Forum has come out with a second poll of the by-elections, here are the results. My one major concern about all their polls has been the very large over sampling of the older population. If I had the time I would try and work out what sort of impact this has on the polls and how it compares to the way people vote.
Victoria Oct 26th Nov 12th change
Murray Rankin NDP 47% 47% 0
Donald Galloway Green 20% 26% +6
Paul Summerville Liberal 16% 14% -2
Dale Gann Conservative 16% 12% -4
(sample size Oct 26th 295, Nov 12th 509)
Calgary Centre Aug 14th Oct 26th Nov 12th change Oct-Nov
Joan Crockatt Cons 44% 48% 32% -16
Harvey Locke Liberal 21% 28% 30% +2
Chris Turner Green 12% 11% 23% +12
Dan Meades NDP 14% 8% 12% +4
(sample size Aug 14th 250, Oct 26th 343, Nov 12th 354)
Durham Oct 26th Nov 12th change
Erin O'Toole Cons 46% 42% -4
Larry O'Connor NDP 24% 26% +2
Grant Humes Liberal 20% 22% +2
Virginia Ervin Green 6% 7% +1
(sample size Oct 26th 379, Nov 12th 583)
I had not expected to see the race tighten up so much in Calgary Centre. The growth is clearly for the Greens, but I do not think it is nearly enough to win. Also this poll is once again a very small sample size. The chances of having serious systemic flaws are high because of the demographic weighting that needs to be done to simulate the population.
Even with his large gain in support, Green Chris Turner is still clearly #3 in the race. I do not see it as realistic that he could win. The question then becomes can Harvey Locke win for the Liberals? I think it will all come down to who decides to vote. I presume that based on these polls all the Liberal effort is going to go into Calgary Centre. The pressure to get NDP and Green voters to back Harvey Locke is going to go through the roof.
If you delve further into the poll, Chris Turner does lead among voters aged 35-54 and is tied with Locke and Crockatt in the 18-34. Where Chris Turner suffers is among the older people where he is far behind the other two. If the voter turn out were to reflect the actual demographics of the riding, I think the results would work in Turner's favour, though still not likely enough to win.
Durham is not changed much, a mild drop in CPC support and milder increase for the other parties. The Greens remain a non issue in the riding. Since the race is a clear win for the CPC, turn out is likely to be low.
In Victoria the NDP retain their commanding lead but what has changed is the race for second. In October it really was a three way competition for second and now the poll shows that the Greens have as much support as the other two combined. The larger sample size of 509 means the poll is less likely to be too far off the mark.
Turnout could make a big difference in Victoria. It is possible that Liberal and CPC supporters stay home in higher than expected numbers and that the NDP does as well as they are expected in a riding they hold a commanding lead. What could then make a difference is extra turn out from Green voters - it is not unreasonable for the Greens to possibly boost the turn out enough so that they come a stronger second.