Friday, February 26, 2016

Week in Polls: Manitoba PCs hold lead, NDP down nationally, margin narrows in Ontario

The latest poll out of Manitoba, which will vote on April 19, shows the Progressive Conservatives under Brian Pallister continuing to hold a wide lead over the governing New Democrats and surprising Liberals.

Mainstreet Manitoba poll
The new poll from Mainstreet Research put the PCs at 50%, down just one point from Mainstreet's previous poll of February 11. Rana Bokhari's Liberals were up three points to 23%, while Greg Selinger's New Democrats were unchanged at 21%. The Greens followed with 6% support, down one point.

These numbers have been holding steady for some time, as Mainstreet has pegged the three parties within a tight band of support over their last three polls: 50% to 52% for the Tories, 20% to 23% for the Liberals, and 20% to 21% for the NDP.

With half the vote and against a divided field, the PCs would win a big majority government with these numbers. And that should come as no surprise: in addition to their 27-point lead in the 'rest of Manitoba', they are also up 18 points in Winnipeg.

Seat and vote projections for the Manitoba election will be launching soon, likely after next week.

Liberals dominating in Forum national poll

New federal numbers from Forum Research show the Liberals with a big national lead as well as some impressive numbers at the regional level. The party suffering, though, is not the Conservatives, but rather the New Democrats.

Forum federal poll
The Liberals led with 49%, up three points from Forum's previous survey of December 6-8. The Conservatives were unchanged at 32%, which matches their election result, while the New Democrats were down three points to 10%.

That is half of the vote they took on October 19.

The Liberals put up some massive leads in this poll: 15 points in British Columbia, 16 points in Ontario, 37 points in Quebec, and 50 points in Atlantic Canada. The Conservatives were only ahead, though by 34 points, in Alberta, while the Prairies were a toss-up.

This is a very bad poll for the New Democrats, who ranked fourth in Quebec with just 11% (behind the Bloc, which was down to 13%). The best result for the NDP in this poll was in British Columbia, and there it only scored 14%.

Margin narrows in Ontario

Despite polls showing Kathleen Wynne being one of Canada's least popular premiers, her party is still running competitively with Patrick Brown's Progressive Conservatives in Ontario.
The new poll from Mainstreet Research (which quickly seems to be supplanting Forum as Canada's busiest pollster) puts the Tories at 36%, down four points from Mainstreet's previous poll of November 1. The Liberals were up five points to 33%, while the NDP was up one point to 26%.

These numbers match quite closely to a Forum poll from the end of December.

Regionally, the province is quite divided. Only in Southwestern Ontario, where the Tories are up by 10 points, and in Toronto, where the Liberals are up by six, does any party have a robust lead. Elsewhere, the margins are much closer: a four-point PC lead in Eastern Ontario, leads of two points for the Liberals in Northern Ontario and the Hamilton/Niagara region, and a lead of just one for the PCs in the 905 area code.

Lethbridge poll provides window on Alberta

Polling by Lethbridge College shows some interesting results in the city of Lethbridge, providing a bit of a glimpse on where things stand in the province today.

At the provincial level, the poll found the Progressive Conservatives leading in Lethbridge with 33% support, followed by the New Democrats at 26%, Wildrose at 22%, and the Liberals at 12%.

Compared with the provincial election results in the two ridings of Lethbridge East and Lethbridge West, this represents a slide of between 26 to 31 points for the NDP — a huge number. All the other parties have taken advantage, though primarily the PCs. They were up between nine and 12 points, the Liberals up between six and seven points, and Wildrose between three and five points.

The NDP also seems to be suffering at the federal level in Lethbridge. The poll put the Conservatives ahead with 50%, followed by the Liberals at 31% and the NDP at 8%.

The boundaries are not contiguous, but to put these numbers into perspective the Conservatives took 57% of the vote in the federal riding of Lethbridge in October. The Liberals took 19% and the NDP took 21%, suggesting that the federal NDP has also taken a tumble in the region.

Trump, Clinton heavily favoured in Super Tuesday

About a quarter of delegates for the Republican and Democratic primaries will be handed out over the next week, and Donald Trump and Hillary Clinton are poised to dominate.

Our go-to site, FiveThirtyEight, gives Trump some big leads in some of the states voting on Tuesday. According to their "polls-plus" forecast, Trump is up by 5.5 points in Virginia, 10 points in Georgia, 10.5 points in Alabama, and 25 points in Massachusetts. Oklahoma is a toss-up while Ted Cruz is up by 13 points in Texas.

On the Democratic side, Clinton is up by eight points in Massachusetts, nine points in Oklahoma, 32 points in Virginia, 33 points in Arkansas, 35 points in Tennessee, 36 points in Texas, and 49 points in Georgia. She is also ahead by 39 points in this weekend's primary in South Carolina.

Only in Vermont is Bernie Sanders forecast to win on Tuesday — by 74 points!

At the national level, there hasn't been much change in the last week. Trump leads with 35%, followed by Cruz at 19%, Marco Rubio at 16%, and Ben Carson and John Kasich at 8% apiece. For the Democrats, Clinton is ahead with 49% to 39% for Sanders.

My CBC articles this week


  1. It is odd that Mainstreet would choose to poll Manitobans only 9 days since their last fieldwork. This far out from election day, one would not expect to find any change in support in such a short period. Yet Mainstreet have the Libs up 3 and the Tories and Greens down 1. I don't think you can simply dismiss this as margin of error. Though the Tory lead is still huge, the gap between them and the Liberals is down by 4 points, which is some narrowing over just 9 days of a pre-campaign period.

    @Eric: Some pollsters say they also ask people who they voted for last time, and then weight their results accordingly. In practice, how would this weighting work? Also, do you think it is going to make much difference to the accuracy of a pollsters work?

    1. I don't see why it is odd for Mainstreet to be polling frequently - the election call is two weeks away.

      The margin of error for the Liberals moving from 20% to 23% in these two polls is 2.8 points, so it is not particularly significant.

      Weighting by past vote would be to calibrate your sample according to the previous election's results. I don't think that is a very good idea, since memories can be inaccurate (confusing with a federal election, for sample) and people move.

  2. The fieldwork was from 20th Feb and the earliest possible day for the election to formally start is the 15th March, so that was 23 days, just over 3 weeks before the election campaign. I might expect pollsters to publish as frequently as weekly in an election campaign, but 3-4 weeks before that campaign has started?

    Surely, the margin of error for a party moving from 20% to 23% in two polls with a 9 day fieldwork gap, is far more significant than if the fieldwork was a month apart.

    I suppose if you ask the historic question you at least give yourself the opportunity to decide if your sample was off

  3. That Lethbridge poll is pretty interesting.

    The NDP literally swept the area in 2015 with a huge win by Phillips and Fitzpatrick tagging along for fun - but with those numbers they would be lucky to hold on to either seat, split vote or not. If they can't hold on to a super-friendly small city like Lethbridge, then they could be looking at a wipe out outside of Edmonton and maybe, *maybe* a Calgary riding or two.

    But they've got until 2019 to turn around and there have been more impressive miracles. Still, has a majority government ever seen its support collapse this quickly? Notley's government isn't even a year old yet and they're trailing third in at least one poll, dropping by the dozens of percentage points in others. This cannot be a good sign.

    1. Both Christy Clark and Gordon Campbell experienced precipitous declines in support after their elections but, British Columbians are notoriously fickle.

      NDP governments do not do as well as other parties getting re-elected; Dave Barrett was a one-termer as was Darrel Dexter and Bob Rae. History is not on Notley's side although prairie NDP governments do have a better record at re-election.

    2. We all hope Mulcair will go gracefully-Although that does not appear to be his style. Old Tom has a bit of Diefenbaker in him methinks.

    3. More than a bit, he ran a conservative cmapaign promising fiscal responsibility and -oh so shocking to "progressives" balanced budgets. I finally understand why Brian Topp (who is not a politician per se) did so well during the leadership race, Mulcair is ideologically closer to John Turner Liberals than Tommy Douglas Dippers, the 2015 campaign demonstrated this in spades.

  4. Eric, I'm surprised you didn't mention this in your article on the Mainstreet Saskatchewan poll: the support levels of
    Sask: 52%
    NDP: 34%
    Liberal: 10%
    Green: 3%

    Is very similar to the 2007 Saskatchewan election result of
    Sask: 51%
    NDP: 37%
    Liberal 9%
    Green: 2%

    Even the regional results aren't that dissimilar with one or two major differences in the cases of Regina and rest of Saskatchewan explaining the differences.

    Unfortunately I don't have my sums of the 2007 regional results with me but I remember the 'rest of Saskatchewan' outside of Regina and Saskatoon were something like
    Sask Party 61% (57)
    NDP: 32% (30)
    Liberal: 6% (11)
    With the 2015 poll results in brackets. As you can see the best explanation for the slight difference is that the Liberal support is relatively much higher now in the 'ROS.'

    The numbers for Saskatoon are almost identical in both 2007 and 2009.

    In 2007 the NDP won the city of Regina by a fairly wide margin over the Sask Party however those results were skewed by the Sask Party candidate in Regina-Walsh Acres stepping down and his name being removed from the ballot. In the 2011 election the Sask Party received 3,679 votes in the riding or 58.2% of the vote. So this was not all that small a skew.

    Also, this Mainstreet poll shows the Green Party at 9% support in Regina vs. the 3% they received in 2007.

    As you know, in that election the Sask Party won 38 of the 58 seats and the NDP took the remaining 20 with the NDP losing another seven ridings by very narrow margins.

    Of course, for this election, the Sask Party will now benefit from the 'power of incumbency,' so, if the polls don't tighten further, I would expect the NDP to win less than 20 seats (even though the number of seats has increased by three.)

  5. It is interesting to compare the Green Party across the provinces of Saskatchewan and Manitoba. In Sask, the Greens were the first and so far only party to announce a full slate of 61 candidates, while in Manitoba, the Greens only have 5 candidates announced. Mainstreet polls the Sask Greens at 3-4% and the Manitoba Greens at 6-9%. Come April, in Sask one can expect the Greens to poll their full opinion poll figure. In Manitoba, that seems very unlikely to happen. During an election campaign, pollsters need to find some way of properly dealing with a party that is running far fewer than a full slate.

    1. The Saskatchewan NDP completed their full slate of candidates over the weekend.

      Eric, these are the numbers from the 2007 election. Keep in mind the totals for Regina and Saskatoon include one or two combined urban/rural ridings.

      There are still two or three combined urban/rural ridings, I don't know if Mainstreet included any sampling in those rural areas in 'rest of Saskatchewan' or part of the cities since that's who they'll be voting with.

      Votes 101,900
      Sask 43,569 42.76% 5 seats
      NDP: 42,410 41.62% 7 seats
      Lib: 12,987 12.74
      Green 2,376
      Other 558

      Total 93,041
      Sask 32,348 35.73% 3 (but 10 running)
      NDP: 44,236 47.54% 8
      Lib: 12,956 13.93
      Green 2,480
      Other 121

      Rest of Province
      Total 258,068
      Sask 153,854 59.62% 30
      NDP 82,058 31.80% 5
      Lib: 16,642 6.45%
      Green 4,272
      Other: 1,242

      The 'rest of Saskatchewan' includes small cities, the 2 northern ridings and rural/small town ridings.

      I count 7 ridings as being small city ridings because they are mainly urban. In addition to the two ridings each in Prince Albert and Moose Jaw, I also include The Battlefords (North Battleford), Swift Current and Yorkton as being small city urban ridings.

      These are the combined results in those seven ridings:

      Total 54,518
      Sask 26,522 48.65% 4
      NDP: 23,202 42.56% 3
      Lib: 3,617 6.63
      Green 909
      Other 268

      These are the results in the other 28 'rest of Saskatchewan ridings' including the two northern ridings that the NDP won.

      Total 203,550
      Sask: 127,332 62.56% 26
      NDP: 58,856 28.91% 2
      Lib: 13,025 6.40%
      Green: 3,363
      Other 974

    2. Why would they have to deal with it? Its Eric that must deal with it, since he's the one doing the projection.

      Besides, how could they deal with it? They're already using demographic and regional skews on their numbers, then asking to apply partisan ones would be a nightmare. Which Green respondents should be cut out and where should they go? Do they end up as DKs or do we ask them for their second party and then transfer the vote? How big of a sample size do you need to accurately reflect a candidate count anyway? On and on, it's impossible to do such a thing and expect accuracy.

  6. Mulcair needs to step down. With the numbers he is putting up even kind hearted Dippers will toss him out. He will likely achieve the minimum (50%+1) but that is far short of the ringing endorsement he needs. I predict Mully will receive 54% support at the leadership review vote.

  7. Eric, these aren't the final aggregate numbers for the 9 U.S presidential primaries held yesterday, but I think they're close enough. I got the results off of either CNN or the state election board websites at around 3 A.M Pacific Time last night. In most cases, the results had at least around 99% of precincts or counties reporting. Off course, none of the results were official.

    The only thing missing that would make a slight difference to the percentages is that I didn't compile the numbers of votes for the 'other' candidates (mostly candidates who have since 'suspended' their campaigns) for most of the states.

    There were 9 states that held primaries and both parties had two caucus states as well: Minnesota for both parties and Alaska for the Republicans and Colorado for the Democrats. Turnout in caucus states tends to be roughly 1/3 to 1/5 the turnout in primary states, so I don't think caucus voters are necessarily reflective of the voters as a whole.

    The only significance I give to any of the caucus vote is that since Donald Trump couldn't win the Alaska caucus even with Sarah Palins' endorsement, I think we can pretty much conclude that she is a spent force in American politics.

    Total Votes: 5,532,751
    Hillary Clinton: 3,410,641 61.64%
    Bernie Sanders: 2,076,449 37.53%

    Total Votes 8,212,931
    Donald Trump: 2,892,314 35.22%
    Ted Cruz: 2,443,413 29.75%
    Marco Rubio: 1,822,988 22.20%
    John Kasich: 531,676 6.47%
    Ben Carson: 479,122

    Total votes: 13,745,682 (My guess is that the actual final number of total votes will be around 14 million)

    Roughly 60% of voters cast ballots in the Republican Primaries. However, 7 of the 9 primaries were held in southern or or south western states which are mostly heavily Republican. The Democrats should be disappointed that the Republican primaries in the 'purplish' leaning 'red' state of Georgia as well as in the Commonwealth of Virginia which narrowly voted for Obama both times had higher turnout, but there are five Republicans stumping for votes versus just the two Democrats. (In Georgia, Ben Carson received nearly 10% of the vote, so he was slightly more than completely irrelevant there.)

    I think it is also at least somewhat noteworthy that Hillary Clinton received around 500,000 more votes yesterday than Donald Trump.

    In regards to comparing these results last night to the national primary polling, the Democratic numbers are completely skewed by so many states being held in states with large black populations who overwhelmingly support Hillary Clinton. There is nothing my simply summing of the primary votes can do about that. On the Republican side though, the only major skew I think is that one of the states holding a primary last night is Texas, Ted Cruz' home state. Not only did Cruz win Texas by about 15%, but roughly 1/3 of all Republican voters yesterday were in Texas, so their votes alter the percentages significantly (About half of all of Ted Cruz' votes yesterday came from Texas).

    Republican primary totals less Texas:

    Total votes: 5,493,524
    Donald Trump: 2,139,474 38.95%
    Marco Rubio: 1,323,450 24.09%
    Ted Cruz: 1,212,851 22.08%
    John Kasich: 7.50%
    Ben Carson: 6.60%

    Again, the percentage results for all the candidates are slightly higher than their actual vote percentages because the votes for 'others' aren't as complete.

    What we see though are results that are quite consistent with the national poll numbers: Trump at roughly 40%, Rubio and Cruz basically tied for second and Carson and Kasich roughly tied for 4th place but well back.

    1. With these numbers I would expect the NDP to win less than 10 seats. Democratic socialism is dead-defeated by the imcompetentence of the left worldwide but especially in France, Brazil, Jeremy Corbyn and labour and our own home grown NDP.

      Mulcair is pulling the NDP down but, the rot runs much deeper than one man. If he stays on the BC NDP will lose their fourth election in a row next year. The sad truth is the best and the brightest in Canada do not support the NDP.


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.