Friday, September 28, 2012

Is McGuinty's support collapsing?

A new poll was released by Forum Research yesterday for the Toronto Star, suggesting that the Progressive Conservatives and New Democrats are in a close race in Ontario. That leaves Dalton McGuinty's Liberals on the sidelines with only 1 in 5 support.
Forum was last in the field in Ontario in a poll released to a media outlet (there was a poll for a consultancy out earlier this month) on Aug. 15, and since then the Tories slipped one point to 37%. The New Democrats, however, were up seven points to 35%. The Liberals were down seven points to only 20%, while the Greens were up one to 7%.

Now, Ontario has been a very confusing province when it comes to polling. Forum is recording a huge shift in support, which is not implausible considering the difficult headlines for the McGuinty government of late. But the other firm active in Ontario, Nanos Research, has been consistently at odds with what Forum has reported. Will the next poll from Nanos show these sort of stellar NDP gains?

The gains and losses for the NDP and Liberals are statistically significant, with the most important shifts in support having taken place in the Greater Toronto Area. There, the Liberals fell 10 points to 20%, with the NDP gaining seven to hit 34%. In the 416 area code, the Liberals were down 15 points to only 22%.

Everywhere in Ontario, the Progressive Conservatives and New Democrats were either tied (southwestern Ontario) or held insignificant leads over the other party (the Tories in eastern Ontario, the 905 area code, and northern Ontario, the NDP in the 416 area code). It would make for a very close election result, with the Liberals shunted off to the side.
And dramatically so. With no pocket of regional strength, the Liberals could only hope to hold on to a handful of seats. The projection model gives them only three, all in Toronto. The Progressive Conservatives win a majority government of 64 seats while the New Democrats win 40.

That two-point edge is rather decisive for the Tories, as the NDP vote in Ontario is pretty inefficient. But the degree of inefficiency that Forum assigns to them is a little ridiculous. Forum's seat projection model would award the Tories 60 seats, the NDP 24, and the Liberals 23. With their own regional results showing the Liberals with no regional concentration of support, it makes very little sense that Forum's model would still give the Liberals that many seats. If the Liberals still had a lot of support in Toronto but had dropped everywhere else, then the sort of result envisioned by Forum could be possible. But with only 22% support in their best region of the province, it just wouldn't happen.

According to Forum, Dalton McGuinty's personal numbers have collapsed. He used to enjoy better approval ratings than the unloved Tim Hudak, but no more. His approval rating dropped nine points to only 20% and his disapproval increased by seven points to 67%. Even among the dwindling number of Liberal supporters his numbers have dropped by six points to 71% approval.

That is not to say that Hudak has taken advantage of this. His numbers have held relatively steady, with an approval rating of 26% and a disapproval rating of 49%. The biggest shift came among PC voters - his approval rating was down five points to only 48% among his party's supporters.

Andrea Horwath's numbers were also rather static, with her approval rating registering 48% and her disapproval rating dropping four points to 23%. Both Hudak and Horwath's net ratings improved slightly (the gap between approval and disapproval shrunk by three points for Hudak, and increased by six points for Horwath), but this seems to be a case of Ontarians turning against McGuinty rather than towards another party or leader.

But as has been the case in the past, it is difficult to know what to believe with this poll. It would not shock me in the slightest if Nanos Research released a new poll showing the New Democrats with less than 25% support and the Liberals and Tories still in a close race. The increase in NDP support has not been replicated at the federal level either, with the exception of a new poll by Environics. Aside from their results, the NDP is still hovering around 30% in the province (including in the most recent federal Forum poll) and hasn't experienced any sort of dramatic leap.

Ontario remains a province to keep a close eye on, but the trends are not exactly clear. The New Democrats have had a good month with their by-election win in Kitchener-Waterloo and the Tories might be benefiting from the Liberals' discomfiture by keeping their heads down. If numbers like these are confirmed in the coming months, the opposition parties might be more emboldened to take the government down in the spring. But the result of such an election could be a surprise.

Wednesday, September 26, 2012

Tories, NDP tied in Abacus poll

Abacus Data released its latest federal poll on Monday, showing that the Conservatives and New Democrats are tied in national voting intentions. But a tie isn't good enough for the New Democrats - a seat breakdown suggests that the Tories can win many more seats with just as much of the vote as the NDP.
Abacus was last in the field Aug. 10-12, and since then the Conservatives dropped two points to 35%. The New Democrats picked up three points to hit 35%, while the Liberals were down three points to only 17%.

That isn't an all-time low for the Liberals - they were polling even lower than that in the immediate aftermath of the May 2011 election - but this is one of the lower results we've seen for the party.

The Bloc Québécois came up fourth with 7%, while the Greens were unchanged at 6%.

The shifts in support for the Conservatives and New Democrats are statistically insignificant, but the drop of the Liberals is outside the margin of error for a party with that much support. Perhaps this is not too shocking, considering that the Liberals are still leaderless (and more so than usual, Bob Rae has not been very visible of late).

It is worth noting that Abacus used Angus-Reid's online panel for their polling. Based on the information available in their report, it would appear that Abacus recorded, or at least released, more demographic results than Angus-Reid ever does: Conservatives lead among evangelicals while the NDP leads among non-religious people, for example. It is unclear whether Abacus's last poll also used Angus-Reid's panel.

Things held very stable in British Columbia and Ontario, with no party losing or gaining more than a point. The New Democrats maintained a five-point advantage in British Columbia, while the Conservatives held firm with an 11-point lead in Ontario. The NDP narrowly edged out the Conservatives in the Prairies, and were up by nine points over the Bloc Québécois in Quebec. The Bloc appears to be benefiting slightly from the PQ's recent election victory.

Alberta and Atlantic Canada had some statistically significant swings. In Alberta, the Conservatives dropped 12 points to 58% while the NDP was up 15 points to 29%. The Liberals fell nine points to 6%. But this is probably a statistical fluke considering the sample size - the same goes for Atlantic Canada, where the NDP was up 18 points to 47% and the Liberals were down 17 points to 19%.
Despite the tie in voting intentions, the Conservatives would come out ahead in seats with 157 to the New Democrats' 140 on the proposed boundaries for the new 338-seat electoral map. The Liberals would win only 26 seats and the Bloc Québécois 14.

The Conservative victory is mostly won in Ontario, where the party takes 84 seats. The Liberals manage only nine in the province, and appear to be somewhat penalized by the new boundaries. On the old map, the Conservatives would win 71 seats, the NDP 25, and the Liberals 10 in the province.

The New Democrats do very well in British Columbia and the Prairies, but also manage to win four of the 34 seats in Alberta. This is a rather shocking result, but with 29% in this poll some shocking things would be bound to happen.

On the leadership front, Thomas Mulcair has the highest favourability rating with 36%, compared to 35% for Stephen Harper and 27% for Bob Rae. He also has the lowest unfavourability score with 22%, while Rae managed 34% and Harper 50%.

Justin Trudeau, a virtual lock for a leadership run, has a favourability rating of 39% and an unfavourability rating of 23%, giving him slightly better overall numbers than Mulcair. Marc Garneau, a likely candidate for leadership, has a favourability of 26% to 15% unfavourable. He has very high "neutral" and "don't know" numbers.

Among Liberal supporters, Trudeau has a 70% favourability rating, with only 9% saying they have an unfavourable opinion of him. At 7%, Garneau's unfavourability among Liberals is lower, but so is his favourability rating: 51%. He starts out with a big hill to climb, as 42% of Liberal voters had a neutral opinion or were unsure what they thought of him.

And now for something (almost) completely different - if anyone watched The Rick Mercer Report last night, they would have seen Rick's Rant about polls. I think Rick is expressing a frustration that a lot of people have with polling right now, but I also think this is being borne out of a bit of a false narrative when it comes to polling. He refers to Alberta, Quebec, and the disappearance of the Bloc as things the polls got wrong. I'll give him Alberta - that was a disaster. But the problem in Quebec was rather minimal: underscoring the Liberals by a few points in polls taken a few days before the vote. That was hardly a catastrophe. It's not as if the polls said that Québec Solidaire would win 20% of the vote and the CAQ would form a majority government. The PQ won with a minority, as most of the final polls suggested was a strong possibility.

And the polls did not miss out on the Bloc - the final four polls of the federal campaign averaged 24.6% for the Bloc Québécois in Quebec, instead of their actual result of 23.4%. Perhaps Rick is referring to a site like mine that did not predict that the Bloc would win only four seats, but I don't think it is fair to say that the polls missed out on the Bloc's demise.

This is something I have heard again and again, that the polls did not anticipate the NDP's rise in Quebec and the Bloc's fall. They absolutely did anticipate these things. They didn't see it coming two or three weeks before the vote, but that is because Quebecers hadn't swung over to the NDP yet. Polls don't predict future events, opinions ebb and flow during a campaign and we have seen many examples of that over the past 16 months. But critics of polls never point out their successes, even their recent ones: very good results in Quebec during the federal campaign, a perfect call in Manitoba's election, good performances in Ontario and Saskatchewan, and decent individual polls in Prince Edward Island and Newfoundland & Labrador. Sure, Alberta was a Dewey-defeats-Truman moment and they could have been better in Quebec, but let's not throw the baby out with the bathwater. And that's my counter-rant.

Tuesday, September 25, 2012

B.C. Conservatives drop amid leadership woes

Late last week, Ipsos-Reid released a new poll on the provincial voting intentions of British Columbians. The poll was taken earlier this month at a time when questions were being raised about John Cummins' leadership. He survived the call for a leadership vote over the weekend, but already the uncertainty seems to have hurt his party.
The B.C. New Democrats led by a wide margin in this poll with 49% support, an increase of one point since Ipsos-Reid was last in the field June 5-11. The B.C. Liberals picked up three points to hit 32%, while the B.C. Conservatives were down four points to 12%. The Greens remained stable at 6% support.

The changes in support for the NDP and Liberals were statistically insignificant - not so for the drop of the Conservatives. With a very difficult weekend for the party (low turnout in the vote on leadership, a less-than-impressive win for Cummins, and the loss of their only MLA, John van Dongen) it will be interesting to see if the Conservatives will lose further support.

As I discussed yesterday, the 32% support for the B.C. Liberals is deceptively high. Compared to Angus-Reid and Forum Research, the two other firms most active in the province, Ipsos-Reid has awarded the Liberals higher scores to the tune of about six points per poll. That is not to say that the B.C. Liberals are not at 32% - Ipsos could be gauging the B.C. electorate more accurately than their rivals - simply that things have remained relatively stable for the Liberals over the summer.

In any case, no matter who is doing the polling the end result is a massive NDP landslide. Voter turnout is unlikely to swing things dramatically either, as the New Democrats have majority support among those aged 35 or older in this poll. Those are the people who are most likely to vote on Election Day.

Interestingly, though, the New Democrats are not doing so well among men. While they held a 33-point lead among women (57% to 24%), they were only up by a single point among men (41% to 40%). This could be a mere statistical fluke, but it does have the potential to be problematic for Adrian Dix.

The New Democrats held leads in Metropolitan Vancouver (52% to 32%) and on Vancouver Island (57% to 24%), having gained three and four points, respectively, since June. The Liberals were steady in these two regions.

In the North and the Interior, however, the New Democrats were down five points to 40% while the Liberals were up eight to 36%. This is the only region of the province where the NDP lead is not statistically significant. We saw similar numbers in the recent Angus-Reid poll, suggesting that this part of British Columbia will be the most competitive. It is also the region where the B.C. Conservatives have the most support. A few three-way races can be expected in the Interior.

With these numbers, the B.C. New Democrats would win a majority government of 64 seats. The B.C. Liberals would form the Official Opposition with 20, and one independent would be elected. The B.C. Conservatives would be shut-out, just short of where they need to be to win a couple of seats.

The problem could indeed be John Cummins. His personal numbers are far below those of Adrian Dix and even Christy Clark. Whereas Clark has an approval rating of 33%, Cummins' is only 23%. He has a much lower disapproval rating at 40% to Clark's 60%, but that is due primarily to the 37% of British Columbians who are unsure of their opinion of him. Remove them from the equation (and the 7% who are unsure of Clark), and you get very similar numbers.

Dix has better numbers than his party, with an approval rating of 51% and a disapproval rating of only 34%. He is also the favourite person to be Premier at 35% to Clark's 22% and Cummins' 9%. But these numbers may not be as telling as they appear - if we remove the "don't knows" we get numbers that are almost identical to the voting intention results.

Less than eight months remain before British Columbians go to the polls and the B.C. Liberals have been unable to move the dial by more than a tick or two (in either direction) for quite some time. Eight months ago, the B.C. New Democrats were enjoying a 14-point lead over the Liberals. That lead is now between 17 and 21 points. Can the Liberals really manage to turn things around and gain two or three points per month on the NDP by May 2013? Unless some sort of game-changing event takes place, it seems very unlikely.

Monday, September 24, 2012

Partisan language, and interpreting polls

If you think that the tone of debate on Parliament Hill has become more partisan in recent years, it’s not just you. An analysis of House of Commons’ transcripts shows how one portion of the daily routine in the House has been increasingly used for partisan ends since the Conservatives came to power in 2006.

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

I also have a column in The Hill Times this morning dealing with the importance of interpreting a poll within proper context, and how interpreting it falsely can lead to false narratives. You can read my column here, but you'll need a subscription to The Hill Times (a smart investment).

In the column, I mention the recent Angus-Reid poll in British Columbia that showed a statistically insignificant increase for the B.C. Liberals, prompting headlines that the party was on the upswing. The column was written before this weekend's Ipsos-Reid poll, but the reaction to that poll is another perfect example of how polls need to be put into proper context.

I will look into the entrails of the poll in more detail later this week, but the important numbers are 32% for the B.C. Liberals and 12% for the B.C. Conservatives. That 32% was heralded as terrific news by Liberals in the province, and added to a narrative that formed over the weekend - that the B.C. Liberals are not quite dead and that the B.C. Conservatives are not a legitimate alternative.

John Van Dongen, the party's only sitting MLA (he crossed the floor from the Liberals) decided to quit the party over the weekend, justifying his departure with John Cummins' inability to make gains in the polls. Would he have felt the same way if the only numbers he had to go on were the ones from Angus-Reid and Forum, showing only a small gap between the two parties? Or was he basing that decision in part on the new Ipsos-Reid poll, that put the B.C. Liberals 20 points ahead of the Conservatives?

Trumpeting the Ipsos-Reid poll as terrific news for the B.C. Liberals is a bit of a problem, as it must assume that the polling firm is acting in a vacuum. This is the kind of poll reporting that Nate Silver criticized last night.

As you can see from the chart above, the B.C. Liberals have consistently over-performed in polls by Ipsos-Reid - on average by about six points. Both Forum and Angus-Reid have had the Liberals at much lower numbers. The good result from Ipsos-Reid is not, then, a sign that the Liberals are on a real upswing from the mid-20s of recent months. They have merely experienced a statistically insignificant increase from a polling firm that generally gives them better results.

That is not to say that Ipsos-Reid is wrong. There is every chance that Ipsos-Reid is closer to the mark and that Angus-Reid and Forum are under-scoring the B.C. Liberals. But placing Ipsos-Reid's numbers within context is absolutely essential in interpreting this poll correctly. Otherwise, the headlines accompanying the next poll from Forum or Angus-Reid have a good chance of falsely giving the impression that the B.C. Liberals are slipping once again.

Friday, September 21, 2012

Confusion in Ontario

If a provincial election were held in Ontario today, there’s no telling who would win. Dalton McGuinty, Tim Hudak, and Andrea Horwath would all have a shot, and not because they are locked in a close three-way race. The polls are just all over the place.
Ontario is the most confusing province in Canada when it comes to voting intentions, as the two firms most active on the provincial scene are in complete disagreement on what is happening and have been for months.

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

Ontario has been a very difficult province to cover. Forum Research and Nanos Research are both active in the province and they both consistently show completely different results. They both had decent showings in the 2011 provincial election, and there really isn't any other firm active in the province to provide a better idea of who has it right. They are so different that a "somewhere in between" reading of the polls is not very useful.

A poll that was released earlier this week, conducted by Forum for the Broadview Strategy Group, further muddied the waters as it gave the NDP a one-point lead over the Tories (36% to 35%), with the Liberals at only 22%. I generally dislike giving much prominence or credence to polls that are not done for the media or released directly by the polling firm (and Forum has yet to post the details to their website).
And somehow, Forum's seat projection model gives the Tories 48 seats with these results, the Liberals 30, and the New Democrats only 29. Perhaps the regional distribution is that great for the Liberals and that horrible for the NDP, but with a province-wide swing my model would divvy-up the seats as follows: 59 seats for the Progressive Conservatives, 42 for the New Democrats, and six for the Liberals.

Even here you can see that the New Democrats do not have very good vote efficiency, but it is very difficult to imagine how the Liberals could win 30 seats and form the Official Opposition with only 22%.

All in all, it makes Ontario - or more precisely, the polls being released for Ontario - very frustrating. The legislature being in a minority situation, and thus an election being always just around the corner, I hope the polls will start to align so that we can have a clue as to what is going on.

Thursday, September 20, 2012

Liberals move ahead in Nova Scotia

Earlier this month, the Halifax-based Corporate Research Associates released their quarterly polling for the four Atlantic provinces. With Nova Scotia on track for an election as early as next year, the results in this province are particularly important - especially since they show the incumbent New Democrats in trouble.
The Corporate Research Associates were last in the field between May 7 and June 4, and since then the Nova Scotia Liberals picked up eight points to lead with 41% support, well ahead of the New Democrats. They were down four points to 31%.

The Progressive Conservatives, at 22%, slipped six points and the Greens were up one point to 5%.

That is a big change in Nova Scotia, as Darrell Dexter's New Democrats have been leading since early 2011. Now is certainly not the time for the NDP to fall behind with an election around the corner.

The New Democrats' drop is within the margin of error, however. The increases and decreases of the Liberals and Tories are not, suggesting that Stephen McNeil is picking the pockets of Jamie Baillie in particular.

McNeil is, by far, the leader seen by Nova Scotians as the best person to be premier. He has 35% support on that score, a gain of eight points. Dexter is unchanged at 23% while Baillie is down four points to 17%. If we just look at the Best Premier numbers as a share of the total (removing the undecideds and "none of the aboves"), we see McNeil at 45%, Dexter at 29%, and Baillie at 22%. In other words, McNeil is a net positive for his party while Dexter is not.

Satisfaction with the New Democratic government is down four points to 37%, with the number of people saying they are "mostly satisfied" decreasing. Dissatisfaction is up two points to 54%, but more importantly it is the number of people who are "completely dissatisfied" that increased, rather than those who are just mostly dissatisfied.
In terms of seats - and using the old electoral map, as a new one is going to be released in the coming weeks - the Liberals would win half of them in the legislature with 26, while the New Democrats would win 21 and the Progressive Conservatives five. If Stephen McNeil could convince an NDP or PC MLA to become Speaker, he would have a workable majority government.

The Liberals win seats in every part of the province, taking 10 in the Halifax Regional Municipality, eight in the Annapolis Valley and on the South Shore, five in Cape Breton, and three around the Bay of Fundy and in central Nova Scotia. The New Democrats win eight in Halifax, seven in Fundy/Central, four in the Valley and the South Shore, and two in Cape Breton.

The Tories are shut out of the HRM, but win two in Cape Breton, two in the Valley and the South Shore, and one in the Fundy/Central region.

This is a close enough result that the Liberals could easily win an outright majority, especially with Stephen McNeil's personal numbers besting Dexter's by such a considerable amount. But it is still a very close race, at least in terms of seats. The current map appears to give the New Democrats more resilience than they might deserve considering they trail the Liberals by ten points. It will be interesting to see how the new map changes things, if it all.

Tuesday, September 18, 2012

Two new polls, two stories

To inaugurate a new sitting of Parliament, polls by Nanos Research and Harris-Decima were released to a curious public. Depending on how one might want to spin the results, there is something for everyone in these polls - not something you want to see when searching for consistency in polling data.
We'll start with the poll by Harris-Decima for The Canadian Press, which was in the field both before and after the poll by Nanos Research.

Harris-Decima's last public poll was conducted between June 7-18 and since then the Conservatives picked up three points to lead with 34%. The New Democrats were down five points to 27%, transforming a one-point edge into a seven-point deficit.

The Liberals were up one point to 24%, while the Greens were unchanged at 7% support.

This is a change of fortunes for the New Democrats. They have not trailed the Conservatives by seven points since Thomas Mulcair became leader of the party at the end of March. The shifts in support for the Tories and the NDP since June are statistically significant, as are the shifts in Ontario. There, the Conservatives were up six points to 39% while the New Democrats dropped nine points to 23%, putting them well behind the Liberals, unchanged at 29%. If that is a true depiction of what is going on in Ontario, the New Democrats are in a lot of trouble.

Shifts elsewhere were within the margin of error. The New Democrats led in Quebec with 31% (-5), followed by the Bloc Québécois at 25% (+4) and the Liberals at 24% (+5). The Tories and NDP were tied in British Columbia with 33% apiece, while the Conservatives led in Alberta and the Prairies. Atlantic Canada is a close race between the New Democrats and the Liberals.
In terms of seats, these numbers would deliver 152 to the Conservatives on the proposed boundaries of the new 338-seat electoral map. The New Democrats would win 96, the Liberals 74, the Bloc 15 (official party status returns) and the Greens one.

The poor showing of the NDP in Ontario is the primary problem for the party in this poll, while their weaker numbers in Quebec open up a swathe of seats to the Liberals and Bloc Québécois. But this poll puts the Conservatives in minority territory, which could be enough to end their time in government. The New Democrats and Liberals could combine for 170 seats, just above the half-way mark.
While the Harris-Decima poll points to a wide Conservative lead, the Nanos poll suggests that the Tories and NDP remain in a neck-and-neck contest.

Nanos was last in the field July 7-12, and since then the Conservatives slipped 1.2 points to 32.4% support. The New Democrats were up 0.1 point to 30.4%, while the Liberals were down 1.9 points to 24.6%.

That Conservative edge is statistically insignificant, and none of the shifts in support suggest any real movement. Status quo, in other words, rather than the changing landscape of the Harris-Decima poll.

No shifts of consequence happened at the regional level, the Conservatives ahead with 35.4% in Ontario (-1.4). They were closely followed by the NDP at 32.6% (+7) and the Liberals with 26.3% (-5). In Quebec, the NDP slipped 5.1 points to 33.7% and were trailed by the Liberals at 24% (-1) and the Bloc at 20.5% (+3.3). And in British Columbia, the Tories had the edge with 34.3% support (-6.1) to 31.1% for the NDP (-3.2) and 21% for the Liberals (+1.5).

The only lead of statistical significance in this poll was in the Prairies, which includes Alberta in Nanos's reckoning. That makes it somewhat more difficult to estimate how to distribute support across the three provinces included in the Prairies.
But with Nanos's numbers the Conservatives would win 148 seats to 114 for the New Democrats and 72 for the Liberals. The Bloc Québécois would win only three seats and the Greens just one.

The difference maker for the NDP in this poll is Ontario and Quebec. While this poll does not give a lot more support to the NDP in Quebec than Harris-Decima does, the poorer showing for the Bloc leaves a lot of Quebec seats in the NDP's hands. The Liberals still manage to take a good chunk. Here again, the Conservatives are in minority territory and outnumbered (186) by the NDP and Liberal opposition.

Nanos includes personal ratings for all of the leaders, and the numbers were quite good for Stephen Harper. Despite his party slipping marginally nationwide, his own numbers were up by five points to 28% on "vision for Canada", six points to 29% on trust, and 10 points to 37% on competence. That gave him an overall score of 93.4 points on Nanos's Leadership Index (just the sum total of the percentages on the three qualities), a gain of 20.7 points.

Thomas Mulcair was up 1.2 points to 48 on the Leadership Index, picking up two points on trust (18%), holding steady on competence (13%), but losing one point on vision (17%).

Bob Rae was down 3.4 points on the index to 38.1, and was followed by Elizabeth May at 21.1 (+5.7) and Daniel Paillé at 10.1 (+3.1).

Paillé's numbers in Quebec are problematic for the Bloc Québécois. He trails Harper, Mulcair, and Rae on all three qualities by significant margins. Only 6% of Quebecers consider him the most trustworthy and only 3% think he is the most competent and has the best vision. Mulcair, by contrast, managed 40%, 30%, and 35% on those three factors. What this suggests is that Daniel Paillé is either a complete unknown or a drag on the Bloc's numbers (considering his low visibility, 'unknown' is a safer bet). Gilles Duceppe always did much better in these Nanos polls, and if the Bloc can't improve its leader's reputation they will have no chance of growing beyond their current base.

These two polls tell different stories. Harris-Decima puts the Conservatives firmly in the driver's seat, while Nanos shows the same close race between the two main parties. Conservatives can delight in the seven point lead or the improving personal numbers for their leader. New Democrats can point to their continued resilience in Nanos's poll, while Liberals can rejoice that they are at 24% in Quebec. Even the Bloc can be happy with Harris-Decima's numbers, and the Greens are looking good with 12%-13% in British Columbia.

But other polls haven't been so positive for the Liberals, or so negative for the New Democrats in Quebec. And while the results of the Harris-Decima and Nanos polls are within each's margin of error, the two stories they tell cannot both be true. Either the Tories are comfortably leading or they are not. Until other polls emerge to tell us which is the case, partisans of every stripe have something to gloat over.

Monday, September 17, 2012

B.C. NDP continues to hold 20+ point lead

Since the beginning of September, two polls have been released on the provincial situation in British Columbia. Both polls show that the B.C. New Democrats hold a lead of over 20 points, a seemingly insurmountable margin with only eight months to go before the next election. And with the B.C. Liberal government deciding not to call a sitting of the legislature until February 2013, it does not seem like there will be much opportunity for anything to change.
The earlier poll by Forum Research found NDP support to be at 45%, down four points since their previous poll of July 31. The Liberals were unchanged at 23%, while the B.C. Conservatives were up two points to 20%.

The Greens were at 10% while 2% said they would vote for another party.

None of these shifts in support were statistically significant, though the leads that the NDP holds throughout the province certainly are.

The NDP led with 44% in Vancouver and the Lower Mainland, down four points, while they were up two points to 51% on Vancouver Island. The party was down 10 points in the Interior/North, however, to 40%.

The Liberals were down two points there to 18%, while they were down three points to 18% on Vancouver Island and up one point to 25% in Vancouver and the Lower Mainland. The Conservatives were up seven points to 18% on Vancouver Island, two points to 20% in Vancouver, and one point to 24% in the Interior/North, where they are in second place.
The newer poll by Angus-Reid tells the same general story. Angus-Reid was last in the field between July 30 and Aug. 1, and since then the New Democrats were down three points to 46%. The Liberals were up three points to 25% while the Conservatives were unchanged at 19%.

The Greens were down one point to 8% while 1% of respondents said they would vote for another party or independent candidate.

Despite how the poll was widely reported, the gain for the Liberals is statistically insignificant, suggesting that nothing much has occurred in the province.

The NDP has 34% in the Interior and leads with 49% in Metropolitan Vancouver, 50% in the North, and 53% on Vancouver Island. The Liberals trail in second in every region with 28% in the Interior, 26% in the North, 25% in Vancouver, and 24% on the island. The Conservatives are third with 25% in the Interior, 19% in Vancouver, 18% in the North, and 11% on Vancouver Island.

While the next set of polls from British Columbia will tell us whether the B.C. Liberals are on an upwards trajectory, the only thing these polls suggest as a potential trend is that the New Democrats may have slipped a little from their 49% of mid-summer. They appear to be down most significantly in the Interior, the only area of the province that appears to be somewhat competitive.

The overall race, however, is far from competitive. The average result of these two similar polls would deliver 69 seats to the B.C. New Democrats, with the B.C. Liberals being reduced to only 10. The B.C. Conservatives win four and two independents would be elected.

The seat projection model for British Columbia is still rudimentary. A new regional model will be launched soon, likely in November.

The two surveys included approval ratings for the leaders and showed similar results.

Christy Clark's approval stands at between 26% and 28%, with her disapproval sitting at between 56% and 62%. Strikingly, those British Columbians who intend to vote for the Liberals are quite loyal to Clark. Her approval rating among Liberal voters is 65%, according to Forum, only marginally lower than the ratings for John Cummins (66%) and Adrian Dix (68%) among their own supporters.

Province-wide, Dix has an approval rating of between 43% and 45%, with a disapproval rating of between 29% and 38%. That is a rather wide spread, with Angus-Reid finding the disapproval rating being higher. The source appears to be those who are not sure - they represent 29% of the population according to Forum but only 16% according to Angus-Reid. That suggests that Dix may have hit a wall in terms of growth potential, since a decrease in "undecided" opinion increases the proportion who disapprove of the NDP leader.

Cummins has a very high unknown factor, at between 33% and 39%. His approval rating sits at between 22% and 23%, while his disapproval rating is between 38% and 45%. Here again, the lower "not sures" in Angus-Reid's poll has the effect of increasing Cummins' disapproval rating. This also occurred for Christy Clark, though to a somewhat lesser degree. We may infer from this that undecideds will be more likely to swing towards the B.C. Liberals, though not significantly.

Based on these numbers, and considering that Dix's NDP has held a wide lead over Clark's Liberals for about as long as time remains before the next election, many of the B.C. Liberals' MLAs will not be working in Victoria in nine months' time. In the end, their break from having to sit in the legislature until February will probably be good practice.

Thursday, September 13, 2012

Provincial party leader approval ratings

Earlier this week, Angus-Reid released its quarterly poll on the approval and disapproval ratings of Canada's provincial premiers (with the exception of Prince Edward Island). Angus-Reid also includes the ratings for the highest scoring opposition leader for each province.

My analysis of the poll for The Huffington Post Canada can be found here.

Angus-Reid was kind enough to provide me with the approval ratings for other opposition leaders. Taken together, this makes a list of 30 leaders. Granted, some of those on the bottom of the list are lesser known leaders with high "Don't Know" scores, but it makes for an interesting nationwide ranking.
Brad Wall tops the list again, with an approval rating of 66%. Alison Redford is ranked second with 55%, while Greg Selinger ranks ninth with 48%. David Alward sits in 10th spot, followed by Kathy Dunderdale at 11th, Pauline Marois at 12th, Dalton McGuinty at 17th, Christy Clark at 25th, and Darrell Dexter at 27th.

For Leaders of the Official Opposition, Adrian Dix is ranked fourth overall, followed by Stephen McNeil (5th), Danielle Smith (6th), John Nilson (13th), Brian Pallister (14th), Tim Hudak (19th), Victor Boudreau (26th), and Dwight Ball (28th).

If we look at it in terms of ideology, New Democratic leaders have the best average score with 41%. Conservative leaders (including Wall, Clark, Smith, and Legault) average 36%, while Liberal leaders average 32%. Of course, it isn't quite as simple as that - Smith would bristle at being lumped together with Redford, while Clark, Charest, and Legault are difficult to pin down in terms of where they are on the political spectrum. Generally speaking, though, it would be appear that provincial politicians from the NDP narrowly edge out the Conservatives of various stripes, while the Liberals are not far behind. But each camp has their bears and their bulls.

Hopefully Angus-Reid will provide me with this information when they release their next quarterly poll in December. It will be interesting to see who moves up and who moves down.

Wednesday, September 12, 2012

August 2012 federal polling averages

We are slowly emerging from the polling doldrums of the summer. With only two national polls and one provincial poll (in British Columbia) having been conducted during the month of August, the polling average does not tell us too much. Nevertheless, as after the month of July, I present the August federal polling averages for the sake of continuity. Hopefully, with Parliament returning in September, we will have a more robust set of data with which to work next month.
The Conservatives averaged 35.4% support in the month of August, an increase of four points over their July numbers. The New Democrats slipped into second but did pick up 0.2 points, and sat at 32.8% support.

The Liberals were down 1.5 points to 20.8%, while the Bloc Québécois was up 0.9 points to 6% and the Greens were down 1.7 points to 5%.

The regional results do not show any major shifts or variations from the July averages, or from the general trends we have been seeing since Thomas Mulcair became leader of the NDP.

The Conservatives were ahead in Alberta, the Prairies, and Ontario, while the NDP had the edge in British Columbia, Quebec, and Atlantic Canada.

You'll notice I have added a pale orange bar to the monthly federal averages chart, indicating the month in which Mulcair won the NDP leadership race. I will do the same when the Liberals choose their leader.

I have updated the seat projection model to reflect the likely results on the new 338-seat electoral map, using the proposed boundaries. Credit for the transposition of results has to go to the website, and thanks go to Kyle at Blunt Objects for compiling the results into an easy-to-use spreadsheet. With this work seemingly already done (and the numbers of Pollmaps do closely align with the transposition I have already completed), I do not think it will be necessary to continue my analysis of the proposed boundaries.
With the August poll averages and the new 338-seat map, the Conservatives would win 160 seats, nine short of a majority government. The New Democrats win 123 seats, the Liberals 49, the Bloc Québécois five, and the Greens one.

The Conservatives rack up big numbers in Alberta (33 of 34 seats) and Ontario, enough to compensate for NDP gains in British Columbia and Saskatchewan, where the new boundaries are particularly helpful for the New Democrats.

It will be interesting to see what kind of results the new map will give if the Conservatives and New Democrats move back into a neck-and-neck race.

Only one poll had any numbers on approval ratings this month, so there is little to report on that front. Nevertheless, the chart to the right shows how the ratings of the various leaders have been shifting since the spring.

The fall political season in Ottawa begins generally where it left off, with the Conservatives and New Democrats trading the lead and the Liberals still well behind. With the Quebec election now over, it will be worthwhile to keep an eye on how the federal voting intentions of Quebecers will (or won't) shift in the coming months. The Liberal leadership race, slated to begin in earnest in November, will then become the focus both on the Hill and here at ThreeHundredEight. The next federal election is still years away, but there will nevertheless be plenty of things to keep us busy until then.

Monday, September 10, 2012

Quebec election: expectations vs. results

Less than a week ago, the results of the Quebec election came as a bit of a surprise to pollsters and pundits. But the surprise was not in the grand strokes of the outcome itself: as expected, the Parti Québécois won with the Liberals forming the Official Opposition, the Coalition Avenir Québec coming third, and Québec Solidaire winning two seats. The surprise instead came in the details.

A minority result for the PQ was always a distinct possibility, but it turned out to be a lot smaller than expected. The Liberals were best positioned to finish second, but they out-performed expectations to a significant degree. And the CAQ, despite a good haul of votes, did not manage to win as many seats as they might have and a few of their star candidates did not win in their ridings.

ThreeHundredEight's projection gave a slight edge to a PQ majority, but this overall result was envisioned. The problem with the projection was that the Liberals were under-estimated in the polls, which meant that the projected results over-estimated both the PQ and the CAQ.

The Quebec election was not a repeat of the Alberta election. Equating the two would be a silly mistake. The polls in Quebec were generally within the margin of error for the PQ, CAQ, and QS and they only under-estimated the Liberals. The polls from the Alberta election gave Wildrose a 10-point lead when the result was a 10-point win for the Progressive Conservatives. Nothing close to that occurred last Tuesday.

In terms of the Alberta election, the question of "what went wrong?" can certainly be asked. The public polls did not capture a swing in the last week that internal polls did manage to identify. The question that should be asked in Quebec is "why were the results different?"

Polls are often erroneously seen as simply a way to predict an outcome. They are, of course, a snapshot and things can change - but voting intentions rarely shift to a significant degree in a matter of days. That is why results usually match the last polls of a campaign. But polls are most useful as a way to understand and contextualize what is going on during an election campaign. After the election is over, they can be a very good tool to understand what happened. And projections based on polling data do the same.

Of course, polls and projections will always have a degree of error and I will take a look at the errors the projection made below. But working from the idea (which has been proven over time, notwithstanding the few exceptions) that in a normal election polls can track the voting intentions of the population accurately and that projections can turn those numbers into accurate seat counts, a comparison of polls/projections to results paints a very good picture, not of how polls and projections failed, but of what worked and what didn't for the parties themselves.

For a full statistical breakdown of the projection's performance, I invite you to take a look at the Electoral Track Record.

How would the seat projection model have worked had the polls been bang-on in each of the model's six regions? Polls for three of them (Montreal island, Montreal suburbs, and Quebec City) were available throughout the campaign while polls for the other three (Eastern, Central, and Western Quebec) were available for some of the polls.

The model would have awarded the Parti Québécois 53 seats, the Liberals 42, the CAQ 28, and two to Québec Solidaire. In other words, a strong result for the PQ and QS and an under-estimation of Liberal strength to the benefit of the CAQ.

The riding accuracy, at 81.6%, would not have changed but the potential winners would have been identified in 94.4% of ridings when using the seat projection ranges.

Those ranges (using the polling volatility in the last week of the campaign as a guide, as the model normally does) would have encompassed the final result: 41-65 for the PQ, 34-55 for the Liberals, 18-38 for the CAQ, and 1-2 for QS. The projection would have called for a PQ minority, a strong Liberal opposition, and a third place finish for the CAQ, while the odds would have been considered very unlikely for a PQ majority or a CAQ second-place finish. This would have been a good call.
So the projection model would have been able to forecast the election properly with accurate polling - but nevertheless the most likely outcome would have still under-estimated the Liberals to a significant degree. Why?

This is where the results of the election and how they differ from polls and projections can tell us something about what went right for the Liberal campaign. In the Alberta election, to take one example, the projection model would have done very well with accurate polls with one exception - the extraordinary resilience of Liberal incumbents in Edmonton and (especially) Calgary. With the Liberal vote dropping so steeply in Alberta as a whole, no projection model that wasn't based on a hunch could have forecast that the Liberal incumbents in these two cities would prove so difficult to topple. While that was an example of the projection model being wrong, it said more about the Liberals in those ridings than it did about the model itself.

Region vote projections vs. results
The chart to the left shows the difference between the projection and the results in each of the six regions. We can see that the Liberals did significantly better than expected in Montreal (though it did them no good in terms of seats), in Quebec City, and central Quebec.

The performance of the Liberals in central Quebec was quite remarkable. Polling data that was available for the region in the mid-point of the campaign, when the Liberals were far from out of the race, showed that central Quebec was not a strong part of the province for the party. It put them at 27%, behind the PQ and the CAQ. As their vote tanked in the polls in the regions of Quebec and among francophones (not to mention the riding polls that showed the Liberals with little support in Nicolet-Bécancour, Saint-François, Saint-Maurice, Sherbrooke, and Trois-Rivières), the projection put the Liberals at between 23% and 29% in the region. They ended up with 32.2% of the vote, beating out the CAQ (30.9%) and the PQ (28.4%).

As a result, this was the region with the most amount of errors and the region that skewed the projection against the Liberals most importantly. Their vote in this highly francophone region turned out in great numbers, voting for their incumbent MNAs in particular. Along with a surprising degree of resilience in the Montreal suburbs, particularly Laval, this was the region of Quebec that 'saved' the Liberals from a far more significant defeat.

Seat projections vs. results
But what happened to the Parti Québécois? They were over-estimated almost across the board, with eastern Quebec being the only region where they out-performed expectations. The seat ranges were only off by three seats for the PQ, however, and that can mostly be attributed to the strong Liberal performance in central Quebec. But it can also be blamed on the PQ's under-achieving results in the suburbs around Montreal.

The PQ was projected to take between 38% and 42% of the vote in this region, but ended up with 35.9%. Neither the Liberals nor the CAQ were outside of their projected ranges in the suburbs, but both were at the higher end of expectations. The 450 was considered the make-or-break region for both the PQ and the CAQ, and in the end this was correct. The PQ won 15 seats instead of the projected 19, putting them at the very bottom of their expected results. They could have won as many as 24 seats in the region, enough alone to have given them a majority government.

For the CAQ, they did as well as expected in and around Montreal and Quebec City. The projection gave them seven seats in Quebec City, which they indeed won, and seven in the suburbs. They took six.

But it was in francophone Quebec outside of the two main centres where the CAQ vote did not materialize. The party won no seats in eastern Quebec instead of the projected one, one seat in western Quebec instead of the projected three, and five seats in central Quebec instead of the projected nine. But, generally speaking, the projection was quite good in estimating the CAQ's vote - and the projection model would have still given them nine more seats than they actually won if the polls had been completely accurate.

Why? The CAQ's under-performance is one case where the error that the projection made tells us something important about the CAQ. The CAQ's vote share aligned quite closely with the polls and the projection, but their seat result did not.

The reason for this is that the CAQ's vote was far more uniform than the ADQ's. If the ADQ had gotten 27.1% of the vote in this election, they very likely would have won around 28 seats. But despite the CAQ having swallowed up the ADQ and its MNAs, the party is not the same. It does use the same populist and vaguely nationalist language of Mario Dumont, but the party and its leader take a more centrist approach. That changed the party's voter profile somewhat, enough to give the CAQ relatively even support - particularly in central and western Quebec. The model expected the CAQ's vote to be more heavily concentrated in central Quebec to the detriment of the western part of the province - instead, the CAQ's vote was more uniform and that meant a lot of second place finishes.

When the next election occurs, which considering the minority government in place could not be very long, the model will undoubtedly perform better because it will be able to use the CAQ's vote as a baseline, rather than the ADQ's.

Another factor which made the projection over-estimate the CAQ was the adjustment made for the floor-crossers in Blainville, Deux-Montagnes, and Sanguinet. While the adjustment would have worked in most elections, for whatever reason it did not work in this one. It over-estimated the support of Daniel Ratthé, Benoit Charette, and François Rebello to a significant degree. Without the adjustment in place, the model would have correctly given Blainville to Ratthé but Deux-Montagnes and Sanguinet to the PQ. The results for each party would have been quite accurate - indicating that the floor-crossers did not personally bring any large number of  votes with them.

If we look closely at the results, the projection did well on some of the minor details. Québec Solidaire's results were projected very well, with the result of 12% on the island of Montreal (their most important region) being off of the projection by 0.1 point. The model also did quite well with Option Nationale, the Greens, and the other parties.

The projected ranges were correct for every party in some tough-to-call ridings as well: Jean-Lesage (correctly forecasting a close three-way race), Groulx, and Gouin, among others. Generally speaking, the model did quite well on an individual basis for each party in the ridings where the winner was called correctly. The ridings that were not called correctly often had very large errors, suggesting that fiercely local circumstances were at play in many of them.

But the turnout adjustments may not have helped matters. While it did correctly assume that QS would be over-estimated and the PLQ under-estimated, and the adjustment did not effect the projection for the CAQ significantly, it was wrong to consider the PQ under-estimated in the polls. The last surveys by CROP and Léger were excellent for the PQ, and the last surveys by Forum and EKOS over-estimated the PQ greatly.

The model made a bet that, in most elections, would have been correct. Going forward, I am not sure if this is a bet that is worthwhile to make. The adjustment seems to have caused a lot of confusion during the campaign, with readers not understanding why I had the CAQ so low and the PQ and PLQ so high, no matter how many times I explained it. Perhaps it would be best for ThreeHundredEight to limit itself to poll aggregation and what those polls would deliver in terms of seats, with the high and low ranges being used to estimate how the polls are expected to be wrong. I welcome suggestions from readers.

With the surprising performance of Jean Charest's Liberals, the Quebec election threw almost everyone for a loop. But the results might have had more to do with the Liberals' stronger organization and the mythical prime à l'urne playing a role than the errors of polls and projectors. Quebec's distinct society does seem to come with distinct elections - no one will be surprised if next time the result is, well, a surprise.

Saturday, September 8, 2012

Final Projection and Results from 2012 Quebec Election

The below is an archived post of the Quebec projection page for the 2012 provincial election.

The numbers below reflect the final projection that was made on September 3, 2012 for the general election held the following day. An analysis of how the projection model performed can be found here.
All seat and vote projections are subject to the margins of error of the polls input into the models, as well as to the unpredictable nature of politics at the local level. The degree of uncertainty in the model is laid out by the inclusion of high and low ranges for both the seat and vote projections. These describe the likely range of results, with the actual vote and seat projections reflecting the most likely forecasted outcome. A full accounting of how the vote and seat projection model is designed can be found here and here, respectively.

The Poll Average reflects the unadjusted aggregation of all recent polls. The projection reflects the adjusted result, taking into account past discrepancies between voting intentions and voting results. The seat projections are based on the vote projections, and not the poll averages.

The following is a full account of the polls included in the model and their relative weighting..
The Riding Projections below were the forecasted results in each riding. The high and low ranges included in the projections are especially important, as the degree of uncertainty is far higher at the riding level. Any party whose high result is above the low result of the party projected to win the riding was considered to have a good chance at victory. Riding Projections are not polls and are not a reflection of the actual voting intentions of voters in each individual riding. They are forecasts based on regional trends.

(For best results, right-click and open the riding projection images in new windows or tabs in order to magnify.)
The following chart compares the final results of the election to the final projection.
The following charts display how the vote and seat projection ranges changed over time, as well as the final result.
These next charts show how the vote projection itself (the most likely outcome, rather than the ranges) shifted since the first projection of July 17, as well as the final results.
The following average of voting intentions by language was for information-purposes only. They were not used in the seat or vote projection models, and were unadjusted for turnout.
The Island of Montreal contains the following ridings: Acadie, Anjou-Louis-Riel, Bourassa-Sauvé, Bourget, Crémazie, D'Arcy-McGee, Gouin, Hochelaga-Maisonneuve, Jacques-Cartier, Jeanne-Mance-Vigier, LaFontaine, Laurier-Dorion, Marguerite-Bourgeoys, Marquette, Mercier, Mont-Royal, Nelligan, Notre-Dame-de-Grâce, Outremont, Pointe-aux-Trembles, Robert-Baldwin, Rosemont, Saint-Henri-Sainte-Anne, Saint-Laurent, Sainte-Marie-Saint-Jacques, Verdun, Viau, and Westmount-Saint-Louis.

Quebec City contains the following ridings: Charlesbourg, Chauveau, Chutes-de-la-Chaudière, Jean-Lesage, Jean-Talon, La Peltrie, Lévis, Louis-Hébert, Montmorency, Taschereau and Vanier-Les Rivières.

The Off Island Montreal Suburbs include Laval and the parts of the Laurentides, Lanaudière, and Montérégie within the Greater Montreal Area. They contain the ridings of Beauharnois, Blainville, Borduas, Chambly, Chateauguay, Chomedey, Deux-Montagnes, Fabre, Groulx, La Pinière, Laporte, La Prairie, L'Assomption, Laval-des-Rapides, Marie-Victorin, Masson, Milles-Îles, Mirabel, Montarville, Repentigny, Saint-Jérôme, Sainte-Rose, Sanguinet, Soulanges, Taillon, Terrebonne, Vachon, Vaudreuil, Verchères and Vimont.

Eastern Quebec includes the regions of Gaspésie-Îles-de-la-Madeleine, Bas-Saint-Laurent, Côte-Nord, and Saguenay-Lac-Saint-Jean. It contains the ridings of Bonaventure, Chicoutimi, Côte-du-Sud, Dubuc, Duplessis, Gaspé, Íles-de-la-Madeleine, Jonquière, Lac-Saint-Jean, Matane-Matapédia, René-Lévesque, Rimouski, Rivière-du-Loup-Temiscouata, and Roberval.

Central Quebec includes Estrie, Mauricie, Centre-du-Québec, and parts of Chaudière-Appalaches and the Capitale Nationale that are not within the Quebec City metropolitan area. It includes the ridings of Arthabaska, Beauce-Nord, Beauce-Sud, Bellechasse, Champlain, Charlevoix-Côte-de-Beaupré, Drummond-Bois-Francs, Johnson, Laviolette, Lotbinière-Frontenac, Maskinongé, Mégantic, Nicolet-Bécancour, Orford, Portneuf, Richmond, Saint-François, Saint-Maurice, Sherbrooke, and Trois-Rivières.

Western Quebec includes the Outaouais, Abitibi-Témiscamingue and the parts of Lanaudière, Laurentides, and Montérégie not within the Montreal metropolitan area. They include the ridings of Abitibi-Ouest, Abitibi-Est, Argenteuil, Berthier, Bertrand, Brome-Missisquoi, Chapleau, Gatineau, Granby, Hull, Huntingdon, Iberville, Joliette, Labelle, Papineau, Pontiac, Richelieu, Rousseau, Rouyn-Noranda-Témiscamingue, Saint-Hyacinthe, Saint-Jean, and Ungava.

Friday, September 7, 2012

NDP wins KitWat, Liberals hold Vaughan

Dalton McGuinty was not given a majority government last night and Andrea Horwath's NDP made a big breakthrough in Kitchener-Waterloo, as some 80,000 Ontarians voted in two provincial by-elections.

The result in Kitchener-Waterloo was the most anticipated, as a win by the Liberals would have awarded McGuinty enough seats to ensure his government's survival. Instead, Catherine Fife of the New Democrats earned 10,000 more votes than her party received in the 2011 general election in the riding and won.
Fife captured 39.8% of the vote (18,559 in all), increasing her party's share by 23.1 points and 10,309 votes - an incredible performance for a party that has not done better than 9,000 votes in recent elections.

Both the Progressive Conservatives and Liberals suffered. Turnout was high in Kitchener-Waterloo, almost equaling the turnout from the (admittedly very low) turnout from 2011. The Tories under Tracey Weiler finished second with 31.8%, a drop of 12 points and 6,842 votes from October.

Eric Davis of the Liberals saw his vote share drop by 12 points as well and 6,633 votes. This means that Fife was able to draw support from both parties in order to emerge as the winner.

The Greens, at 3.3%, actually increased their vote haul by 208. Not a bad showing for a party that usually does horribly in by-elections.

The final poll by Forum Research suggested that the New Democrats could win this, while ThreeHundredEight's forecast also gave the NDP a shot. It is worth noting, though, that both the New Democrats and Liberals were at the extreme limit of what was expected, suggesting that the NDP did a fantastic job in this by-election campaign - and that the Liberals did very badly.
Not so in Vaughan, where voters cast their ballots in about the same way as they did in October.

Steven Del Duca won with 51.2% of the vote, a bare 1.8 points behind Greg Sorbara's performance in last year's election. Tony Genco of the Progressive Conservatives increased his vote share by 2.2 points but, at 33.4%, he was still far behind Del Duca.

The New Democrats replicated their 2011 result with 11.3%. That might appear unremarkable, but it is difficult for a party that will clearly finish third in a by-election to get their voters to the ballot booths. The NDP had an abysmal result in the 2010 federal by-election in this riding, one that was a clear face-off between the Liberal (Genco, actually) and Conservative candidates. That their vote did not tank is a good sign - and the Greens increased their vote share in this riding as well.

This result was forecast both by the polls and ThreeHundredEight, so it went just about as expected. This highlights how the result in Kitchener-Waterloo was very specific to Kitchener-Waterloo. If the Liberal drop and NDP gain in that riding was supposed to be indicative of a provincial trend, then Vaughan should have experienced something similar. It did not.

Nevertheless, the night was a good one for the New Democrats. Their vote held firm in Vaughan and soared in Kitchener-Waterloo. They picked up a seat and were last night's clear winners. The Liberals held on to Vaughan with no significant drop in support, but fell sharply in Kitchener-Waterloo where they had high hopes. It was a mixed night for them. That makes the Progressive Conservatives the party that ended up on the bottom, as they lost one of their 'safe' ridings and made no gains in Vaughan, the kind of riding they will need to win in the next election if they are to form government.

The results of these two by-elections mean that Dalton McGuinty's government could still be defeated at any time. But the results may also extend its tenure, as they give Tim Hudak no reason to pull the plug.

Thursday, September 6, 2012

Fateful by-elections in Ontario

Two by-elections are taking place in Ontario today. Dalton McGuinty's Liberals can cobble together a pseudo-majority government if they win both, and it would go a long way towards ensuring the survival of the Ontario government until the end of McGuinty's four-year mandate. The Liberals have a very good chance at winning one of the two by-elections, but the other is truly too close to call.

That by-election is the one in Kitchener-Waterloo. Formerly a safe Progressive Conservative seat under Elizabeth Witmer, Kitchener-Waterloo has transformed into a three-way race where each of the major parties has led in a poll.
The fundamentals of Kitchener-Waterloo make this riding one that the Progressive Conservatives should win. Witmer won it by respectable margins in 2007 and 2011, and the provincial trends suggest that the Tories should have no problem holding the riding. The party is either steady or up in the polls while the Liberals are down.

But Tracey Weiler will have a tough fight on her hands. The first poll by Forum Research shortly after Witmer's resignation gave the Liberals a seven-point lead. A few months later, Forum gave Weiler a four-point lead and suggested that the New Democrats were making gains. Their last poll, taken only two days ago, put Catherine Fife of the NDP ahead by 16 points!

That doesn't mean Liberal candidate Eric Davis is out of the running. Another poll, this time by Oracle, taken at the end of August put the Liberals and Tories in a tie, with the NDP only four points behind.

Taking the fundamentals of the riding into consideration as well as the riding specific polling, ThreeHundredEight can only forecast that this riding is a toss-up between the Progressive Conservatives (25% to 46%), the Liberals (25% to 38%), and the New Democrats (14% to 39%).

The winner in Kitchener-Waterloo will be the party that gets its voters to the polls. But with the high stakes, this should be a high turnout by-election. Will voters go with the party they have in the past? Will they give McGuinty a majority? Or will they try out the NDP?
The by-election in Vaughan is much more clear cut. The Liberals have dominated in this riding, winning by more than 40 points in 2007 and 20 points last year. The fundamentals give the Liberals the edge despite their sagging poll numbers, and polls done specifically for the riding have given Liberal candidate Steven Del Duca a double-digit lead over the Tories' Tony Genco. A poll done in early August did give the PCs a one-point advantage, but this was before the candidates for the riding were nominated.

This makes Vaughan a Strong Liberal riding, and one that the Liberals should hold. But Vaughan can swing - it went over to the federal Conservatives in 2010 in a close by-election and then voted strongly for Julian Fantino in 2011. So an upset could always happen, but Vaughan should stay Liberal. The forecast is for the party to take between 40% and 58% of the vote, with the Tories taking between 28% and 39% and the New Democrats between 9% and 15%.
Lost in the hubbub over the Quebec election on Tuesday was a by-election in Manitoba, where Brian Pallister (newly minted leader of the Progressive Conservatives) was making his attempt to enter the Manitoba legislature in place of former leader Hugh McFadyen.

As expected, Palliser easily won the by-election in Fort Whyte with 55.2% of the vote, a small drop of seven points from the 2012 election. He won by 23.6 points, virtually identical to the average margin of 24 points expected by the forecast model.

What was unexpected was the strong performance of the Liberals. Bob Axworthy was a decent candidate for the Liberals and put up a hard fight, while the governing New Democrats held back. As a result, the Liberals made a gain of 23.7 points and finished with 31.6%, a great result for a party that had 7.5% support province-wide in the last election (and 7.9% in Fort Whyte). The New Democrats accordingly dropped 18.4 points to only 11.4%, but considering the circumstances I would not take this as a knock against the government.

What the results do show is that the Liberals are not quite dead yet. If they are seen as the best alternative to the Progressive Conservatives in a riding, Manitobans will vote for them. Will the Liberals be this competitive in Fort Whyte in 2015? Probably not, but they need the little bit of good news more than the New Democrats, safely ensconced in Year 13 of their reign and with a majority government, do.

Wednesday, September 5, 2012

Quebec election aftermath

Last night's election results in Quebec were surprising, but the event that took place in Montreal after the results were known overshadowed to a great degree what happened in the voting booths. The attack perpetrated by an undoubtedly deranged individual was an assault not only on his victims, but on all Quebecers - and all Canadians - and upon our democratic values. My thoughts are with the families of the two victims. Tragically one has passed away.

This sort of heinous act puts a lot of things in perspective, but it is important not to let the crime of a single man carry more weight than the act that millions of Quebecers took part in yesterday, an act that hundreds of millions of people worldwide are forbidden from doing and for which they have given their lives.

I will have some analyses of last night's results on the websites of The Globe and Mail and The Huffington Post Canada later today and I will update this post with links. UPDATE: The Globe article can be found here. The article for The Huffington Post Canada can be found here.

It will take a little while to pore over how the projection performed, especially as I will have to calculate the regional results of last night's vote in order to see how the model would have done with the correct numbers. That was a big problem last night. No poll was very accurate, especially in terms of the Liberals. The highest number they had in a late-campaign poll was 29%, and that was in Forum's election-eve survey that also had the PQ at 36%. The poll that had the best result for the Liberals also had the worst result for the PQ. In other words, it was not very helpful.

The projection did make a few good bets, forecasting that the polls would under-estimate the Liberals and over-estimate the support of the CAQ (edit: at least in terms of Léger and CROP), Québec Solidaire, Option Nationale, and the Greens. It was wrong to believe that the polls would also under-estimate the Parti Québécois and that was the difference between projecting a bare-minimum majority government and the minority government the party actually won.

A preliminary analysis suggests that the projection made the right call in about 82% of ridings (with the projected ranges, that is bumped up to 86%), a decent performance considering the faulty inputs of the polls. There were a few surprises, and like in the Alberta election the resilience of the Liberal incumbents was extraordinary.

Once I plug the actual regional results of the election into the model, I'm very confident that the projection will provide a close result. But that just proves once again that the challenge is being able to forecast results with the polling data that is available. More and more, that data is looking less and less useful. The two polls conducted on the eve of the vote were both fatally flawed (one suggested a clear PQ majority, the other a Liberal catastrophe). It is unfortunate that Léger and CROP were not in the field later in the campaign, but it is impossible to know if they would have performed any better.

In the end, ThreeHundredEight's vote projection turned out to be closer than any of the polls, with a total error of 8.1 points (compared to 8.4 for Léger and 10.4 for Forum/CROP). That is not quite good enough, and more work will have to be done to determine how to best forecast the difference between voting intentions and voting behaviour. Wider confidence intervals and the recognition that there is only so much that polls can tell us about what will happen tomorrow may be in order.

Monday, September 3, 2012

Final Quebec Projection: Parti Québécois victory

Tomorrow night, projects that Pauline Marois and the Parti Québécois will win the Quebec provincial election with a plurality or majority of the National Assembly's 125 seats. Though the projection leans slightly towards a majority government, the margin is too close to definitively project with confidence what form the government will take.

This post was updated on September 4, 2012 - see below.

Similarly, either the Liberals or the Coalition Avenir Québec will emerge as the Official Opposition. It is more likely that the Liberals will be given the job, but the CAQ is also capable of jumping from third to second spot in Quebec City.

With 34.1% support, the Parti Québécois is projected to win 63 seats tomorrow night. That is the bare minimum for a majority government. The Liberals, with 27.9% support, are projected to win 33 seats. That narrowly edges out the CAQ, who are projected to take 26.3% of the vote and 27 seats.

The PQ's wobbly majority (if the Speaker is from the party, the Assembly would be split straight down the middle) is likely to get some support from Québec Solidaire. The party is projected to win 7.1% of the vote and two seats.

Option Nationale is projected to take 2.5% of the vote, while the Green Party of Quebec and other parties are expected to get 1% of the vote each.

This campaign has shown a remarkable amount of stability for all of the noise that has come out of the province. Since the campaign began, neither CROP nor Léger Marketing have pegged PQ support at anything other than 32%, 33%, or 34% support, with no discernible change in support since the beginning of the campaign. The Liberals have dropped from around 30% to around 27%, but aside from the Forum poll the Liberals have been steady since before the series of four debates. The CAQ has also been very steady in the last bit of the campaign. They did experience a bump from the early days, but it was only an increase of a handful of points (Léger had them at 28% in their final poll of the campaign and at 27% in their poll from Aug. 6-8).

This leads me to believe that we should not expect any sort of late swing over the final four unpolled days of the campaign (Sept. 1 to 4). If things weren't moving with a debate and the myriad of rather intense stories over the course of the election, why would they suddenly swing in this last stage?

UPDATE (Sept. 4, 2012): As they did at the end of the Alberta provincial campaign, Forum Research conducted a poll last night. They released it through The National Post. Frank Graves from EKOS Research also published the results of their own polling on his Twitter feed before midnight.

I will not be updating the projection to include this new information (it would be impossible to include EKOS's numbers as they only reported the topline results). Federal electoral law and the law in many provinces prohibits the publishing of new public opinion polls on Election Day. Whether projections would also be prohibited could potentially be up to the interpretation of a judge. But in addition to not wanting to run that risk, I believe ThreeHundredEight should try to follow the spirit of the law and not publish new projections on Election Day.

The Forum poll, in any case, backs up the final projection. Like the projection, Forum says that Léger and CROP have under-estimated the PQ and the Liberals and have over-estimated the CAQ and QS. If Forum has it, then the PQ will win a majority government and almost certainly a number of seats within the projection's upper range for the party. Similarly, the Liberals will become the Official Opposition and the CAQ the third party. I am confident that the projected seat ranges will hold true when the ballots are counted tonight.

EKOS's poll is another kettle of fish. While they have identical numbers for the PQ and the CAQ as does Forum Research, they have the Liberals at 23% instead of 29% and QS at 11% instead of 6%. That is a very important difference. Granted, EKOS published these numbers as part of a longer poll and as a means of calibrating their surveying model. But the firm can't put money in the pile and expect to get it back if their bet loses. Forum and EKOS are going for the glory, while Léger and CROP are safely ensconced in virtually identical numbers a few days back from the vote. It will be interesting to see who will turn out to be right. And in the post-mortem, I will post what the projection would have shown had I included the numbers from Forum and EKOS.

Perhaps the long weekend, one of those mythical opportunities for families and friends to discuss politics and make a collective decision, will play a role in tomorrow night's vote. But there is nothing to suggest that we should expect a wholesale swing from one party to another. The number of undecideds is not unusually high and though the CAQ's supporters seem the softest, this comes as no surprise considering they are a 'new' party.
Because of this lack of volatility in the last stage of the campaign, the projected seat ranges are relatively tight.

The PQ is projected to win between 57 and 75 seats, straddling the line of a majority government but not overlapping with any of the other parties: the Liberals are projected to win between 25 and 39 seats and the CAQ between 20 and 31 seats. The fight, then, is between these two parties for the role of the Official Opposition. As most of the battles are not between the Liberals and the CAQ, however, if the CAQ over-achieves and manages to place second there is a greater chance that the PQ will be reduced to a minority. On the other hand, if Liberal support tanks the PQ is most likely to take advantage and form a majority government.

The question of how voting intentions will differ from actual votes has been an important one for this campaign, as the projection model bets that the PQ and PLQ will be under-estimated and the CAQ and QS over-estimated.

But if that does not occur, and if the polls are bang-on, it would not change much. Removing the turnout adjustment from the projection and instead using only the poll average, the PQ would still win with 61 seats - just below a majority government. The Liberals would still form the Official Opposition with 33 seats and the CAQ would win 29 seats.

The ranges would put the PQ at between 53 and 75 seats, the Liberals at between 24 and 38 seats, the CAQ between 21 and 36 seats, and Québec Solidaire between one and three seats. In other words, the projection would move from leaning towards a PQ majority to leaning towards a PQ minority, while the Liberals and CAQ would have almost even odds at forming the Official Opposition. Hardly paradigm-shifting.

Now that we have gone over the general outline of the projection, let's look at it in further detail by going through it party-by-party.

Click to zoom
Though the Parti Québécois is almost certain to finish on the top of the pile tomorrow night, the campaign will not have been one of their shining moments. A win is a win, of course, but the PQ is likely to emerge victorious with less of the vote than they earned in their 2008 defeat (35.2%), not to mention their defeats of 1985 and 1989.

The big unknown is whether the PQ will be able to win a majority government. They are on the cusp of one, meaning that a handful of votes in a handful of ridings could make the difference. Will supporters of Option Nationale and Québec Solidaire swing over to the PQ to give them a majority? Or will the PQ slip through on their own?

The PQ is projected to win between 32.8% and 35.5% of the vote, which runs the gamut of a middling minority to a strong majority government. They appear to have the support of 37.3% of francophones, precariously close to minority territory, while they also have the support of 9.2% of non-francophones.

Their best regions are in eastern Quebec and the Montreal suburbs, where the party is expected to win half of their seats. Eastern Quebec is particularly strong for the PQ, as they should sweep the Gaspésie, the Saguenay, and the Côte-Nord with 43.4% of the vote. They are also looking good in the Montreal suburbs with 39.5% support, but with between 15 and 24 seats it is the region where their majority government will likely be won or lost. They are fighting the CAQ north of the city and the Liberals in Laval, primarily.

On the island of Montreal, however, the PQ is likely to re-elect all of their incumbents except Nicolas Girard in Gouin, though the projection does consider a Girard victory possible. How the QS vote will turnout could determine whether the PQ is able to win any new seats on the island. The PQ should do well in the Mauricie region of central Quebec, while they should win two seats in Quebec City (but could win between one and three).

With the lone possibility of a single seat in Quebec City, if the PQ wins a majority government they should be well-represented in every part of the province. If they are given a minority, it will likely be because the Quebecers who live in the suburbs of Montreal defect to the CAQ, like they did in 2007 when they swung to the ADQ. That will be the region to watch.

Click to zoom
Jean Charest's Liberals started the campaign in a surprisingly good position, considering the problems the party and government has had in its third mandate. The Liberals were within a couple of points of the PQ and with a good campaign could have managed to be re-elected to a minority government. François Legault got off to a running start with the candidacy of Jacques Duchesneau, which brought the corruption issue to the forefront, and from that point the Liberals saw their re-election chances drop to nil.

But they still have some solid bases of support. The projection gives them between 26.9% and 28.9% of the vote, enough to keep them in the running for the job of the Official Opposition. They have the support of 61.4% of non-francophones, which gives them a good chunk of seats automatically, but only 19.6% support among francophones. That means a lot of defeated incumbents off the island of Montreal.

That is their best region, where 20 of their 33 seats are likely to come from. They lead with 36.8% but are in danger of being reduced to as few as 16 seats on the island, depending on how the PQ vote turns out.

Elsewhere, the party is in dire straits. After leading at the outset, the Liberals have dropped precipitously in Quebec City. They are projected to win 27.1% of the vote but only between zero and four seats. In the suburbs around Montreal they could win three or six, thanks in large part to residual support in Laval.

In eastern and central Quebec, the party might only win three seats. The four seats in the Outaouais projected to go Liberal, meanwhile, is the only pocket of concentrated Liberal support outside the West Island.

But I think the Liberals should not be under-estimated. The possibility exists where support for the Liberals will simply tank as voters decide that they can't support this nine-year-old government and will cast their ballot for a party that can win. However, the Liberals have a good organization and are the go-to option for convinced federalists, so I would not be surprised to see the Liberals over-achieve expectations. That could, for example, play out in Sherbrooke. Jean Charest is projected to lose, but it is always risky to bet against Charest. If one Liberal riding is going to be missed by the projection, I expect it to be this one. On the other hand, a gap of 12 or so points like the one recorded in the last Segma poll is also easy to imagine. The survivability of Liberal incumbents could be one major factor tomorrow night that prevents the CAQ from finishing second.

Click to zoom
The projection believes the CAQ will finish third, however. The CAQ has decent support but it lacks efficiency and concentration. They are lagging behind the PQ among francophones by seven points, and are likely to finish a strong second in a lot more ridings than the Liberals will. But the CAQ is the big unknown of tomorrow night - will they over-achieve or under-achieve? If anyone has momentum, it is the CAQ - and they also have the 'change' card in their deck. But they lack an organization and the kind of enthusiasm needed to make-up for that deficient ground game. And will their support emerge in unexpected places?

With between 24.8% and 27.8% projected support, the CAQ does have a chance of finishing second in the popular vote. Even so, the odds are against them finishing second in the seat count, due to the Liberal advantage in Montreal. They have the support of 30.3% of francophones and 18.1% of non-francophones, a very respectable score for François Legault. But neither of those numbers are high enough to put the CAQ over the top.

Their two strongest regions are Quebec City and central Quebec. The CAQ is projected to win seven of 11 seats in the capital, and lead with 37.4%. In central Quebec, particularly in Chaudière-Appalaches and parts of Estrie and the Centre-du-Québec, the CAQ is well placed to win a swathe of seats as well.

Whether they form the Official Opposition could be decided in the Montreal suburbs, where they are projected to win seven seats (including those of Legault, Duchesneau, and Gaétan Barrette). They could bump that up to nine, but the polls have not moved in their favour in this region enough to give them many more seats.

The party is weaker in eastern and western Quebec, where only a handful of seats might go their way. They are expected to be shut out from the island of Montreal, though with 17.1% support they have done much better than the ADQ.

The CAQ is the hardest party to project in this election because of its status as a "new" party. The polls indicate that the CAQ is following in the footsteps of the ADQ, but there is a strong possibility that their support could be higher or lower than expected in certain ridings. The three ridings occupied by the floor-crossers (Blainville, Deux-Montagnes, and Sanguinet) are particularly mysterious. The model awards bonuses to floor-crossers as they take their old votes with them, but it is hard to say if that will happen with these three CAQ incumbents. These are three seats that I suspect could go a different way tomorrow night than expected, particularly Blainville where the PQ has a strong candidate. If that occurs, though, the odds of a CAQ Official Opposition diminish considerably.

If any party is going to cause a surprise in the polls, it is the CAQ. How their vote turns out will likely decide the result of the election.

Click to zoom
How the votes of Québec Solidaire will turnout will also be very important. From less than 4% in 2008, QS has done fairly well in this election. They are projected to almost double their support, and if things go very well for them they could more than double their vote haul.

The party has 7.6% support among francophones and 4.7% support among non-francophones, a decent split for a party that has their best shot in some of the most multi-cultural parts of Montreal.

That is where their election will play out, as the projection gives them 12.1% on the island. They could also take as much as 16.1%, but the projection does not award QS the possibility of more than two seats.

Personally, I would not be shocked to see QS win three or four. They have a good shot in Laurier-Dorion and Sainte-Marie-Saint-Jacques, two ridings where Québec Solidaire's maximum result is three points or less from being within range of making the ridings potential pick-ups. Amir Khadir was not expected to win in Mercier in 2008, so we should not be surprised if QS is able to over-achieve in these two ridings.

Outside of Montreal, QS is not in the running for any other seats. They are not expected to get more than 6% support in any other region of the province. They could put up some good scores in Hull and Taschereau, but in neither of those ridings is the expected winner low enough to put QS within range.

If QS manages to get both Khadir and Françoise David elected, they should consider the election a success. Having their two strongest personalities in the National Assembly will make it much easier for them to win three or four seats in the next election. If they manage to pull that off this time, the election will be a smashing success.

Meanwhile, the survival of Option Nationale depends on whether their leader is re-elected in Nicolet-Bécancour. The projection does not expect that to happen, but it also does not rule it out. It is very difficult to estimate what will happen in this riding. The polls have shown that either Jean-Martin Aussant or the CAQ's candidate could win, but the samples have been so small that the riding can be considered a four-way race. If Aussant is not elected, ON will not survive as a viable party. If he is re-elected, it has a shot.

But its support is very small and the party is unlikely to hit double-digits in any riding aside from Aussant's. That they managed to put 121 candidates on the ballot is remarkable for such a young party, but while its youth gives it some of the most enthusiastic support it also provides a handicap. Young people don't vote in the same numbers as older people.

People who support Option Nationale would likely vote PQ or QS if the party was not on the ballot. Aside from seeing whether Aussant gets re-elected, it will be interesting to see whether ON will end up costing either PQ or QS any seats.

This was a horrible campaign for the Green Party. Their leader, Claude Sabourin, was unable to get himself in the news aside from a story in which a reporter from La Presse was able to get himself nominated as a Green candidate without any sort of background check.

With little more than half of a full slate, this is the worst performance for the Greens since 2003. They are not expected to take more than 2% of the vote, and are more likely to get 1%. Sabourin might manage double-digits in Notre-Dame-de-Grâce, but any shot the Greens had at strong performances were lost when the CAQ emerged as the alternative choice of anglophones.

One long night, many questions to answer

Let's be honest - no one will look at polls and projections the same again after the Alberta campaign. Never mind that it was an extraordinarily exceptional event, there will always be some (including Jean Charest) who will point to that result as a reason to never believe polls ever again. Tomorrow will determine whether he has a point. If something like that happens, the race is close enough that any of the three main parties could win. I really don't think something like that will happen, but let's take the opportunity to recognize that no one can ever be 100% certain of anything when it comes to democracy.

Putting that aside, Quebec appears to be almost certainly headed towards a PQ government. The question that will be answered tomorrow night is whether they will be awarded a minority or majority government. If the PQ does not get a majority, their minority is likely to be so large that a scenario where Jean Charest remains in power is unimaginable. So that is really what is at stake, and it will be up to those swing voters hesitating between QS/ON and the PQ, the CAQ and the PQ, and the Liberals and the CAQ to decide.

Then there is the question of who will form the Official Opposition. If the Liberal vote turns out, their vote efficiency virtually assures they will finish second. If it doesn't, or if the CAQ gets a boost in support in the last days of the campaign, then François Legault will become the Leader of the Official Opposition. If Jean Charest is not elected in his riding of Sherbrooke, Legault will likely become the de facto head of the opposition either way.

In addition to those two big questions, there are many small ones to answer. Will Françoise David be elected? Can QS win a third or fourth seat? Will Jean-Martin Aussant return to the National Assembly? Will voters be scared off from the PQ because of the threat of a referendum? Will Jacques Duchesneau, Gaétan Barrette, Léo Bureau-Blouin, the CAQ transfuges, and even François Legault win in their ridings?

It will be a long and fascinating night.