Site Administrator Of:

Supporter Of:


Interesting poll predictions for Ontario

If you plug the poll numbers from the Canadian Press/Decima Poll yesterday for Ontario into the Ontario Election Predictor that Hill and Knowlton are running, you get an interesting predicted seat count.

– Liberals with 41% popular vote would still according to the predictor get 66 seats – which is only 2 down from 2003 and almost 62% of the seats in the legislature.
– The PC’s at 33% would only gain 2 seats to 31 – 29% of the seat distribution.
– The NDP at 13% would stay the same at 10 seats. – 9.3% of the seats in the legislature.
– The “Other” which in this case would be the Greens at 11%, would gain 0 seats.

No guaranteee that would happen of course – the predictor does a uniform swing shift in seats and doesnt take into account specific ridings (where there might be a super popular candidate for 1 particular party as an example) but if it were to happen, this would just be another of your “quirks” in our lovely FPTP system we currently use here.

That tells me the predictor is thinking the Green surge in popularity – even if its sustained at this level to the election – is stealing votes from either NDP or PC disgruntled voters as a protest vote of sorts, and won’t be concentrated enough in any 1 riding to win seats.

UPDATE: On the other hand, the predictor appears to be setup so that it entirely discounts “other” parties from winning. I decided to ratchet up the “other” popular vote to 20%, and dropped the NDP to 10%. In that scenario, all the seat swings went to the Liberals. I’m sorry.. but if the Greens were to get 20% in a provincial election, as remote as that may seem, you can’t tell me they aren’t winning a single seat, even with this FPTP system we use for voting.


2 comments to Interesting poll predictions for Ontario

  • Hey Matt,
    We had some less than complete data to work with from Elections Ontario when we were re-building our 2003 database of previous polling information (which was onerous with the new boundaries in many ridings).  If you check out our 'how it works' page on the H&K Predictor site it should give you a better idea of how we dealt with missing polling/vote data for parties other than the NDP, PCs & Liberals.  Please feel free to contact me anytime through the site with you have any additional questions.

  • I wouldn't be surprised if the seat predictor just ignored "others", since how does it know if that 20% is the Green Party vote, or if it's 4% each for 5 different smaller parties.

unique visitors since the change to this site domain on Nov 12, 2008.