider it likely that the projection model will ever provide a better result than it did in Manitoba.

Liberals up in Ontario, but outcome still uncertain

Before getting to what will likely be the last Ontario projection of the campaign, a few words about last night's election in Manitoba.

The polls nailed it. Angus-Reid and Probe Research put out the last polls of the campaign and both agreed that the NDP had 46% support and the Tories 43%. They put Liberal support at between 7% and 8%.

The end result was 46% for the NDP, 44% for the Tories, and 7.5% for the Liberals.

And when the polls are right, the seat projection model should be right as well. And, with the one error being the riding of St. Norbert, decided by less than 200 votes, the seat projection model called 56 of 57 ridings for an accuracy rating of 98.2%. If only the good people of St. Norbert had voted for the PCs, the projection would have been 100% correct. But I'll take 98.2%.

With this boost of confidence, I can now look to the Ontario election. But it is going to be trickier, because two polls released at the end of the day yesterday ruined the consensus that was building.

Nanos Research set the stage, saying that though they pegged the gap at about five points from October 1 to 3, the last two days of polling showed the Liberals over 40% support and well on their way to a majority.

EKOS Research, in a poll put out yesterday morning, agreed, with the Liberals at 39.1% support to the PCs' 29.1%. A poll they released this morning through iPolitics, with extra data from yesterday, corrobated their findings. Their IVR poll taken from October 2 to 4 and surveying 2,081 people has the Liberals at 39%, the Tories at 29.7%, and the New Democrats at 23%. In what EKOS considers likely voters, the Liberals still lead with 40% to 32% for the Tories and 21% for the NDP.

Regionally, EKOS has the Liberals ahead in Toronto (49% to the NDP's 23%), the GTA (38% to the PCs' 33%), southwestern Ontario (35% to the PCs' 31%), and northeastern/central Ontario (36% to the PCs' 28%). The Tories and Liberals are tied at 36% apiece in eastern Ontario, while the PCs lead in northwestern Ontario (37% to the NDP's 29%).

Ipsos-Reid then sealed it with a poll for the Ottawa Citizen, giving the Liberals a 10-point lead. In a telephone poll taken between September 30 and October 3 and surveying 1,020 people, Ipsos-Reid found Liberal support standing at 41%, with the Tories trailing with 31% and the NDP with 25%. Considering that Ipsos-Reid has generally put Conservative support higher than most other polling firms (to, perhaps, their credit), this kind of lead for the Liberals was doubly important.

The Liberals led in almost every part of the province in this poll: in the GTA (44% to the PCs' 28%), central Ontario (36% to the PCs' 35%), eastern Ontario (50% to the PCs' 35%), and southwestern Ontario (36% to the NDP's 34%). The NDP led in the north with 40%, ahead of the Liberals and Tories (both at 28%).

Ipsos-Reid also found that Dalton McGuinty was the preference to be premier at 37% to 27% for Tim Hudak, and gave the Liberals a 10 point lead among those aged 55 or older, the people in Ontario most likely to vote.

Then Abacus Data came along and muddied the waters in a poll for Sun Media. In an online poll of 1,015 Ontarians taken between October 3 and 4, the Liberals led with 37% to the Tories' 34%. A lead, but nowhere near as comfortable as Nanos, Ipsos-Reid, and EKOS found to be the case. The New Democrats came in third with 24%.

Regionally, Abacus gave the Tories the lead in eastern Ontario (42% to the Liberals' 35%), southwestern Ontario (35% to the Liberals' 27%), central Ontario (51% to the NDP's 24%), and the Hamilton/Niagara region (32% to the Liberals' 31%). The Liberals lead in Toronto (49% to the NDP's 26%), while the NDP led in the north (39% to the PCs' 34%). The GTA was tied with 38% apiece for the Liberals and Tories.

Abacus gave the leadership edge to McGuinty, with 30% to Hudak's 21%.

Finally, this morning Angus-Reid in a poll for the Toronto Star completely demolished the consensus of a Liberal lead. In an online poll taken between October 3 and 4 and interviewing 2,223 people, Angus-Reid pegged Progressive Conservative support at 36%, with the Liberals trailing at 33% and the NDP at 26%.

The Tories led in the 905 (41% to the Liberals' 29%), the southwest (38% to the Liberals' 29%), and the east (43% to the Liberals' 30%). The Liberals led in the 416 (46% to the NDP's 27%) and the Hamilton/Niagara region (36% to the PCs' 35%). The NDP led in the north (39% to the Liberals' 36%).

Perhaps significantly, though, Angus-Reid had McGuinty at 24% on who would be the best person to be premier, slightly ahead of Hudak (23%).

So what to make of this? We aren't talking margin-of-error issues at this point. Some of the pollsters will have egg on their face on Thursday night. One might consider Angus-Reid the outlier, but they nailed the Manitoba election and their data is newer than that of Nanos and Ipsos-Reid, and partly that of EKOS. It's also the largest poll in this last week. Perhaps it is a methodological issue, as the online polls are showing more favouable results for the Tories than those done over the telephone. Angus-Reid shows the Tories gaining, everyone else has them falling.

There are a few points of consensus, however. Dalton McGuinty is beating Tim Hudak on the all important leadership question, while regionally the Liberals are dominating in Toronto and the NDP is good in the north. Outside of these two regions, there is no consensus but it should be pointed out that in only one poll do the Tories have the kind of lead they need in the GTA to make a run at the premiership.

So, with all of these conflicting numbers the projection is falling, as it should, somewhere in the middle. The Liberals are projected to take 36% of the vote, with the Progressive Conservatives netting 33.2% and the NDP 24.8%. The Greens take 4.6% of the vote.

Since yesterday, that is a 0.1 point gain for the Liberals and a 0.5 point gain for the Greens. The Tories are down 0.2 points and the NDP is down 0.5 points. Nothing major, then.

The Liberals have picked up one seat, and are now projected to win 57. The Progressive Conservatives are unchanged at 30, while the New Democrats are down one to 20 seats.

The seat change is in the southwest, where the Liberals are now projected to win 13 seats, the Tories seven, and the New Democrats one.

There are still a number of close races, but not enough to put the Tories in range of even a minority government.

The Liberals lead in six close races and trail in four, putting their seat range at between 51 and 61 seats. Only 30% of their range is in minority territory.

The Progressive Conservatives lead in four close races and trail in five, putting their range at between 26 and 35 seats. They are no longer in danger of being supplanted by the NDP, as was starting to look like a possibility yesterday.

The New Democrats lead in one close race and trail in two, putting their range at between 19 and 22 seats.

But with the polls as confusing as they are, Thursday night is far from a sure bet. The Liberals were clearly going to win PEI and the electoral geography of Manitoba pointed to an NDP victory.

Tomorrow's final projection report will also make mention of the poll error adjustments I have done for Manitoba and PEI. In addition, though, I will put up projections for the kind of lead that Nanos, Ipsos, and EKOS are showing, as well as the result that Angus-Reid predicts. It will be a long night, and anything could still happen. But even the Tory lead in the Angus-Reid poll is not large enough to give the PCs much hope - which means we're looking at a Liberal re-election.

Final Manitoba Projection: Selinger's NDP wins majority

With hours to go before the polls close in Manitoba, ThreeHundredEight.com projects that Greg Selinger's New Democrats will win a majority government, with Hugh McFadyen's Progressive Conservatives forming the Official Opposition. Jon Gerrard's Liberals will be the third party in the Manitoba legislature.

This comes after the addition of Angus-Reid's Manitoba poll released yesterday, conducted between September 30 and October 2, and surveying 800 people online. The poll was identical to Probe Research's numbers, with the New Democrats at 46% and the Progressive Conservatives at 43%.

This is a margin that, however, makes me quite uncomfortable, particularly in light of yesterday's results in Prince Edward Island.

The New Democrats are projected to take 45.6% of tonight's vote, with the Progressive Conservatives taking 42.9%.

The Liberals are projected to take 8% of the vote, while the Greens take 3.1%.

That is, in any case, what the polls are suggesting.

Since the projection of September 30, this is a gain of 0.9 points for the New Democrats. The Tories and Greens have each lost 0.4 points, while the Liberals are down 0.1.

The New Democrats are projected to win 36 seats, with the Progressive Conservatives winning 20 and the Liberals one.

The New Democrats are projected to win 25 seats in Winnipeg, four in the north, four in the southeast, and three in the southwest. The Progressive Conservatives are projected to win 10 seats in the southeast, five in the southwest, and five in Winnipeg. The Liberals win their one seat in the provincial capital.

Four polls were released in this campaign, all in the last two weeks. Only one of them showed a Tory lead, while the last three all put the margin between the NDP and the PCs at three points, inside the margin of error.

Averaged out, however, the New Democrats have been tied or have led the Tories for the entirety of the campaign. But the race remains incredibly close, and dangerously close for a projection.

The New Democrats are leading in one close race and trailing in two others, all implicating the Tories. This puts the New Democrat range at between 35 and 38 seats, and the Tories between 18 and 21 seats.

I suspect, however, that more races will be quite close.

But it should not be forgotten that projections are not a crystal ball looking into the future - they are based on the available polling data, which currently shows an NDP lead.

Though the ridings in Manitoba are larger than that of Prince Edward Island, they are still small and local factors will undoubtedly skew some of the riding projections. Also problematic is that the boundaries have changed, and this has the potential to greatly change how voters behave. For many voters, their "incumbent" ran in a different riding in the 2007 election, so there is a good chance that there will be a few surprises tonight.

Adjusting the projection for the polling error in the 2007 Manitoba and 2011 federal elections, however, changes little in the outcome.

If the error is the same as it was in the 2007 provincial campaign, the New Democrats should take closer to 49% of the vote and win 37 ridings.

But if the error is the same as it was in the 2011 federal campaign, and as we have seen in many cases the Conservative vote is often under-estimated, the Tories would win more of the popular vote - but only just. The projection model would still give the NDP 36 seats to the Tories' 20, but the PC range would be extended to as many as 26 seats, with the New Democrats taking the 31 others.

If we use the riding projection accuracy rating of last night's projection, we could expect as many as eight ridings to be called incorrectly tonight. That is a potentially significant number. If the projection tonight is as accurate as last night's, the New Democrats could win between 28 and 44 seats and the Progressive Conservatives between 12 and 28 seats. In other words, the legislature could be split down the middle at 28 seats apiece for the NDP and the Tories.

If the polls have it, though, the edge the New Democrats have in Winnipeg (roughly 20 points) should be the clincher. The city contains 54% of the province's ridings, and the New Democrats are strong enough in the north, Brandon, and a few other rural pockets to make up for any Winnipeg ridings that side with the Progressive Conservatives.

But the outcome is far from pre-determined. With the margins of error of the polls, the Tories could actually come out with more votes. And if the changes in riding boundaries has a particular effect, they could come out with more seats as well. I suspect it could be a long night in Manitoba, as it will be for myself here in Ottawa.

Liberals gain in Ontario, still low in Newfoundland

Before getting to today's polls, a few words about Prince Edward Island's election last night. A full post-mortem will come after the Newfoundland and Labrador election, when there will be some time to digest the results.

There were only two polls conducted in Prince Edward Island during the campaign and none in the last six days. But they nevertheless under-estimated Tory support and over-estimated that of the New Democrats. Had the polls been accurate, ThreeHundredEight's projection model would have had the Liberals at 25 seats and the Tories at two. Still off, but a little better.

My projected seat range with those numbers would have been 1-3 seats for the Tories, but if I extend that range to ridings projected to have been won by six per cent or less (instead of five per cent), then the high seat range for the PCs would have been five. I will be using the margin of six per cent to determine whether a race is close or not for subsequent elections.

Now, projecting a slightly larger overwhelming Liberal majority than the overwhelming Liberal majority that actually was elected last night is not a huge problem. It is Prince Edward Island, a province with a population smaller than a few ridings in Ontario. When two PC ridings were won by about 30 votes and 425 votes separated my high seat range of two from their result of five, I don't consider the model to have failed.

But the under-estimation of Conservative support, the third consecutive time this has happened in an election I have done projections for, is vexing. I will take a close look at it before making my final Manitoba projection later today.

Ontario

One poll was released this morning by Nanos Research, as part of their rolling three day polling. Taken between October 1 and 3 and surveying 900 people, Nanos finds Liberal support standing at 37.7%, followed by the Tories at 33.2% and the NDP at 25.8%. This matches EKOS's latest findings in that the Liberals are beginning to pull ahead. Though the sample sizes are small, Nanos even has the Liberals over the 40% mark over the last two days.

With this Nanos poll added to the projection, the Liberals now stand to take 35.9% of the vote, with 33.4% going to the Progressive Conservatives.

The New Democrats are projected to take 25.3% while the Greens take 4.1%.

This is a gain of 0.7 points for the Liberals and 0.1 points for the NDP since yesterday's projection. The Greens have dropped 0.2 points and the Tories 0.6 points.

One seat has changed hands as a result, with the Liberals now up to 56 seats and the Progressive Conservatives down to 30. The New Democrats remain at 21 seats.

The Liberal seat gain comes in southwestern Ontario.

The Liberals are now projected to win 12 seats in the region, with seven going to the Tories and two to the New Democrats.

As mentioned, the bar for the seat ranges has now been moved to six per cent as a result of the PEI election. This doesn't change too much in Ontario, however.

The Liberals are leading in eight close races and trailing in five, putting their seat range at between 48 and 61. This is significant, as the Liberals are now at about even odds of pulling out a minority or majority government on Thursday.

The Progressive Conservatives lead in four close races and trail in six, putting their seat range at between 26 and 36. Things need to change in order for the Tories to be back in the race to form government.

The New Democrats lead in two close races and trail in three, putting their seat range at between 19 and 24. If things continue to go badly for the Tories, a scenario in which they finish third is not out of the question.

Newfoundland and Labrador

When MQO Research put Liberal support at 13% last week, Kevin Aylward, leader of the Newfoundland and Labrador Liberals, put out a press release accusing the Tories of having bought and paid for the poll. It was an unusual thing to do, as most political leaders brush poll results aside. It was a risky gamble that would pay off if the next poll put the Liberals up.

Well, it didn't. The latest poll by NTV and Telelink, conducted between October 1 and 2 and surveying 511 people by telephone, puts Progressive Conservative support at 58.9% of decided and leaning voters. The New Democrats follow with 28.4% while the Liberals stand at 12.5%.

In other words, this plainly independent poll is even better for the Tories and worse for the Liberals than the poll by MQO Research.

Kathy Dunderdale comes out even better on the leadership question, with 55% saying she is the best person for the job. Lorraine Michael of the NDP follows with 19% while Aylward is at 7%.

With this poll added, the Progressive Conservatives are now projected to take 56.7% of the vote. The New Democrats follow with 30.3% and the Liberals come up in third with 12.8%.

This represents a gain of 2.9 points for the Tories and a drop of 2.5 points for the NDP since the last projection. The Liberals are down 0.4 points.

This has not caused any changes in the seat projection.

The Progressive Conservatives are still projected to win 42 seats, the New Democrats four, and the Liberals two.

This has nudged a few seats out of being considered close races. The Liberals and NDP each lead in one close race, trailed by the Tories.

This puts the PC seat range at between 42 and 44 seats. The New Democrats are projected to win between three and four seats and the Liberals between one and two seats.

There is still a week to go before Newfoundland's vote, and I expect two or three polls before then. But Newfoundland and Labrador will pose some of the same problems as Prince Edward Island. The ridings are about twice as large, but they are still very small and subject to local factors to a degree that other provinces, particularly Ontario, are not. The polls seem to be relatively consistent, though, so the big question will be how they translate into seats.

I will post the final projection for Manitoba shortly, incorporating the latest poll from Angus-Reid.

NDP gains in Manitoba and PEI, Selinger and Ghiz on track for re-election

Yesterday, I was alerted to a poll that was released by MarketQuest Omnifacts Research for Prince Edward Island. This is the same firm which broke the ice in the Newfoundland and Labrador campaign. Few media outlets picked up the poll, but I have the details and so the Prince Edward Island projection has been updated.

But before getting to the MQO poll and PEI, let's take a look at the new poll by Probe Research for the Winnipeg Free Press.

The telephone survey, taken between September 21 and 28 (and so dipping its toe into the survey dates of both the Viewpoints Research and Environics polls) and interviewing 1,000 Manitobans, found that 46% support Greg Selinger's New Democrats, with another 43% intending to vote for the Progressive Conservatives.

The Liberals stood at 7%, while the Greens are at 4%.

Considering the margin of error, this is no change at all since Probe's last poll from the end of June which put the NDP and PCs tied at 44% and the Liberals at 9%. But coupled with the polls from Viewpoints and Environics, we can say that the NDP is ahead.

They are particularly ahead in Winnipeg, with 53% to 35% for the Tories. Outside of the city, the vote intentions are flipped: 54% for the Tories and 34% for the NDP. A big problem for the PCs is that in Winnipeg they only lead in the southwestern part of the city, and there by a statistically insignificant three points. It doesn't seem that the Tories will have much luck in the city.

It's worth noting how the polls have changed over time in Manitoba. PC support has been remarkably consistent, hovering between 40% and 47% for at least a year. But the NDP has taken off, standing in the mid-40s after being in the mid-30s in the spring. And the Liberals have sunk below the symbolic 10% mark.

With this poll, which I suspect will be the last of the campaign, the New Democrats are projected to win 44.7% of the vote, with the Progressive Conservatives taking 43.3%.

The Liberals take 8.1% of the vote while the Greens take 3.5%.

This is a gain of 1.6 points for the NDP since the September 27 projection, and a loss of 0.6 points for the Tories. The Liberals are down 1.5 points while the Greens are up 0.6.

This has not resulted in any seat changes. The New Democrats are projected to win 36 seats, the Tories 20, and the Liberals one.

With the NDP gain, several seats have dropped out of the PCs' range. Only one Tory seat and one NDP are now considered close, putting the NDP range at between 35 and 37 seats and the PC range at between 19 and 21 seats.

The NDP's advantage in Winnipeg is extraordinarily important, as the city has over half of Manitoba's ridings.

Now to Prince Edward Island. The poll by MarketQuest Omnifacts was taken both over the telephone and online between September 23 and 25. Surveying 411 people, the poll is smaller and somewhat older than The Guardian's poll by CRA. But the results are not much different.

MQO found that Liberal support in PEI stood at 53%, exactly the same as the CRA poll. It differed on PC support, pegging them at 32%, and NDP support, finding them at 12%. The Greens stood at 3%.

The NDP level of support is problematic, as they aren't running a full slate of candidates. And this is where we run into a problem - how to adjust for something like that? Should we assume that survey respondents don't know whether or not an NDP candidate is running in their riding? Certainly many, maybe the majority, are aware of this and so the voting intentions number may not inaccurate. It is a bit of a puzzle, but in any case it is unlikely that the NDP will play a big role in Monday's election.

With this poll, the Liberals are now projected to win 52.9% of the vote, down 0.2 points since yesterday's projection.

The Progressive Conservatives are down 1.9 points to 34.2%, while the New Democrats are up 3.1 points to 8.1%. The Greens are down 0.9 points to 4.1%.

The seat projection has not changed: 26 seats for the Liberals and one for the Tories. The seat ranges are also unchanged: 25 to 26 seats for the Liberals and one to two seats for the Progressive Conservatives.

Little change in Ontario, race narrows in Manitoba

A bevy of polls in two provinces were released over the last 24 hours, so we have a lot to get to. Let's start with Ontario before moving on to Manitoba.

Last night, Abacus Data released the results from its latest poll, taken between September 23 and 25. It showed the Progressive Conservatives leading with 37%, compared to 33% for the Liberals, 23% for the New Democrats, and 6% for the Greens. Regionally, Abacus had the Tories leading in eastern Ontario (51%), the GTA (37%), and southwestern Ontario (38%). The Liberals led in Toronto (45%), while the New Democrats were tied with the PCs in the north (37%).

Abacus used the standard voting intentions question for this poll, but also used the sliding scale they had experimented with earlier in the campaign. The results of that method of polling were not very different: 38% for the PCs, 30% for the Liberals, and 25% for the NDP. Compared to their last poll using this method, that is a drop of three points for the Tories, a drop of two for the Liberals, and a gain of five for the NDP.

Appearing this morning is a poll from EKOS Research, which hasn't waded into the provincial campaign yet and has been silent since the federal election. The EKOS poll, taken between September 21 and 25 using the IVR method, found that the Liberals lead with 34.9% of the vote, with the Progressive Conservatives at 31.4%, the NDP at 24.7%, and the Greens at 7.6%.

EKOS has the Liberals leading in Toronto (47%), while the New Democrats are ahead in "northeastern and central" Ontario (33%). The Tories hold a slight edge in the suburban GTA (35%), southwestern Ontario (34%), eastern Ontario (41%), and northwestern Ontario (34%).

Certainly, these polls differ. But they were taken over a few different days and we have no trend to look at in the EKOS poll for comparison, since this is their first.

With these polls added to the model, the Progressive Conservatives hold the slimmest of leads with 34.6% of the vote. The Liberals trail with 34.5%, while the New Democrats stand at 23.7% and the Greens at 6.0%.

This is a drop of 0.8 points for the Tories and 0.4 points for the Liberals. The NDP is up 0.3 points and the Greens are up a full point.

This does not result in any seat changes, so the Liberals are still projected to win 55, the Progressive Conservatives 32, and the New Democrats 20.

The ranges for the Tories and the Liberals have also not changed, though the New Democrats are now in range of one more seat.

That puts the NDP range at between 18 and 22 seats, while the Tories and Liberals are unchanged at between 28 and 43 and between 44 and 60 seats, respectively.

Now let's move on to Manitoba, where two polls were released yesterday.

The first, by Viewpoints Research for CJOB News and the Manitoba Real Estate Association, put the New Democrats ahead by a healthy amount.

After being provided the full details, I can say that with decided and leaning voters included, the New Democrats lead in the poll with 48.9%. The Progressive Conservatives follow with 40.3%, while the Liberals stand at 8.2% and the Greens at 2.6%.

This poll was taken by telephone between September 14 and 21. Some criticism of the poll has come from the fact that Viewpoints has a relationship with the provincial NDP. However, the poll was not ordered by the party.

In Winnipeg, the New Democrats lead with 54% to the Tories' 31%. The Liberals are third with 10%. Outside the provincial capital, the Tories are ahead with 53% to the NDP's 40%.

The second poll was by Environics, and was conducted online btween September 20 and 26, so this poll only overlaps with the Viewpoints poll on a few days. Environics found that the Tories are leading with 45% to the New Democrats' 42%, with the Liberals at 10% support.

Interestingly, on who would make the best premier it is NDP leader Greg Selinger who comes out on top with 33% support to Hugh McFadyen's 29%.

These polls tell a different story, but only at first glance. They were mostly taken over separate weeks, and with their respective margins of error (assuming the Environics poll had a random sample) are not actually at odds. We can take from these two polls that the race is very close, and that the PCs may have a little momentum.

With these polls, the Progressive Conservatives are back in front with 43.9% of the vote. The New Democrats are not far behind with 43.1%, while the Liberals are at 9.6%. The Greens trail with 2.9%.

This is a gain of 4.4 points for the PCs since yesterday, and a 7.1-point loss for the NDP. The Liberals are up 3.6 points, while the Greens are down one.

In terms of seats, the New Democrats are projected to win 36, down three from yesterday. The Progressive Conservatives win 20 (up two) and the Liberals win one (up one).

The Tories have picked up one seat in southwestern Manitoba and two more in Winnipeg, all from the New Democrats. The Liberals pick up one seat in Winnipeg from the Tories.

There are only four projected close races, and at this stage the New Democrats lead in all of them, with the Tories trailing. This means that the NDP is at the top of their seat range, while the Tories are at the bottom of theirs.

This means the NDP is on track to win 32 to 36 seats, the PCs 20 to 24 seats, and the Liberals one. With the new polls, they are comfortably ahead in Jon Gerrard's riding.

Only one week remains before the vote is held in Manitoba, and I am told that we can probably expect two more polls to come out of the province. It really is setting up to be a close race in the popular vote, but the NDP seems to have the geographic advantage.

Prince Edward Island is voting in less than a week, and a poll is supposed to come out sometime before the vote is held.

In Ontario, the debate is tonight and less than two weeks remain in the campaign. A lot is at stake, and with things as close as they are tonight's debate is extraordinarily important for all three parties. McGuinty needs to solidfy his support, Hudak needs to make gains, and Horwath needs to do well enough to ensure a minority government.

McGuinty and Selinger in majority territory

With less than two weeks to go in Ontario’s election campaign, voters remain split on who should form the province’s next government. But that is good news for the incumbent Liberals, since were an election held at the end of last week Dalton McGuinty would have likely won a razor-thin majority.

To read the full details of today's Ontario projection update, check out The Globe and Mail website here. The riding projection in the right-hand column is also updated.

A few words on the Forum Research poll that came out this Saturday. The size of the poll should not be over-emphasized. The +/- 0.5 point margin of error only applies to a truly random sample, and of course no poll is taken in a vacuum. As many pointed out during the census debate, increasing the sample size doesn't erase errors in the methodology. That is not to say there is a problem with Forum's methodology, only that a poll of 40,000 people isn't automatically the be-all and end-all of polling.

Because the poll was so large, riding level numbers were also released. Aside from a few cases, often in ridings that the firm itself called "anomalies", the riding polls in the GTA matched comfortably well with my own projections. I haven't seen many numbers outside of the GTA yet, however.

But the big news today is out of Manitoba. CJOB News is reporting on a new poll by Viewpoints Research. As it is radio, they are stretching out the release of details throughout the day, so for now all we have is the decided vote: 41% for the New Democrats, 32% for the Progressive Conservatives, and 5% for the Liberals. Another 19% are undecided, which means we can extrapolate that 51% of decided voters are supporting the NDP, 40% the Progressive Conservatives, 6% the Liberals, and 4% the Greens.

As I am starved for new information out of the province, I will use these numbers for the projection until the full decided/leaners details are released later today. Another poll is apparently scheduled for release tonight, so the situation in Manitoba is finally going to be a little clearer.

The New Democrats are now projected to take 50.2% of the vote, up 10.3 points since the last projection from the end of July. The Progressive Conservatives are down 4.8 points to 39.5%.

The Liberals have sunk 5.5 points to only six per cent support, while the Greens are down 0.1 point to 3.9%.

This results in the New Democrats winning 39 seats in the 57-seat legislature, a comfortable majority. The Progressive Conservatives win 18 and the Liberals are shut-out (though just barely). That is a gain of six seats for the NDP since the last projection and a loss of two for the Liberals and four for the Tories.

The New Democrats have picked up two seats in the southeast and four in Winnipeg, giving them 27 of the 31 seats in the capital. They are running close to the PCs in the rest of the province, however, with 12 seats to 14. The Tories have lost two seats in southeastern Manitoba and three in Winnipeg, but gained one seat from the Liberals in the capital. The New Democrats took the other.

There are only a few close races in the province, and the Liberals are implicated in one of them. Otherwise, the seat range for the NDP and Tories is between 16 and 20 seats for the PCs and 37 and 41 seats for the NDP.

The ranges chart gives a visual representation of the distance between the parties.

With the new numbers being put out by CJOB today and the rumoured other numbers appearing tonight, I should be able to update the Manitoba projection tomorrow as well. Prince Edward Island is going to report soon, and Manitoba is supposed to have at least one more poll at the end of the week. We'll undoubtedly have some new Ontario numbers soon as well.

Also, check out my column in today's The Hill Times. It requires a subscription to read online (the paper edition is available in many downtown kiosks in Ottawa), but is well worth it (the online subscription, that is).

Incumbents leading in provincial elections

Change is a major theme in world politics lately. But not in Canada, where 2011 is setting up to be the Year of the Incumbent.

Canadians from the Atlantic Ocean to the Rocky Mountains are in the mood for the status quo, as five and possible six of the provinces headed to a fall election are on track to re-elect their incumbent governments.

Even this year’s federal election was an exercise in smooth sailing. Though the New Democrats soared to second place and the Liberals and Bloc Québécois fell by the wayside, Stephen Harper’s five-year-old government was re-elected for another five years. In Quebec, scene of much of the election’s change, it could even be argued that the province’s voters simply exchanged one social democratic opposition party for another.

Popular governments in Saskatchewan, Newfoundland and Labrador, and Prince Edward Island are safely on course to be re-elected, with their incumbent governments leading the main opposition parties in their respective provinces by about 30 points. Brad Wall and Robert Ghiz, the country’s two youngest premiers, will almost certainly be re-elected for a second term in Saskatchewan and PEI, while Kathy Dunderdale is expected to extend the lifespan of the Progressive Conservative government in Newfoundland and Labrador, first established by Danny Williams in 2003.

You can read the rest of the article on The Huffington Post Canada website here.

I'll be updating the projection later this morning. There is a new poll from Abacus Data for Ontario out, though it does use an unusual methodology.

Though they won't be added to the projection, there are also some polls to discuss from Forum Research on the Toronto vote and from the Corporate Research Associates on how Prince Edward Islanders view the parties in their province.

I've also heard word of another poll lurking about, with a release later this morning. I'll probably hold off until that one is available.

In Manitoba race, odds could tip in Greg Selinger’s favour

On October 4, Greg Selinger will be asking Manitobans to hand his New Democratic government its fourth mandate in his first election campaign as leader. Until recently, it appeared that the days of Mr. Selinger’s tenure as premier were numbered, but newfound popularity could extend the life of Manitoba’s longest running government since the 1950s.

You can read the rest of the article on The Globe and Mail website.

Manitoba will be an interesting race, and at the moment it is the election whose result is most difficult to predict. The latest polls put the New Democrats and Tories in a tie, and the two parties have been running neck-and-neck almost since the 2007 election.

But the odds seem to be stacked in the NDP's favour. My projection model gives them a good shot at winning a majority of seats even with a larger gap between the PCs and the NDP than currently exists in the vote projection.

The NDP wins a lot of close races in Winnipeg, while the Tories pile up huge majorities outside of the city. This is the key reason for the inability for the Progressive Conservatives to win without a large lead in the popular vote.

This has actually been the case for a little while now. Using the UBC's uniform swing model for the 2007 election, the New Democrats would have still won a majority of the seats with 41.1% of the vote to the Tories' 44.6%. The re-districting does not seem to have reversed the situation. Indeed, it seems to have entrenched the NDP even more reliably. We shall see how it plays out.

NDP wins fourth mandate in first Manitoba projection

The Manitoba projection model is now completed, and projects a New Democratic majority government in the province, with the Progressive Conservatives forming the Official Opposition.

However, the NDP wins this fourth mandate with fewer votes than the PCs.

The Progressive Conservatives are projected to have the support of 44.3% of Manitobans, compared to 39.9% for the NDP.

The Liberals stand at a projected 11.5% support, while the Greens are at 4.0%.

But because of greater vote efficiency, the New Democrats manage to elect 33 MLAs to Winnipeg, compared to 22 for the Progressive Conservatives and two for the Liberals.

This represents a loss of three seats for the NDP, a gain of three seats for the PCs, and a gain of one seat for the Liberals compared to their current standing in the Legislative Assembly.

Outside of Winnipeg, the Progressive Conservatives take the majority of the seats with 16 to the NDP's 10.

The PCs take 10 seats in the southeast, six in the southwest, and none in the north, while the NDP takes four seats apiece in northern and southeastern Manitoba and two in the southwestern part of the province.

In Winnipeg itself, the New Democrats still dominate with 23 seats to the PCs' six and the Liberals' two.

Understandably, as the NDP is down by 4.4 points several of the seats they hold in the projection are marginal.

The NDP has a five-point or less lead in five seats, while the PCs are leading by such a margin in two seats and the Liberals one.

The PCs are trailing by five points or less in five seats, while the NDP is behind by five points or less in three.

This puts the NDP range at between 28 and 36 seats. At the lower end of the scale, the NDP is in danger of being defeated by the PCs while at the higher end it puts them on par with their current caucus. For the PCs, they could win as few as 20 seats or as many as 27, indicating that at this point they stand a very good chance of improving upon 2007's results.

The Liberals range at between one and two seats. This is the difference between the party holding on to the successor to the riding of Inkster after Kevin Lamoureux's departure or not.

Now, Manitoba's ridings have been re-drawn. Some of the ridings have changed completely while others have changed only a little bit. Many thanks go to Kirk Vilks, who calculated the transposition of votes using the poll-by-poll breakdowns. In the chart to the left, incumbents are determined either by the sitting MLA who is running for re-election in a given riding, or by the winning party in each riding after the transposition of votes.