Tuesday, October 30, 2012

Forum poll on the three Federal by-elections

Forum Research conducted a poll for the Globe and Mail for the three federal by-elections.   The results are not really very surprising, each of the incumbent parties have a strong lead.

What has to be kept in mind with this poll is that the sample sizes are very small making the odds of the sample being an accurate reflection of the public much less likely.   The samples were also weighted to reflect  the census demographics of the riding.   It also means the poll is a reflection of the whole electorate and not a reflection of the public that will be voting on election day.

Size of the samples of people that expressed an opinion
  • Durham 379 respondents
  • Calgary Centre 343 respondents
  • Victoria 295 respondents
One odd thing I found in the poll was that the samples found everyone that had an opinion for this by-election had voted in the last federal election.   The survey should have had a lower number of people answering this question if it was an accurate reflection of the public.

Party   Durham Calgary Cen Victoria 1st 2nd  3rd  4th 
CPC       46      48          16     2   0   1/2  1/2
NDP       24       8          47     1   1    0    1
Greens     6      11          20     0   1    1    1
Liberals  20      28          16     0   1  1 1/2 1/2

I am surprised, but based on this poll the closest race is Calgary Centre.   This latest Forum poll in Calgary Centre is not wildly different that the one they conducted in August.

Party   2011 elect Aug 14 Oct 28
CPC      57.7%      44%    48%
Liberals 17.5%      21%    28%
Greens    9.9%      12%    11%
NDP      14.9%      14%     8%
Others    0.0%       8%     5%

The data would certainly indicate that the one candidate with any hope of defeating the Conservatives is the Liberal Harvey Locke.   Though the margin between him and Crockatt is still huge and outside of a super human effort insurmountable.

I am surprised the NDP is doing as well as it is in Victoria.  Denise Savoie was a popular local MP, popular beyond the NDP.    I had honestly expected to see a bigger drop in NDP support, though the sample size is small.  When I took a first pass at estimating the by-election in Victoria I got the order of the candidates broadly correct.   My biggest difference from the poll was the level of NDP support, 37.5% is what I estimated versus 47% in the poll.   At the same time my estimation of Paul Summerville was outside of the margin of error, I am high on his support.

Party    2011 Elect Oct 28th
NDP       50.8%       47%
Green     11.6%       20%
CPC       23.6%       16%
Lib       14.0%       16%

In Durham I see nothing of interest (2011 results corrected, I accidentally entered the 2011 Ontario election results intially)

Party   2011 Elect Oct 28th
CPC       54.6%      46%
NDP       21.1%      24%
Lib       17.9%      20%
Green      5.4%       6%


DL said...

Your numbers for Durham from the 2011 election are wrong

Bernard said...

Thanks for catching that, I put in the numbers from the 2011 Ontario election by mistake

Ari said...

Hey Bernard,
If you add up the responses for each party for the last federal election you discover that 18 people did not vote in the last federal election. They left out the did not respond numbers in their tables. Same with income for instance, 258 people responded to the income question 295-258=37 non-responses.