Wednesday, October 8, 2008

Sixth projection

There is a slight revision to the projection method starting today and going forward. In ridings were a party is not fielding a candidate; rather than ignore the polling fluctuations for the non-contesting party, they will be distributed in the same way that party's 2006 votes were distributed.

For instance, in Central Nova, the Liberals are not running and have endorsed the Greens. To create a baseline for the projection, I distributed 60% of the Liberal vote to the Greens and the remaining 40% proportionally among the Conservatives and New Democrats. Today's polling average shows the Liberals down 3.5% in Atlantic Canada from their 2006 result, so I shave 60% of 3.5% off of the Greens and the balance proportionally off of the other two parties. I will use this method for the other 3 ridings where a party has opted to endorse another candidate today and in future projections.

With that business out of the way, on to today's update. We are using the following polls:

Angus Reid - Oct 6-7 (1.00 weight)
EKOS - Oct 5-7 (0.93 weight)
Ipsos-Reid - Sep 30-Oct 2 (0.36 weight)
Nanos Research - Oct 5-7 [ON, QC, Atl only] (0.93 weight)
Harris-Decima -Oct 4-7 [ON, QC, Atl, BC only] (0.86 weight)

National results (change vs. 2006, change vs. last projection)
CPC 130 (+6, -6), Lib 84 (-19, +3), BQ 49 (-2, n/c), NDP 42 (+13, +4), Ind 3 (+2, n/c), Grn 0 (n/c, -1)

Atlantic (change vs. 2006, change vs. last projection)
Lib 22 (+2, +4), NDP 6 (+3, +2), CPC 3 (-6, -5), Ind 1 (+1, n/c), Grn 0 (n/c, -1)

Quebec (change vs. 2006, change vs. last projection)
BQ 49 (-2, n/c), Lib 13 (+2, n/c), CPC 11 (+1, n/c), Ind 2 (+1, n/c), NDP 0 (n/c, n/c)

Ontario (change vs. 2006, change vs. last projection)
CPC 46 (+6, +2), Lib 41 (-13, -2), NDP 19 (+7, n/c)

Saskatchewan and Manitoba (change vs. 2006, change vs. last projection)
CPC 19 (-1, -3), NDP 5 (+2, +2), Lib 4 (-1, +1)

Alberta (change vs. 2006, change vs. last projection)
CPC 28 (n/c, n/c)

British Columbia (change vs. 2006, change vs. last projection)
CPC 23 (+6, n/c), NDP 11 (+1, n/c), Lib 2 (-7, n/c)

North (change vs. 2006, change vs. last projection)
Lib 2 (n/c, n/c), NDP 1 (n/c, n/c)

A few remarks...

Polls show a huge Liberal resurgence in Atlantic Canada, particularly in the Nanos poll which shows them at 50%! Not sure if this is real or a rogue poll combined with a series of polls with high margins of error. We'll see in the coming days.

Polling has otherwise stablized, with the Conservatives moving up from their low yesterday in Ontario and no change at all in Quebec, Alberta, British Columbia or the North.

Thanks to our revised formula, and strong numbers for the NDP in the Atlantic, the Greens are now shown to be losing Central Nova by 2.7% to the NDP with Peter MacKay finishing a distant third.

I'll continue to update daily until October 14. I welcome your suggestions if you think any aspect of the formula needs to be adjusted.

NOTE: These projections do not reflect the wishes, nor predictions, of the author. They are based on simple math using the criteria outlined here and here.


nuna d. above said...

What is your take on the Telegraph-Journal front page story about the large Conservatve lead in NB? I'm sure your Tory numbers for Atlantic Canada today are a little low.

Anonymous said...

Well, My take is that the T-J made a HUGE attempt to try to influence voting in NB. That's all it is. I believe the survey was done in the last few days of Sept. before the debates and before things started turning around for the Liberals.

Just look at the national polls in the last couple of days--
Nanos--Con 33, Lib 29, NDP 20, BQ 11, Gr 7.

"The media is the message."

nbpolitico said...

Nuna, I am not sure how comfortable I am with them being called "my numbers" as they are wholly dependent on what the polls are saying.

The polls used yesterday had the following Atlantic results:
Angus-Reid Lib 50, NDP 22, CPC 12
EKOS Lib 32, CPC 30, NDP 29
Nanon NDP 32, Lib 31, CPC 31
Harris-Decima Lib 32, CPC 31, NDP 29

If those polls were to be believed - partiuclarly the Angus-Reid one - then certainly the Tory numbers would be bordering on a shut out in the Atlantic. I think the Angus-Reid poll is likely a rogue poll as the other numbers are fairly consistent, and it will be replaced by either a new A-R poll or be weighted out over time.

nbt said...

I wondered your opinion on that myself, nbp? I wasn't all that impressed with how the Telegraph Journal did this before when they came out with a similar type article using (days old) CRA data indicating that Shawn Graham's Liberals were riding high (they had a picture of him on a horse).

Also, it would have been interesting to see what the numbers were for other weeks they missed?
Much like this recent poll which displays the exact opposite of what most of the other national polling companies are saying (who poll on a weekly and sometimes daily basis).

A little hackish, don't u think?

Howard said...
This comment has been removed by a blog administrator.