Monday, April 29, 2013

B.C. race tightens, but NDP still in control

Though the margin between the B.C. Liberals and B.C. New Democrats has tightened since the last projection update of Apr. 19 (based on polling data running up to Apr. 14), the NDP remains the odds-on favourite to win the election of May 14.

The New Democrats are now projected to take 47.6% of the vote, down 0.3 points from Apr. 19. The Liberals gained 2.8 points to hit 32.4%, the highest they have been since mid-January, while the B.C. Conservatives dropped 3.3 points to 8.5%. The B.C. Greens fell 0.7 points to 8.2% in the projection, while support for independents and other parties increased to 3.4%.

The NDP is now projected to win 59 seats, down six, while the Liberals are pegged to take 25. The ranges have widened considerably, to between 45 and 75 seats for the New Democrats and between eight and 39 seats for the Liberals. The Greens are now projected to be capable of winning as many as three seats with current polling support.

There are some important reasons why the projection and forecast ranges have increased so dramatically, and support for the Conservatives and other parties has changed so much.

First, the methodological shift in the projection. As the official list of candidates was released over the weekend, the projection model has been updated to accurately reflect the field of candidates that British Columbians will be able to vote for on May 14. In the case of the independents and fringe parties, this allowed me to estimate how much support they will get in each riding and, therefore, in each region of the province. This is why support for independents and other parties has increased from 1.8% to 3.4%. 

More importantly, an adjustment has been added to the vote projection model to take into account the fact that the Conservatives have put up only 56 candidates. That means they are running in only 66% of ridings, making it extremely unlikely that they will be able to take as much of the vote as the polls are awarding them. Most respondents who intend to vote Conservative likely have no idea whether or not there will be a Conservative name on the ballot, and if they find out between now and the election their change in support will not be reflected. And, in the past, there has been a strong correlation between polls over-estimating a party's support and a less-than full slate. Accordingly, the Conservatives will now have the "no seat in the legislature" adjustment applied to their polling numbers, which means a reduction by a factor of 0.73.

Lastly, because of the updated candidates list, the projection has to be tweaked to ensure accurate distribution of support. For example, prior to this update Green support was spread over the 14 ridings in the Vancouver Island region. But since they are only running 11 candidates in the region, that same support needs to be distributed over 11 ridings instead. All else being equal, that increases their support in each riding where they have a candidate. The same has had to be done to the Conservatives.

As for the wider ranges, this is due to the campaign having started and the numbers of polls having dropped. Prior to an election campaign, the projection model reduces the weight of a poll with each passing week. During a campaign, that weight is reduced by the same amount each day. The projection has had only two polls added that were taken during the campaign (Angus-Reid, Apr. 24-25 and Justason Market Intelligence, Apr. 15-23) meaning that the pre-campaign polls now have an extremely low weight in the model. And as the projection dates a poll by its median date, that means the Justason poll is considered six full days older than the Angus-Reid poll. That reduces its weight considerably, and results in the Angus-Reid poll taking up roughly 93% of the aggregation right now. Due to the Angus-Reid being almost the only poll being considered by the model, the uncertainty of where the parties currently stand is very high.

This will undoubtedly change as more polls are released to capture the effect of tonight's debate. But keep that in mind when looking at the numbers in this latest update.

The Justason poll, released today via The Tyee (surveying 600 via telephone and internet), is interesting as it shows the NDP with a 22-point lead. That is unchanged from Justason's last poll from the end of January, suggesting that little has changed since then. Justason gives the NDP 49% to 27% for the Liberals, 12% for the Greens, and 11% for the Conservatives. Like other surveys, it shows the Greens doing very well on Vancouver Island.

The Angus-Reid poll was released on Friday via CTV and The Globe and Mail (surveying 812 via their online panel) but was conducted more recently than Justason's survey. The poll gives the NDP 45%, the Liberals 31%, the Conservatives 11%, and the Greens 10% support.

Compared to Angus-Reid's last survey taken just before the campaign began, that represents a gain of three points for the Liberals, a drop of one point for the Conservatives, and a decline of three for the Greens. The NDP held steady.

A sign that the Liberals are making up ground on the New Democrats? We will have to see what other surveys show, as the three-point increase in Liberal support is within the margin of error (or would be, if the sample was probabilistic). There is good reason to suspect a statistical wobble, as the Liberals have been hovering between 28% and 31% in Angus-Reid's polls going back to November 2012, and the Justason survey suggests no reason to believe there has been a big change in voting support. 

Also of note are two riding polls that were released by the Prince George Citizen last week for the ridings of Prince George-Valemount and Prince George-Mackenzie. Conducted by Oraclepoll on Apr. 17-20 and Apr. 15-18 (respectively) and surveying 300 people in each riding, the survey found the Liberals narrowly ahead. In Prince George-Valemount, incumbent Liberal Shirley Bond had 46% to 41% for the NDP's Sherry Ogasawara and 10% for the Conservative candidate. In Prince George-Mackenzie, rookie Liberal Mike Morris had 44% to 37% for the NDP's Bobby Deepak, 14% for the Conservative candidate, and 5% for the Greens.

In part due to these two polls, but also the better Liberal numbers in the Interior/North in both the Justason and Angus-Reid polls, the two Prince George ridings have gone over to the Liberals from the NDP in the projection. But the results of these surveys are similar to the ones that were done for the two Kamloops ridings, showing stronger Liberal support than expected. Perhaps this is a methodological bias on Oraclepoll's part, or perhaps this is a sign that the Liberals will not be easy to defeat in the Interior and North. It is something to keep an eye on. 

Thursday, April 25, 2013

Quebec Liberals move ahead under Couillard (and Trudeau)

Earlier this week, La Presse released the latest results of CROP's political polling in Quebec. The results show a striking uptick in Liberal support, both at the provincial and federal levels.
CROP was last publicly in the field in February, and since then the provincial Liberals picked up eight points to move into the lead with 38%. The Parti Québécois dropped five points to 25%, while the Coalition Avenir Québec also fell five points to 22% support.

Québec Solidaire was up two points to 11%, while Option Nationale had 3% support (this is the first CROP poll to list them specifically, they used to be included with the "other parties").

Clearly, the arrival of Philippe Couillard has boosted Liberal numbers. Though the results weren't put out publicly, CROP was already showing a small uptick in Liberal support in the days just before his leadership win. Between March 13-18 (Couillard was named leader on the 17th), the PLQ had increased from the 30% of February to 31%, while the PQ was down from 30% to 29% (the CAQ had dropped to 25%). Those changes were all within the margin of error (or would be, with a comparable probabilistic sample), but the trend was heading in the same direction.

Despite their 13-point lead provincewide, the PLQ was not leading in every region in this poll. They were in the lead on the island of Montreal with 46% (+12 from February), while they were narrowly ahead in the regions of Quebec outside Montreal and Quebec City with 39% support (+10).

The Parti Québécois, however, was narrowly in front in the 'couronne' of Montreal (Laval and the surrounding shores) with 34%, while the CAQ was ahead in Quebec City with 38%. Though Couillard does not seem to have boosted the Liberal numbers in the couronne, he did bump the PLQ up by 10 points in Quebec City to 36%. The PQ had to drop nine points to 19% to make that happen, and the PQ also paid the price in the rest of Quebec with an eight point drop.

Amazingly, considering where the Liberals were polling under Jean Charest, Couillard has put them in a tie with the PQ among francophones with 30% apiece. That is a nine point gain for the Liberals since February, and a five point drop for both the PQ and the CAQ (who fell to 25%). That opens up a lot of possibilities to the Liberals.

Couillard has become the most popular provincial political leader in the province, with 28% considering him the best person to be premier. That compares to only 16% for Pauline Marois and 14% for François Legault. In March, Marois and Legault were tied at 20% apiece, while Jean-Marc Fournier had 11% (he was interim leader at the time, and now heads up the PLQ in the National Assembly as Couillard does not hold a seat).
But he would have an easy enough time winning one if an election were held today, as with these CROP regional numbers the model gives the Liberals 69 seats and a majority government. The Parti Québécois takes 35 seats while the CAQ is reduced to 16. Québec Solidaire wins four seats, while Option Nationale takes one.

(Note: Jean-Martin Aussant is projected to win his riding of Nicolet-Bécancour in this projection. The problem is he won't be running there next time. He has decided to try his luck in a Montreal seat, but until that seat is known the model will continue to assume he is running in Nicolet-Bécancour.)

The Liberals manage to win seats throughout Quebec, taking less than half in and around Montreal. The PQ wins the bulk of their seats in the suburbs of the metropolis, while the CAQ keeps most of its seats in Quebec City, central Quebec, and the Laurentides/Lanaudière region.

But Couillard is not the only popular Liberal in Quebec - the poll shows that Justin Trudeau has made huge inroads in the province as well, primarily at the expense of the NDP.
CROP found the Liberals to have the support of 38% of Quebecers, up 19 points since March. The New Democrats dropped nine points to 30%, while the Bloc Québécois fell three points to 18% and the Conservatives dropped five points to 10%.

These are not easy numbers for the New Democrats to ignore. CROP was the first pollster to record the party's surge in the 2011 election in the province, and they have routinely given them the best numbers of any of the pollsters inside or outside of Quebec.

Between October and March, the Liberals averaged only 20% in Quebec in CROP's polling, while the NDP averaged 38%. In that same time period, all other polls done in the province gave the Liberals an average of 25% and the NDP only 32% support. For CROP to have the NDP trailing the Liberals by eight points, when in the past they gave them a positive spread more than twice the size of other firms, is something significant.

We'll see how long it lasts, however, as the regional numbers in this survey point to either a very unusual result or a complete change in how Quebecers view federal politics (of course, it would not be the first time that has happened).

The federal Liberals were ahead in every part of Quebec, with 44% on the island of Montreal (+22 since February), 34% in the couronne (+9), 32% in Quebec City (+16), and 37% in the rest of Quebec (+13). Seeing the Liberals ahead on the island is not too shocking and having them in front in the suburbs is believable, but beating the Tories in Quebec City and the Bloc (and the NDP) in the regions of Quebec? Trudeau will have had to completely transform the perception of the Liberal Party of Canada in the province for this to be true. But isn't to say that it isn't true.

The New Democrats had relatively uniform support throughout the province, which could be dangerous if it means losing a lot of seats by narrow margins (they did the opposite to the Bloc in the 2011 election). The Conservatives had horrific results everywhere but Quebec City, while the BQ was only competitive in the Montreal suburbs (where they currently hold zero seats).

Among francophones, the NDP barely edged out the Liberals with 33% to 32%, while the Bloc came in third with 22% (du jamais vu...). That represented a drop of seven points for the NDP since February and a gain of 13 for the Liberals.

Among non-francophones, Trudeau's arrival has boosted the Liberals by 16 points since February to 63%, most of that coming from the Conservatives, who dropped 17 points to 13%. The NDP has been sliding among this demographic for some time, and in fact only dropped one point since February.

On who would make the best prime minister, Trudeau narrowly edged out Thomas Mulcair with 29% to 27%, while Stephen Harper rounded out the list with just 10%. CROP's polling also showed that Trudeau stacked up well against Mulcair on many issues, but that generally Quebecers felt the NDP leader to be a better adminstrator while Trudeau had better personal qualities (rassembleur, trustworthy, etc.).

In terms of seats, the model might be a little hamstrung. It is not set up to incorporate sub-regional data at the federal level (something that will have to be rectified by 2015). But with the province-wide results, the New Democrats would win 41 seats, the Liberals would take 31, the Conservatives would hold four, and the Bloc Québécois would take two. However, considering that the Liberals were ahead of the NDP in every region of Quebec, it is likely that the model is under-estimating the number of seats the Liberals could win. Incumbency is a factor, though, that will play into the NDP's hands. Of course, if Trudeau is on track to win the next election he might also be able to recruit a strong field of candidates to negate the NDP's intrinsic advantage.

Quebec is the place where the Liberals can do the most damage to the New Democrats. The NDP is much better placed, even against Trudeau, in the West while the Atlantic region has been in the bag for the Liberals for awhile. Ontario is where the Liberals can put a big dent in the Conservatives' numbers, but it is Quebec that the Liberals will need to win if they are to supplant the New Democrats as the main alternative to the Stephen Harper. At this stage, it seems they are well on their way. For now.

Wednesday, April 24, 2013

Wildrose leads Tories (no, this isn't 2012)

Over the weekend, Léger Marketing released new numbers for the Alberta provincial scene for the Calgary Herald and Edmonton Journal. The numbers are not good for Alison Redford, just as they haven't been in the latest polls on her approval ratings as premier.

Yes, this is Alberta and we all remember what happened in Alberta last year. But that wasn't just a case of polls being wrong. Other things were at play, including an undeniable late swing in voting intentions. Arguably, some pollsters did miss the swing in their final polls when they should have been capturing at least some of it. But Léger Marketing is free from blame - their last poll was out of the field a full week before the vote.

I wrote about this poll for The Huffington Post Canada, and I go into more detail about the numbers there. But let's take a look at the topline and regional results here.
Léger was last in the field in January with a telephone poll, but have changed methodologies to use an internet panel this time. Comparing the two polls, then, is a little bit of an apples to somewhat different apples exercise.

Nevertheless, since then Wildrose picked up nine points to lead with 37.3% support. The Progressive Conservatives dropped 11 points to 28.5%, while the Liberals were up five points to 16.7%. The New Democrats were virtually unchanged at 13.7%, while support for other parties dropped a little to 3.8% (this would include the Greens and the Alberta Party).

The race is very tight in Edmonton, where the Tories and Wildrose were tied at 27% apiece (Léger's January poll had no regional results). That represents a drop of around 17 points for the Tories since the election, and a gain of four points for Wildrose. The Liberals made the biggest gain, picking up 10 points to hit 23% while the NDP was up three points to 20%.

In Calgary, Wildrose held firm at 37% but the PCs dropped 15 points to only 30% in the city. The Liberals and NDP were each up six points, to 17% and 11% respectively.

And in the rest of Alberta, Wildrose was up marginally by five points to 48% while the Tories dropped 14 points to 29%. The NDP was up four points to 11% while the Liberals were up five to 10%.

The Progressive Conservatives have taken a big hit everywhere. They are still relatively competitive in Edmonton and Calgary as voting intentions are divided four ways, but outside of the two cities the yawning gap between Wildrose and the PCs means that Redford loses a lot of seats.
With these numbers (and recall that the seat projection model was quite accurate with the actual regional results plugged into it), Wildrose would win around 54 seats and form a majority government.

The opposition would be slightly more robust than it is now, however, with 17 Tories, 10 Liberals, and six New Democrats.

Wildrose wins half of its seats outside of Edmonton and Calgary, but still wins the majority of seats in Calgary and the plurality in Edmonton. The Tories take 14 of their 17 seats in the two cities, while Edmonton is responsible for 12 of the 16 seats that the Liberals and New Democrats win.

What is notable about this poll is not that Wildrose is taking off - their gains were relatively modest and they gained no ground in Calgary - but rather the PC vote has collapsed. A silver lining for Redford is that a lot of  it is going to the Liberals and NDP, where it might be liable to come back by election day in order to stave off a Wildrose victory (similar to what occurred last year). But Redford's personal numbers have tanked, whereas they were still pretty good even in the middle of the campaign when the Tories were trailing Wildrose by double digits. A party can rebound more easily when it is led by a popular leader. When it isn't, it is doubly hard to make up lost ground. But Redford has time: the next election won't be held until 2016.

Tuesday, April 23, 2013

35 years of polling

There was a flurry of polls released just prior to Justin Trudeau's leadership victory last Sunday, but since he took over it has been virtual silence. Forum ventured out with a post-announcement poll that, putting the Liberals at 43%, has to be considered on the high side. Unfortunately, no other pollster has waded into the pool since then, so we don't know if Trudeau was indeed giving the Liberals that much of a boost.

So if we can't find any recent polls, let's find some really old polls.

When I was poking around the various public opinion poll archives that exist, I came across a table that compiled all of the voting intentions polling from Environics stretching back to 1978. I used the data for my article today for The Globe and Mail looking at leadership boosts. Check it out.

But let's take a look at this polling from Environics stretching back 35 years. Note that I calculated the share of decideds only in the chart below, corrected what appeared to be a few transcription errors in the chart, and used my own monthly averages after 2010. The chart is not calibrated to give equal space to equal time, but it is close enough.
Environics federal polling, 1978-2010 (Click to magnify)
And there you see the vagaries of public opinion over a handful of decades.

What is most striking about these numbers to me is that the political landscape of the 1990s and 2000s that we have come to consider as the norm in Canada were actually pretty anomalous. The race was a close three-way race for much of the 1980s and the very early 1990s, making the current situation more reminiscent of those days than anything particularly new.

Throughout the 1990s and early 2000s, the Liberals were way ahead of the pack. They weren't Canada's dominant party only because of a split of the vote on the right, either. The very pale blue line combines the support of the Progressive Conservatives and the Reform/Canadian Alliance parties throughout that period, and only at the end of 1993 before the PCs bottomed out was the combined support for these parties anywhere close to where the Liberals were (true, the race was closer during elections but outside of them the Liberals were coasting).

It is interesting to see that the recent surge in NDP support is nothing that this country has not seen before. The New Democrats were routinely polling above 20% in 1981-1982 and continuously between 1985 and 1992 - a very long period of time politically. The New Democrats were second in the country, or at least tied for it, for a brief time in 1982 and then from around 1986 to 1992, when the New Democrats were leading nationally in a few individual polls in 1987, 1990, and 1991.

The chart does suggest, however, that opinion was much more fluid in the 1980s than it was throughout most of the period starting in the 1990s and since. In a relatively short period of time the Liberals or PCs would surge or drop 10 to 20 points or more. The Tories were under 30% at the end of 1979, and then rose all the way to over 50% by 1983, while the Liberals were at 50% in 1980 before slipping to 20% by 1984.

It demonstrates how every party has had its day in the sun and its catastrophic drops.

The New Democrats were gaining consistently after 1984 until they moved into first in 1987, while they dropped from first in 1991 to fifth and under 10% in 1993. They slowly gained from then until their boost in 2011.

The Liberals were up-and-down under Pierre Trudeau and then John Turner, before gaining again under Jean Chrétien and dropping under Paul Martin. Much of the damage was done at that point of time, it seems, in large part due to the Gomery Inquiry. Stéphane Dion and Michael Ignatieff could only give the party fleeting blips of support before dropping again.

And then there are the Tories, struggling before Brian Mulroney became leader, struggling again just before his first mandate was over, and then dropping like a stone shortly after his 1988 election victory as the Reform Party ate into their support. The Progressive Conservatives in the 1990s they were vying for runner-up status in the mid-teens with the Bloc Québécois, New Democrats, and the Reform/Canadian Alliance parties. The Canadian Alliance took off a little when Stockwell Day was named leader, but they never really challenged the Liberals.

Only after the merger of the parties did the Conservatives become competitive, and by 2006 they were leading in national voting intentions, something they have not yet given up for any long period of time. But their recent polling is among their worst as a united party. The last time they dropped this low was around the time that Reform and the Bloc started pulling their coalition apart.

I think this chart shows how difficult it is to predict anything beyond a few months into the future with any accuracy. The creation of the Bloc, and then its destruction at the hands of the NDP, was unexpected. That the Liberals would poll well over 40% and easily lead a crowd of small parties for a decade could hardly have been seen in the days when the New Democrats were pushing the Liberals into third in the late 1980s. That the Progressive Conservatives would be reduced to two seats within a decade of polling at almost 60%, and that they would eventually be taken over by a Western protest party, would have seemed like some fantastical alternative history.

So how might Justin Trudeau do in 2015? A better question, considering the peaks and valleys of this chart, might just be how he will do next month. And then we'll see.

Friday, April 19, 2013

Friday update: By-elections in Labrador and close race in Ontario

In this week's Friday update: a new poll for the Labrador by-election, a forecast for the provincial by-election in Cartwright-L'Anse au Clair in Newfoundland and Labrador, and a look at a very close race in Ontario. Are the Greens about to breakthrough? (No, but some interesting numbers nonetheless.)

A second poll confirms Jones' lead

A poll was released earlier this week by Abacus Data, done for VOCM (a radio station in the province). The results confirmed both what Forum Research's IVR poll (Abacus did one with live-callers) and what the regional swing would have us believe to be a likely outcome.
The poll gives an even wider lead to Yvonne Jones, with 63% to 20% for Peter Penashue and 17% for Harry Borlase of the NDP.

Abacus broke it down regionally, and you can see that the Liberals are way ahead throughout the province. To give it some context, their best performance in the poll was in eastern Labrador, where the Liberals did best in 2011. The NDP seems to have dropped a little in Labrador City, which had given the NDP good numbers in the last federal election, while the Conservatives are only slightly competitive in central Labrador, which had been more or less split between Penashue and Todd Russell of the Liberals in 2011.

There are no good numbers for the Tories in this poll, as only 29% of respondents said they would consider voting for the Conservatives (and only 58% of those who voted for them in 2011), compared to 60% for the Liberals (and that doesn't include the "don't knows").

The poll does suggest that the Conservative strategy for the by-election is reasonable, however, with Penashue emphasising his ability to deliver for the riding as a cabinet minister. Fully 60% of respondents in poll agreed that Labrador is better served with a government MP. Nevertheless, they don't seem willing to vote for one.

Jones provincial seat likely to remain Liberal, probably

Due to Yvonne Jones running as the federal Liberal candidate in Labrador, her provincial riding of Cartwright-L'Anse au Clair in eastern Labrador is vacant.

This should be an easy win for the provincial Liberals. The party is still trailing the PCs and NDP by a wide margin province-wide, but the Tories have dropped significantly and they have traditionally been the main opponent to the Liberals in the riding.

Jones has represented Cartwright-L'Anse au Clair since 1996, and has won with huge majorities taking more than 70% of the vote in the last two elections.

The Liberals have held the riding since 1975, after the Labrador Party held it for one term. The riding would seem to be a Liberal stronghold, but Jones has represented it for so long that it is difficult to know for certain whether it is still a primarily Liberal riding or just a Jones riding. She was, after all, first elected as an independent by defeating the incumbent Liberal MHA.

It seems unlikely the Tories are in the running. Their support has plummeted in recent months. The New Democrats, however, have been surging and have led in a few of the recent provincial polls. The NDP only took 2% of the vote in 2011, the first time they had run a candidate in the riding since the 1990s. But now that Jones won't be on the ballot, how many voters would consider going with the NDP this time around? If the New Democrats find a great candidate - and the Liberals don't - they could have a serious shot. They will certainly be gunning for the riding, as a win would vault them ahead of the Liberals in the seat count in the House of Assembly (and potentially land them Official Opposition status).

But because of the numbers that Jones put up in the last election and the favourable trends, the model considers this riding to be a Strong Liberal seat. I don't suspect anything will happen to change the model's mind.

Close race in Ontario

EKOS released a provincial poll for Ontario earlier this week, and I wrote about the implications in an article for The Huffington Post Canada on Wednesday.

The poll does not show much difference from the other surveys we've seen since Kathleen Wynne became premier and Liberal leader. It gave the Tories 31.7% to 30.8% for the Liberals and 25.5% for the New Democrats. The Greens managed 9.7%, the only really unusual number. Most surveys have been showing a close race between the PCs and the OLP with the NDP in the mid-20s, so there is nothing much new here.

Adding it to the By-Election Barometer, however, changes the status of Windsor-Tecumseh from a Strong NDP riding to a Likely NDP win. The Liberals had a bit of a better result in southwestern Ontario in this survey.
In terms of seats province-wide, though, the Progressive Conservatives would barely eke out a plurality with 41, while the Liberals would take 39 seats and the New Democrats 24.

And the Greens would win three seats with 9.7% support, which rises to as much as 17% in the northern part of the province (which, in my model and in the estimations of EKOS and several other pollsters, includes the area stretching all the way down to Lake Simcoe).

Would the Greens actually win three seats if an election were held today? Almost certainly not - EKOS only gives the Greens 5.7% support among "likely voters", and at that level their seat count would drop to zero. But it is interesting to see that the Greens aren't too far away from potentially being competitive in a handful of ridings. If the provincial Greens make a splash in British Columbia's election on May 14, I would not be surprised if Green parties throughout Canada gain a little lustre.

Wednesday, April 17, 2013

Resiliency for B.C. Liberals in the Interior?

The newest projection for the now on-going election campaign in British Columbia continues to have little good news for the B.C. Liberals, as the B.C. New Democrats remain well in front and with a roughly 98% chance of winning on May 14. But a new riding poll for the two seats in Kamloops suggests the B.C. Liberals may be more difficult to knock off in some ridings than expected.

The projection now incorporates all of the new polling that was released for the start of the campaign, from EKOS, Angus-Reid, and Ipsos-Reid. But there have been no major changes: the NDP is projected to be between 46.1% and 49.7% support and capable of winning between 57 and 73 seats, based on current polling data. The B.C. Liberals trail with between 28% and 31.2% support and between 10 and 27 seats. More precisely, the NDP stands at 47.9% and 65 seats to 29.6% and 19 seats for the Liberals.

The other parties trail at some length, with the B.C. Conservatives between 10.6% and 13% support and the B.C. Greens between 7.9% and 9.9%. Neither are currently projected to be in the running for any seats, but the forecast gives the Greens a high of four seats and the Conservatives a high of one. Unfortunately for Christy Clark, her high forecast is only 37 seats - short of a majority.

For a broader analysis of the trends and the projection, please read my article this morning in The Globe and Mail.

Before going into the three province-wide polls, a new survey by Oraclepoll commissioned by The Daily News in Kamloops is worth a look. The poll was conducted between April 12 and 14 and surveyed 300 people apiece in the ridings of Kamloops-North Thompson and Kamloops-South Thompson via IVR. Unfortunately, the polls did not include a Green candidate in either riding as any have yet to be named.

In Kamloops-North Thompson, the NDP's Kathy Kendall was given 49% to 43% for the Liberals' incumbent MLA Terry Lake. The Conservatives got only 8% in the riding. In Kamloops-South Thompson, rookie Liberal candidate Todd Stone took 48% to 34% for the NDP's Tom Friedman and 18% for the Conservative candidate.

This poll has been included in the projection, but both ridings had been considered to be relatively easy victories for the NDP by the model. Both are now a bit closer with the inclusion of this poll, but it suggests that the B.C. Liberals may not be so easy to knock off in the Interior. We saw something similar happen for the Alberta Liberals in Calgary and Edmonton and the Quebec Liberals in central Quebec. In both cases, the Liberals were able to hold on to a lot more of their support than expected, while their vote tanked where they did not hold a seat. Could the same thing happen here?

How a Green candidate would change things up is uncertain. It might drag the NDP's vote down a bit, but also the Liberals'. There is some indication that much of the Green support is coming from disaffected Liberals who do not see themselves in the Conservatives but won't support the NDP.

The support for the Conservatives in these two ridings should be of some concern for the party. Not that these were serious targets for the party - they weren't - but it should be disappointing to them that they are not even close to being in the running in two B.C. Interior ridings. In Kamloops-South Thompson, they haven't been able to take greater advantage of the lack of a Liberal incumbent, the sort of situation that should give the Conservatives more of a chance. Their support in the Interior would need to double before the Conservatives could start thinking of winning that riding.

For the NDP, their score in Kamloops-South Thompson is virtually unchanged from 2009, despite their uptick in support. That suggests that the NDP may not be able to make inroads throughout B.C., and that more than a couple ridings where they are pegged to be ahead by a narrow margin could be difficult ones for them to win.

Hopefully we'll see more riding polls from the Interior to shed some more light on this.
For the provincial polls, we'll start with EKOS. They are doing something interesting, reporting their results for the entire population but also just likely voters. Likely voters are primarily those who voted in the last federal election, but EKOS does some other calculations to model what the voting population will look like as opposed to the general population. As the purpose of the projection is to predict the outcome of the election, and not what all British Columbians think, the projection will only be using the 'likely voter' numbers from EKOS.

The big question with these numbers is whether other firms are already doing some of these calculations. It seems that some of them are, so EKOS is merely showing a little bit more about how the sausage is made.

But in terms of the general population, EKOS shows little change from their last poll from February. And their likely voter numbers are well within the norm of other surveys. Of note: the NDP leads by four points among men but 20 points among women. They also lead in every age group except among those over the age of 65, where the Liberals have the advantage.
The new Angus-Reid poll also shows little change from their last survey, conducted in mid-March (NDP down three, Liberals unchanged, Greens up two). They do show a drop in Adrian Dix's approval rating, however, which could be something to keep an eye on.

This poll had the much ballyhooed result of the Greens in second on Vancouver Island, with 22% to the Liberals' 19%. Considering the sample sizes, it is not much of a lead. And it is counter to what Ipsos-Reid found (see below).

Angus-Reid shows the same gender split as EKOS, with the NDP up by 11 among men but 25 among women. Perhaps most importantly, their respondents rated Dix more highly than Clark on the economy (27% to 22%).
Ipsos-Reid has shown a little more change since its last poll of mid-March, but it is still within the margin of error (or would be, if this was a probability sample). Ipsos has the NDP and Liberals down three points apiece, and the Greens and Conservatives up two each.

Again, we see the gender divide: the NDP up by seven among men, but 31 among women. And they lead by 17 points among those aged 55 or over (i.e., voters).

The approval ratings (30% for Clark, 51% for Dix) show no real change, but both Jane Sterk and John Cummins had increases (to 28% and 19%, respectively). Note that in the rolling three-poll average of approval ratings, Sterk now rates more highly than Clark.

Also problematic for Clark is that 43% of British Columbians strongly disapprove of her. That is a huge number. And 58% of respondents expect the NDP to win a majority (they must be ThreeHundredEight readers), compared to only 8% who expect the Liberals to be re-elected to a majority. Perception can be everything.

The regional result on Vancouver Island is worth noting, as it differs from Angus-Reid's. The two polls put the Liberals and Conservatives at roughly the same level of support, but the big difference seems to be between the NDP and the Greens. But if this were a random sample, the margin of error would be over eight points - so maybe this is much ado over nothing.

What is perhaps most remarkable about these polls is that they have been more or less identical for the last seven months. The opinions of British Columbians seem pretty solidified. With such unusual consistency, it may be too much to expect the Liberals to overcome such a huge margin in just four weeks.

Monday, April 15, 2013

March 2013 federal polling averages

Considering that we're already closer to May than we are to March, and that the Liberals just changed things up quite a bit by choosing Justin Trudeau as their leader yesterday, now may seem like an odd time to take a look at the March polling averages. It is. My apologies for it being so late, but I've been burned before - Nanos once released a poll from the previous month on the 11th.

In any case, the number of polls released in March was rather low: only three, compared to seven in February and five so far in April. For that reason, in addition to it being so late in the month, I will present the information for continuity purposes and just briefly go through the highlights.
The Conservatives averaged 31.3% in March, virtually unchanged from February when they averaged 31.5%. The New Democrats averaged 27.1%, also not much different from their 27.5% of February. However, they fell to third for the first time in the monthly averages and their lowest point since April 2011.

The Liberals averaged 28.3% to place second, up 2.8 points from February.

The Greens dropped 0.9 points to 6.5%, while the Bloc Québécois was down 0.3 points to 6.1%.

Compared to the last time Abacus, Forum, and Léger were in the field within a 22-day period (Dec. 3-18), both the Conservatives and NDP have dropped while the Liberals have gained considerably.
The Tories fell two points and the NDP 2.7 points in polls by these three firms since December, while the Liberals were up 5.7 points.

Click to magnify
Regionally, the Conservatives led in British Columbia (barely), Alberta, the Prairies, and Ontario, while the Liberals were ahead in Atlantic Canada and the NDP in Quebec.

Of note: the Liberals were at their highest point in Alberta since December 2010 and their highest in Atlantic Canada since August 2010. The NDP was at its lowest in March in the Prairies since April 2011, while this was the first time the New Democrats fell below 30% in the monthly averages under Thomas Mulcair in Quebec. For the Conservatives, their result in Atlantic Canada was the lowest on record going back to January 2009 (their previous low had been 25%).

In terms of seats, compared to February the Conservatives were down two and the NDP was down 14, while the Liberals were up 12 and the Bloc gained four on the proposed boundaries of the 338-seat map.
The Conservatives would be reduced to a minority, and could easily be outvoted by the combined 187 seats of the NDP and Liberals. Unlike February, however, the Liberals would have formed the Official Opposition in a March election.

Approval ratings
It will be much more interesting to look at the April averages, as the number of polls is greater and the changes taking place are more dramatic. Already, the Liberals are averaging 31.3% in April (unweighted) compared to 30.6% for the Conservatives and 25.2% for the New Democrats. I imagine any other polls that will be released in April will show striking results as well.

Leadership races tend to have a big effect on the polls - it was an interesting time to be a poll watcher after Michael Ignatieff and Mulcair took over their respective parties, and it is interesting again with the newly-minted Trudeau. And now that all five parties (probably) have the leaders that will lead them into 2015, the real race begins.

Friday, April 12, 2013

Friday update: Final Liberal leadership endorsement rankings

For this week's Friday update, a brief word about the Labrador by-election, how approval ratings relate to re-election chances, and the final update to the Liberal leadership endorsement rankings.

Trudeau first, Murray or Hall Findlay second

The leadership endorsement rankings weren't much fun this time around, as Justin Trudeau got most of the endorsements that were handed out from Liberal luminaries. By my count, 84 current or formerly elected federal or provincial legislators and senators from the Liberal Party (and its provincial cousins) gave out an endorsement (not including the leadership contestants themselves), and Trudeau got 66 of them (or 79%). Joyce Murray received ten (12%) and Martha Hall Findlay received four (5%).

The endorsement rankings were a lot more interesting to follow during the NDP leadership race, as Thomas Mulcair, Brian Topp, and Peggy Nash battled it out for top spot. In the end, Mulcair was ranked first and he won. Daniel Paillé was also ranked first in the rankings for the BQ's race.

Of course, one hardly needs a sophisticated model to predict the outcome of the Liberal leadership race. Every single indicator points to Trudeau winning, and winning by a large margin. But who will come second?

The endorsement rankings - unweighted by region - assume that Murray will finish second. She has more endorsements than Hall Findlay, and the ones she did receive are more highly valued than those that have been awarded to Hall Findlay.

Speaking of which, Hall Findlay is finally on the board. New endorsements added to the model (they may not actually be new, but they are new to the rankings) include those from Kent Hehr, an Alberta Liberal MLA, and former provincial MLAs James Doyle and Aldéa Landry (New Brunswick) and Weslyn Mather (Alberta).

New Trudeau endorsements include Marcel Proulx, Pablo Rodriguez, and Maryse Gaudreault - all from Quebec.

No new ones were added for any of the other contestants.

Murray does have a good chance of finishing second because of all the non-Liberal support she might get from groups like LeadNow and supporters of people like David Suzuki and Bruce Hyer. There is good reason to think that she will finish second, over and above the endorsement rankings. Her fundraising has been better than Hall Findlay's, though it is still far behind Trudeau's.

However, after Marc Garneau dropped out Hall Findlay became the most viable non-Trudeau (and anti-cooperation) option on the ballot. She may be able to get more support than Murray because of that, particularly since Murray's non-Liberal supporters may not bother voting in what is considered a race that will be an easy win for Trudeau. The weighted endorsement rankings (weighting the endorsements by province, similarly to how the votes will be weighted by riding) put Hall Findlay in second.

That may prove to be a happy accident as the results will almost certainly be much closer to the mark than the unweighted results. I call it an accident as Hall Findlay is the only leadership contestant that has been endorsed by any Albertan Liberal (at least as far as I can find). That means Hall Findlay is awarded 100% of the points in Alberta, and that propels her from under 1% of the weighted endorsements points to just under 10%. Murray remains at about 8%, while Trudeau is reduced to 82%. That seems like a slightly more plausible result.

But this will almost certainly be a case where the endorsement rankings do not end up very close to the official tally. Obviously, Deborah Coyne and Karen McCrimmon won't be completely shut-out, and Martin Cauchon will probably get more than 1% of the vote. The race was simply too lop-sided, and few with any skin in the game wanted to go up against the Trudeau juggernaut.

If I had to guess, I would give Coyne and McCrimmon about 2% of the vote each, and Cauchon another 5% (he could potentially get more, as he has also had good fundraising). I'd give another 15% apiece to both Murray and Hall Findlay (I won't guess who finishes second, as I have no clue) and that leaves about 60% of the vote to Trudeau. I think that is a reasonable guess - Stephen Harper took 69% in his leadership win (56% of weighted ballots), and I'm not sure if he dominated that race as much as Trudeau has.

Sunday update: In the end, the weighted rankings turned out to be quite close to the final result. Actual point share for each candidate was 80.1% for Trudeau, 10.2% for Murray, 5.7% for Hall Findlay, 2.6% for Cauchon, and 0.7% apiece for Coyne and McCrimmon. Quite close.

One note on turnout. As of writing, turnout among registered voters is 66.7%, or 84,803 of 127,173 registered voters. That easily beats the NDP's turnout of just under 50% in their leadership race. Or does it? While the NDP did have some 130,000 eligible voters, many of them were provincial NDP members or lapsed members who had no interest in voting for the federal leadership in the first place. If the NDP had required registration, how many would have been eligible? It almost makes the 294,000 Liberal members and supporters more comparable, meaning the turnout is under 30%. But then again, many of those supporters might have been even less interested than the provincial and lapsed NDP members from 2012. It makes it an apples to oranges comparison, in the end, at least for turnout.

It is easier to compare the total number of actual votes, and in that sense the Liberals have the NDP beat. Even that comparison is limited, though, since the Liberals have that new supporter category. But if the NDP had had a similar category, or if the Liberals had not introduced it, would the totals have been all that different? I suppose we can't know.

The Liberals also have a shot at beating the number of voters in the 2004 Conservative leadership race, as they are only about 13,000 short of doing so. They are about 21,000 short of beating the PQ's number in 2005, though, and that might be tougher to top. And they'll need almost perfect turnout to beat the Canadian Alliance's vote total in 2000 (some 120,000).

It will be interesting to parse through the results once they are released on Sunday. Here's to hoping that the Liberals are generous with the data. (Also, if I have missed any endorsements please leave them in the comments (with a link, preferably) and I will add them to the rankings throughout the weekend.)

Liberals still favoured in Labrador

The by-election forecast for Labrador has been updated, and it remains a Strong Liberal riding. The two new polls added to the forecast, by Nanos and Abacus, have somewhat stronger Conservative results for Atlantic Canada, but that probably has nothing to do with the race there. Labrador is a hyper-local by-election, and the results could be surprising. Will voters punish Penashue or vote to have a cabinet minister as their MP?

The Liberals are also favoured in New Brunswick's Kent by-election, which will be held on Monday. But there has been no new polling data for the province since the last Corporate Research Associates poll.

Approval ratings and re-election

My article in The Globe and Mail today compares approval ratings to re-election chances, and finds that approval ratings have been a decent predictor over the last few years. The last approval ratings reported in an Angus-Reid quarterly report before an election was within two points of the result for that premier's party in the 2011 Newfoundland and Labrador, 2011 Saskatchewan, and 2012 Quebec elections, and wasn't too far off in Manitoba, Alberta, and Ontario (considering they were conducted a month or more before).

Schedule during the B.C. election campaign

The campaign in British Columbia effectively starts this weekend. My workload will be greatly increasing during the campaign (I'm like a seasonal worker, really), so new posts to the site may often be limited to projection updates for the B.C. election. I will try to keep up with updating the by-election forecasts and the aggregations of federal, Ontario, and Quebec polls, but I apologize if anything falls through the cracks. And let's hope Darrell Dexter waits until after May 14 to launch that province's campaign.

Thursday, April 11, 2013

Approval ratings for 30 premiers and leaders

Angus-Reid's quarterly polls on premier approval ratings are always interesting, but this week they made it even better by releasing the approval ratings for all major party leaders in nine provinces (poor PEI was excluded, as even with a 7,000+ national sample the number of respondents would be tiny on the island).

It isn't the first time that these numbers have been released, as Angus-Reid was kind enough to provide me with them back in September. But hopefully they will continue to do this every quarter, as it provides a great deal of information.

Let's rank the 30 premiers, opposition leaders, and major party leaders in the nine provinces by approval rating.
Three of the top five have not changed since September: Brad Wall, Lorraine Michael, and Adrian Dix. Only Wall, Michael, and Danielle Smith have a better than 50% approval rating, while all others have under 50%.

Quite a few more leaders have a disapproval rating of 50% or more: David Alward, Tim Hudak, Pauline Marois, Dexter, Alison Redford, Amir Khadir, Christy Clark, and Kathy Dunderdale.

Note that the leaders of Québec Solidaire and the Quebec Liberals are not correct in this poll. That is understandable for the Liberals, who chose Philippe Couillard as their new leader (replacing interim leader Jean-Marc Fournier) in the midst of this polling. Khadir, however, should have been replaced by Françoise David. Khadir is no longer the male spokesperson for the party, and in effect the female spokesperson (David) is the de facto leader. I suspect she would have had a much higher approval rating than Khadir, and might have cracked the top ten.

It is hard to find any partisan advantage in these numbers. Conservatives like Wall, Smith, and Brian Pallister are ranked highly, while Clark, Dunderdale, and John Cummins are at the bottom. Stephen McNeil, Brian Gallant, and Dwight Ball score well for the Liberals, while John Gerrard and Raj Sherman do not (and you might include Clark with them as well). The NDP has three of the top five in Michael, Andrea Horwath, and Dix, but also Dexter in the bottom tier.

There does seem to be an anti-incumbent bias in the numbers, as nine of the top ten are third party or opposition leaders, while five of the nine premiers are ranked 20th or lower. There doesn't seem to be a gender bias, though: Redford, Dunderdale and Clark are at the bottom of the list but Michael, Smith, and Horwath are at the top.

Who has made the biggest gains since the September poll? If we look at it only in terms of ranking, we see that Ball made the biggest gain, jumping 18 spots to 10th overall. Gallant is ranked 17 spots higher than interim leader Victor Boudreau was, while Dominic Cardy gained ten spots to move into the top half of the list. Atlantic Canadian leaders were on the move.

The leader that had the largest fall was Redford, who was ranked second in September but is now ranked 25th. Her approval rating has just plummeted in the last few months. Dunderdale also fell, by 17 spots, to 28th, while Marois and Cam Broten (compared to John Nilson) each fell eight spots to 20th and 21st overall. Wall, meanwhile, held on to top spot while Cummins and Victor Lau were unchanged as the bottom two on the list.

But is an approval rating the best way to measure these leaders? Broten, for example, was just recently named leader of the Saskatchewan NDP and his approval rating of 32% is actually quite decent, considering his disapproval rating is only 19% (49% have yet to make up their minds).

The chart to the left ranks leaders by their net approval rating (approval minus disapproval) and it changes things quite a bit. Edit: An earlier version of this post had incorrect net ratings for Wall and Sterk.

Broten makes the biggest move, increasing by 15 spots to sixth overall. Sterk also jumps 13 spots, as she has a large "not sure" total as well.

The bottom four are all premiers, meanwhile, with Dexter, Redford, Clark, and Dunderdale at the bottom of the list. Marois is also in the bottom tier. Wall remains the only premier in the top tier, but now becomes the only premier in the top half, too. Wall also has the distinction of being the only premier ranked more highly than all of his main opponents.

Most other premiers are ranked below two opponents: Dunderdale, Clark, Redford, Dexter, Marois, and Alward are all third or fourth in their province. Selinger is ranked just a little more highly than Gerrard, while Kathleen Wynne is ranked much more highly than Hudak (though she trails Horwath).

Approval rating is a pretty important indictor of ability to win. If an election were held today, based only on voting intentions polls, the approval ratings leaders (not net ratings) in every province except Ontario and Quebec would win (and we can't even say that with certainty about Quebec, as Couillard might have ranked more highly than François Legault).

These rankings aren't anything to sneeze at - if the elections in B.C. and Nova Scotia go as expected, premiers will take up three of the top six spots in the approval ratings rankings in a few months. And if these ratings continue, one has to wonder whether some of the other leaders at the bottom of the list (Redford, Dunderdale, Cummins, maybe Hudak) will be on their way out in the next year or two. Gerrard is already leaving and is just waiting for his replacement, while Khadir shouldn't have been included at all. That accounts for seven of the bottom 10 spots in the approval ratings ranking, and eight in the net ratings ranking.

Are approval ratings the kiss of death? Possibly - three of the four names that have disappeared from the list since September were ranked 17th or lower.

Tuesday, April 9, 2013

B.C. New Democrats still in control

There have been few major changes in the projection and forecast for the British Columbia provincial election, but all that means is that the time remaining for the B.C. Liberals to turn things around has continued to shrink. The odds that the party would be able to comeback and beat the B.C. New Democrats in the popular vote are still less than 50-to-1.

The New Democrats continue to lead with a projected 48.7% of the vote, down 1.2 points from the projection of Mar. 25 (which used polling data running to Mar. 19). The Liberals are down 0.5 points to 29.9%, while the B.C. Conservatives remain steady at 10.4%. The B.C. Greens picked up 1.7 points and are now projected to have 9.2% support.

Due to Liberal gains in the B.C. Interior, the party has picked up one seat since Mar. 25 and is now projected to win 21. The New Democrats dropped one to 63, while one independent is projected to win. In terms of the ranges, the New Democrats would be expected to win between 51 and 73 seats if an election were held today, compared to between 9 and 33 seats for the Liberals and 0-4 for the independents.

Problematic for the Liberals is that the high forecast for May 14 has fallen from 54 to only 43 seats - the bare minimum for a majority government. As time runs out, even the absolute best case scenario is precarious for the party.

Regionally, things are relatively steady in Metro Vancouver and in the Interior/North. In and around the city, the NDP leads with 50.9% (-0.4 points) to 30.9% for the Liberals (-0.7) and 8.9% for the Conservatives (-0.5). The Greens made a 1.5-point gain to reach 6.9%.

In the Interior and North, the NDP is down 1.6 points to 41.3%, costing them one seat. The Liberals made their only regional gain here, picking up a bare 0.7 points to reach 32.7%. The Conservatives are down 0.6 points to 14.7% while the Greens are up 1.6 points to 9.7%.

There was more movement on Vancouver Island, where the Greens are up 2.4 points to 13.8%. The New Democrats are down 3.7 points to 51.7% and the Liberals 1.9 points to 23.2%. The Conservatives made the largest gain anywhere here, with a 3.1-point uptick to 10.5%.
The update was brought about by a new poll from Insights West, a polling firm based in British Columbia that was launched last year and whose president had previously been with Ipsos-Reid. This is their first foray into provincial politics, as far as I can tell, and the results are well within the norm of what other surveys have shown. Insights West also has very good disclosure, with its report containing both weighted and unweighted samples. They use their own panel for their polling.

The survey shows 45% for the NDP, 28% for the Liberals, 15% for the Greens, and 10% for the Conservatives. They are neither the highest nor lowest recent results for the NDP, Liberals, or Conservatives - but they are on the high-end for the Greens. We will have to see what the other firms say about the Greens as the campaign kicks off.

The Conservatives being at 10% is interesting, in part because it is within two points of where 10 of the last 11 polls have pegged the party's support to be. Why is that interesting? Because the UBC prediction market has consistently had the Conservatives at or just under 20% since mid-March. Whatever the market is recording, the polls have yet to see.

One in five respondents were undecided in this survey. The undecideds did lean slightly to the B.C. Liberals (25% to 20%), but not nearly enough to make much difference. The gender gap remains, with the NDP enjoying an 11-point advantage among men but a 24-point edge among women. They also lead by 18 points among voters aged 55 and over, a key demographic.

At the regional level, there are few surprises in this poll. The New Democrats are well ahead in both Vancouver and on Vancouver Island, while the race is closer in the Interior and North. The strong Green result on Vancouver Island has been noted before in other polls, though this is the highest result I have on record.

Steady as she goes, then. The campaign will need to shake things up radically in order to put the result in any doubt, but it will be interesting to see where the Green numbers go from here. For many British Columbians, voting for the Greens might become a very plausible option if the NDP looks like it is set to easily win - their vote might not be wasted or inadvertently send a Liberal to Victoria. Too much of that sentiment in NDP-Liberal races, though, and the New Democrats could be in trouble in a few individual ridings.

Friday, April 5, 2013

Friday update: By-election polling and Liberal leadership

This week's Friday update takes a look at new by-election polling in Labrador, new forecasts for the pending provincial by-elections in Ontario, updates to the Liberal leadership endorsement rankings, and discussion of the newest national poll.

Ipsos-Reid puts Liberals first

The polling has been coming fast and furious this week, but it has also been remarkably consistent. The latest missive is from Ipsos-Reid. I'll try to give the poll a deeper look next week, but it gives the Liberals 32% to 31% for the Conservatives and 27% for the New Democrats. It is yet another poll to put the Liberals at or above 30%, the Conservatives below 32%, and the NDP at or below 27%. It makes for a very tight grouping of polls.

As a result, the Liberals now lead by the tiniest of margins in the weighted average, with 30.7% to 30.5% support. Some of the regional numbers are interesting as well: the Liberals are almost tied with the Tories in Ontario, and are now ahead by more than three points in Quebec. That Ipsos gave the Liberals 32% in the province is interesting, as prior to this survey it had mostly been Forum that was bullish on the Liberals in Quebec.

Note that the reason the newer Ipsos-Reid poll has a lower weight in the average than the Forum poll is that the model considers the Ipsos-Reid poll to be older. As it was in the field between March 28 and April 3, it has a median date of March 31, whereas Forum was just in the field on April 2.

Also note that, despite what was reported in articles by Global and Postmedia, this Ipsos-Reid poll is not a hypothetical "what-if" Trudeau poll. Ipsos-Reid did not mention Trudeau's name in the voting intentions question.

Improving Liberal fortunes in Labrador and London West, worsening ones in Windsor-Tecumseh

With all of the new polling, the by-election forecast for Labrador has been updated and it remains rough for the Conservatives. The polling in Atlantic Canada has been extremely consistent, showing the Liberals at over 40% and the Conservatives at under 20%. That means the swing is strongly in favour of the Liberals, who average a victory margin of 43.6 points when the regional swing from the last month's worth of polls is applied to the riding.

Backing that up is a poll from Forum Research for Labrador. The short of it is that it gives Yvonne Jones a 36-point lead, with 57% to 21% for Harry Borlase of the NDP and 20% for Peter Penashue of the Tories.

I lay out the caveats for this poll in today's article for The Huffington Post Canada. Check it out.

Overall, it keeps Labrador a Strong Liberal forecast.

The numbers are also good for the Liberals in London West, as the polling picture has brightened for the Ontario Liberals in southwestern Ontario. The riding is now a Likely Liberal win, up from a Lean Liberal.

Conversely, the situation has gotten better for the New Democrats in Windsor-Tecumseh, as it is now a Strong New Democrat forecast, up from a Likely NDP win. Why the difference? Simply that there has been a drop in the number of polls in the last month in Ontario, making it more difficult to judge the volatility between polls. When that happens, the ranges tighten up and the NDP's low forecast pulls ahead of the high forecast for the Liberals.

Justin Trudeau's endorsement edge remains enormous

Also updated today were the endorsement rankings for the Liberal leadership race. Quite a number of endorsements have been added to the rankings, but it doesn't change much.

Joyce Murray did get the nod from Peter Milliken, while Gordon Wilson (former B.C. Liberal leader) and Judi Tyabji (former B.C. MLA) improve her numbers in British Columbia. She now has 33% of the endorsement points in the province, the only one where the race between her and Justin Trudeau is competitive (at least from an endorsement stand point).

But the list of Trudeau endorsers grew even larger, with a long list of senators added and former provincial leaders David Peterson and Yvonne Jones, among others.

Overall, it meant a marginally worsening share of the endorsement points for Trudeau, who dropped below 90% on the weighted rankings (to 89.9%). He is still miles ahead of Murray and the other contestants, however, who have not been able to attract a single endorsement that is counted in the rankings (all of Martin Cauchon's and Martha Hall Findlay's points come from their self-endorsements). Only a gimmicky turnout shock would prevent Trudeau from winning the race at this point - the greatest danger he faces (and it is a small one) is that too many people will consider his victory so inevitable that they don't bother to vote. Not a bad position to be in.

Thursday, April 4, 2013

Liberals move into first in poll average

Before getting too excited after reading this headline, Liberals should know that they are only ahead by 0.1 percentage points and that this is using this site's weighted poll aggregation methodology. And they have only been put in first place in a single poll (the aggregation does not take into account polling that includes Justin Trudeau as Liberal leader). And if you do a straight-up average without taking anything else into account, you can only go back two polls to give the Liberals the lead (going back a third puts the Conservatives ahead by 1.7 points).

But with those caveats laid out, that the Liberals have moved ahead in the polling aggregation for the first time since the election - a methodology that has been consistently employed - is rather remarkable. It could also be a blip.

I wrote a full analysis of the polls and calculated a seat projection for The Globe and Mail. Please take a look at it, as the analysis goes into far more detail than I will here.

Since I didn't go into the regional breakdown in my Globe article, here is how the current aggregation was translated into seats using the 338-seat electoral map.
The Liberals have moved ahead primarily on the back of the Forum poll that was released yesterday evening. It gave the Liberals 33% support to only 29% for the Tories and 25% for the New Democrats. When Trudeau was included as Liberal leader, the party's support increased to 40% while the Conservatives dropped to 28% and the NDP to 21%.

A flash in the pan? Maybe not - Léger also recently placed the Liberals in a close race with the Conservatives using a different methodology (online rather than Forum's IVR). And even these two polls aren't the first ones to show a transforming political landscape.

Here are the last seven federal polls that have been conducted. They show a great deal of consistency.

These seven polls from five different firms using three different methods of polling (Forum and EKOS use IVR, Nanos uses live callers, Abacus and Léger use an online panel) have all put the Conservatives at between 29% and 32% support. With the exception of the Abacus poll, the New Democrats have also been within a three-point band of 24% to 27% support in these seven polls, while in five of them the Liberals have been in a four-point band of between 29% and 33% support.

When you consider the numbers as a whole, it is hard to argue that the Liberals have not made big gains in the last two months and that the New Democrats and Conservatives have both taken a big hit.

"So what?", you ask. There is a long list of polls showing this or that leader poised to sweep this or that election in the past, only for them to fall apart and lose. Kim Campbell, François Legault, and (barring an unlikely comeback) Christy Clark are three easy examples.

That should provide a good reason to consider these sorts of poll results with caution (though, it must be pointed out, the Liberals are doing well already without mentioning Trudeau and that is harder to dismiss). But just because some leaders in the past failed to deliver on their promise does not mean that every leader will. Though the numbers are drooping lately, Thomas Mulcair did salvage the New Democrats' support in Quebec just as the pre-Mulcair leadership polls said he would, and he remains their best asset in the province. Conversely, after a boost throughout most of 2012 after his leadership victory, Mulcair's NDP is now just about where the pre-convention polls suggested a Mulcair-NDP would be.

In the same way that no one should be readying 24 Sussex for the return of the Trudeaus based on these numbers alone, no one should also assume that they will lead to nowhere. What should be assumed is that things will probably change, that we don't know how they will change, and that at best the numbers are a decent reflection of what Canadians probably think now. "The only poll that matters is on election day" may be correct in terms of determining who forms the next government, but it is completely wrong when it comes to public opinion. As I have said many times before, there is some meaning to public opinion between elections - we don't just get to express our views once every four years.

However, it is hard not to believe that the Liberals are setting themselves up for a disappointment. They did the same with their announcement of having 294,000 members and supporters, only to have some 130,000 register to vote. (As an aside, there is a difference between the two, which many have over-looked or ignored. The NDP did have some 130,000 members for their leadership race, but they didn't have to register to vote in the same way as in this Liberal race and in the end only 60,000 or so did cast a ballot. And if some quick research is correct, the Conservatives had some 250,000 eligible to vote in their leadership race in 2004 but a little less than 100,000 did. The 160,000 or so Liberal members/supporters who did not register to vote are not gone from the party forever, so it is apples-to-oranges to compare the 130,000 registered voters to the NDP's membership list in 2012 or the Conservatives' in 2004.)

Though the numbers are already quite good for the Trudeau-less Liberals (though of course many respondents are assuming Trudeau will win), the numbers the party is putting up with Trudeau in the mix (around 40%) are certainly unrealistic. Jean Chrétien managed those numbers during his tenure, but that was at a time when the New Democrats were lucky to break double-digits (though another five national points that used to be the Bloc's in Quebec is now up for grabs). Measured against those highs, Trudeau is certainly to disappoint. Measured against where the party was on 2 May 2011, he is certainly to do well. In the short term, he just needs to keep the party in the middle of those two extremes. That seems possible, but it will certainly not write the same headlines.

Tuesday, April 2, 2013

Strong Liberal results in poll, with or without Trudeau

A new poll from Léger Marketing shows very strong results for the Liberal Party of Canada, even without Justin Trudeau's name being mentioned. When he is added to the mix, the poll continues the trend of a big boost in Liberal fortunes, over-and-above the usual honeymoon that new leaders can usually expect to enjoy.
We don't often hear from Léger at the national level. The last coast-to-coast poll from the firm was released in December. The Conservatives slipped four points to 31% since that poll, putting them well within the range of where other polls have put the party since the beginning of February.

The Liberals picked up an enormous 12 points, however, putting them at 30% and only one point behind the Tories. The New Democrats managed 24%, a drop of six points and their worst result in any poll since the 2011 federal election campaign. Is this a fluke or a sign of things to come?

The Greens fell two points to 7%, putting them even with the Bloc Québécois. All of these changes in support would be outside the margin of error of two polls of this size with random samples.

These are remarkable numbers for the Liberals, but not altogether unusual. This is the third of the last four polls to show the Liberals at either 29% or 30% support. But to give them a six-point lead over the NDP is certainly a change of pace.

At the regional level, the Liberals were up throughout the country compared to Léger's December poll, with double-digit bumps in British Columbia, Alberta (+12 in both), Quebec (+11), and Ontario (+15). The poll actually puts the Liberals ahead in that last province, the first non-Nanos poll to do so since April 2011.

The numbers should be of great concern to both the Conservatives and NDP, as they come without a mention of Justin Trudeau's name. The numbers in Quebec are particularly problematic for the New Democrats, as this represents an 11-point drop from Léger's December poll, but also from their February poll of 1,024 Quebecers. Léger has generally given the NDP higher numbers in Quebec than their non-Quebec competitors, and that always suggested that the lower numbers recorded by Ontario-based firms might have been off the mark. For Léger to show such a poor result should be very worrying to the New Democrats.

Unusual numbers can generally be seen as outliers until proven otherwise, however, so hope is not lost for the NDP. We will have to see what CROP says in their next poll, which we should probably see before Léger reports again. It would be helpful, though, if CROP reports before Justin Trudeau becomes Liberal leader.
And this is why. Léger shows that under Trudeau, the Liberals surge to 37%, pushing the Conservatives down one point to 30% and the NDP down four points to 20%. This compares quite favourably to Léger's Trudeau poll in December, which put the Liberals and Tories even at 31% under his leadership and the NDP at 24%.

Regionally, Trudeau boosts the Liberals only slightly in the western provinces but pushes the party up by 10 points in Ontario and 15 points in Atlantic Canada. He puts the party narrowly ahead in Quebec as well (though, interestingly, the NDP's numbers improve slightly there).

But without Trudeau, the Liberals are still in a strong position and would be well-placed to form the Official Opposition.
With Léger's non-Trudeau numbers, the Liberals take 109 seats to 131 for the Conservatives, 55 for the New Democrats, 42 for the Bloc Québécois, and one for the Greens (all on the 338-seat map).

This makes it difficult for the Liberals and NDP to combine for a coalition government, however, as it puts them five seats short of a majority. The poor showing for the NDP in Quebec (and the small improvement for the Bloc) tips the balance and puts the opposition in the same awkward position of needing Bloc support as in 2008.
There is no such problem with Trudeau's numbers, however. His party takes 146 seats to only 124 for the Conservatives, 35 for the New Democrats, 32 for the Bloc, and one for the Greens. Together, Trudeau's Liberals and the NDP command 181 seats, more than the 169 needed for a majority.

But Trudeau's arrival makes things easier for the Conservatives in the west. Though they drop two points in British Columbia, they win 28 seats instead of 20 as the New Democrats take a big hit that the Liberals cannot turn into seats. They also win two more in the Prairies. Overall, the Conservatives win 10 more seats in the west while the Liberals only win one more and the NDP drops 11.

But the Liberals win their plurality in Ontario, Quebec, and Atlantic Canada (as they used to). With 71 seats in Ontario, they replicate what the Conservatives did in 2011. They are not able to supplant the Bloc in Quebec, however, as the Liberals have trouble making gains outside of the Montreal area. Most of those battles are between the NDP and the Bloc, but the New Democrats have been hamstrung by Trudeau. 

It makes for a very different political landscape, one that seems to have been constantly transforming over the last two years. Can these hypothetical numbers be dismissed? It would be easier to do so if they were not so consistent. In 10 polls since the end of October from Forum, Angus-Reid, and Léger, the Conservatives have always registered between 29% and 33% against Trudeau, while the NDP has always been between 19% and 24% (and only as high as 22% since December). In seven of the polls, the Liberals under Trudeau have managed between 37% and 42%. 

Maybe this will all disappear once Trudeau actually becomes leader, but he has been increasingly seen as the obvious winner for several months now and he has been getting plenty of media coverage. Things should not change too much once it becomes official, at least at first. But where will things go after that?