Tuesday, March 13, 2012

Alberta Tories slip as Liberals make gains

A new poll by Abacus Data for the QMI Agency indicates that the Progressive Conservatives are losing ground in Alberta, but unlike some other polls they appear to be losing some of their support to the provincial Liberals.

The projection has been updated to include this poll, and as a result the Tories have dropped 3.2 points in the projected popular vote to 38.7%, their lowest so far in the projection and the first time they have been pegged below 40%. Their range has tightened to 12 from 16 points as volatility as been reduced, but the range has moved in a negative direction: to roughly 33% to 45% from 34% to 50% on February 25.

The Progressive Conservatives have also dropped two seats and are now projected to win 65, thanks in large part to decreases in support in Calgary and the Rest of Alberta. Again their range has tightened, to 40 seats from 47, with the floor increasing to 36 from 34 but their ceiling decreasing to 76 from 81.

The beneficiary is not the Wildrose Party, which has gained only 0.3 points to hit 27.6% in the projection. That is their highest so far, but their range has decreased to the point where it no longer overlaps with the Tories. The highest projected range is 30.6%, 2.1 points below the low range of the PCs. Wildrose has picked up a seat, however, in the Rest of Alberta where the Tory drop has been most severe (it happens to be the seat that Danielle Smith will be contesting).

Instead, the Liberals have made the biggest step forward, increasing their support by 2.6 points in the projection. They now stand to take 16.3% of the vote, with their range now between 13% and 19%, up from 12% to 15%. They are now projected to win two seats, but their high range has doubled to 10 seats from only five. The Liberals have had their support increase across the board: 3.4 points in Edmonton to 19.7%, putting them in second, 3.4 points in Calgary to 16%, and 2.6 points in the Rest of Alberta to 12.6%. Though they are currently only projected to win seats in Edmonton and Calgary, their ranges put them in play in pockets throughout the province.

The New Democrats are up by 0.4 points to 12.7%, with their range standing at between 12.2% and 13.2% support. They remain projected to win three seats, but their high range has increased to six from four on February 25. But it is not only good news for the NDP, as they have dropped a point in Edmonton to 14.8%. However, they also gained 1.1 points in Calgary and 2.2 points in the Rest of Alberta.
The most important movement is taking place in the two cities. You can see that the Tories have been on a slow decline in Edmonton and that they have fallen more sharply in Calgary.

But aside from an initial uptick in mid-February, Wildrose has been holding steady despite the slip in PC support.

The New Democrats have been relatively steady as well in both cities, but it is the Liberals who have made gains, surpassing Wildrose in Edmonton and continuing their consistent increase in support in Calgary. As these are the two cities where the Liberal and NDP opposition have the most to gain, this is important to watch.
This is the first time that Abacus has reported on the situation in Alberta, and they are using a new method for this campaign. The polling firm usual uses an online panel, but this time they used the IVR method, as they will be doing throughout the Alberta campaign.

The Edmonton and Calgary numbers in this poll are generally what we've seen elsewhere, though the New Democrats are looking weaker in Edmonton than the norm and the Liberals are looking stronger in both cities. But the dead heat between the Tories and Wildrose in Calgary is nothing unusual.

The results in the rest of the province are drawn from smaller samples, but are also somewhat standard. The Tories are running low in northern Alberta, but some of their usual support may have been drawn off by that large "Another party" result. The New Democrats are looking surprisingly competitive outside the two main cities, while the Liberals and Wildrose are both doing best in southern Alberta, as should be expected. That the Tories are doing so badly, however, is somewhat of a surprise and a potential problem for Alison Redford.

If we meld the three regional results together, we get 33% for Wildrose, 30% for the Tories, 16% for the NDP, and 15% for the Liberals in the Rest of Alberta. That means a big drop in support for the PCs to the benefit of all three other parties.

This could be due to Redford's negative favourability numbers. Abacus found that 29% of Albertans have a favourable opinion of the PC Leader, compared to 31% who have unfavourable opinion. Another 31% were neutral.

Raj Sherman of the Liberals and Brian Mason of the NDP also had net negative numbers, with Sherman scoring 22% favourability to 27% unfavourability and Mason splitting 20% to 25%. Only Wildrose Leader Danielle Smith had net positive numbers: 31% favourable to 23% unfavourable.

Smith also scored best among her supporters: 77% of Wildrose voters have a favourable view of her compared to only 3% who don't. Mason also does well among his supporters with 68% favourability, but Redford (57%) and Sherman (56%) should be doing better within their own voting pool.

That is not to say that Redford and Sherman are seen unfavourably by their own supporters. No party leader had a higher than 5% unfavourability rating among their own voters. But their lower favourability results could have a negative effect on the enthusiasm of their voters.

This Abacus poll would result in a much larger opposition than my current projection forecasts, with 22 Wildrose MLAs, six Liberals, and five New Democrats. But the Progressive Conservatives would still win a majority with 54 seats.

The Tories dominate in Edmonton with 22 seats to four for the NDP and three for the Liberals, while they also do well in the Rest of Alberta: 18 PCs to 10 Wildrose, two Liberals, and one NDP.

But the real contest is in Calgary, where the Progressive Conservatives win only 14 seats. Wildrose wins 12 and the Liberals one, the governing and opposition parties splitting the city in half.

Abacus and, by extension, ThreeHundredEight's projection model show a much closer race than might have been expected only a few months ago. But the Progressive Conservatives still hold a huge advantage, even with a lead of only five points. Wildrose is, so far, unable to break into Edmonton where one-third of the legislature's seats are located, and as neither the Liberals nor the NDP can alone corner the centre-left opposition vote, the Tories are able to cruise well over 20 seats in the capital. That makes the competition in Calgary and the Rest of Alberta less important than it might be, as the Tories can get away with winning little over one-third of seats in these parts of the province and still form a majority government.

Though the polling volatility has somewhat calmed, there is enough of it that, coupled with the uncertainty programmed into the model, the possibility for a Wildrose minority does exist. It is very remote, however, as the chances of a PC victory are something along the lines of 87% at this stage. But it can also go very wrong for Wildrose, as their is a slim chance that the Liberals could hold on to the Official Opposition role by the skin of their teeth over Danielle Smith. And to complicate matters further, the Liberals could still fall behind the NDP to sit as the fourth party in the legislature, if they sit at all.

In other words, there is still a lot that has to play out and the order of the four major parties is far from certain. And with the news recently turning sour for the Progressive Conservatives (Gary Mar's fundraising, Alison Redford's pay for a non-sitting committee), the race could become even closer.


  1. By now it is obvious the PCs have been in government for too long. In three years, they will become the longest serving government of Canada with 44 consecutive years in office (compared to the NS Liberals between 1882-1925). However, it is doubtful that a majority government will form in this election given all the vote splitting and the drop in PC support everywhere to all parties.

  2. Why do you project the NDP at 12.7% when they have been higher than that in every single solitary poll that has been released in the past year. Like clockwork they are 13 or 14 or 15%.

    1. As explained in the link in the right hand column about Vote Projection Methodology, I have moved away from only averaging out the polls. I have also added an adjustment based on expected differences between voting intentions and voting behaviour, based on an analysis of voting vs. poll results in 15 recent federal and provincial elections.

      In short, the analysis indicated that the first and second parties in the legislature tend to be under-estimated by polls while third and fourth parties are over-estimated. Parties without a seat in the legislature (i.e., the Greens) are over-estimated to a significant degree.

      This adjustment based on position in the legislature acts as a proxy adjustment for organization, fundraising, and voter enthusiasm.

      To use Alberta as an example, both the Tories and Liberals were under-estimated in polls for the 2008 election, while the NDP and WRA were over-estimated.

      The average of NDP support in polls is, as you say, between 13% and 14% (not 15%, they haven't been higher than 14% since October 2011). Averaging out the six polls in the projection gives the NDP an average of 13.3% support.

      But I am projecting that they will actually get 12.7% on voting day.

      In any case, if we are rounding these off they both come to 13%.

  3. Is West Yellowhead considered part of the "North" region? 'Cause I wonder if that 12% spike for others is support for Glenn Taylor. West Yellowhead's been fertile ground for the opposition parties in past.

    1. I'm not sure. It is more likely central, but in any case one riding shouldn't account for such a spike.

    2. Yah, at best it would be ~3% I think? (50% above the mean in one out of 15 ridings). IIRC though this wasn't the only poll with a high-ish others number for the north.

  4. Redford should have called an election as soon as she had that public fight with Dalton McGuinty. That was her best chance to win.

    Every day she waits she increases the risk of becoming the next Harry Strom (the last Alberta Premier to lose an election).

    1. I partially agree, however Redford had pledged to pass the budget in Leg before going to the polls. For better or worse, she has been remarkably straightforward.

      You're right about the baggage weighing on the party, pre-campaign, but I would note that the incumbent parties tend to poll lower during session. I expect to see more positive polls once the writ is dropped.

      It will definitely be interesting/revealing to look at the polls DURING the campaign.


  5. "It will definitely be interesting/revealing to look at the polls DURING the campaign."

    Which is WHEN ???? Damned if I can find a date anywhere here !!

    1. The election has not been called yet. Redford has to schedule the vote before May 31, which means she must call it by the beginning of May, but people expect the call to come in the next two weeks, with the election then 28 days later.

    2. Peter,

      According to the Premier, the election will be called immediately following the budget. I think this is important since it allows the public to vote on the budget (approve or reject it), with advanced debate in Legislature.

      Further to Eric's comment, conventional wisdom (and most AB journalists) are suggesting that the election will be April 23rd or 30th. Given that the budget is likely to pass soon, my money would be on the 23rd.


  6. Thanks Eric. That makes a lot more sense.

  7. Dear Santa,

    All I really want for Xmass is for the PCs to lose power. 40 years is long enough.


    Wild Rose Supporter

  8. A question and maybe you've answered it but how do you factor in age based voting patterns with turn out on election day for a projection? What I mean is that older people are more likely to vote right of centre while younger, especially university students, are more likely to vote left of centre. Actually I should say support rather than vote. On election day though the numbers shift a bit in elections from the polls as more older people vote than younger people. Younger voters have the lowest turn out rates. Do you account for that or have you tried?

    1. Not directly, but adjustments are made to the polling average based on past divergences between poll results and election results, which would include the differences in turnout of difference age groups.

  9. Hi Eric,

    Quick question re: polling. I found this criticism of the latest Sun poll:

    "For example, the regional breakdown of party support includes only 81 respondents surveyed from southern Alberta, which results in an unreliably high 11% margin of error (via @calgarygrit). The optimistic results for the Wildrose Party, which already receives daily enthusiastic editorial support from right-wing Sun media, leads me to take with a grain of salt any political polling produced by this media network.

    Most legitimate polls, including those conduced by Leger Marketing and Environics, have shown the Tories with 45%-55% support province-wide and the three main opposition parties – Wildrose, NDP, and Liberals – grouped together in the mid-teens. A number of recent polls, including one conducted by Return on Insight, have produced results suggesting that the Wildrose Party has begun to break from the pack of opposition parties, which is not unbelievable at this point."

    This was from Alberta blogger Dave Cournoyer. Mr. Cournoyer's blog is centre-left, so even this commentary has political bias. However, I do think he makes an astute point. Sun Media editorials are hugely aggressive to the incumbent Premier and are unabashedly pro-WR.

    Given the politicization of polls, is it possible for you account for perceived biases, or is this too subjective?


    PS- keep up the great work!

    1. Hi Tom,

      The model does, in an abstract way, take this sort of thing into account, as polls are weighted by the accuracy track record of the polling firm.

      But a few comments - in terms of the sample size, the small number of respondents in southern Alberta, for example, is about as large as what we usually see for Atlantic Canada in federal polls. It is somewhat unreliable, but it is perhaps more important to look at the "rest of Alberta" sample, which is a more robust 262, and the data set that was used for this projection.

      As for the politicization of polls, while editorial commentary can spin the results of a poll, I tend to trust pollsters themselves. Pollsters would not be in the business very long if they were cooking their data, there are just too many ways to get caught. Pollsters aren't motivated by the idea of providing faulty data, they are motivated by the work that they do.

  10. Hi Eric, a question on your methodology here for individual ridings. I didn't see the answer to this one on your seat projection methodology section.

    What accounts for serious variations in areas with similar demographics, such as Calgary-Varsity and Calgary-West? As far as voters go, both of the ridings have similar demographics. In Calgary-Varsity, the PCs have a lead over the Liberals and the Wildrosers (the retiring MLA is a Liberal). In Calgary-West, the Wildrosers have a lead over the PCs, and the Liberals are far behind. In the previous election, the Wildrosers were a non-factor in both areas and the Liberals placed a strong first or second between both. There is no star candidate in either riding.

    How do you arrive at conclusions so different for those two ridings? There are a few other examples, but this is the first one that caught my eye on the seat projection.

    Merci, Ben

    1. Ben,

      Several factors are at play in these two ridings, and the same goes for other cases.

      The most basic reason is that the election results of 2008 were quite different. In Calgary-Varsity (adjusting for the boundary changes, of course), the results were 47% LIB, 37% PC, and 7% WRP. In Calgary-West, they were 49% PC, 32% LIB, 13% WRP.

      Already, those are very different. CV is more Liberal (but the Liberals have lost a lot of support in Calgary) and CW is more Tory. But CW also had an almost twice as large WRP result as CV, which explains why the WRP can jump to 27% in CV and 43% in CW.

      Also, in both of these cases there is a retiring incumbent. In CV, that further reduces Liberal support, making it possible for the PCs to come out ahead and makes more votes available for Wildrose to gobble up. Though it is simply happening because of the math, one can imagine that Liberals are going to the Tories in this riding and Tories are going to Wildrose in some degree.

      In Calgary-West, the retirement of the PC incumbent means that their vote is more depressed, allowing WRP to gain even more of an advantage.

      Demographically the two ridings might be similar, but politically they are not.

  11. I have lived in Alberta all my life, and I can tell you, that both me and many of my co-workers and friends, have had enough of the Tories.
    This government has become stale, and outlived their usefulness and my patience. I could handle Klein, hell I kinda liked him, he told it as it was. With this current Redford government its just one thing after another, the phantom commitee's,a minister bullying a school board-because they arent being "nice" to the Tory's, telling Albertans that doctor bullying doesnt occur, but then suddenly admits its a problem, the Mar suspension......when does it end? If you MLA's pulled this crap with a private employer, you would be fired.

    I guess honesty and integrity is expecting too much from our elected representatives.

    I want my government to be dynamic, and foremost accountable, to those that elected them. I will never vote Liberal (after the NEP), thats why I will putting a Wildrose sign on my front lawn.

    Sorry if my comments sound "redneck" Allison, but you can now remove me from the Tory Christmas card list.


    1. Preach on, brother S!

      40 years is long enough!

      Wild Rose Supporter


COMMENT MODERATION POLICY - Please be respectful when commenting. If choosing to remain anonymous, please sign your comment with some sort of pseudonym to avoid confusion. Please do not use any derogatory terms for fellow commenters, parties, or politicians. Inflammatory and overly partisan comments will not be posted. PLEASE KEEP DISCUSSION ON TOPIC.

Note: Only a member of this blog may post a comment.