Sunday, December 11, 2005


Northern Ontario pollster predicts Conservative majority

Sudbury pollster Paul Seccaspina is predicting that the Conservatives will win a majority government in the Jan. 23 federal election. This notion goes against all the national, regional and local polls and pundits who are predicting a minority for either the Conservatives or Liberals.

But, it got Gregory D. Morrow thinking about what exactly would it take to form a Conservative majority government, and whether that is even possible. So, he ran a few simple scenarios through my election forecast model. I simply asked, holding all other parties constant, at what percentage shift from Liberal to Conservative, applied equally across all regions (including Quebec), would result in the Conservatives getting 155 seats? It is a theoretical calculation, of course, but it is a rough approximation.

The answer: The Conservatives would have run the table outside Quebec and see almost a 9 point pure shift from Liberal to Conservative across every region from the 2004 election. This is going to be virtually Impossible.

This would produce the following seat count:
Conservative 155
Liberal 71
Bloc Quebecois 59
NDP 23

This would correspond to the following support levels:

Conservative 38.1%
Liberal 28.2%
NDP 15.7%
Bloc Quebecois 12.4%

Conservative 40.0% - 58 seats
Liberal 36.2% - 35 seats
NDP 18.1% - 13 seats

Bloc Quebecois 48.9% - 59 seats
Liberal 25.4% -12 seats
Conservative 17.3% - 4 seats
NDP 4.6% - 0 seats

Conservative 44.8% - 30 seats
NDP 26.5% - 3 seats
Liberal 20.1% - 3 seats

Conservative 70.1% - 28 seats
Liberal 13.5% - 0 seats
NDP 9.6% - 0 seats

Conservative 48.8% - 21 seats
NDP 23.5% - 3 seats
Liberal 21.9% - 4 seats

Conservative 39.0% - 15 seats
Liberal 37.7% - 14 seats
NDP 19.8% - 3 seats

These seat numbers are not credible in Quebec - the Bloc is way stronger and the Conservatives are still in single digits and will not win a single seat. The numbers in both the Atlantic and BC are just comical.

Two other election prediction sites to read:
Democratic space - Leger polling results in Saskatchewan for early December predicted 7 Liberal seats is - insert descriptive here - laughable.

You should be careful to differentiate between 'prediction' and 'projection'. At, I 'project' the seats based on polls - so if a poll shows an unusual result, it shows up as strange seat counts. That is a projection, not a prediction of what will happen. At, I do a running average to arrive at my 'current' projection. Only on the day before the election, will the 'projections' turn into 'predictions'.
-Gregory D. Morrow
Post a Comment

<< Home