Friday, November 30, 2012

Ontario Liberals gain at expense of NDP

Forum Research is out with a new Ontario provincial poll today, published in the Toronto Star, showing that the Liberals have made big gains over the last four weeks, almost all of it at the expense of the New Democrats. Is it an outlier or a sign that the Ontario Liberal leadership race is improving the fortunes of the party?
Forum was last in the field Oct. 30-31, and since then the Progressive Conservatives have dropped two points to 35%. That still gives them a statistically significant lead over the Liberals, who were up seven points to 29% support. The NDP was down five points to 27%.

The Greens were up one to 8% and 1% of respondents said they would vote for other parties.

The Liberal gain and NDP drop is outside the margin of error, so this is not due to a statistical wobble. Whether it is an outlier result is hard to say - with the leadership race in the news it is not surprising to see the Liberals making gains. If this sort of result had come from Nanos Research, which has been more bullish on the Liberals in Ontario, then it would potentially be less significant. But coming from Forum, which has had the Liberals at 22% and 20% in their last two polls, this is perhaps more meaningful.

A few notes on methodology: Forum conducted this poll over two days, which they also did the last time they polled in Ontario. I applaud this, as their usual one-day snapshots are more susceptible to error. But, as usual, the report contains no information on unweighted samples. It does contain information on how respondents voted in the last provincial election, with that breaking down to about 40% for the Liberals, 33% for the Tories, and 19% for the NDP. The actual results are within the margin of error of the poll, but this does suggest that there is a potential that the sample is slightly more favourable to the Liberals (and Greens and others) than to the Tories and NDP. Whether that is due to lapses in memory or methodological bias is impossible to say.

The Tories were ahead in most parts of the province, with 37% support in the Greater Toronto Area, compared to 31% for the Liberals and 26% for the NDP (down six points). In particular, the PCs were strongest in the 905 area code with 41%, followed by the Liberals at 28% and the NDP at 25%.

The Progressive Conservatives were also ahead in eastern Ontario with 37% to 31% for the Liberals (+12) and 22% for the NDP, while they were in front in southwestern Ontario with 36% to the NDP's 31% and the Liberals' 23%.

The Liberals had the advantage only in the 416 area code, with 36% support to 31% for the Tories and 28% for the NDP. The New Democrats, however, were ahead in the north with 32% to 28% for the Liberals, 25% for the Tories, and 15% for the Greens (+7).

It is worth noting that the Liberals were up in every region of Ontario, while the New Democrats were down across the board. The Liberals had the edge among women by a margin of two points over the NDP, while the Tories were ahead by 15 points among men.

44% of respondents favoured calling an election, compared to 49% who said they were against another election right now. Sadly, 44% might be a good estimation of the next election's turnout.
With these numbers, the PCs would win a minority government of 49 seats, with the Liberals winning 30 and the New Democrats winning 28. But that is a close enough margin to the magic number of 54 seats that a majority government isn't out of the question for Tim Hudak. A continuation of the Liberal government is also theoretically possible, if they could get the support of the NDP.

The Tories win 34 of their seats in eastern, central, and southwestern Ontario, with a smattering of other seats in the Golden Horseshoe. The Liberals win 22 of their seats in and around Toronto, while the New Democrats win their seats in and around Toronto, Hamilton, and northern and southwestern Ontario.

Forum also included some polling on the Ontario Liberal leadership race, showing little change in what the numbers have indicated so far.

Gerard Kennedy topped the list of favourite choices at 16%, followed by Sandra Pupatello at 10% and Kathleen Wynne at 8%. The other candidates scored 3% or less. Fully 30% of respondents said that none of these were their favourite choice, while 27% were undecided.

Kennedy led in and around Toronto and in northern Ontario, while Pupatello led in eastern and southwestern Ontario. That jives pretty well with the regional distribution of Pupatello's caucus support.

Among Liberal supporters, Kennedy managed 25% to 16% for Pupatello and 13% for Wynne. Removing the undecideds and "none of the aboves" gives 38% to Kennedy, 25% to Pupatello, and 20% to Wynne. The other candidates were well behind.

Though Kennedy beat out his competitors on questions of "who cares most about people like you", the economy, likability, and trustworthiness, his numbers were very similar to his overall support. In other words, that tells us little about why Kennedy is favoured.

It will be interesting to see how the numbers fluctuate as the race firms up a little more. Kennedy is the favourite choice of NDP voters, suggesting that he has the greatest potential to draw supporters away from the New Democrats. If Pupatello is chosen, however, she may have more luck attracting Tories to the Liberal fold, keeping the NDP in the race. Combine that with how the PCs threw away their last chance and the next election is wide open.

Wednesday, November 28, 2012

Vote efficiency returns to the PQ

On Saturday, Le Devoir and the Montreal Gazette released the results of the latest Léger Marketing poll on the voting intentions of Quebecers, as well as their views on the new budget. The results show very little change in support for the major parties, but at the regional level the vote is back to working out to the advantage of the Parti Québécois, an advantage the party had lost in the last election.
Léger was last in the field Oct. 15-16, and since then the Parti Québécois was up one point to 33%, putting them narrowly ahead of the Liberals, who were also up one point to 31%. The Coalition Avenir Québec was down three points to 22%.

Québec Solidaire was unchanged at 8%, and Option Nationale was up two points to 4%. The Greens and other parties were unchanged at 2% and 1%, respectively.

None of these changes of support are statistically significant, though it is worth noting that, at 22%, the CAQ is back down to pre-campaign levels of support. Also, ON has never done better than 3% in any poll.

The PQ has traditionally enjoyed a much higher vote efficiency than the Liberals, due to their lead among francophone voters. The classic example is the 1998 election, where the Liberals actually won more votes while the PQ formed a majority government. Their victory by less than one point would have, by that measure, ensured far more than a four-seat edge in the last election. It was the closer race among francophone voters between the PQ and the CAQ that changed things.

But this poll gave the PQ a 15-point lead among francophones, with 40% support to 25% for the CAQ. The Liberals were down to 20%. That is very important for the PQ, as it makes it possible for them to win a majority government even with only 33% province-wide support.

The Liberals led among non-francophones with 68%, followed by the CAQ at 12% and the PQ, QS, and Greens at 6% apiece.

Regionally, the PQ managed a lead in the regions of Quebec with 37%. The Liberals were down to 24% and the CAQ was unchanged at 23%. The Liberals were ahead in and around Montreal with 37% support, a gain of eight points, while the PQ was down to 31% and the CAQ was down eight points to 18%. In Quebec City, however, the CAQ was in front with 34% to 26% for the Liberals and 24% for the PQ.
This would result in a rather comfortable majority government for the Parti Québécois, with 70 seats to the Liberals' 41 and the CAQ's 11. Québec Solidaire would win two seats and Option Nationale would win one.

How is this possible? The francophone vote makes all the difference. The CAQ has fallen quite a bit in and around Montreal, which makes no difference on the island but means a lot of PQ pick-ups off of it. And as the Liberals and CAQ are both well down in the regions, the PQ moves ahead in a lot of close contests. QS and ON have picked up support outside the main cities, but that has not come at the expense of the PQ.

There is no reason to expect the sort of under-estimation by the seat projection model of the Liberals' resilience as occurred in the last election, as that was in large part due to the more unpredictable distribution of the CAQ's support. In other words, there would be very good reason to expect that the PQ would indeed win a majority government with these regional numbers. Based on past performance, the probability of a PQ majority would be about 77%.

The poll had a few other interesting results. Satisfaction with the government stands at 37%, with dissatisfaction down two points to 54%. That is not horrible compared to the dismal numbers of the Charest government, but also a sign that Quebecers still remain quite polarized. In any case, in this divided electorate 37% support is enough to win a huge majority.

The budget that Minister of Finance Nicolas Marceau recently introduced gets a pass by 33% of Quebecers, which Christian Bourque of Léger Marketing said was actually a good result, considering that budgets are generally unpopular (Raymond Bachand's last budget had 20% support). Only 22% wanted the opposition to defeat the budget, which would result in another election (the CAQ has ruled out any idea of governing in coalition with the Liberals).

Support for sovereignty stood at 39% in the poll (49% among francophones), generally where it has been for some time. The Léger report shows that, since April 2011, sovereignty has been at between 36% and 43% support, a band narrow enough to say that it has held rather steady over that period. A recent Angus-Reid survey put support at only 32%, but had not removed the undecideds as is done by CROP and Léger. Among decideds, support was at about 37% in the Angus-Reid poll.

What is interesting about this question is that support for sovereignty is very strong among PQ and ON supporters, but is split 50-50 among supporters of Québec Solidaire. The party is sovereigntist but is apparently identified more as a left-wing party and a safe one for left-wing federalists to vote for. This would suggest that QS is not eating into the PQ's support as much as might be thought. In fact, at these numbers, Option Nationale is taking away about as much potential sovereigntist support from the PQ as QS is.

The poll also included a look at where the four candidates for the Quebec Liberal leadership stand. The result is generally the status quo, but with a big uptick in undecideds from 32% to 42%.
Philippe Couillard remains the favourite, but he has taken a hit. Among all Quebecers, his support was down six points to 31%. It was down 14 points to 34% among PLQ supporters. However, that still puts him well ahead of Raymond Bachand, who was down one point to 13% among all Quebecers and two points to 14% among Liberals. Pierre Moreau and Jean David are well behind. But while Couillard appears to have a wide advantage, reports suggest that among delegates the race is far closer between Couillard and Bachand.

The biggest challenge facing the next Liberal leader, aside from the potential fallout of the on-going Charbonneau Commission, will be to regain the sympathy of Quebec's francophones. They also prefer Couillard to Bachand by 32% to 15%, but whoever does win will have a steep hill to climb among this election-deciding electorate.

Tuesday, November 27, 2012

Three holds, two close races

The by-elections last night ended up as most by-elections do, with the incumbent parties holding on to their seats. The results in Durham and Calgary Centre went as scripted, with the Conservatives winning the first by a wide margin and the second by a narrow one. Victoria, however, was supposed to be a cakewalk for the New Democrats - instead, it almost gave the Greens their second MP in the House of Commons.

My detailed analysis of the results can be read at The Globe and Mail website. This post will focus more on the forecasts that were made and what each party can take away from the results.

Calling for the incumbent to win is never a risky bet, but the By-Election Barometer continued to have success with its ninth consecutive correct call. The results fell within the forecasted vote ranges for all parties in Durham and Calgary Centre (except for the NDP in the latter), but not in Victoria where the Greens performed unexpectedly well. That is itself a telling result.
Durham was the easy call last night, as there were few indications that the race was going to be much different than the one in 2011. Erin O'Toole easily won with 50.7% support, down less than four points. That's a performance that is generally in line with what most parties do when they lose an incumbent MP.

The New Democrats did quite well, though, picking up more than five points and distancing themselves from the Liberals, who did about as well as in 2011. While the result of 4.1% for the Greens may not look very good, that is actually not so bad in a by-election - the Greens tend to do very poorly in by-elections when they are not in the running.

The results all fell within the forecasted ranges, and quite close to the median forecast for both O'Toole and Larry O'Connor. The last poll by Forum gave the Liberals more of the Conservative vote, suggesting that some Liberals stayed home or opted for O'Toole when it became clear he would be the winner. Turnout was somewhat low, even for a by-election.
It was not as low as it was in Calgary Centre, however, which was supposed to be the most competitive race of the three. There were even news reports that turnout was brisk. Instead, turnout was less than 30%. That is a quite bad result, and before any accusations of vote-splitting are made there needs to be some questions concerning why neither Harvey Locke nor Chris Turner were able to turn out more of their vote.

In the end, Joan Crockatt won by a margin of 4.2 points, almost identical to the five point margin that was measured by the final polls of Forum and Return on Insight. While the scale is very small, this is a comeback of sorts for Alberta polling. Both firms were quite close, with Forum under-estimating Crockatt and Locke's support to the benefit of the NDP and RoI under-estimating Turner's support, again to the benefit of the NDP.

In fact, the New Democrats did quite poorly in this by-election. The forecast expected a minimum of 5% for the NDP, while the lowest poll result had them at 8%. Instead, they ended up with less than half of that. It would appear that many of the voters who cast their ballot for the NDP in 2011 went elsewhere or stayed home. Only the Conservatives lost more support, much of that going to Turner instead of Locke.

The results were again near the median forecast for Crockatt, Locke, and Turner, with Locke somewhat out-performing expectations and Crockatt under-performing. This is a good result for the Liberals, and it is a sign that they can be competitive in Alberta. This is also a very good result for the Greens, but it is hard to imagine that the party would be able to retain this level of support in a general election. The NDP lost out primarily due to the party not being seen as in the running compared to Locke and Turner, but the New Democrats are nevertheless the Official Opposition. While this is too isolated a case to call this a rebuke of Thomas Mulcair's western strategy, there certainly isn't any silver lining in Calgary Centre for the NDP.
There is in Victoria, however, where despite a steep drop in support the New Democrats held on to the riding.

Murray Rankin must have been somewhat nervous looking at the numbers coming in, as Donald Galloway of the Greens held the lead for much of the early counting. The last poll of the race was done two weeks before the vote and showed the NDP with a comfortable lead, though it also suggested that the Greens were making a push. But with generally poor polling numbers in British Columbia for the Greens of late, it was very surprising to see such a strong showing for Galloway in Victoria.

But this was something I had called a wild-card in the By-Election Barometer. Elizabeth May's riding of Saanich-Gulf Islands is next door and the Greens have generally done well enough in Victoria in recent provincial and federal elections. Turnout was about average for a by-election (though, at almost 44%, stellar compared to Durham and Calgary Centre), but the Greens certainly got their vote out to the polls. The party picked up over 6,000 new voters in this riding alone.

Rankin's result fell just below the forecasted range, while Galloway was just above it. A late poll might have shed more light on the race, but because there wasn't one the forecast had to rely more on province-wide data. That Galloway took so much more of the vote from Rankin than expected is an indication of just how local this race was. The Conservatives dropped a lot of votes as well, but did a little better than expected: instead of finishing just behind the Liberals, they placed just ahead. But both Dale Gann and Paul Summerville's results fell within the forecast ranges.

In sum, it was a decent night for polling as the results in Durham, Calgary Centre, and (to a lesser extent, but certainly in terms of the prospects of Gann and Summerville) Victoria generally aligned with expectations. The forecast was accordingly good, with only three candidates of the 12 forecasted falling outside their expected ranges (and that by an average of only 2.8 points), and most of the results within a few points of the median forecast. I think this is about as good as can be done in by-elections with the information available and the amount of unknowns at play.

It is difficult to choose a winner or a loser in last night's results. The Greens did about as well as possible without winning a race, making the by-elections a big moral victory for the party. That is only worth so much, however. The real test will be how Green candidates will do in Victoria and Calgary Centre in 2015.

The Liberals showed they can be competitive in Alberta, which is a strong signal from a party that has been increasingly written-off, both provincially and federally, in western Canada. But they slightly under-performed their 2011 results in Durham and Victoria. As that federal election has to be a low point for the party if they have any hope of a future, that is nothing to crow about.

The New Democrats held on to their riding in Victoria, so they have more to show from last night than either the Greens or Liberals. They were never in the running in either Durham or Calgary Centre. But they almost lost Victoria and their result in Calgary was very low for the Official Opposition. Growth in Durham, on the other hand, suggests that the NDP can make inroads in Ontario, an absolutely essential component to any winning NDP map in 2015.

As the Conservatives won two of three by-elections at stake, they have to be last night's big winners. But that is more due to the bullet they dodged. Had Crockatt lost the by-election in Calgary Centre (and she almost did), the Conservatives would have had a very bad night. Questions would have been asked about whether the party was shifting too far to the right, or if the split between Wildrose and the PCs could poison the party in the long-term. If these are indeed problems, the winning result could lead to the party ignoring them, but a win is a win and the Conservatives have plenty of time to rebuild whatever bridges have been burnt with a segment of the Albertan electorate. Durham was a strong result with over 50% support and only a slight dip in support (as should be expected when an incumbent retires), and though Victoria was a very poor showing, it isn't the sort of riding the Tories need to win.

Mixed results for all four parties, then. Silver-linings for the Greens and Liberals in their losses (which, in the end, would have been upsets anyway) and wins for the Conservatives and NDP despite their setbacks. All four parties have something to be proud of and something to give them pause. But with their huge increase in vote totals and vote share, the Greens are the only party that may have changed things last night.

Monday, November 26, 2012

Two similar federal polls

Last week, both Nanos Research and Forum Research released their latest numbers, and both polls gave a very similar picture of the federal race. That picture suggests a roughly five-point Conservative lead, with the Liberals and New Democrats neck-and-neck for second place. Further along in this post is a breakdown of tonight's by-election races.
We'll start with the Nanos poll. The firm was last in the field Oct. 4-11, and since then the Conservatives picked up 0.5 points to reach 33.8%. The Liberals were down 1.1 points to 29%, while the New Democrats were down 0.7 points to 27.2%.

The Bloc Québécois was up 0.2 points to 4.9%, while the Greens were up 0.8 points to 3.7%.

None of these shifts appear to be outside the margin of error, while the Conservative edge is just inside of it.

The Conservatives led in the Prairies, which in Nanos's polling includes Alberta, with 48.4%, followed by the NDP at 28.1% and the Liberals at 19.2%.

The Tories were also given the edge in Ontario with 39.7% support, compared to 33.3% for the Liberals and only 19.9% for the NDP. That is a very low score for the New Democrats in Ontario, but it is not unusual for Nanos to have these sorts of numbers.

At 39.4%, the Conservatives were ahead of the New Democrats (29.3%) in British Columbia, while the Liberals were down 14.8 points to 25.8% in the province. The last Nanos poll had the Liberals at 40.6% in B.C. It looked like an outlier at the time, and it is quite clear now that it was.

The Tories narrowly edged out the Liberals in Atlantic Canada with 34.9% to 34.7%, with the NDP down to 26.4% support.

The New Democrats did place first in Quebec with 33.6%, just ahead of the Liberals who were up to 31.6%. The Bloc Québécois was third with 19.3% while the Conservatives were down to only 11.5%.
Forum's national numbers were very similar to the ones put out by Nanos, seemingly confirming the state of the race.

Since Forum's last poll of Oct. 27, the Conservatives picked up two points and led with 33% support. The NDP was down four points to 28%, putting them tied with the Liberals, who were up one point. The Bloc and Greens were unchanged at 6% and 4%, respectively.

The drop by the NDP is noteworthy, while the Tories and Liberals were statistically unchanged.

The Conservatives led in Alberta with 61%, followed by the Liberals at 20% and the NDP at 15%. The Tories also led in the Prairies with 50%, a gain of 15 points, while the NDP was down 24 points to 24% and the Liberals placed third with 20%.

The Conservatives also had the advantage in Ontario with 34%, followed closely by the Liberals at 31% and the NDP at 29%. The Conservatives were also ahead in British Columbia with 42% (+15). The NDP was down to 34% and the Liberals were down 10 points to 16%.

The Liberals edged out the NDP in Quebec with 33% to 29%, while the Bloc had 22% and the Conservatives 14%. The Liberals were also in front in Atlantic Canada with 36% to the NDP's 34% and the Tories' 26%.

It is perhaps worth noting that neither of these polls included details on unweighted sample sizes.
The Forum poll would have resulted in the Conservatives winning about 145 seats on the proposed boundaries of the new 338-seat map. The Liberals would win 98 seats and the NDP 84 seats, with 10 going to the Bloc and one to the Greens.

For the NDP, falling behind the Liberals in Quebec and placing third in Ontario scuttles their chances of finishing ahead of the Liberals in the seat count. The close race in Ontario is, in large part, what deprives the Tories of a majority.
The results of the Nanos poll are quite similar, with the Conservatives winning 146 seats, the Liberals 97, the NDP 93, and the Bloc only two.

Again, Ontario and (to a lesser extent) Quebec is the reason for the third-place finish for the NDP. The Conservatives are hurt more in this poll, however, by their lower results on the Prairies.

Forum also included some polling on a Justin Trudeau-led Liberal Party, finding that his leadership would vault the party into first place with 39% support, dropping the Tories to 30% and the NDP to only 23%. The Trudeau Liberals would lead by significant margins in Ontario, Quebec, and Atlantic Canada and edge out the NDP in Alberta and the Prairies. In terms of seats, this would deliver 159 to the Liberals (10 short of a majority), 122 to the Conservatives, and 46 to the NDP.

More realistically, both Forum and Nanos had some personal ratings for current party leaders. Forum found that Stephen Harper's approval rating stood at 36%, compared to 55% disapproval. That is virtually unchanged from their last poll. Bob Rae's were also relatively steady, at 35% approval to 33% disapproval.

Thomas Mulcair's numbers worsened, his approval rating falling five points to 33% and his disapproval rating increasing four points to 34%. His approval rating among NDP voters stands at only 59%, compared to Harper's approval rating of 86% among Conservative voters (Rae sits at 57% among Liberals).

Nanos also registered decreasing numbers for Mulcair in their leadership index, but not to a significant degree. Harper managed 30% on trust, 41% on competence, and 33% on vision, giving him a total score of 104.2 points, a gain of 8.3 since Nanos's last poll. Mulcair scored 16%, 12%, and 18%, respectively, putting his leadership index at 43.6, down 4.1 points. Rae was up 4.6 points to 37.7, while Elizabeth May was Nanos's biggest gainer, jumping 7.6 points to 23.7 points.

As a share of points, Harper was well ahead with 48% to 20% for Mulcair, 17% for Rae, and 11% for May.

By-Elections tonight

Federal by-elections will be taking place tonight in Durham, Victoria, and Calgary Centre. The By-Election Barometer has a breakdown of the by-election races. Durham is forecast as a Strong Conservative riding, giving the Tories a 95% chance or higher of winning. Victoria is a Strong NDP riding, while Calgary Centre is a Likely Conservative win. The polls have indicated a close race there, but they have all put Joan Crockatt in the lead and the riding itself leans heavily towards the Tories.

In terms of specific vote share forecasts, I have seen criticisms that my confidence intervals are quite large. This is a recognition that, when it comes to by-elections, we know very little. In fact, it is quite unusual that we have had two polls done in Victoria and Durham and four in Calgary Centre during their respective campaigns. But these races are quickly evolving, and two weeks separates Forum's last polls in Victoria and Durham from today's vote. The last polls in Calgary Centre were done about a week ago, or even earlier.

During the Republican primaries, which were heavily polled, Nate Silver at FiveThirtyEight still had 20-point confidence intervals in his forecasts in Iowa for the leading candidates, and that kind of forecasted range was common throughout the primaries. The results were usually close to his median forecast, but not always. With more information going into these by-elections, it would be much easier to make tighter forecasts.

The race in Durham seems the most straightforward, and the forecast gives Erin O'Toole between 40% and 55% support, to between 19% and 30% for Larry O'Connor of the NDP. Grant Hume of the Liberals is expected to finish with between 15% and 25% of the vote, while the Greens could get between 4% and 10%. O'Connor does have some name-recognition in the riding so the potential for a stronger finish is there, but it seems a near certainty that O'Toole will win. The rolling average margin of victory, based on the swing from province-wide polling, stands at 16.8 points. In a normal election campaign, a call based on that size of margin would be right 81% of the time.

Victoria is another riding that shouldn't have any surprises. Murray Rankin of the NDP is forecast to take between 40% and 62% of the vote, while Donald Galloway is expected to place second with between 10% and 30%. The Liberals should finish third with between 13% and 25%, with the Conservatives taking between 10% and 25%. The rolling average margin of victory of 25.8 points would result in a correct call in a general election 90% of the time.

Calgary Centre is the real race to watch tonight. The polls have been close and the campaigning has been fierce. How will it all shake out tonight? What role will turnout play? Will strategic voting decide the result? This is, without a doubt, the most difficult by-election to call of the three and one of the most interesting in recent memory. The forecast is for Joan Crockatt of the Tories to take between 30% and 55% of the vote, with Harvey Locke of the Liberals placing second with between 24% and 37%. Chris Turner of the Greens could also do quite well, with a forecasted result of between 10% and 30%. The NDP should finish fourth with between 5% and 20%. The average margin in the provincial polls is 14 points, which would result in the right call 77% of the time. But the margin was five points in the last two riding polls - the kind of call that the model would only have 60% confidence in making. The forecast still expects a Conservative victory, but it would also not be surprised if Crockatt does not win.

Friday, November 23, 2012

New Calgary Centre by-election poll

Late last night, Return on Insight released a new telephone survey of voting intentions for the upcoming Calgary Centre by-election race. The results generally fall in line with what Forum Research has been reporting, but the poll was heavily criticized on Twitter due to its source. After trading emails with Bruce Cameron, President of RoI, and after being provided with a copy of the raw data report, I believe the results of the poll are credible.
The live-caller survey was conducted November 20-21, so a few days after Forum's last poll of November 17 but just before the brouhaha over Justin Trudeau's newly rediscovered comments from 2010 about Alberta. However, the poll was conducted just as the comments made by David McGuinty were breaking, along with the lightning-quick apology and resignation as critic.

The poll found Joan Crockatt of the Conservatives ahead with 37% support, five points up on Harvey Locke of the Liberals, who was at 32%. That gap is within the margin of error, but again we see that in every poll Crockatt has had the advantage. The odds that all of these polls have been wrong in the same direction is quite low, making a Crockatt lead a very strong possibility.

Chris Turner of the Greens was at 17%, while Dan Meades of the New Democrats was at 12%. Another 2% said they would vote for other parties.

As this poll was done with live-callers, the undecided number is more reliable than what we get from IVR or online surveys (the last IVR poll had about 7% undecided). RoI finds the number of undecideds to be at 16%, while 12% of respondents said they wouldn't vote.

The cross-tabs of the poll has one interesting similarity with Forum's last survey. Though the sample sizes are very small, both Forum and RoI found that Crockatt was leading among voters aged 18-34, while the race was much closer among older voters. That is counter-intuitive, as Conservatives tend to do better among older voters. But there is some indication that younger voters in Alberta are far more conservative than their fellow Canadians, based on some other polling I have seen of Millennials. So, this may not be as unusual as it looks. If it is indeed the case, that does not bode well for Crockatt as this tranche of the electorate does not turnout in large numbers.

In terms of the gender gap, however, the polls are not in agreement. Forum's last poll found Crockatt with a three point edge among men and a seven point advantage among women, while RoI gave Crockatt a nine point lead among men and a two point edge among women.

As to the criticisms that Bruce Cameron is tied too closely to the Liberal campaign, this is what Cameron has to say:

"I am not doing any polling or paid work for the Locke campaign, despite a three second video clip of Harvey [Locke] ripped out of contest pre-writ by the Greens claiming I am ... Harvey is a good friend but I made it clear to him and the campaign manager Donn Lovett that I could not work on the campaign. Nevertheless, Harvey and I talk regularly as we have done often over the past 20 years."

At my request, Cameron sent me the poll's raw data. Having looked over the report, I have little reason to believe that this poll is any less credible than any other. The amount of weighting that was done was quite small (if only this sort of information was made available by some other firms), with the usual corrections being made to account for the low number of younger respondents. There are no major problems with the wording or ordering of the questions asked or anything of the sort. Though I do believe that pollsters should not get involved in politics or even give the impression of being supportive of or opposed to any party or candidate, there are many examples of this happening and I do not believe that it is necessarily a reason to discount a pollster's work without any other justification.
That being said, there is the potential that the poll is under-estimating Crockatt's support as the number of respondents who claimed to have voted Wildrose in the April 2012 election was about three-quarters of what it should be. This could be the "shy Tory"  effect at play, bad memory, or confusion between provincial and federal politics. We will find out on Monday. EDIT: An earlier version said that Wildrose support was registered as 1/3 of what it should be. After verifying with Cameron, this was apparently a glitch in coding. Nevertheless, Wildrose support was still somewhat under-represented, but perhaps not significantly so.

The poll included a few other interesting questions, including certainty to vote. The chart in the media report has some errors in terms of the certainty of all respondents to vote (and also contains a typo in the question, the actual question asked did indeed identify November 26 as election day), but the numbers for each of the parties appears to be correctly recording the certainty of decided voters. It says that 64% of Locke's supporters are certain to vote, compared to 67% for Crockatt, 72% for Turner, and 79% for Meades.

If we just use those findings and assume that it will represent turnout, Crockatt's advantage increases to 38% against 30% for Locke, 18% for Turner, and 14% for Meades.

The survey also asked non-Conservative-supporting respondents who they would choose if they could only select one 'progressive' candidate to defeat Crockatt. The result seems to be very favourable for Locke (40% to 17% for Turner and 14% for Meades), but if we remove the undecideds from the equation we get 56% for Locke, 24% for Turner, and 20% for Meades. That looks better for Locke than it actually is: on the regular voting intentions question, Locke is already getting 52% of the non-Conservative vote, compared to 28% for Turner and 20% for Meades. In other words, when asked this question most people were sticking with their original choice.

This poll does not disagree with Forum's last two surveys to any significant degree. Crockatt and Locke's support is well within the margin of error of Forum's last poll, while Turner's and Meades' is at the outer edges of it. We will have to wait and see if any more numbers will emerge over the weekend (Forum does enjoy the election-eve release). At this stage, Crockatt is still the odds-on favourite but both Turner and Locke could put up some impressive numbers. Will it be enough?

Thursday, November 22, 2012

Tight provincial race in Toronto

Lost in the weeds of a poll on the fortunes of Mayor Rob Ford of Toronto, Forum Research's latest poll on the feelings of Torontonians included provincial voting intentions, as well as thoughts on the on-going Ontario Liberal leadership race. The results show that the vote is splitting three-ways in Toronto, and that - shockingly - Torontonians favour a local boy for the leadership.
The poll found the New Democrats ahead in Toronto with 34%, putting them narrowly up on the Liberals, who were at 31%. The Progressive Conservatives came up third with 29%, while the Greens were well back at 5%.

Forum broke down the city into four regions, with each party leading in at least one of them. The New Democrats were ahead in Toronto/East York and in Etobicoke/York, while the Liberals were ahead in North York. The Tories had the edge in Scarborough.

But each part of the city was also split: the Liberals were in the race for Toronto/East York and the Tories in Etobicoke and North York, while all three parties were competitive in Scarborough.

That the Liberals are leading in North York should not be too surprising, they won all of the area's seats in October 2011. The NDP also did quite well in that election in their two pockets of support in this poll. But the PCs being ahead in Scarborough makes it quite possible that Tim Hudak could make a mini-breakthrough into Toronto. This could be an anomaly of the poll, as the Tories did not do very well in the Scarborough ridings in the last election, but it does suggest that the electoral map of Toronto could be a mish-mash of colours when Ontarians are called back to the polls.

Forum's look at the Ontario Liberal leadership race in Toronto tells us little about how the convention will unfold, but it does give an indication of what Torontonians, the last bastion the Liberals can still (mostly) count upon, would like to see as the outcome.
Not surprisingly, Gerard Kennedy and Kathleen Wynne, both from Toronto, topped the list. Kennedy came out well ahead, however, with 22% support as the best choice among all polled residents of the city. Wynne came up second with 11%, while Sandra Pupatello (Windsor) was third with 6%. Eric Hoskins (5%), Glen Murray (4%), and Charles Sousa (1%) rounded out the list.


Among Liberal voters in Toronto, Kennedy does better: 32% to 12% for Pupatello and 10% for Wynne.


It is worth noting that 27% of all respondents said they had no opinion and 21% said "none of these", leaving a lot of points on the table. If we remove them, along with the 3% who said "someone else", we get Kennedy at 45% to Wynne's 22% and Pupatello's 12%.

Word is that Kennedy's campaign is not as well organized as those of some of his main competitors, but having an edge in Toronto could be important. If most of the delegates from ridings in the city vote for him, he will have a solid base of support on the first ballot. Whether he will be able to grow that support, however, will be the big question at the convention - as it was in 1996.

Tuesday, November 20, 2012

Likely close contest in Calgary Centre

The last poll by Forum Research in Calgary Centre spilled a lot of ink last week, as it showed a mere two point lead for Conservative candidate Joan Crockatt over Liberal Harvey Locke in what is supposed to be one of the safest regions for the Tories in all of a Canada. But it had the look of an outlier poll - that is, until Forum did another survey this week and showed that the race was looking no different.
When originally faced with such an outlandish result, the smart money was to consider the poll more likely than not to be misreading the race. To have another poll showing no statistically significant difference adds a lot of weight to the argument that Calgary Centre is indeed a horse race, which no longer seems so unthinkable considering how things have been unfolding in the riding. Nevertheless, the possibility that Forum is still misreading the race for methodological reasons cannot be dismissed out of hand, but there is less reason for caution than there was a few days ago.

The poll found Crockatt at 35%, up three points since Forum's Nov. 12 poll. Locke was unchanged at 30%, while Green candidate Chris Turner was up two points to 25%. Dan Meades of the NDP was down four points to 8%, while 2% of respondents said they would vote for another party (-2). As mentioned, none of these shifts of support are outside the margin of error.

While the evidence leans in favour of a Crockatt lead (she has not trailed in any of Forum's polls) the gap between her and Locke is within the margin of error. But so is the gap between Locke and Turner, though again the most likely situation is 1-2-3 for the Tories, Liberals, and Greens.

But this poll is not completely beyond questioning. Though I highlighted Forum's recent successes in riding polls last week, there are a few points to consider. Firstly, the poll was conducted on a Saturday, generally a bad day to do a poll.

Secondly, the findings in the report on how respondents voted in the last federal election do not match up entirely with the real results. The totals for the Tories and Greens are within the MOE, but it seems possible that the sample is slightly (maybe by a point or two) over-sampling Liberals and under-sampling New Democrats (by two or three points). This can be due to lapses in memory or a changing population (2% said they voted for the Bloc, so they are either displaced Quebecers or fibbers), and is probably not enough to cause a major problem. We also don't know how Forum weighs these responses, if it all. If they don't, there is the possibility that Crockatt's lead is slightly under-estimated.

But it is hard to tell with sample sizes of about 400 people or less in each of the polls Forum has conducted during the by-election campaign. It is actually theoretically possible that the poll is gauging things correctly within the margin of error and Crockatt's support has not dipped below 37% over the last few weeks, or that Locke's has never been higher than 25% since the campaign began.
That is, however, more unlikely than the true support levels being closer to Forum's estimation. Theoretically at least, the odds that the real situation is at the extreme of the margin of error is quite low. But recent elections have demonstrated just how much theory can depart from reality.

The chart above shows support for each party and the bands of support that is accounted for by the margin of error. The band is narrower, for example, for the NDP than the Conservatives because margin of error is calculated differently depending on the support a party has. To put it simply, a party can't have a margin of error of +/- 5% if its support is gauged at 4%.

The chart shows how it is possible for there to be little movement from poll-to-poll. The Liberals being stagnant at 26% is possible throughout Forum's polling. Conversely, it is also possible (though far less likely) that the Liberals have gone from 16% to 35%. What this chart does show quite clearly, however, is that a drop in Crockatt's support has almost certainly occurred, with almost all of that going to Turner of the Greens.

Despite the close race, a lot of factors weigh in Crockatt's favour. Most importantly, she has yet to relinquish the lead in any of these polls. She is the closest thing to an incumbent, and her support is likely to be drawn from those older voters who are more likely to turnout. It is possible that the enthusiasm the Turner and Locke campaigns have created (simply because it doesn't look like they'll be trounced) could boost turnout among their supporters and that Crockatt's performance in this campaign will depress her own turnout, but the fundamentals of the riding still point to her having the advantage. Nevertheless, this last week could be decisive.

Monday, November 19, 2012

Conservatives lead by four in latest Ipsos

Late last week, a new poll from Ipsos-Reid was released further confirming that the race remains a three-way contest, though the Conservatives and New Democrats still hold the advantage over the third-place, but gaining, Liberals.
Reports that have accompanied this poll emphasize just how much the New Democrats have lost to the benefit of the Liberals. That is not inaccurate, but the comparison needs to be put into context: Ipsos-Reid was last in the field June 20-21.

June was the high-watermark of NDP support, so that they have dropped so much from a June poll is not much of a surprise.

Specifically, they were down eight points to 30%, putting them four points behind the Conservatives. They slipped one point to 34%. The Liberals were up eight points to 26%, while the Bloc Québécois was at 7% and the Greens at 2%. The shifts in support for the Liberals and NDP appear statistically significant.

It should be noted that this Ipsos-Reid poll was actually in the field just before last week's Abacus Data poll, and it also has to be noted that the poll Ipsos did in June was conducted online only. This poll was a hybrid telephone/online poll, so any shifts in support since their last survey could be partly the fault of the different methodologies that were employed.
But the changes Ipsos recorded since their June poll fall well into line with other recent surveys. On average, the Liberals have gained 5.4 points in polls by the same firms taken in June/July and October/November. The NDP has dropped 4.9 points and the Tories have slipped 1.1 points.

Clearly, the last four or five months have good for the Liberals. Their gain has come primarily at the expense of the NDP. That plays into the hands of the Conservatives, but at current support levels they would only likely win a minority government - which could spell the end of their time in power.

The result of 2% for the Greens ties their worst result in any poll taken since the May 2011 election, and even well before that. However, there are two reasons for this low number for the Greens. When Ipsos-Reid does their telephone polling, they do not "prompt" for the Green Party as they do for the other parties. Respondents are allowed to say they will vote for the Greens, and that is recorded, but they are not listed along with the other parties in the survey questions. Nanos Research also does not prompt, though they do not prompt for any of the parties.

This has the effect of giving lower scores for the Green Party. If we look at the average results of pollsters who have been in the field at least three times since March 24 (when Thomas Mulcair became leader of the NDP), we see that Ipsos-Reid and Nanos each average 3.8% Green support over that time.

Forum, which uses an IVR method that does prompt for the Greens, comes in third with 4.1%. Abacus's online panels yield an average result of 5.8%, while the prompting live-callers of Environics and Harris-Decima get the highest results for the Greens. Recent elections suggest that not prompting will get a more accurate result for Green Party support, but that still probably under-scores its actual support among the general (including non-voting) population.

But there is a second reason for Ipsos's low number. While unprompted support for the Greens is recorded in their telephone polling, respondents to their online poll only get the option to support "Other". The average that was calculated above included both Green and Other support. In Ipsos's last online-only poll, support for other parties was 4%. Generally speaking, it is safe to assume that most of this is support for the Greens, but in this case it explains why the party scored so low. That 2% for other parties comes largely from the Green supporters who responded to the online survey. That makes Green support likely closer to 4% than to 2%.

But back to the poll itself. The Conservatives led in Alberta with 63%, followed by the Liberals at 18% (+13) and the NDP at 13% (-11). The Tories also led in the Prairies with 46%, while the NDP was second with 28% and the Liberals third with 19%.

The New Democrats led in Quebec with 34%, followed by the Bloc Québécois at 26%, the Liberals at 25% (+10), and the Conservatives at 14%.

The Conservatives had the edge in Ontario with 36%, putting them six points ahead of the NDP (-10) and the Liberals (+8), who were tied at 30%. They also had the advantage in British Columbia with 43% to the NDP's 34% and the Liberals' 19%.

The Liberals were narrowly up on the New Democrats in Atlantic Canada with 35% to 34%, while the Conservatives were third with 25%.
With these numbers, the Conservatives would win 144 seats on the proposed boundaries for the new 338-seat electoral map. The New Democrats would win 98, the Liberals 84, and the Bloc Québécois 12.

This puts the Tories in a minority situation, whereas the New Democrats and Liberals have enough seats to command a majority. The NDP's lower-than-usual results in B.C. and the Prairies hurts their numbers, while the Conservatives are hamstrung by the closer race in Ontario. But at 26 seats short of a majority, they would also need to do better on the two coasts.

This Ipsos-Reid poll confirms some recent trends, particularly the large Liberal gain in Quebec. While that is the major reason the NDP has slipped further from the Conservatives (put them back to 43% in the province and the national margin shrinks to one or two points), the Liberals have stolen enough points here and there from the party to make things even more difficult. But that does not absolve the Conservatives, who are down six points from the 2011 election. They are keeping their head above water, but they are far from being in a comfortable position, and have been for the last year.

Thursday, November 15, 2012

A close race in Calgary Centre?

Polls by Forum Research in each of the ridings holding by-elections November 26 caused quite a bit of a stir yesterday. Though two of them showed the incumbent parties well ahead in their traditionally safe ridings, the third showed the Conservatives in danger in their own backyard. What to make of this?
The survey for Durham is the least controversial. It showed Erin O'Toole of the Conservatives well ahead with 42% to 26% for the NDP's Larry O'Connor. Grant Hume of the Liberals was third with 22%. That generally fits in with expectations. Forum last polled Durham (as well as the two other ridings) on October 26, but no shifts of statistical significance have taken place.

The results for Victoria are also relatively unsurprising. Murray Rankin of the NDP led with 47%, while Donald Galloway of the Greens was up six points to 26%. Paul Summerville of the Liberals was third with 14%, while Dale Gann of the Tories had 12% support. The gain for Galloway is outside the margin of error, and certainly an interesting aspect of this race.

But the results in Durham and Victoria don't raise any alarm bells. The results in Calgary Centre do.

Forum found that Joan Crockatt of the Conservatives had dropped 16 points to 32%, putting her only two points ahead of the Liberals' Harvey Locke. His gain of two points is within the margin of error. Chris Turner of the Greens, however, gained 12 points to reach 23%, putting him in a solid third place and in the running. The NDP's Dan Meades was in fourth with 12%.

While I'll grant Thomas Mulcair his criticism of polling in Alberta in his dismissal of Meades' low numbers (he's actually up four points, though that is within the margin of error), his statement that the polls indicated the NDP was in trouble in Quebec in the last federal election doesn't hold water. The NDP was in a very comfortable position in the final week of the campaign in the province.

Is it possible that the Conservatives are neck-and-neck in Fortress Alberta? It is not shocking to see the Liberals at 30% in Calgary. Their provincial counterparts do better in that city and it isn't outlandish to see the Liberals at this level of support considering their province-wide polling at the moment. Having the Greens so high and the Conservatives so low is unusual, and would mean that local factors are seriously at play.

This is plausible, considering some of the stories that have appeared in local media lately. There is the split between the PCs and Wildrose and the fact that Turner is a relatively good candidate for the Greens. Nevertheless, this is Alberta.

It is worth noting the difference in sample sizes between Calgary Centre and the other two. Surveying only 376 people is not huge, and the lower number of respondents suggests that the response rate was quite low in the riding.

Forum uses the IVR method, which can have mixed results. Nate Silver recently calculated that "robodial" polls performed worse than their online and live-caller counterparts in the recent American election, finding an average error of five points for IVR polling compared to 3.5 points for telephone and 2.1 points for online surveys. However, some IVR firms performed well while others were near the bottom. This is actually to be expected - robodialling is the cheapest way to do polling so it is more likely that incompetent polling firms would still take a stab at it. There is more chances that the men will be separated from the boys when it comes to IVR polling than for online and telephone polling, as those methods are much more expensive and non-serious players are less likely to use them.

So where does Forum lie on the spectrum? We actually have some recent by-elections to use as a guide. Forum was active in both Kitchener-Waterloo and Vaughan when those ridings held by-elections in early September. Forum actually did quite well, with an average error of 3.3 points per party (PC, NDP, Liberals, and Greens only) in Kitchener-Waterloo and 1.8 points in Vaughan. In both ridings, Forum made the right call, including the upset NDP victory.

Forum also did well in the three riding polls they released during the provincial Quebec election. They had polled Orford, Saint-François, and Sherbrooke for a local English-language newspaper, and had an average error of two points per party (PQ, PLQ, CAQ, and QS only) in Orford and Sherbrooke and 1.5 points in Saint-François. So far, so good.

But a general rule in polling is that if something looks like an outlier, it probably is. Does the Calgary Centre poll look like an outlier? From the outside looking in, considering the riding's history, it looks like it could be. Locals may have a better clue.

However, while some IVR firms did well in the recent presidential election and Forum has had a string of good riding level polls of late, the method they use appears to be more vulnerable to wild variations between polls.

The Quebec provincial election is a good example of that. Let's look at the variations from poll to poll for the top four parties in the polls done by CROP (using live-callers), Léger Marketing (using an online panel), and Forum (using IVR). For example, if the PQ gained two points and the Liberals lost three points in a poll, while the CAQ and Québec Solidaire held steady, that would be a total variation of five points.

Léger had the most consistent polling, with an average total variation from poll-to-poll of 4.4 points. In one of their polls, all four parties had the same level of support as they did in their previous survey. In their most volatile poll, they had a total of eight points changing between the four parties.

CROP's polling was also consistent, with an average total variation of 5.8 points. Their lowest total change between polls was three points, their highest was nine.

Forum, however, had very inconsistent polling. On average, their numbers shifted by 14 points from one poll to the next. On two occasions the total variation was over 20 points between polls, and only once was the total variation lower than Léger's and CROP's average.

This is not to impugn Forum's accuracy; it is certainly possible (though less likely) that Forum was accurately depicting a volatile race while Léger and CROP were inaccurately suggesting a more stable electorate. But it does raise alarm bells when we see that the vote in Calgary Centre has shifted so radically, considering Forum's recent experience with IVR polling.

In the end, Forum could turn out to be right on the money and Calgary Centre could indeed be a very close race. If another poll emerges we will have a better idea of whether that is the case or not. For the moment, though, it might be better to exercise caution.

Wednesday, November 14, 2012

NDP support drops in Abacus poll

Yesterday, Abacus Data released their latest national survey suggesting that the New Democrats have taken a big hit over the last two months, almost entirely to the benefit of the Liberals. But the Liberals were already starting from quite low in Abacus's polling, meaning that the net effect is a wider Conservative lead rather than a three-way contest.
Abacus was last in the field Sept. 14-18, and since then the Conservatives picked-up one point to lead with 36% support. The New Democrats fell six points to 29%, while the Liberals were up five points to 22%. These latter two are statistically significant shifts in support, and the margin between these two parties has been narrowing in most recent polls.

The Bloc Québécois and Greens were unchanged at 7% and 6%, respectively.

Abacus conducts its polling using the online panel run by Angus-Reid. Their report includes both weighted and unweighted sample counts for the regional distribution, though not for other demographic categories. Nevertheless, Abacus is always very transparent about their polls when approached.

The gender gap is not nearly as wide in this survey as we have seen in some others: the Tories lead by eight among men and six among women.

It also worth noting that a majority of respondents now think that the country is headed in the right direction, up 10 points to 51% since September.

The Conservatives led in this poll with 43% in Ontario, followed by the New Democrats at 28% and the Liberals at 23%. That 43% number is a very big one for the Tories - they have been pegged at or above 40% in only six of the last 36 polls in the province.

The Tories also led in Alberta with 60%, followed by the NDP at 18% (-11) and the Liberals at 17% (+11). In the Prairies, the Conservatives led with 49% to 27% for the NDP (-16) and 16% for the Liberals.

In Quebec, the Bloc Québécois actually moved ahead with 31%. The NDP fell nine points to 30%, while the Liberals were up eight points to 21%. The Tories brought up the rear with 14%.

The province appears to be going through (another) period of flux. If we look at the last 10 polls done in Quebec (stretching back to the end of August), we see that the NDP has averaged about 33.8% support. But in the 10 polls before that (stretching back to mid-June), the party averaged 39.5%. In contrast, the Liberals averaged 17.4% in the older polls but 22.4% in the newer ones. They appear to have gained about five points, almost all of it at the expense of the New Democrats. That should be worrying to Thomas Mulcair.

But the party was ahead in British Columbia with 40% to the Tories' 34% and the Liberals' 19%, and edged out the Liberals in Atlantic Canada with 36% to 31% (+12). The Conservatives were third with 25%.
With these numbers, the Conservatives would likely win around 167 seats on the proposed boundaries of the 338-seat map. That puts them three short of an outright majority, but that margin is close enough that the Tories could easily pull it off.

The New Democrats win 82 seats and the Liberals win 50, while the Bloc takes 38 seats and the Greens keep their one.

That swing in Quebec is hugely important, as though the Bloc did not make any real gains they move ahead in many ridings simply because of the NDP's slip. A lot of them would be incredibly close, however, so the potential for a dozen or so seats to go back to the NDP is not out of the question. But that is somewhat besides the point - the Conservatives aren't at play in Quebec.

The real problem in this poll for the opposition is Ontario, where the Conservatives win 85 seats. The Tories are about where they were on election night in this survey, but the Liberals are several points down - reducing their seat haul further.

All around, the numbers are quite good for Stephen Harper in this poll. In addition to the increase of people who think the country is headed in the right direction, his job approval rating increase by three points to 39%, while his disapproval dropped to 44%. His favourability was unchanged at 35%, but his unfavourability was down four points to 46%. Mulcair has better overall numbers (29% favourable to 24% unfavourable), but this poll suggest a steep decline in Mulcair's favourability rating (and an increase of his neutral rating by five points to 36%).

The Conservatives need to rebuild a few bridges in British Columbia and Atlantic Canada, but have a far less formidable task ahead of them than either the Liberals or the NDP. Being this far behind in Ontario is a big problem, and neither party has a chance of unseating the Tories (at least, on their own) unless one of them dominates in Quebec like the NDP did in 2011. The Bloc's vote in the province seems somewhat tied to the support for the provincial Parti Québécois, so how Quebec's provincial politics will shake out over the next three years (which will almost certainly feature another election) could be very important.

Monday, November 12, 2012

Alberta Tories still in control

A few days before this weekend's annual general meeting of the Progressive Conservatives in Alberta, Environics released the details of their latest survey of provincial voting intentions. They show that the Progressive Conservatives remain comfortably in the lead, and have made inroads in rural Alberta.
Environics was last in the field Aug. 10-22 and since then the PCs picked up two points, leading with 45% support. Wildrose was up three points to 29%, while the Liberals at 13% and the New Democrats at 12% each lost one point.

None of these shifts in support were outside of the margin of error of this telephone poll. Environics' report on the poll is compact, but thankfully includes the unweighted and weighted samples of their survey.

Alison Redford's Tories led in both Edmonton and Calgary with 43% and 45% respectively, putting them ahead of Wildrose (23% and 25%). The Liberals did better in Calgary than they did in Edmonton, with 20% to 14% support, while the opposite was the case for the New Democrats (19% to 9%). All in all, this is generally what we've seen in Alberta for some time.

The Progressive Conservatives also led in Alberta's small cities with 49%, while Wildrose picked-up nine points to hit 31%. The New Democrats trailed with 13%, while the Liberals were down 12 points to 6% in the province's smaller urban centres (where they hold no seats).

In the rural parts of the province, the Tories were up 10 points to 47%, putting them ahead of Wildrose. Danielle Smith's party fell to 43%. The NDP managed 6% support and the Liberals were down five points to 4%.

As this poll was taken during the month of October, any of the fallout from recent controversy over MLA compensation or the decisions made at the annual general meeting concerning the party's links with the federal Conservatives was not recorded. It is unclear whether any of this will have any effect - the Tories have been comfortably leading in every poll since the last election and, so far, do not seem in danger of losing many of the centrist supporters they attracted in April back to the Liberals.

Wildrose, meanwhile, has a lot of ground to make up. They've lost about 10 points in Calgary and, whereas they were neck-and-neck with the Tories in the last election outside of the two main cities, this poll puts them about 10 points behind. There is no reason for Wildrose to hit the panic button just yet, however, as the next election is scheduled for 2016.

Thursday, November 8, 2012

October 2012 federal polling averages

October was a quiet polling month, at least on this side of the border. Only two national and one provincial (Quebec) polls were conducted in October, down quite a bit from the eight polls that had been done during the month of September. Though the monthly averages are not drawing from as large of a sample as is usually the case, here they are for the sake of continuity.
The Conservatives averaged 32% in the month of October, down 1.5 points from their average result in September. The New Democrats were down 1.1 points to 30.2%, while the Liberals were up six points to 28.3%.

The Bloc Québécois averaged 5.4% support and the Greens 3.5%, and an average of 0.5% of respondents said they would vote for another party.

Regionally, the Tories had the edge in Alberta and Ontario while the New Democrats were ahead in Quebec and the Prairies. The Liberals were in front in Atlantic Canada and British Columbia.

This represents a huge jump in support for the Liberals nationwide, but also in most parts of the country as well. This is in large part due to the big number the party put up in October's Nanos poll, but is there anything we can discern from the trends across both polls that were conducted this month?
Not particularly, at least compared to September. Both Nanos and Forum were in the field in both months, but they had the Tories and NDP heading in different directions. Nanos saw a small uptick in Conservative support while Forum saw a drop, and Forum recorded an NDP gain and Nanos a slip. They both agreed, however, that the Liberals were gaining - by an average of 3.8 points.

If we look at the previous month in which both Nanos and Forum were in the field, we see some clearer trends. Since July, the Conservatives have held relatively steady while the New Democrats dropped an averaged of 2.7 points. The Liberals made large gains in both polls since July.
But compared to September, the October numbers knock the Conservatives down 24 seats to 128 on the 338-seat map. The New Democrats would win 115 seats, up six since September, while the Liberals would win 92 (up 29).

The Bloc would win only two seats (down 11) and the Greens would win one.

Approval ratings
The poor performances of the Conservatives in British Columbia (due in large part to the Liberals) and in the Prairies (due to the NDP) makes it very difficult for the party to win anything but a bare plurality of seats. The New Democrats would need to make up some ground in Ontario in order to move ahead of the Tories, but a Liberal gain in the province is more efficient at whittling down the Conservative number.

With only two polls in the field in October, it is impossible to determine if some of the wilder results (particularly in British Columbia) are outliers or not. It does seem that the Liberals are eating into the support of both the Conservatives and the New Democrats, and it is difficult to separate that from the on-going leadership race. I suspect that we will continue to see some strange fluctuations until the next leader is chosen.

Tuesday, November 6, 2012

Nate Silver and the trials of a forecaster

Election night update: Congratulations to Nate Silver on a job well done!

Morning after update: After a long night and early morning, there won't be any posting today. But I invite you to read my article for The Globe and Mail in which I assess how the various forecasters did. The polls did quite well last night, though it is worth noting that Silver's forecast was better than that of RealClearPolitics, which does a simple averaging of the polls. It all makes me lament for the polling situation north of the border - how jealous I am of the amount and quality of data that is available in the United States!
-----
There is an election today. Millions of Americans will be heading to the polling booths today - if they haven't cast their ballots already - and will be glued to their televisions as the results start pouring in. Millions of people around the world will also be paying close attention to the election results, including a good number of Canadians.

Polls have taken up a large space in coverage of the election, and in the last week or so that focus has been especially directed at Nate Silver, who runs the FiveThirtyEight blog for The New York Times. As Silver's forecasts a few weeks ago still showed that Barack Obama was favoured to win, despite the President trailing by a narrow margin in most national polls, he was heavily criticized for being a partisan hack or simply wrong. Some of these criticisms were well-argued and thoughtful, but most were innumerate or downright personal. The backlash against these criticisms resulted in some of the best analysis of what forecasters do that I have ever read, and even more attention was thrust upon Silver (though I don't think he welcomes it).

As long-time readers of this site know, and as anyone who can look at the URL in their browser and put two -and-two together will figure out, I am an admirer of Silver's work. I was inspired to launch this site because of his own work during the 2008 presidential election, and many of the methods I employed were inspired by his. The idea of weighing the pollsters by their track record, for instance, comes directly from FiveThirtyEight. But after being in operation for more than four years (almost as long as 538 itself), ThreeHundredEight has developed its own distinct character.

But despite very different methods and backgrounds (not to mention track records, though I suspect Silver would struggle just as much with our tricky electoral system and lack of good and plentiful polling), we work in the same field. As a result, a lot of the criticisms that have been aimed at Silver over the last week have hit close to home. Though I have (thankfully) only rarely received the same kind of treatment in national media, and the scale is greatly different, I have seen some of the same kind of criticisms (and insults) on the Internet, in my inbox, and on my Twitter feed, and know what it is like to have your motivations and competence unfairly called into question.

Though this site and others like FiveThirtyEight, and there are several of them in the United States, are thought of mostly as prediction sites, that only scratches the surface. I have always found Silver's analyses of polls to be far more interesting than his forecasts, which have always been based on what the polls are saying anyway. This site tries to do the same thing - providing analysis of polls that goes beyond what can be found in most media coverage, and providing more than just a guess as to what could happen.

Polls take up a lot of place in media coverage in Canada as well as in the United States, and there is a corresponding need for responsible poll reporting in this country. Readers need to understand what polls are able to tell us and what they are not. And they can only gain so much from reading about a single poll in their favourite newspaper. As the US election has made quite clear, different polls can say very different things.

Ten polls released in a short time span each tell 10 different stories. Some of those stories will be an accurate depiction of what is happening, and some of them will not. Focusing on one poll or another will lead people to miss the forest for the trees.

The usefulness of a site like ThreeHundredEight or FiveThirtyEight is in telling the story of what those 10 polls are saying, and not just in terms of aggregation and prediction. Silver takes a macro look at all of the polls and describes a narrative that the polling data - and only the polling data - backs up. I try to do the same here. Our projections are about what the polls are saying will happen based on the information available now and what has happened in the past, not what we think or hope will happen, and both of us are always quick to include plenty of caveats.

And those caveats are incredibly important. As of writing, Silver's model indicates that Obama has a 91.6% chance of winning. That means Mitt Romney could still win, but the odds are heavily stacked against him. With an eye towards the limitations of public opinion polling, Silver is forecasting that the odds that the polls will be wrong enough to give Romney the win are very low. There is little valid argument to make against such a statement.

Looking at our recent electoral history in Canada, the fact of the matter is that in only 44.5% of cases has the polling error been large enough to erase the 2.6-point margin that Silver current gives Obama in the popular vote. And as those errors have an equal chance of going one way or the other, that would give Obama a 77.7% chance of winning the popular vote, based on how often the polling margin between two parties in Canada has turned out to have been proven wrong on election night. Add Obama's electoral college advantage to that, and you can see why Silver gives Obama such good odds. But that still doesn't mean Romney can't win.

I had my own set of caveats in my final projection for the Alberta election, which was very open to the possibility of the Progressive Conservatives winning a majority government. I was heavily criticized at the time for that openness - until the results started pouring in. (Then I was criticized for not giving the Tories a thumping majority to begin with.) In the recent Quebec election, I forecast that the Parti Québécois would win, but the question was whether it would be a majority or minority government. In future elections I intend to emphasize more of this uncertainty.

These sorts of caveats and confidence intervals might seem like hedging, but it is in fact the responsible thing to do. There always needs to be a recognition of what we do not know - and the potential error in polls is one of those things. But there also has to be a recognition of what is most likely to happen based on the information that is available to us, and that the information is going to be reliable more often than not.

Criticisms of the specific nature of Silver's model might be warranted. It is a very complicated model that will probably perform only slightly better than a simpler model most of the time. But there are a half-dozen well-known forecasting models operating in the United States, and dozens of lesser known versions. Each has their pros and cons, but to limit the appreciation of the work that Silver does to the percentages in his charts is to miss out on the real value of his work: objective, fact-based analysis of polling data.

Forecasters are not fortune-tellers - our work is based on the data that is available to us and is only as good as the information we are provided. Some critics expect the sort of accuracy that cannot be achieved by models that are based on the polls. When the polls are wrong, models based on them can only do so much. I would have been raked over the coals if I had claimed that the polls were under-estimating the Tories by 10 points in the Alberta election. Expecting any statistical model to foresee and emphasize those 19-times-out-of-20 outcomes is unrealistic, and reduces their usefulness (if the confidence intervals are wide enough, a model will never be wrong). There are plenty of other places to find predictions based on intuition and opinion alone - and anyone who has studied those kinds of predictions can tell you how valuable they are.

Most frustratingly from this side of the argument, a lot of the critics that have leveled their guns at Silver are well-placed to do so. It is very easy to criticize from the sidelines when nothing is risked, like a heckler at a comedy club who would die of stage fright if he was forced on stage. As long as a critique is reasonable (and not based on something as ridiculous as the sound of Nate Silver's voice), the critic is in a win-win situation. He or she is just showing a healthy journalistic skepticism (especially if the forecast ends up being good). If Silver's forecast turns out to be wrong (and a Romney victory will be seen as such, even if Silver's forecast remains open to the possibility), he or she will have been a clairvoyant.

If Romney wins, Silver will have been proven foolish for having relied too much on those increasingly inaccurate polls. Those who are still calling the race a toss-up also risk nothing, even if Obama ends up winning by a relatively comfortable two or three points. "The race was never as clear-cut as Silver claimed, the undecideds just broke towards the president because of Hurricane Sandy in the last days of the campaign" - you know that argument is coming (and it is already being made).

This no-lose gamble is worth taking if it means "I told you so" gets to trend on Twitter. Everyone wants to be smarter than the egg-head.  Silver's reputation is riding on tonight's result. The reputation of those Doubting Thomases won't be dented by their dismissal of his work - but they'll take the credit if he ends up being wrong.

On top of this, Silver can be dismissed as having been "lucky" if his model proves successful. Colby Cosh of Maclean's made the implication in the last paragraph of his otherwise mostly fair assessment of Silver's track record. But focusing on those occasions when Silver was wrong (how his baseball forecasting model couldn't accurately predict a few exceptional individuals, for example, or his struggles with the UK election) seems to miss the point, particularly from someone who wrote a piece the day before the Alberta election in which it was argued that I was under-estimating Wildrose. I don't mean to pick on Cosh, but merely to point out that, sooner or later, everyone trips over something like the 2012 Alberta election or Ichiro Suzuki. And while Cosh does a good job of showing that Silver is not infallible, Silver would be the first to admit that himself.

But all of Silver's hard work, his thousands of words of analysis, his painstaking attention to detail - and if he's right? Luck. The same people who level criticisms at him for providing probabilities to a one-off event will claim that, if he is right, it doesn't matter because it could have gone either way. It could, of course, but there is a reason why his forecast says one outcome is more likely than the other.

Silver and others in the forecasting business don't have the luxury of chirping from the cheap seats. We bring this upon ourselves, of course, but the best of us are willing to admit when we were wrong and also accept that we could be wrong. A lot of our critics assign a degree of brash, arrogant certainty that most of us simply don't have.

In the end, Silver's forecast is all about the odds. I hope that the cards fall well for him tonight, and if they don't I will still look to him for the best polling analysis anywhere. For my part, this will be the first election in four years that I will watch as a mere observer, and I'm looking forward to a fun night.

Monday, November 5, 2012

Ontario race remains between PCs and NDP

A new poll was released on Friday showing that the Ontario Liberals are in dire straits, no matter who takes over the party. And that means that, for the time being, the real contest is between the Progressive Conservatives and the New Democrats.
Forum was last in the field in Ontario on Sept. 25, and since then the Tories were unchanged at 37% support. The New Democrats were down three points to 32% while the Liberals were up two points to 22%.

The Greens were unchanged at 7% support.

Whereas the last poll from Forum showed a gap between the PCs and NDP that was within the margin of error, this poll gave the Tories a statistically significant lead. But that significance does not extend to any region of the province except eastern Ontario - all the others are close enough to give us an indication of only who is probably ahead, rather than definitively.

The disparity between men and women in this poll is telling. The Tories held a 43% to 29% lead over the NDP among men but the NDP was up five points (35% to 30%) among women. The New Democrats will need to close the gap among male voters in order to put themselves back in a dead heat with the Tories.

But there is something to note about this poll, similar to what I highlighted in Forum's last federal poll. The real problem is that Forum does not included unweighted and weighted samples in their reports (and they are not alone), which makes it difficult to determine what is actually going on. For instance, Forum says on one of its charts that the number of undecideds in this poll was 13%. But it also says that the total sample was 1,102 Ontarians and that, on the voting intentions question, the sample of respondents was 1,047. If the numbers of undecideds is really 13%, then the number of decided/leaning respondents who answered the voting intentions question should have been 959.

And on the question of how respondents voted in the last election, the numbers are off of the actual results - most strongly for the New Democrats, who were five points below their election result. By my rough calculation of Forum's numbers, if the sample was weighted by past voting behaviour the PC lead would be reduced to three points instead of five. But perhaps Forum is already taking this into account in their final numbers. When I asked if that was the case, I was told that this information is proprietary. That is certainly their prerogative, but it doesn't clear things up much.

Another interesting thing to note is that if you add up the sample sizes of how people said they voted in the last election, you end up with 1,008, or 96% of 1,047. In other words, 96% of respondents said they voted in the last election. That means that the sample Forum compiled either has a lot of fibbers (turnout was 49%) or a lot of forgetful people, and that it is probably not representative of the entire population. It might be representative of the voting population, though.
With the numbers in this poll, the Progressive Conservatives would likely win a majority government of around 60 seats, with strong results in rural Ontario but also a few pick-ups in Toronto as well. The New Democrats would win 37 seats and the Liberals only 10, nine of them in and around Toronto.

The poll also included some information on how Ontarians would vote depending on who was leading the party. Forum reported their numbers with the undecideds still included, but if we remove them we get the following results:

Gerard Kennedy - 24%
Eric Hoskins - 19%
Kathleen Wynne - 18%
Glen Murray - 18%
Sandra Pupatello - 17%
Deb Matthews - 17%
Charles Sousa - 15%

The poll also included Laurel Broten, but she has ruled herself out. What the poll suggests is that only Kennedy would improve the Liberals' current numbers, while they would fall with all of the others. Undoubtedly, this is due to Kennedy being a higher profile candidate. The others on the list are not nearly as well known, but their numbers would likely improve somewhat if they actually became leader.

This leaders question actually shows that the real swing voter in Ontario right now is on the fence between the New Democrats and the Liberals. The numbers hardly budged for the Tories no matter who was on the ballot, but those lower-performing Liberals added to the NDP's tally. It would seem to suggest that the Liberals would do better with a left-wing candidate (like Kennedy or Wynne) rather than one from the right (like Pupatello). But, in the end, if the Liberals get themselves back into a competitive position they will need to win votes from the Tories as well.

As of writing, Glen Murray is the only candidate officially in the race for the Liberal leadership. That list will likely get much longer very soon. How it all plays out between now and January, and then how the new leader will do in the short time before the next election, will be interesting to see.