Friday, September 23, 2011

NDP may hold balance of power in Ontario

In the next Ontario legislature, the New Democrats could very well hold the balance of power.

With that comes both opportunity and risk for the party.

You can read the rest of the article on The Huffington Post Canada website here.

It has certainly been a very light polling week. There has been nothing since Monday. I suspect we'll get a flurry of polls soon, however, especially with the Ontario debate scheduled for Tuesday. Manitoba might have to wait until the last week or so of the campaign before we get any indication of what support looks like in that province. I've put a caution sign with the projection at the top of the page, because three months is a long time.

But there was one Ontario riding poll that appeared recently, in Peterborough This Week. It's a little short on the details, but it was apparently conducted by telephone sometime in the week of September 11 to September 17 for the riding of Peterborough, surveying 629 people.

The poll found that Liberal MPP Jeff Leal is on track to be re-elected with 45.5% of the vote, well ahead of PC candidate Alan Wilson, who clocked in at 28.5%. The NDP's Dave Nickle came up third with 22.0%.
This isn't much different from the 2007 result in the riding, where Leal took 47.8%, the PCs 25.6%, and the NDP 16.6%. This is Nickle's fourth kick at the can for the NDP.

It also isn't too different from the current projection for the riding, which has the Liberals with 40% of the vote to 29% for the PCs and 24% for the NDP, so that is a positive sign for the projection.

I've updated the riding projection image, as the candidate information for Ontario, Manitoba, and PEI should now be up to date.


  1. I lived in Peterborough for four years, and that riding poll sounds about right to me. The Ontario Liberal organization is strong there, and both the Liberals and New Democrats will benefit in the vote from the fact that Trent University is in session and the students are in town.

  2. I haven't been polled for many years, but surprisingly already twice in the last two weeks I got automatic pollster messages in my answering machine (I live in Hamilton). I wonder if these are Ontario-wide polls, or my local riding polls (like the one reported for Peterborough)?

    Also, the fact that both times there was no one available to get the phone suggests that they are calling at bad time (during business hours), so probably get pretty biased picture?

  3. Not necessarily, they were probably polling after working hours as well. People who work from home, housewives/husbands, people on sick leave, the unemployed, etc. all vote as well!

    I received a polling call at around 7:30 PM earlier this week, so I expect an Ontario-wide poll soon.

  4. I'm in Peterborough. Nickle received 25% in the last federal election too so I can't imagine him getting anything less than that and probably a good deal more.

  5. Bit curious about the projection for Ottawa Center in the Ontario provincial election. I haven't personally seen any formal polling done for the riding to suggest the NDP have a large lead here. In spite of the federal Paul Dewar landslide the liberal MPP Yasir Naqvi is a pretty strong incumbent. By all accounts the riding is actually going to be pretty close from what I've heard.

    By the way I got a call from forum research the other night, when I was out at the office no less. Not sure why they called the office number.

  6. Actually the Forum poll is out and it is massive, having over 40,000 respondents.

    The results are amazingly close to Eris's (to give credit where it is due), with PC and Liberals at 35% each and NDP at 23%. Their riding projections are 47 seats for the first two and 13 seats for the NDP although some riding projections are very strange.

    Looking forward to how Eric will integrate this data.

  7. Looking at the seat projection for PEI and it scares me. For years AB had little opposition and it was not a great government. You need an opposition.

  8. The Forum poll is unprecedented! I wonder why Eric's model predicts comfortable Liberal minority, even with 2% lower popular vote, whereas the Forum's model has identical (fifty-fifty) prediction for both popular support and seat numbers?

  9. I think Forum used their riding polls to come up with seat wins. I think the report said 17 seats were within the MOE, so Forum's results don't contradict my projection.

  10. The Star report on the poll says 28 ridings were within the MOE, so Eric's projection is definitely not contradicted.

  11. Also, I'd note they have some results that are clearly wrong. They have the NDP losing Parkdale - High Park and that just isn't going to happen. They have Elizabeth Whitmer losing her seat for the PCs. Another thing that just isn't going to happen. They have the Liberals losing St. Catherines, when that seat would probably be around the fiftieth that the Liberals would lose.

    Those riding level results really need to be taken with a grain of salt.

  12. Well, when one riding is within one MOE, it's obviously nothing to talk about. If 17 (or even 28) ridings are within MOE but all differ IN THE SAME DIRECTION (either pro-Liberal or pro-Conservative), this is definitely something to worry about. Assuming of course that the random sampling error is larger than any systematic error inherent in the model or survey.

    Eric - what are you going to do about this poll? If you add it to your model, using the usual rule of giving more weight to larger sample, then your prediction will be pretty much only based on the Forum poll, until the election day (unless your model completely discards polls older than say 1 week). This is because this one poll is probably 10x larger than the total of all the rest of polls.

    You probably wouldn't want to do that - because of the unknown systematic error of this survey, which does not go down with increasing sample size.

  13. T.S.
    Yeah I find the Parkdale-High Park prediction strange. DiNovo is really popular and the Liberal is an unknown-to-the-community parachute from a tony Toronto neighbourhood. To suggest that she would win by more than 5% makes it even more unbelievable. The signs in the riding, for what its worth, doesn't suggest such a movement, although who knows after the May 2nd Quebec phenomenon. I took part in that Forum poll, and was asked who I would vote for, and who I voted for last provincial and last federal. It would be interesting to get those numbers, especially for P-HP, to see if there is a trend away from the NDP.

  14. First,

    All 28 ridings that are within the MOE don't have to go the same way, the difference between the Forum seat total and my seat projection does not require all 28 seats to go the same way. And, of course, that isn't what has to happen to all. Seats can go to and fro, some polled to go Liberal going Tory, others polled to go Tory going Liberal, etc., only the net result is important.

    This poll isn't the be all and end all. Nowhere near every riding poll ends up being correct, especially two weeks out, and Forum was not the most accurate firm in the 2011 federal campaign.

    Sample size weighting is based on the MOE, not the total sample size. For example, this poll has an MOE of 0.5%, which would normally give it six times the weight of a standard poll of 1,000 people, which has an MOE of 3.1%. This poll isn't 40 times as accurate because the sample size is 40 times larger.

    There is a cap on MOE weighting, though, as it makes no sense for the sample size to be six times as important as, say, the accuracy rating of the firm or the age of the poll.

  15. For whatever reason, during the federal election riding polls generally tended to be quite inaccurate. There were riding polls in Quebec just days before the election that had the NDP wayyy back in ridings they ended up winning easily!

  16. A small correction - I didn't suggest that you'd give 10x more weight to this poll (I said "more weight"). And I am with you that there should be a cap on MOE when used in a prediction model, because of the unknown (and perhaps substantial) systematic bias of an individual survey.

  17. Anyone know where the full report for that poll can be found?

  18. BTW Eric,

    Were you planning on doing something on this poll? (

    I'm curious how the actual seat totals change under each leader...

  19. And what about that massive poll in today's Star ???

    40,000+ polled. The result?? A dead tie PC-Lib !!

    Let's get up to date here !!!

    And if you think the NDP will support Hudak you're smoking the wrong stuff !!!

  20. Peter is right !

    NDP means NO CHANGE, more McGuinty !

    Time for Hudak to crank up the coalition boogeyman ! Just like in the federal election its time to get the Liberals all tripped up on where they stand.

  21. Say that Forum poll is nearly accurate, and the Liberals are re-elected (even to a minority) while Dalton McGuinty loses Ottawa South, what happens?

  22. I believe one of the Liberal MPPs would offer to step down so that McGuinty would be able to run and get elected.

  23. After reading the comprehensive riding-by-riding projections in the Star, I think there are three obvious conclusions:

    1. PCs are not going to be elected in Toronto (Eglinton-Lawrence and Don Valley West are the 3rd and 4th strongest ridings for the Liberals in Ontario, and not the seats the PCs will likely take) and they will have a tough time taking seats from the Liberals in the GTA.

    2. If Dalton McGuinty wants a majority government, he'll have to recover lots of lost support in Southwestern and Northern Ontario.

    3. The NDP vote must be just as inefficient as the PC vote because they have 23% support but are once again limited to 13 seats in Toronto, Hamilton & Niagara and Northern Ontario.

    But on election day, Dalton McGuinty will prevail once more because Tim Hudak won't upload the costs that Mike Harris downloaded to municipalities and the fact that more prominent figures in society (such as David Suzuki and Hazel McCallion) believe the Liberal platform will address not only the short-term (next 4 years) issues, but also the long term issues (such as healthcare and the environment) that would extend far past McGuinty's rule as Premier.


  24. If McGuinty loses his Ottawa South (which is hard to believe), then I think he will step down as premier and we will go into a leadership race. McGuinty is not a popular leader, and the Liberals are only strong due to a weak opposition from the Tories.

    I'd predict a quick leadership race where Dwight Duncan or Chris Bentley becomes Liberal leader and in turn, premier.

    Or perhaps the NDP would prop up the Tories in a minority government (stupid move, but can happen), claiming that the voters do not want any more of McGuinty.

    - Maple

  25. But what if the Liberals did generally well in the election, say won a slim majority, and McGuinty was still defeated? I don't think he would step down as leader in such a situation.

  26. In the TOTALLY unlikely event that McGuinty lost his seat AND there was anything other than a Tory majority government - it would depend on a lot of things. If the Liberals had the most seats - I suspect that McGuinty would get someone to step down so he could come in in a byelection. If the Tories had the most seats but no majority and the Liberals were tempted to try to stay in power on the basis of the being the incumbent party would complicate matters enormously if the incumbent premier had no seat.

    But this is all academic. The only way I can see McGuinty losing in Ottawa South would be if there was a huge PC landslide - in which case he would be out of the picture regardless.

  27. If that poll is right, such is not academic, since it would appear the Liberals have collapsed outside the GTA yet solidified within and the PC's have collapsed there. Unfortunately for Hudak, he would have to pretty much run the table in the rest of Ontario and win over 80% of the remaining seats - including a historic breakthrough in northern Ontario - to win a majority with the current numbers in the GTA and adjusting for PC dead zones like Windsor and downtown Ottawa.

    Why is there such an enormous divide between the GTA and the remainder of Ontario?

  28. I think the "foreign workers" thing was a bit of wedge issue between the GTA and the rest of Ontario personally, Craig.

  29. I think the reason is is that the GTA is mostly an urban, and yes, there are many immigrants living and/or working in this area. But besides the GTA and some scattered cities like Ottawa and Windsor and London, most of the ridings in Ontario are rural and and many are highly religious conservative areas.

    I'll be Captain Obvious for a moment and say that this is the reason why you rarely see conservative parties break into cities, and you rarely see more left leaning parties break into rural areas.

    For these reasons, Hudak has his work cut out for him to win a majority. He could still win a minority but I think that it will be a close election. I am sure that this election will result in a minority for Hudak or McGuinty

  30. Hi Eric,
    I have a special request. I have searching for the forum research poll data online to no avail for the whole day. It is not posted on their website (maybe it is, but i can't find it for sure) and nowhere online. All i can find is news articles about the poll with a handful of ridings (where they don't mention numbers, but only winner or loser). Is there a special trick for searching?

  31. Nazar,

    Nope, I have as much as you do!

  32. Forum posts it on their webpage but you have to pay through the nose to actually see it apparently. Odd given how all their competitors are so much more open.

  33. Thanks Eric, at least you proved that I'm sane :)

  34. Eric,

    I've got a link for you of GTA ridings and their results from that Forum poll. I dunnp if you can use it to tweak things or not, but it's interesting to see nevertheless:


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.