Monday, January 11, 2016

PCs lead comfortably in new Manitoba poll

A new poll from Mainstreet Reseach for Postmedia shows the Progressive Conservatives with a wide lead in Manitoba, settling a bit of a dispute from recent polls out of the province as to whether the Liberals really are making a serious move. While their support is the highest it has been in decades, they still do not appear to be in a position to seriously challenge the Tories — at least yet.

The PCs led in the poll with 44%, followed by the Liberals at 27% and the NDP at 23%. This puts Mainstreet in line with the Probe Research poll released over the holidays that had the split as 43-29-22. The Insightrix poll putting the gap between the PCs and Liberals at three points seems to be the odd poll out.

Another 6% supported the Greens, and 24% was undecided. (The infographic above is courtesy of Mainstreet.)

The single-election projection model (the three-election model is still under construction) would deliver 41 seats to the PCs with these numbers, with the Liberals winning nine seats and the NDP taking seven.

The strength for the PCs would seem to be largely due to the split between the Liberals and NDP. The PCs, recall, took exactly 44% of the vote in the 2011 election. The party looks poised to virtually sweep the rural parts of Manitoba, and benefit from a split in Winnipeg. Like the Probe poll, Mainstreet shows the Tories ahead of the other two parties by a much less significant margin in Winnipeg than it does province wide: 37% to 28% for the NDP and 27% for the Liberals. But it should be enough to tip the balance quite strongly in favour of the PCs.

I talked about the political situation in Manitoba (prior to the release of this Mainstreet poll) with provincial columnist Deveryn Ross on the latest episode of the Pollcast. You can listen to it here.

15 comments:

  1. The contention seems to mostly be related to within Winnipeg... the Insightrix poll showed Bokhari far away in first while this one has the Liberals in third. This discrepancy is not a good look for the pollsters...

    ReplyDelete
    Replies
    1. It's not good for the pollster who is wrong.

      Delete
  2. While the Mainstreet decided voters data is similar to Probe's December numbers, Mainstreet reveal that they polled 25% of voters who were Undecided. Probe do not seem to record undecideds. While some polled as undecideds may be non-voters, one might expect the overwhelming majority of these voters to be undecided between NDP or Liberal. I would be hesitant about concluding that Probe and Mainstreet are singing from the same songsheet until I better understood their varying polling methods.
    @Eric: Can you shed any light on this subject?
    Another point to be careful about is the 8% Green poll in Winnipeg. In 2011 the Greens only ran candidates in 70% of Winnipeg ridings and there is little indication that they will match that figure in 2016. Therefore you can assume that a quarter to a third of that Green support will be forced to go elsewhere, probably Liberal or NDP.

    ReplyDelete
    Replies
    1. Probe also has undecideds, but did not report what the number was. Their previous poll had undecideds at 18%. I don't expect this to have caused any differences.

      Delete
    2. You may be right though Probe don't seem to report these publicly.
      A 7% point difference on undecideds is not insignificant.

      Delete
  3. Another way to compare a difference between Probe and Mainstreet is to look at the last time Mainstreet polled on this subject. Back in November 2014 they ran a poll that gave the PCs 53%, 4 points more than any other poll taken since the 2011 election. That figure was 11 points up on Probe's poll of the previous month and 5 points up on what Probe had them at a month later. Perhaps Mainstreet January 2016 numbers giving the PCs a comfortable lead should be treated with caution.

    ReplyDelete
    Replies
    1. Mainstreet also consistently had higher polling averages in the federal and Alberta elections as well sometimes extremely off every other polling firm. They were the only poll that had the PC's leading in Alberta, when every other poll had Notley breaking into a large lead in the crucial first polling after the leadership debate. They were actually 10 - 15 points off in that poll, and in retrospect, it's quite embarrassing. I have yet to find a poll in any province where they had the Liberals ahead of a right wing party compared to other polling firms. (There may be one, but out of the 50 I've seen so far, not one) This is type of polling that makes you question if there is something else in play with this company. It also doesn't help polling's reputation in general.

      Delete
    2. I think you may be mis-remembering the Alberta election, as I don't see what you are saying in terms of Mainstreet's polling.

      Delete
    3. Mainstreet was the last pollster to show a Wildrose lead, but they saw the Tory collapse at about the same time as everyone else.

      Might I suggest someone check a source before making factual claims?

      https://en.wikipedia.org/wiki/Alberta_general_election,_2015#Opinion_polls

      Delete
    4. @Fred G I think their 'bad' Alberta poll was based on fieldwork done on the day of the debate, so may not have reflected it.
      However, your point about their federal polls showing the Tories doing much better than other pollsters did is a valid one, which begs the question "Have they changed their polling methods since October to iron out the problem?"

      Delete
    5. I apologize Fred; I have just read the report and the fieldwork for that Alberta Mainstreet poll was conducted immediately after the debate, which means you are not only right but have a good memory.

      Delete
    6. Thank you Greg. Eric I got the information from your website, and then read the report from there. Yes, Mainstreet's Poll was completely off, helping out the PC's by a huge margin compared to The Forum Poll taken around the same time no less. And yes, Ira, we have to question what's happening there if it's consistently happening with the same company.


      Even in Ontario polling lately, the PC's have been polling higher there too in Mainstreet's polling data. I have been noticing this for a while now, but now it's becoming quite distinct in the last few months. And now there's this in Manitoba. The main problem is this: What happens if their poll said yes, the Liberals are very close to the Conservatives right now. Then we have a completely different outlook on the election, and maybe national news because there are two polls confirming this. That's why looking into poll bias is so important - even if the bias is just the way the poll is taken. It hurts all pollsters reputations, even if it's just a perception problem.

      You would know polling methods better than I do Eric, what are the chances that they are always the highest pollster with the right wing parties, especially at crucial points for polling in Alberta and Federal elections? Or is there something in how the questions are asked or the type of people they poll?

      Delete
    7. House effects like this do exist. For years, EKOS had the Greens way higher than any other pollster did, and always at the expense of the Liberals.

      At the same time, Nanos always showed very low Green results, likely because Nanos they used live callers who didn't prompt respondents with options, so responses were more likely to be for top-of-mind parties.

      Delete
    8. At 538, they do a correction for "House Effect" when a firms polling consistently is off base from the pack and/or actual results. They also are much more transparent on the ranking of relative polling agencies and their historic performance.

      Delete
  4. From the company media release that accompanied the poll; “It’s unexpected that the Liberals are doing so well” said Quito Maggi, President of Mainstreet Research.
    Firstly if I had a polling company I would never publicly admit to being surprised at what I had found out, as to do so conveys an impression that I had a predetermined idea of what I would find out. A predetermined impression could also be described as a bias which is not a good attribute for a pollster.
    Secondly, you have to wonder about the understanding of a pollster who overlooks the fact that because of the Trudeau Government, the Liberal brand is on the up at the moment. Even if that slipped by unnoticed, surely he would have been aware that there had just been two other polls conducted that showed the Liberals "doing so well" or better?
    Very odd.

    ReplyDelete

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.