Tuesday, December 29, 2015

Satisfaction with Trudeau remains high; Manitoba NDP in third

Nanos's weekly 'best prime minister' tracking continues to wobble back and forth, with Justin Trudeau now back up to 53.1%, compared to 12.7% for Tom Mulcair and 11.3% for Rona Ambrose. Early days, of course, but that wobble seems to only be happening for Trudeau. Mulcair is holding steady but Ambrose has yet to experience her numbers rising from one week to the next.

Elizabeth May had 5.1% and Rhéal Fortin had 1.2%.

More relevant poll than who would be the best person for a job that Trudeau has for the next four years is a Nanos survey asking about Canadians' satisfaction with what the Trudeau government has been doing. The results are good for the Liberals: 67% of Canadians are satisfied or somewhat satisfied with what the Trudeau government has done so far. Just 28% were dissatisfied or somewhat dissatisfied.

Satisfaction was highest in Atlantic Canada (79%) and British Columbia (74%), and was lowest in Alberta and the Prairies (57%).

Manitoba NDP drops to third place

Greg Selinger's New Democrats got a lump of coal on Boxing Day when the Winnipeg Free Press published the latest results from Probe Research. The numbers were ugly.

The PCs, who have been holding steady pretty much since the last election, were leading with 43%, followed by the Liberals at 29% and the NDP at just 22%. Those are the best and worst numbers for the Liberals and NDP respectively for a very long time — Probe Research, operating for the last two decades, has never had the NDP in third.

And it gets worse from there. The NDP is trailing in Winnipeg with 29% to 35% for the Tories (the Liberals also have 29%). Only in the core part of Winnipeg does the NDP have a lead. Everywhere else around the city the PCs are in front, and in most parts it is the Liberals that are in second.

The PCs also look set to virtually sweep the area outside of Winnipeg (53% to 29% for the Liberals and 13% for the NDP).

So things are looking good for the PCs in the upcoming April election. Coupled with Brad Wall's likely victory in neighbouring Saskatchewan in that province's April election, right-of-centre parties might have their first victories in three years in 2016. I wrote about what could be a fleeting conservative comeback for the CBC today.

Thursday, December 24, 2015

Ontario PCs narrowly lead unpopular Wynne; plus a little Christmas cheer

A new poll from Forum Research and the Toronto Star shows that the Progressive Conservatives under Patrick Brown enjoy a narrow lead over the Ontario Liberals, despite Kathleen Wynne's disapproval ratings being sky-high.

The poll (only the Star report is online at the moment) gave the Tories 34% and the Liberals 31%, virtually unchanged from the 36% to 30% spread that Forum recorded in early November. The New Democrats had 26%, also unchanged from that previous poll.

Brown seems to have a profile problem. He has a net positive approval rating at 25% approval to 24% disapproval, but 51% of Ontarians still don't have an opinion of him. That is a lot of people to impress — or repel.

Wynne has an approval rating of just 23%, with 61% disapproving of her performance as premier.

Andrea Horwath of the NDP had a 40% approval rating and 26% disapproval rating.

It is a problem for the PCs when they are not doing better against such an unpopular premier. A similar situation in Manitoba is taking place, but the numbers are completely different. In Manitoba, the New Democrats have been in power since 1999 though the premier, Greg Selinger, is relatively new. He is deeply unpopular and trailing in the polls by about 20 points.

In Ontario, the Liberals have been in power since 2003, though Wynne is a relatively fresh face. She is also unpopular. But she is trailing in the polls by just three points.

Twas the night before Christmas on Parliament Hill

I deal with numbers on a daily basis, but for this holiday season I thought I'd try my hand with words. Here's my rendition (along with a dramatic reading!) of "A Visit from St. Nicholas", adapted for the Ottawa bubble.

Happy holidays!

Tuesday, December 22, 2015

Trudeau still tops Nanos PM poll; Watson approval ratings high

The latest installment of the Nanos weekly tracking poll still has Justin Trudeau well ahead on who Canadians think is the best person to be prime minister.

He topped the poll with 50.2%, followed by Thomas Mulcair at 13% and Rona Ambrose at 12.9%. Elizabeth May scored 5.1%, while Rhéal Fortin was at 1.3%.

Compared to the last completely independent sample (recall that Nanos employs a four-week rolling poll with weekly updates), Trudeau is down a little more than three points. Ambrose is also down, but that previous sample included some numbers for Stephen Harper, so it is an apples to oranges comparison.

Trudeau's drop is within the margin of error, but it will be interesting to track these independent samples rather than just the week-to-week variations. I'll put together a chart for that once Nanos's poll chugs along for a little longer.

Ottawa Mayor Jim Watson boasts high approval rating

Forum Research put out a poll on Ottawa issues today. They are virtually all of very local concern, so I invite you to peruse Forum's report.

But Mayor Jim Watson's approval rating is of note. Forum pegs it at 73%, with just 27% of Ottawa residents disapproving of Watson. His numbers are high across the board, particularly among Ontario Liberal and Ontario NDP supporters. But even a narrow majority of PC supporters in the city support Watson, who in a former life was an Ontario Liberal MPP and cabinet minister.

Note that Watson took 76% of the vote in Ottawa's 2014 municipal election.

Monday, December 21, 2015

Divided vote boosting PC chances in western Manitoba

A poll published in the Brandon Sun by Probe Research shows just how much trouble the Manitoba New Democrats are in, with four months to go before voting day in the province.

The poll shows that the NDP's support in western Manitoba has dropped by more than half, from 41% to 18%. Virtually all of that has gone to the Manitoba Liberals, who have risen from 4% to 22%. The Progressive Conservatives, meanwhile, have not seen their vote change at all: they took 53% of the vote here in the 2011 provincial election, and are now polling at 54%.

This is the developing story of the Manitoba election, that the PCs will almost win by default as the NDP loses a huge share of its vote to the Liberals. Look at the regional breakdown in the Probe poll.

Brandon West and Brandon East were relatively close PC-NDP races in 2011, with the NDP prevailing in Brandon East. Now, with hardly any movement on the part of the PCs, the Tories are well in front in Brandon West and narrowly ahead of the Liberals in Brandon East. Considering that the Liberals are not likely to have the sort of organization they will need in this election (they won 7.5% of the vote and one seat in 2011), it will be difficult to get out all of the vote to win a riding like Brandon East — further boosting the PCs chances.

In the rural parts of western Manitoba, the PCs are even less likely to be challenged. But even there, in Parkland the margin was just 14 points in 2011. Now, with the PCs down two points, the margin is 29 points.

Tuesday, December 15, 2015

Trudeau continues to have majority support as PM

With detailed posts becoming less frequent here — at least until the provincial elections in Manitoba and Saskatchewan kick-off — I thought it might be useful to do shorter posts to alert readers of and summarize new polls on various topics that have come out in the previous 24 hours. I'd appreciate your thoughts in the comments section about whether this is worthwhile.

The latest poll comes from Nanos Research, as part of its weekly rolling poll. The latest numbers show Justin Trudeau at 51.5% on who Canadians see as the best choice for prime minister, followed at length by Rona Ambrose at 13.7% and Thomas Mulcair at 12.8%. Elizabeth May and Rhéal Fortin (interim leader of the Bloc) scored 5% and 1.2%, respectively. All seem to be holding steady after the post-election re-alignment.

You can read the full report from Nanos here.

Ambrose is suffering from a lack of familiarity (41% say they do not know if she has the qualities of a good leader), but Mulcair is clearly at a low ebb, as Nanos has not had him this low since the tracking began in 2013.

Thursday, December 10, 2015

Premier approval ratings: cracks in the Wall?

The Angus Reid Institute has published its quarterly review of premiers' approval ratings from coast-to-coast (excluding the territories and Prince Edward Island). The numbers show some significant shifts in opinion since Angus Reid's summer poll, and diverging fortunes for the two premiers headed for an election in April.

But while Brad Wall still topped the list, some (small) cracks are beginning to show in his dominance in Saskatchewan.

Wall's approval rating stood at 60 per cent in the poll, with 35 per cent disapproval. That puts him as the only premier with a majority approval rating, but it is his lowest approval rating recorded by the Angus Reid Institute since these quarterly polls began at the end of 2012.

The Saskatchewan premier was the only one with a strong net rating, but two other premiers also had positive scores. Stephen McNeil of Nova Scotia had the approval of 46 per cent and the disapproval of 43 per cent, while Rachel Notley in Alberta had an approval rating of 45 per cent and a disapproval rating of 44 per cent.

(Paul Davis was included in this poll, as he lost the Newfoundland and Labrador election on November 30.)

Brian Gallant of New Brunswick was the only other premier with a generally even rating, scoring 34 per cent approval and 39 disapproval. His 'not sure' rating of 27 per cent was, by far, the largest.

Four premiers put up some very poor numbers. The best of them was Philippe Couillard of Quebec, with an approval rating of 35 per cent and a disapproval rating of 57 per cent. He was followed by British Columbia's Christy Clark (34 per cent approval, 60 per cent disapproval) and Ontario's Kathleen Wynne (30 per cent approval, 60 per cent disapproval). For Wynne, that was her lowest rating since becoming premier in early 2013.

At the bottom of the list — again — was Manitoba's Greg Selinger, with an approval rating of 22 per cent and a disapproval rating of 65 per cent.

Wall and Selinger, book-ending the table, are also the two premiers headed to an election in the spring. However, in terms of who has positive momentum and who has negative momentum, the roles are reversed.

Since the last quarter, Selinger's net approval rating has increased by three points, from a woeful -46 to a still woeful -43. But since this time last year, in the final quarter of 2014, Selinger's net approval rating has actually improved by 13 points. That is the greatest improvement recorded by any premier in the last year. Admittedly, though, it was from a very low base.

Wall, on the other hand, has seen his net approval rating drop by seven points since the last quarter, and 14 points since last year. Only Wynne had a worse year-on-year drop in support, with her net rating dropping by 20 points to -30.

McNeil and Gallant, perhaps buoyed by the federal Liberals' strong performance in Atlantic Canada, saw their net approval ratings increase by eight and 30 points, respectively. Clark was also up three points since the last quarter.

Notley, however, has seen her net approval rating slide by 13 points over the last three months. That has decreased her net rating from a respectable +14 to a break-even +1. Couillard's rating has also dropped by 13 points over the last three quarters, though he was in a worse position a year ago.

If approval ratings can act as a proxy for potential electoral performances, a few of these premiers are in a very safe position: Wall, McNeil, and Notley. Couillard could prevail in a divided political landscape in Quebec. Clark and Gallant would be in much more trouble, while Wynne and Selinger would be defeated.

Luckily for most of these premiers, they do not have to face the electorate for quite awhile — the provincial elections on the schedule for 2017 are in British Columbia and Nova Scotia, and Clark has shown her resilience before. Wall is a virtual lock for the April election in Saskatchewan, but Selinger looks like he will be in tough.

Of course, four months is an eternity in politics, so neither Wall nor Brian Pallister, leader of the Manitoba Progressive Conservatives, can take anything for granted. But these will be hard numbers to mess up, in Wall's case, or, for Selinger, overcome.

Monday, December 7, 2015

November 2015 federal polling averages

The post-election period begins, and the federal polling will remain light for a little while. But two polls were conducted last month, and they show the Liberals in the midst of quite a little honeymoon.

The Liberals averaged 52% in those two polls conducted by Forum Research and Abacus Data. That is an increase of over 12 points from the election results.

The Conservatives averaged 24.5%, a drop of over seven points, while the New Democrats were down almost six points to 14%.

The Bloc Québécois averaged 4.5%, while the Greens averaged 4%.

The Liberals led in every region except Alberta, with 44.5% in the Prairies, 51.5% in British Columbia, 52.5% in Quebec, 55% in Ontario, and 68% in Atlantic Canada. The Liberals were in second with 33.5% in Alberta.

The Conservatives led in Alberta with 50%, and were second in the Prairies (38%), Ontario (27%), and British Columbia (24%). The were third in Atlantic Canada with 10.5% and fourth in Quebec with 13%.

The New Democrats were in second only in Atlantic Canada, where they stood at 15%. They had third place showings in the rest of the country, with 14.5% in British Columbia, 9% in Alberta, 15.5% in the Prairies, 14% in Ontario, and 15% in Quebec.

The Bloc Québécois held the second spot in Quebec with 17.5% support. The Greens had their best result in British Columbia, where they were at 9.5%.

I haven't yet updated the three-election projection model to incorporate the 2015 federal election results, but based on the 2011 three-election model these numbers would deliver the following seat ranges:

Liberals: 231-278 seats
Conservatives: 51-92 seats
New Democrats: 5-10 seats
Bloc Québécois: 0-6 seats
Greens: 1 seat

Early days, of course, and the polls have little practical impact. But it does suggest that Canadians are reacting favourably to the Liberals' first weeks in office.

Wednesday, December 2, 2015

Newfoundland and Labrador election post-mortem

In the end, the results were a little more interesting than people were expected. The Liberals still won a big majority in Newfoundland and Labrador (roughly as big as the PCs won in 2011), but the opposition was slightly more robust than some had feared it would be.

The Liberals won 57.2% of the vote, and captured 31 of the 40 seats that were on offer. The Tories took 30.1% of the vote and seven seats, while the NDP took 12.1% of the vote and two seats. I went over the regional breakdown of the results in my CBC piece here.

But how did the projection do? Broadly speaking, it did a decent job.

The Liberals were projected to within one percentage point and one seat and the Tories within two percentage points and one seat. The NDP was over-estimated by a little more than three percentage points and two seats.

But overall, with the emphasis having been placed on the minimum/maximum projections, the forecast was good.

And the new three-election model proved its worth. Had the old model been used, the Liberals would have been projected to win 27 seats, with nine going to the Tories.

The district-level projection was not as good as I would like it to be, making the right call in 33 of 40 ridings for an accuracy rating of 82.5%. The potential winner was identified by the likely ranges in one more district, bumping that accuracy up to 85%.

In the maximum ranges, the potential winner was identified in all but one district — the result in Fortune Bay–Cape La Hune was the one that bucked all the trends.

Of the seven errors, one was projected with 50% confidence and four with 67% confidence or less. The errors were not consistently on one side or another, with the PCs winning three seats the projection model awarded to the Liberals and the Liberals winning four seats awarded to either the NDP or the PCs.

With the actual results plugged into the model, the district-level accuracy did not change: 33 of 40 would have been called correctly, though the potential winner would have been identified in 35 of 40 ridings. But the errors would have been more consistent in that the model would have systematically over-estimated the PCs, getting more NDP seats right but more PC seats wrong. The projection with the actual results would have been 27 Liberals, 12 Tories, and one NDP. The old model would have done even worse, with 25 Liberals, 13 Tories, and two New Democrats.

How did the polls do? It might be better to ask how the poll did. Only Forum was in the field in the last five days of the campaign. In the table below I've included all of the pollsters that released data during the campaign, but I think it is worth considering the amount of time between the polls conducted by Abacus and CRA and voting day. This is particularly the case for CRA, which was in the field for more than two weeks.

Forum's election-eve poll was the closest, with an average error of 2.33 points per party.

The performance of the other pollsters appears to have been directly related to the gap between their final poll and election night. I'm not sure if there is much to read in this — though Forum did have a poll done on Nov. 24 that was much closer to their final estimation than it was to the Abacus and CRA polls released during the same week. But there just wasn't enough data to do much comparison between the different pollsters.

Nevertheless, Forum's last poll of the campaign was close to the mark, following on their successful final poll of the federal campaign.

Next up, Manitoba and Saskatchewan in April.

Tuesday, December 1, 2015

Liberals win in dramatic fashion in Newfoundland and Labrador

That the Newfoundland and Labrador Liberals won big in last night's provincial election came as no surprise. If anything, many expected their victory to be even bigger.

But despite those over-heated expectations, Newfoundlanders and Labradorians still delivered the Liberals and Progressive Conservatives their respective best and worst performances in half a century — and the most significant swing in a party's fortunes in the province's history.

You can read the rest of my first run-through of the Newfoundland and Labrador election results here. I'll be back with a more detailed post about how the projection and the polls did this week.

In the meantime, I thought I'd post the updated monthly provincial polling averages chart, which I had let lapse during the federal election campaign. You can click on the chart to the left to magnify it, and you can always find it in the right-hand column of this site.

If we consider the B.C. Liberals a conservative party, Liberal parties are leading in five provinces (Quebec and Atlantic Canada) and are in second place in one. Conservative parties, including the Saskatchewan and Wildrose parties, are leading in three provinces (from Saskatchewan to Ontario), and are in second place in five. The New Democrats are leading in two provinces (B.C. and Alberta), and in second place in three. The PQ is in second place in Quebec.

Monday, November 30, 2015

Final Newfoundland and Labrador projection: Liberal majority

The Liberals under Dwight Ball should be elected with a majority government in today's provincial election in Newfoundland and Labrador, with Paul Davis's Progressive Conservatives most likely to form the Official Opposition. The New Democrats under Earle McCurdy should be the third party in the House of Assembly.

The Liberals are projected to take between 53.5% and 60.8% of the vote, with 56.3% being the precise projection. At the maximum and minimum ranges (the most appropriate measure in a province with such small constituencies), the party is projected to win between 25 and 36 seats, more than enough to put it over the threshold of 20 seats required for a majority government.

The PCs are projected to take between 26.7% and 30.9% of the vote, or 28.1% more precisely. The Tories are projected to win between three and 13 seats.

The New Democrats are projected to win between 13.9% and 16.5% of the vote, awarding them between one and six seats.

In the broad strokes, the polls have been generally consistent in that they have pointed to a majority government for the Liberals, the PCs in second, and the NDP in third. But they have been less consistent in terms of the size of the Liberal lead over the Tories.

Early polling in the campaign was more consistent, giving the Liberals between 65% and 74% of the vote, against 17% to 21% for the Tories and 9% to 15% for the NDP.

But polling done in the later part of the campaign has been less consistent, with the Liberals ranging between 52% and 67%, the PCs between 22% and 31%, and the NDP between 10% and 19%.

Forum has been the biggest source of disagreement, after showing almost identical numbers to Abacus's first poll of the campaign. Forum's two polls done on Nov. 24 and Nov. 29, the two most recent polls of the campaign and the only two done entirely after the final leaders' debate on the CBC, have put the Liberals at either 52% or 54% and the PCs at either 29% or 31%, while the two last polls from Abacus and CRA have pegged the Liberals at either 64% or 67% and the PCs at 22%.

The Forum poll of Nov. 24 was not weighted regionally, with an over-weighting for responses on the Avalon Peninsula. But the effect was negligible, by my calculations expanding the Liberals' lead from 23 points to 25 points. The last poll was weighted regionally, however.

We'll see if Forum or Abacus/CRA is on the mark, though in a campaign where the Liberals are leading by such a wide margin and the PCs are so unpopular we might expect Forum's IVR polling to get somewhat more honest responses. The difference could be that shy Tory effect, or the idea that the Liberals need an opposition, one that has only taken hold in the last days of the campaign. Or it could be that Forum is wrong. We'll see soon enough.

It does make a big difference in the potential outcomes. The chart below shows what the model would be projecting if it was only taking into account each individual poll.

As you can see, Forum's polls still give the Liberals a majority but a much smaller one. The polls from Abacus and CRA would deliver a landslide, with the maximum ranges topping out at 38 seats with Abacus and 40 seats (all of them) with CRA.

It will be interesting to see what the results will be tonight. It is a difficult election to predict down to the seat level, with a limited number of polls having been done in the final week and the potential for local factors and candidates to be hugely important. And what impact will the seeming inevitability of the Liberal victory have on turnout? Lots of potential for some surprises — but just around the edges. Anything but a Liberal majority would be a shock.

Sunday, November 29, 2015

For the record

At an event put on by the MRIA last week, four pollsters went over what they saw happening in the last federal election campaign.

According to this report by iPolitics, Darrell Bricker of Ipsos Reid was quite critical of the work done by polling aggregators, such as myself and Barry Kay of the Laurier Institute for the Study of Public Opinion and Policy (who was partnered with Global News, Ipsos Reid's media partner). He said:

“I would argue, quite frankly, that these models are as likely to maximize error by putting polls together as they are to minimize it…There’s work to do here and there’s value associated with doing this. But it’s not happening well right now…Sorry, CBC.”

Bricker may very well have a point. Aggregation can certainly be done better, and I and others are always working towards improving it.

But for the record, the error in Ipsos Reid's final poll totaled 1.2 points per party, compared to 1.3 points per party for ThreeHundredEight.com and the CBC Poll Tracker. Hardly a case of maximizing error. And the aggregation out-performed Ipsos Reid at the regional level in British Columbia, the Prairies, Ontario, Quebec, and Atlantic Canada. Ipsos Reid only did better than ThreeHundredEight.com in Alberta.

Ipsos Reid did out-perform ThreeHundredEight.com in the 2011 federal election campaign. But since then, when changes were made to the aggregation model, ThreeHundredEight.com has out-performed Ipsos Reid in the provincial campaigns in Ontario in 2011 and 2014, in British Columbia in 2013, in Quebec in 2014, and in Alberta in 2015, as well as in the municipal election in Toronto in 2014.

More broadly, ThreeHundredEight.com's aggregation has out-performed the polls in most elections. This site has had a smaller average per party error than the average error of all the pollsters in the field for 15 consecutive federal, provincial, and municipal campaigns, and 16 of 18 campaigns overall.

Aggregation is a way to minimize error in the vast majority of cases.

In terms of seat projections, there is certainly more work to be done and I have already taken some steps to address some of the issues with the model. But I should also take this opportunity to point out that the model itself, divorced from how the polls do, has identified the potential winners in 90% of ridings, and its error in the more narrow likely ranges has been four seats or less — in total, all parties combined — in 12 of 14 elections where ranges were given.

Bricker is right to point out the amount of polling data that Nate Silver has to work with. Having that much data would certainly make things a lot easier. But our first-past-the-post system and multi-party democracy is also much more complicated. This was shown in 2010 when Silver had just as much trouble as everyone else in projecting the outcome of the election in the U.K. It happened again this year when FiveThirtyEight affiliated itself with a British outfit.

American elections, with their two parties and, at the presidential and Senate levels, very big jurisdictions, are much more predictable. In this past election, the Liberals jumped 20 points and the Conservatives dropped eight from the previous vote. The last time there was a change of government in the United States, the swing between the Republicans and the Democrats was just five points. Almost half of the seats on offer changed hands in the 2015 Canadian federal election. In the 2012 U.S. presidential election, only two states out of 50 changed hands.

There will always be elections that for one reason or another are very difficult to predict without making unreasonable assumptions or leaps, and the 2015 federal election campaign was one of them. Rather than see that as a reason to abandon everything, this unpredictability is something fascinating that tells us a lot about what happened. That's part of the process — experiments that go wrong can be just as informative as those that go right.

Thursday, November 26, 2015

Riding results with a preferential ballot

In my column for the CBC today, I looked at what impact a preferential ballot might have had on the 2015 federal election. Take a look at the full article.

The calculation was based on second choice polling data at the end of the campaign, which had been broken down to the region. So, with that information I was able to estimate which party a Conservative voter in British Columbia, for instance, would mark as his second choice. With that data, I was able to re-run the 2015 federal election in every riding.

Of course, this assumes that all else is equal, including how the campaigns would have unfolded. It also assumes that the regional-level second choice preferences apply in each riding, which of course would not always be the case. So a few assumptions have had to be made. But I think the exercise is nevertheless indicative of what impact a preferential ballot might have had, and what impact it would have in the future if the parties do not change their electoral strategies.

Since I ran this exercise down to the riding level and had the numbers on hand, I decided I might as well share the data with readers here. In the tables below, I've shown the results after going through the exercise. Blank results for a party mean it was dropped at an earlier ballot and its support redistributed to other parties. 

Thursday, November 19, 2015

How the cities went in the federal election

In many ways, Justin Trudeau's victory on Oct. 19 was an urban one, with the Liberals virtually sweeping the urban ridings and winning just enough rural ones to put them over the top for a majority government.

The Elections Canada results page, in addition to showing the breakdown of the results by province, also has the option to show the results in major centres. I'm not quite sure how those are defined (likely any riding within a CMA), and these results are the preliminary results, but they do give us an indication of how each of these major centres voted in the federal election.

UPDATE: The definitions for each major centre can be found here. As you can see, Elections Canada used a wide brush to define these.

In the chart below, I've included all of the major centres that Election Canada defined as having at least three ridings. None of this information is original — but considering the way that Elections Canada has the data organized (you can only see four at a time) I thought it would be useful to compile it all in one place.

The Liberals won big in each of the largest three cities, beating out the Conservatives by 14 points in and around Vancouver and 16 points in and around Toronto, while also edging out the New Democrats by 15 points in and around Montreal. 

But the Liberals also did very well in other big centres, topping 40% in Winnipeg, Hamilton, Kingston, Kitchener-Waterloo, Ottawa, Gatineau, Halifax, Moncton, and St. John's. They also managed to eke out victories in Sherbrooke and Trois-Rivières, the last two some real signals of how the Liberals broke through in francophone Quebec.

Outside of Atlantic Canada, the Ottawa-Gatineau region was the major centre that voted in the largest numbers for the Liberals.

The party also did quite well in the southern Ontario cities like London, Oshawa, and St. Catharines-Niagara, centres they lost to the Conservatives by narrow margins. 

Surprisingly, the Liberals placed second in Quebec City and Regina. Only in Victoria, Saskatoon, and Windsor did the Liberals finish in third place. And it is something to see the party sitting at just under 30% in Calgary.

The Conservatives won the cities in Alberta and Saskatchewan by very wide margins, and managed to squeeze out the Liberals in London, Oshawa, and St. Catharines-Niagara. The party did very well in Quebec City, winning it by 15 points.

The Tories also had strong second-place showings in Winnipeg, Kingston, and Kitchener-Waterloo. Its numbers in Montreal and Gatineau were unsurprisingly low, but the party also put up poor numbers in Atlantic Canada and in both Vancouver and Victoria, though their Vancouver numbers were enough to place them second.

And that is because the New Democrats were pushed out of the urban centres in big ways. The only city they won was Windsor, with just under 39 per cent, and in only St. John's and Victoria (which the NDP narrowly lost to the Greens) did the party manage more than 30% support. 

In addition to those two cities, the NDP also finished second in Saskatoon, Gatineau, Montreal, Sherbrooke, Trois-Rivières, and Halifax. 

The Bloc had poor numbers in the five major centres in Quebec, putting up their best results in Trois-Rivières and Montreal, where they took over 20% support.

The Greens won Victoria by the skin of their teeth (and thanks to Elizabeth May), but nowhere else did better than 4.5% support.

Tuesday, November 17, 2015

District-level polling in Newfoundland and Labrador, and some news

The province-wide race may no longer be very interesting, but that doesn't mean there aren't a lot of fascinating races at the district level in Newfoundland and Labrador. A new poll by Abacus Data for VOCM in the district of St. John's West, for example, shows that former MP Siobhan Coady is in good form, while NDP leader Earle McCurdy is in danger of losing.

There will also be a number of polls out from Abacus and VOCM in the coming week for Newfoundland and Labrador, so keep an eye out for those.

You can read my analysis of this latest poll on the CBC's website here.

And that leads me to the next bit of news. Though the federal election is now over, my work for the CBC is continuing. In addition to the kinds of analyses I have written for the CBC (and, in the past, other media outlets) I will also be writing some articles for the CBC that would have normally found a home here on ThreeHundredEight.com. This article about a district poll in Newfoundland and Labrador is a good example.

Work here on ThreeHundredEight.com will continue, but the pace will be reduced. It will still be a place to host things like the Newfoundland and Labrador projection and other tracking data that would not have a place with the CBC. But since most of my original analyses will be on the CBC's website, I have added a list of links to my latest articles at the top of the right-hand column. This list will be constantly updated.

You can also bookmark my author page at the CBC to keep abreast of all my articles.

I'm also happy to announce that the Pollcast podcast will continue! We'll have the latest episode up hopefully this week, with regular weekly episodes from then on. The focus will still be on polls and polling, but it will cover some other topics. Think of it as the Canadian political geek's podcast. It should be fun!

Thursday, November 12, 2015

First post-election poll is not bad for the Liberals

The first post-election poll is out, and it is a doozy.

Does a poll conducted four years before the next federal election mean a lot? No. Of course, it says nothing about the next election and, really, has no impact on where things stand today. But it is an indication of how Canadians are reacting to the new Liberal government. And before you say "It's only Forum", remember that Forum nailed the election results almost exactly.

Oh, and it marks the first salvo in what will undoubtedly be a barrage of Conservative leadership polling that we will see between now and the date that the next leader is actually chosen.

The poll gives the Liberals the support of 55% of Canadians, an enormous number that the Conservatives never managed in any poll throughout their tenure. My records only go back so far and are incomplete the earlier they go, but even back in the days of 2002 and 2003, when the Liberals faced a divided opposition and the coming Paul Martin juggernaut was poised to deliver Liberal rule for the rest of time, the party was only polling at around the 50% mark.

The Conservatives have dropped down by seven points to 25%, a number they were flirting with in the dark days of the Mike Duffy scandal. The New Democrats have also dropped by a similar amount, down to just 12%, a score that brings them back to the earliest days of Jack Layton.

The numbers in this poll are just astounding: 61% for the Liberals in British Columbia, 56% in Ontario, and 58% in Quebec (the Conservatives, remarkably, are holding firm in Alberta and the Prairies). The NDP takes the brunt of the hit in most parts of the country, down to just 11% in Quebec. All told, these numbers would likely deliver around 245 to 280 seats to the Liberals, with the Conservatives taking 55 to 90 and the NDP less than 10.

This poll does not look like a normal honeymoon poll, or at least not like one the Conservatives ever saw after their election victory. The two post-election polls in 2011 taken at about the same time after the vote as this Forum poll (by Abacus Data and Harris-Decima) gave the Conservatives between 38% and 40% support, identical to the 39.6% they managed on election night. Instead, the NDP was up a little to 33%, dropping the Liberals to 15% to 16% support. There was no honeymoon — just a confirmation of where people stood a few weeks earlier, with maybe a little NDP uptick at the expense of the Liberals.

Polling averages, 2009-2015
The significance of this new poll should not be exaggerated, as its significance is minimal. But there are some worrying signs in these numbers for the New Democrats. Thomas Mulcair's approval rating is now a net -5, after being a robust +17 even in the last week of the campaign. Justin Trudeau has gone from a +9 to a +40, with Canadians approving of him by a margin of 3 to 1.

More problematic for the NDP is that a lot of their supporters seem perfectly fine with the Liberal victory. Fully 53% of current NDP supporters approve of Trudeau, compared to 62% who approve of Mulcair. And 18% of NDP voters say they are very satisfied with the election's outcome. That increases to 72% when we include those who say they are somewhat satisfied.

Of course, much of this satisfaction could merely be with the defeat of Stephen Harper. And Trudeau is at a very high risk of losing much of this new support if he doesn't deliver entirely on his progressive rhetoric. Nevertheless, the NDP has some work to do to convince its traditional supporters that while Trudeau might sound good to them, Mulcair and the NDP remains the real deal.

MacKay leads the name-recognition primary

Perhaps more significant (and we're still talking low significance here) were Forum's numbers on the Conservative leadership race. As he has been in Conservative leadership polling for the entirety of Stephen Harper's time in the job, Peter MacKay was at the top of the list.

Among all Canadians, MacKay was the choice for Conservative leader of 29%, more than double the next most popular candidates.

These were John Baird and Rona Ambrose at 14% apiece, despite the fact that Baird has ruled himself out and Ambrose has been officially ruled out due to her becoming the interim leader.

Next, at 11% each, were two Alberta MPs: Jason Kenney and Michelle Rempel. Scoring below 10% was Kellie Leitch (9%), Tony Clement (7%), and Rob Nicholson (6%).

Among Conservative supporters, MacKay was still well ahead of the pack at 32%, followed by Baird at 18%, Kenney at 16%, and Ambrose at 12%.

Do these numbers mean much? Not really — as the former leader of the Progressive Conservatives and one of the Tories' most high-profile cabinet ministers he should be expected to have the most name recognition. In fact, it is somewhat impressive that Rempel and Leitch, who were both only elected for the first time in 2011, scored as well as they did, beating out such fixtures of Conservative politics over the last decade or more like Clement and Nicholson.

There was some regionalism in the numbers, with the former Nova Scotia MP MacKay scoring best in Atlantic Canada and Kenney and Rempel doing significantly better in Alberta than they did nationwide.

And in terms of gains or losses, if any can be determined, it does seem like Ambrose's rise to the interim leadership of the Conservative Party has helped her. The last Forum poll asking the question in May 2014, if we exclude those who were undecided in that survey (as that was not an option in this one), had Ambrose at 8% among all Canadians, compared to 14% today. MacKay dropped from 38% to 29%, while Kenney, Baird, and Clement (the only other names to be in both of these polls) had no movement of any significance. So, possibly MacKay has lost a little lustre in the eyes of Canadians, or there is just too much apples-to-oranges comparisons here to make anything of it. I'll lean towards the latter.

Tuesday, November 10, 2015

Warning signals for all three big Quebec parties in four by-elections

As expected, the Quebec Liberals held their three ridings and the Parti Québécois its one in the four by-elections that occurred across the province yesterday. But the results were mixed for both parties, with the Liberals putting up big drops in support in two of their three wins while making a big gain in their one loss, while the Parti Québécois had its own drop in support where it was the incumbent and respectable increases in two of the ridings in which it came up short.

Moral and actual victories for both parties, then, if they want to look for them. But nothing for the Coalition Avenir Québec, except a new logo.

Turnout was poor in the two Montreal-area ridings, at just 24% in Saint-Henri–Sainte-Anne and 23% in Fabre. It was more within the norm for by-elections in René-Lévesque on the Côte-Nord (40%) and Beauce-Sud (43%).

The Liberals increased their vote share in Beauce-Sud, and widened their margin of victory over the CAQ by almost 14 points. At 55.9% support, Paul Busque's score was the best performance for the Liberals in this riding in 18 years. Tom Redmond captured 29.9% of the vote, the lowest number for the CAQ or its predecessor ADQ since 1998. Of note is the performance of Quebec's Conservative Party, which at 3% of the vote finished in fourth place ahead of Québec Solidaire.

Monique Sauvé of the Liberals took 44% of the vote in Fabre, a significant drop since 2014 but better than the party's performance in the riding in 2012. The PQ's Jibril Akaaboune Le-François captured 28.6% of the vote, a big increase from 2014 and a little better than 2012, but well below the 36.8% score the party managed in Fabre in 2008.

In the riding of René-Lévesque, the PQ's Martin Ouellet took 49% of the vote, the worst showing for the PQ in the riding since 2003. The Liberals' Karine Otis surprised with 39% of the vote, the biggest increase any party experienced in any of these four ridings. The last time the PLQ did that well was in 1989. Otis was also the only candidate to have an increase in raw vote totals, as she picked up 848 votes over the party's performance here in 2014. Either Liberal voters turned out, or Otis took a lot of the vote away from the CAQ, which experienced its only double-digit drop in vote share here.

Dominique Anglade managed to hold on to Saint-Henri–Sainte-Anne in what turned out to be the closest race of the night. At 38.6%, Anglade nearly matched the Liberals' performance here in 2012 — the last time the Liberals lost an election. The PQ's Gabrielle Lemieux jumped in support to 29.9%, though that was still below the party's performance here three years ago. The big surprise was Marie-Eve Rancourt of Québec Solidaire, who captured 20.7% of the vote, almost double the party's vote share in the last election. Nevertheless, her raw vote count was still down from 2014. The PQ will undoubtedly make some noise about the split of the vote in this riding, as it is the one in which the combined totals of the PQ and QS (two sovereigntist parties) would have been enough to defeat the Liberal candidate.

In my last analysis of these by-elections, I pointed out how these four ridings have, on average, tracked the province wide vote totals quite closely. If these four ridings are still bellwethers when combined, the results do not suggest that Philippe Couillard's Liberals are in much trouble. The party averaged 44.4% across these four ridings, compared to 28.6% for the PQ. That still means a Liberal majority government if those sorts of numbers were repeated on election night. The biggest change, then, would be in the CAQ's tremendous drop in support to just 13.8%.

The CAQ had the worst night, averaging a loss of 6.9 points across the four ridings. By-elections are often difficult for the CAQ, particularly in ridings where the party is not a factor. The ADQ also used to have this issue, so an argument could be made to shrug off their performances in three of these four ridings. But for the party to drop over eight points in Beauce-Sud, where the CAQ was the only other party with a chance to win the riding, is very problematic. And whereas the CAQ has been dropping support to the PQ at the provincial level, in this case it seems to have been primarily to the benefit of the Liberals. If the CAQ is not competitive in a riding like Beauce-Sud, they have few prospects for gains.

The drop in support for the Liberals in the two Montreal-area ridings was significant, and is perhaps something that could be a sign of a deeper problem for the Quebec Liberals in urban ridings. But in both cases the turnout was anemic, so it could have merely been the case that voters did not bother turning out to vote in by-elections that everyone considered a foregone conclusion.

The PQ's increase in both Fabre and Saint-Henri–Sainte-Anne is certainly nothing to sneeze at, but again the low turnout lessens the impact of the party's performance. More concerning for Pierre-Karl Péladeau should be the performance of the PQ in René-Lévesque, a riding the PQ routinely won with over 50% of the vote in bad elections. This is just the kind of riding that should be embracing the PQ if the party is heading in the right direction. Instead, the Liberals put up some big numbers in a riding that has traditionally not been friendly to them. There are some echoes of the federal Liberals' performance in Quebec in this: they took 29% of the vote and finished second in the riding of Manicouagan, of which René-Lévesque forms a part at the provincial level.

This is why the results are a mixed bag for both parties. The Liberals can be happy to see that their vote share, overall, hardly budged from their big victory last year. And the strong performance in the two rural ridings is a sign that the party is doing well among francophones. But their losses in the Montreal area, where perhaps the politics of austerity resonate more, show some underlying weakness in a traditionally safe area.

For the PQ, modest increases in the Montreal area is a positive sign for the party, but losses in a stronghold region of their own, particularly under a new leader, also suggests some underlying weaknesses.

Though François Legault certainly has reason to be concerned, both Couillard and Péladeau can breathe a sigh of relief with these results. They held their ridings and can each point to some strong second-place showings. But Fabre and Saint-Henri–Sainte-Anne indicate that Couillard can't get too comfortable, while René-Lévesque shows that Péladeau is far from following in the footsteps of that riding's namesake.

Friday, November 6, 2015

Liberals start Newfoundland and Labrador campaign with huge lead

The provincial election in Newfoundland and Labrador was kicked off yesterday and will come to a close on November 30, but it is not setting up to be as exciting as the federal campaign that just came to an exciting finish. Instead, the campaign in Newfoundland and Labrador starts with the Liberal Party under Dwight Ball enjoying a 47-point lead over his nearest rival, the incumbent Progressive Conservative government of Paul Davis.

This is from the latest survey from Abacus Data. And with the start of the campaign comes the launch of the projection model for the Newfoundland and Labrador election, the first to use the three-election swing model I outlined earlier.

This campaign poses a few problems for the projection model. Before taking a look at the Abacus poll, let's go through them.

The first is the redistribution of the province's 48 districts into just 40. That represents a great deal of change and the combination of old districts into these smaller new ones. Because of the enormity of the changes, it makes it a tricky election to call with this three-election swing model.

Luckily, Kyle Hutton provided me with the transposition of the 2011 results onto the new boundaries. And the spreadsheet he designed helped me calculate the transposition for the 2003 and 2007 elections as well. So thanks to Kyle for that. You can follow him on Twitter here.

The second and third difficulties are compounded by the first: there were a large number of by-elections in the last legislature, and they featured some incredibly massive shifts in support. There were also a number of floor-crossings, which is doubly complicated in Newfoundland and Labrador due to the importance of the local candidate.

Now that these by-election results and floor-crossing incumbent MHAs have been split among multiple ridings, it makes it difficult to know what impact they will have within the new district boundaries. I have tried to estimate what the results would have been throughout the new ridings when a by-election took place within its boundaries, and have applied the floor-crossing factor as I would normally.

Then there is the matter of Newfoundland and Labrador's small population. It makes it much more difficult to project outcomes when a riding has a smaller population that can more easily swing in one direction or another, particularly when we're seeing such a massive shift in support as we are in this election.

Taken together, this leads me to encourage readers to exercise a great deal of caution with the seat projections, and to rely more than ever on the seat ranges. And as there are a relatively small number of ridings considered at play in the likely ranges, I'd also suggest concentrating on the 95% confidence interval — so the maximum and minimum ranges.

My urging of caution is doubly so for the riding-by-riding projections, which are primarily there to show how I come to my province-wide totals.

So where does that put us at the start of this campaign? It suggests the Liberals could win between 29 and 39 seats, putting them well over the 20-seat mark required for a majority and putting a near-sweep within their grasp (even the likely ranges top out at 38 seats).

The New Democrats are projected to win five seats, but that is up against their very maximum: instead, their range puts them between one and five seats, within the Tories' band of between zero and eight seats. As the incumbent government, however, the PCs have a better chance of winding up in the average to maximum range (one to eight seats) than does the NDP. So despite the average projection, I'd still call the PCs the favourites to form the Official Opposition at this stage.

The Abacus poll gives the Liberals 66% support, almost exactly where the federal Liberals ended up on October 19 and higher than any other result the NL Liberals have recently managed (their previous high was 60% to 62% a year ago).

At 19%, the PCs are at their lowest ebb, while the 15% for the NDP puts them back to where they were for much of 2014.

There is no good news at the regional level for any party but the Liberals, who scored their lowest support levels on the Avalon Peninsula and still managed 59%. The PCs had their best result, 26%, in central Newfoundland, where they are the incumbents in every riding, and 21% on the Avalon Peninsula. The New Democrats had their best result on the Avalon Peninsula and in St. John's, where all of their seats are located. If their vote is concentrated enough, they can win a few seats with these numbers.

Across the board, the numbers in this Abacus poll were stellar for the Liberals. Fully 73% of Newfoundlanders and Labradorians would consider voting for them, compared to just 37% for the NDP and 32% for the Tories. A majority of voters would not consider voting for either the PCs or the NDP. And three-quarters of people in the province think the Liberals will win, including half of PC voters and two-thirds of NDP voters.

While Dwight Ball has very good personal numbers, he isn't beating Paul Davis because the premier is unpopular, who has a positive rating of 32%, with just 25% having a negative impression of him. Earle McCurdy also has a net positive score, with 26% having a positive impression of him against 22% who have a negative one.

But Ball is head and shoulders above both of these, with 50% of Newfoundlanders and Labradorians holding a positive impression of the Liberal leader. Just 10% hold a negative view.

The trump card for Ball, though, is that 60% of voters think it is "definitely" time for a change of government, while another 25% think a change would be good but is not necessarily important. Only 9% think the PCs should be re-elected.

So it will take some doing for the Liberals to lose this one. For now, the thing to watch is whether either the PCs or NDP can do something to avoid the election turning into a complete landslide. The federal Liberals went 7-for-7 on election night in Newfoundland and Labrador. Something very close to 40-for-40, with these sorts of numbers, is likely for the provincial Liberals on November 30.

Thursday, November 5, 2015

Four Quebec by-elections a mid-term test for Couillard and Péladeau

Quebecers in four ridings will be heading to the polls yet again on Monday, as four by-elections are being held in the province to fill vacancies in the National Assembly. And based on the history of these four ridings, the results may provide a telling glimpse of where voters stand in Quebec about 19 months into Philippe Couillard's mandate.

Two of the by-elections are being held in the Greater Montreal region, while the other two are being held in the eastern part of the province. They each have something for the four major parties in Quebec.

In Beauce-Sud, a riding hugging the border with the United States and anchored by the town of St-Georges, the Liberals are hoping Paul Busque can hold the seat after the departure of Robert Dutil, a cabinet minister during Jean Charest's government. They will face their biggest challenge from Tom Redmond, a municipal councilor for St-Georges and the Coalition Avenir Québec's candidate.

At the west end of Laval lies the riding of Fabre, vacated by the Liberals' Gilles Ouimet. Monique Sauvé will be looking to hold it for the party, which should not prove difficult.

At the other end of the province in the Côte-Nord, including the towns of Forestville and Baie-Comeau, is the riding of René-Léveseque. The Parti Québécois's Marjolain Dufour gave up the seat, and the PQ has put up Martin Ouellet to retain it. Though it is unlikely the PQ would lose this riding, the biggest challenge will likely come from Baie-Comeau municipal councilor Karine Otis of the Liberals.

The last riding, the working class Montreal riding of Saint-Henri–Sainte-Anne, was left vacant by the departure of former cabinet minister Marguerite Blais of the Liberals. The only real 'star' candidate of these by-elections is Dominique Anglade, former president of the CAQ, who the Liberals are hoping can keep the riding for their party. The PQ's Gabrielle Lemieux will likely put up the biggest fight, though Québec Solidaire is gunning for a strong performance with their candidate Marie-Eve Rancourt, who ran for the party here in 2008.

These four ridings, individually, are not particularly interesting. The smallest margin of victory in these four in 2014 was 12.3 points in Beauce-Sud, while the other three were won by margins of 30 points or more. Only in Beauce-Sud is an upset a serious possibility, as the margin has averaged just 5.4 points in the riding over the last three elections. The average margin has been 15 points or more in the other ridings.

Together, however, these ridings are quite interesting. That is because they have actually managed to be ridings that, combined, matched the province-wide outcomes quite closely.

As you can see in the table above, the average result in these four ridings has never differed very greatly from the overall results in the province. They have been very close for the Parti Québécois in particular, which means these by-elections may serve as a good indication of how Pierre-Karl Péladeau is doing as the PQ's new leader.

The average result across these four ridings may be the most interesting number to look at on Monday night. But let's return to each of the four races.

Though attention has primarily been on the race in Saint-Henri–Sainte-Anne, the by-election in Beauce-Sud could prove the closest. It was a very tight race between the Liberals and the CAQ (represented here by the ADQ's results in 2008) in 2008 and 2012, and was actually won by the ADQ in 2007. The CAQ's vote here has slipped a little over the last few elections, but it is nevertheless robust. 

The Liberals have dipped in the polls since the 2014 election, which may open up an opportunity for the CAQ here. But the CAQ has also dropped in support, primarily due to the arrival of Péladeau on the scene. That still gives the Liberals the best odds of holding on, but the CAQ's hopes are lying in Beauce-Sud.

There is little doubt that Fabre is a riding the Liberals should have no trouble retaining. They have easily won the riding over the last three elections, and saw their vote increase significantly in Fabre in 2014. The PQ's share has dropped consistently here over the last few campaigns, while the CAQ's strong showing in 2012 was due in part to the candidacy of Anglade, who is no longer available to the party. The ADQ did come relatively close here in 2007, but even that was when the Liberals were in a much poorer state than they are currently.

Polls suggest the Liberals are doing very well in the Montreal region, so this riding should hold firm for them. The question may be to see whether the PQ's Jibril Akaaboune Le-François can make some inroads in order to demonstrate that Péladeau's PQ can woo voters around the island of Montreal.

If Fabre is safe territory for the Liberals, then René-Lévesque is a stronghold for the PQ. With the brief exception of an ADQ by-election victory in 2002, the PQ has held this riding without interruption since 1994. In the 1995 sovereignty referendum, almost three out of every four voters in the riding opted for independence.

Support for the PQ has been over 50% in each of the last four elections, including the 2007 and 2014 campaigns when the party had the worst provincial performances in its history. Even in 2014, Dufour won the riding by just over 33 points. Considering that the Liberals are not in a better position in the polls today than they were a year ago, it is extremely unlikely that they can overcome that gap. In fact, it will be a disappointment for the PQ and Péladeau if they do not improve upon the 55% that Dufour managed in 2014.

Since its creation in 1994, the riding of Saint-Henri–Sainte-Anne has only ever voted for the Quebec Liberals. Any change from that would be a tremendous upset.

The Liberals won this riding by just over 30 points in 2014, and still managed to win it by just over six points in 2012. Even under the best of circumstances, it is a stretch to imagine the PQ being able to wrest this riding away from the Liberals — particularly with Anglade on the ballot. 

But the riding does serve as a bit of a test for the PQ and QS. For the results to be good news for the Parti Québécois, they would like to see their numbers be back over 30%. More interesting might be the numbers for Québec Solidaire, which has managed double-digits in this riding over the last two elections. The party is also doing quite well in the polls, registering as high as 13% province-wide and nearly 20% in the Montreal region. A strong sign that QS is heading in the right direction would see its results top 20% in Saint-Henri–Sainte-Anne, though that might be a high bar to meet.

The safe money in these four by-elections would be on the four incumbent parties holding on. The CAQ could potentially pull off a victory in Beauce-Sud, while if things go well for the PQ they could get the Liberals nervous in Fabre and Saint-Henri–Sainte-Anne. On the whole, however, the defeat of any incumbent party would be big news.

So instead the interest in these four by-elections will be on the overall performance of the parties, and whether they can out-perform expectations. They will primarily serve as tests to Couillard and Péladeau. Is the Liberal government in trouble or on the right track? Has Péladeau's leadership of the PQ led to any real gains for the party? Real ballots, and not just polls, will give us a clue on Monday night.

Monday, November 2, 2015

A different approach to seat projections

One lesson learned from the 2015 federal election is in the difficulty in projecting outcomes when the voting base of a party has dramatically shifted from one election to the next. In this case, it was the Liberal Party that attracted new and lapsed voters to the polls. Many of these voters appeared in unexpected numbers and in unexpected places, which is why virtually all seat projectors estimated that the Liberals were on track to win only a minority government, with a majority government being possible but unlikely.

Turnout in most elections does not dramatically shift from one campaign to the next, making it much easier to project outcomes at the seat level based on regional or provincial changes in support. It is no accident that the model has performed best when changes in turnout were lowest, while it has struggled more in elections with big changes in turnout.

But this is not all about new voters. More people voted in 2006 than in 2008 or 2011, meaning that a lot of people who voted in the 2006 election stayed home in the next two elections as the Liberals plummeted in public support. The Liberals did attract a lot of first-time voters in this campaign, but they were also successful in bringing back a lot of voters that just stayed home in 2008 and 2011.

Though there were a few surprises, the vast majority of the ridings the Liberals won in this campaign were ridings they had held in the past. Many of their pick-ups in Quebec, for instance, were won by the Liberals in 2000.

Taken together, this suggests that looking further back into a riding's history than just the most recent election can tell us a lot about how likely a riding is to go one way or another in the next election. If the basic principles of the swing model can work when looking at just one election, then they should be able to work when applied to elections further back in time, swinging the results from two or three elections ago according to where the polls are today.

This is what I tested with the 2015 results, swinging the results in each riding over the last three elections. I tried various weightings (looking at just the last two elections, giving all three elections equal weighting, trying proportions of 4/2/1, etc.), but the one that worked best was also the one that was most intuitive. It weighted the three elections with a proportion of 3/2/1, or 50% for 2011, 33.3% for 2008, and 16.7% for 2006.

As shown below, the results were very positive.

Using the actual provincial and regional vote shares (which is the real test of the projection model), the old model would have delivered the Liberals 154 seats, the Conservatives 120 seats, and the NDP 56 seats. The actual results for the Liberals and Conservatives would have fallen outside of the likely ranges, and into the maximum and minimum ranges.

But with the new model, the Liberals would have been projected to take 166 seats, the Conservatives 116, and the New Democrats 45. It would have been almost right on the mark for the NDP, and exactly on the mark for the Bloc Québécois. Most importantly, the Conservatives and Liberal results would have fallen within the likely ranges — the Liberal result falling well within that mark.

The regional projections would have been improved significantly, particularly in Quebec. The actual result there was 40 seats for the Liberals, 16 for the NDP, 12 for the Conservatives, and 10 for the Bloc. With the new model, it would have given 38 seats to the Liberals in Quebec, with 18 going to the NDP and 12 to the Tories. Only in Atlantic Canada would the results have not fallen within the maximum and minimum ranges, with the Liberals topping out at 30 seats and the NDP bottoming out at two. Peter Stoffer and Jack Harris would have still been projected to survive.

Even considering where the vote projection model stood on the eve of the election, this new seat projection model would have given the Liberals a range of between 128 and 177 seats — meaning their final tally of 184 seats would have fallen outside the likely range, but the potential for a Liberal majority victory would have been a big part of the final analysis (rather than the marginal part it was actually given).

Accuracy would have been higher at the riding level, increasing from 81.4% to 82.5%. But most significantly, the potential winners would have been identified in the high-low ranges in 91.4% of cases, meaning in only 29 ridings would the potential winner have been missed, compared to 43 with the old model.

This new model requires a few other changes, including some judgement calls in terms of how to handle special cases. This involves ignoring the results of some past elections, such as the 2006 and 2008 elections in Saanich–Gulf Islands, where Elizabeth May did not run, or cases where an independent took an out-sized portion of the vote (as in 2006 in Portneuf–Jacques-Cartier, when the Conservatives ran a candidate against independent André Arthur, or in 2008 when Bill Casey ran as an independent in what is now Cumberland–Colchester).

In addition, the incumbency effect built into the model is now a redundancy, as the results of the previous elections already take into account an incumbent's ability to withstand wider trends. Losing an incumbent, however, is still applicable.

Note that I do intend to run this new model through more tests as time permits to ensure that its improved performance isn't due to a fluke related just to the 2015 federal election. But time is an issue, because the next provincial election is just weeks away.

The next test: Newfoundland and Labrador

This new model may never be tested for real at the federal level, as the next election may not be decided according to the first-past-the-post system. Elements of it could be part of a model designed to project the outcome of a ranked ballot election, however.

The real tests will come in the provincial elections that will still use the first-past-the-post electoral system. And the next one is coming very soon: Newfoundlanders and Labradorians are heading to the polls on November 30.

How would the new model have performed had it been used back in 2011? Again, it would have done a better job.

The actual results of that election delivered 37 seats to the Progressive Conservatives, six seats to the Liberals, and five to the New Democrats. This despite the NDP finishing ahead of the Liberals in the popular vote by over five percentage points.

With the actual results plugged into the 2011 model, it awarded 41 seats to the PCs, four to the Liberals, and three to the NDP. Not a bad showing, considering it still would have given the Official Opposition nod to the Liberals. But the riding-level accuracy of just 75%, or 36 out of 48 ridings, left a lot to be desired.

The new model would not have improved upon the 2011 performance, at least in the top-line numbers: it still would have been 41 Tories, four Liberals, and three New Democrats. But the riding-level accuracy would have increased to 81.3%, with correct calls in 39 out of 48 ridings.

The election in Newfoundland and Labrador will pose a few problems, in that the number of seats has been reduced from 48 to 40. There have also been a large number of floor-crossings and by-elections since 2011, further complicating matters. These are the sorts of things that can throw any seat projection model for a loop.

But I will put the new model to the test nevertheless and see how it does. The same principles behind what has been, I believe, a very effective model (which has been used in 17 provincial and federal elections) are still in place, so I consider this a refinement rather than a wholesale change. We'll see how it does soon.

Tuesday, October 27, 2015

How the riding polls and riding projections did

One last go around on the 2015 federal election polls and projections before we can put this long campaign to rest.

At the riding level, the projection model was certainly not as close as I would have liked it to have been. Overall, the model called 269 ridings correctly and identified the potential winner (as defined by the parties considered capable of winning the riding by the high and low projections) in 291 ridings. That adds up to an overall accuracy of 79.6% on the calls, and 86.1% for identifying the potential winners.

Where did the model do better? It identified the potential winners in 94% of ridings in Alberta, 91% of ridings in Ontario, and 89% in the Prairies. It performed worse in Atlantic Canada (84% of winners identified), British Columbia (83%), and Quebec (76%).

Not surprisingly, the biggest misses were in terms of the seats that the Liberals ended up winning. The largest group of misses were ridings in which the New Democrats were projected to win, only for the Liberals to pick them up. There were 24 of these ridings, located primarily in Quebec, urban and northern Ontario, and in Atlantic Canada. This is where Liberals unexpectedly defeated New Democrats.

The next largest group were the 19 ridings in which the Conservatives were favoured but the Liberals actually won. These were largely in the Greater Toronto Area, in Vancouver and the Lower Mainland, and in New Brunswick. This is where Liberals unexpectedly defeated Conservatives.

There were seven ridings projected to go NDP that actually went to the Bloc Québécois (mostly north of Montreal), and seven ridings projected to go Conservative that actually went NDP (in the Prairies and the B.C. Interior).

On average, the misses were called with just 65% confidence and the average margin of actual victory in these ridings was about 6.7 points. So, they were modestly close races.

One aspect of the seat projection model worked very well. The assigned probabilities of victory turned out about as expected, though at the lower levels of confidence they were somewhat more confident than they should have been. This is likely due to many instances of three-way races, when the model is designed for two-way contests.

As you can see, the calls were generally as correct as they were expected to be.

At the 50% to 64% level, where the calls performed significantly below expected levels of confidence, Liberal victories were missed in 58% of them, or in 22 ridings. That alone gives an indication of how the Liberals were winning close ridings they were not expected to win. Add those 22 ridings to the final projection of 146 for the Liberals, and you have them knocking on the door of a majority government, rather than apparently coming up well short.

With the actual results plugged into the model, the accuracy of the riding level projection increases to 81.4% (or 275 out of 338 ridings), and to 87.3% (295 ridings) for identifying the potential winners.

As discussed in my analyses of the projection model's performance, we're looking primarily at the Liberals picking up new voters in unexpected places, with strategic voting apparently helping the New Democrats out-perform expectations in Western Canada. The Liberals' vote efficiency in Ontario and elsewhere was also above expectations, though well in line with what the Conservatives were capable of with a similar amount of support in 2011.

Riding polls

Now that we've dissected my performance at the riding level, how about the pollsters?

In the charts below, I've included only the polls done within the last two weeks of the election campaign, and compared the riding-level polling only for the parties that finished in the top three slots on election night. The actual results are in the gray areas, and the date refers to the last day the poll was in the field. Let's start in B.C.

Across the board, you can see that in every riding the riding-level polling under-estimated where the Liberals ended up. In two cases, it turned a third-place showing into a win: Burnaby North–Seymour and Coquitlam–Port Coquitlam. Depending on the riding, either the New Democrats or Conservatives found themselves over-estimated as a result.

The results in Vancouver Granville were particularly interesting, as Mainstreet's final poll came very close, whereas Environics' poll for LeadNow did not. There had been a lot of controversy in the riding due to LeadNow's endorsement of the NDP, despite the edge given to the Liberals in their final poll.

A few riding polls came quite close to the mark, considering the margins of error. For example, in Courtenay–Alberni, Nanaimo–Ladysmith, and South Okanagan–West Kootenay. Not coincidentally, these were NDP-Conservative races in which strategic voting might have kept the Liberal surge at bay.

Now to Alberta.

Again, in Alberta we see the riding polls under-estimate the Liberals significantly in every riding. Unlike in B.C., however, the polls were quite good at gauging Conservative support. Instead, it seems that NDP support collapsed in the final days and went to the Liberals. This appears to be another indication of the role strategic voting played in these races. That especially appears to have been the case in Calgary Centre and Edmonton Centre.

Ontario was slightly different.

In Ontario, the Liberals were under-estimated in most riding polls, but not all of them. In the ridings of Brampton North, Bruce–Grey–Owen Sound, Hamilton West–Ancaster–Dundas, Kanata–Carleton, Kitchener Centre, Perth–Wellington, and Peterborough–Kawartha, the results for the Liberals were within the margin of error of the final riding polls.

This should not come as a surprise. Unlike Alberta, in which the NDP did worse than expected in the popular vote, the NDP's support had largely already collapsed in Ontario well before election day. There was no surge that sunk the NDP's chances at the last moment in Ontario (as there might have been in Quebec). The Liberals were already riding high in the province in the week before the vote.

And in the ones the polls did miss, it wasn't always the same party that took the hit at the expense of the Liberals. In Timmins–James Bay and Nickel Belt it was the NDP, but in Flamborough–Glanbrook, Nepean, Kenora, and Sault Ste. Marie it was the Conservatives.

There were fewer riding polls done in Quebec in the final days, but they did moderately well.

But here again we're looking at the Liberals being under-estimated, and significantly so in Chicoutimi–Le Fjord, one of the most surprising Liberal wins of the night. That vote came primarily from the Conservatives, but also the NDP. In Jonquière and Lac-Saint-Jean, the polls did quite well.

There were only a few polls done in Atlantic Canada as well, but they were generally poor. The Liberals out-performed these polls by nine to 15 points, with the NDP taking the hit where they were most competitive and the Tories taking the hit where they were competitive.

More lessons to be drawn from the discrepancies, then.

Who did best? In terms of average error per party (only the top three) of those riding-level pollsters in the field in the last two weeks, Segma Recherche did the best with an average error of 3.4 points per party. Next was Environics at 4.6 points per party, followed by Insights West and Mainstreet at 5.5 points per party each. MQO Research had an average error of 7.0 points per party, while Oraclepoll had an average error of 7.8 points per party. Of note is the performance of ThinkHQ in Edmonton Centre, off by just 1.7 points per party among the top three.