Friday, July 03, 2009

Seat Projection

I'm going to start doing a somewhat regular seat projection. I'll use aggregate poll data from here (I don't feel compelled to duplicate good work) but my own methods for predicting results. Briefly, my methodology results in a riding-by-riding projection based on current polling and historical data but I will save a riding-by-riding prediction (it makes for a long and boring post otherwise) for an actual election period. My first projection is at right above the archive. Here's the national and provincial (yes, provincial not regional) breakdown:

National: CPC 118 LPC 121 NDP 29 BQ 39 IND 1

BC: CPC 18 LPC 8 NDP 10
AB: CPC 28
SK: CPC 13 LPC 1
Territories: CPC 1 LPC 1 NDP 1
ON: CPC 37 LPC 58 NDP 11
PQ: CPC 8 LPC 27 BQ 39 IND 1

My finger was only on the scale twice in these projections, both times in favour of the Conservatives. First to hand the seat vacated by Bill Casey to the Tories which I think only makes sense. The second time was to give the Tories Nunavut. It's very hard to do accurate readings on the North for a bunch of reasons and I think that Leona Aglukkaq is stronger than my number's tell me. The Quebec numbers I don't entirely trust (may have to work with my formulae). If I was just taking educated guesses I'd say Mulcair holds on to Outremont and the Tories don't save themselves in 8 seats but the data says otherwise. I was close to giving Mulcair his seat in spite of my data and I may in a future projection.

Long story short is we are in a dead heat at the moment. We shall see what the summer brings.


Anonymous said...

not sure about the rest, but you're being far too generous with manitoba.

simard will absolutely not win his seat back, steeves doesn't stand a chance, neville's is up for grabs (not to mention she might retire), and the libs MIGHT pick up churchill, IF fontaine runs.

at the absolute best i would say the libs have two MB seats next time around. my bet would be on one.

Ian said...

In Alberta I see a tight race (again) for Edmonton-Strathcona (NDP Linda Duncan's seat), and we'll see if people swap from NDP to Liberal (giving a newbie Con the seat) or if fewer Liberals stay home, but Linda's lead increases (while the NDP vote across Alberta may drop, it will likely ride the same in this riding).

Edmonton-Centre and East will also be interesting. The NDP and Liberals will be putting in strong candidates, so I see a vote split and the Cons winning it again, perhaps an election after (if the NDP overtake the Grits in the riding), it could fall. Ed-East is also the next AB hope for the NDP and may be close (but likely not close enough) to fall.

Other than that, most of Alberta will be swept from current numbers (of course it also depends on how many Cons are disenfranchised by deficits and stay home).

Anonymous said...

These numbers seem unrealistically optimistic. What is the formula you are using?

Aaron Ginsberg said...

The numbers are based on the aggregated polling data at So for instance, in Manitoba, my subjective opinion is that the LPC is dead in the water there. However, if we are actually polling in the 23% range in the prairies that would logically translate (and my calculations reflect that logic) that certain seats would turn Liberal. I think there is a chance Duncan carries Edmonton-Strathcona but it requires Liberal voters to continue backing her, something that is difficult to mathematically predict (at least with my limited math skills). Ridings like Edmonton-Strathcona and Outremont where there is a sudden shift in support to a party are very hard to understand or project going forward. If the NDP are a couple of seats short, that's the explanation.

I'm just plugging in the polling numbers available. I don't think my formulae all that biased one way or another.

All views expressed in this blog are those of the author and the author alone. They do not represent the views of any organization, regardless of the author's involvement in any organizations.

All comments are the views of the individual writer. The administrator reserves the right to remove commentary which is offensive.

The author is not responsible for nor does he support any of the advertisements displayed on the page