View Single Post
Old 10-18-2015, 11:12 PM   #617
MarchHare
Franchise Player
 
MarchHare's Avatar
 
Join Date: May 2004
Location: YSJ (1979-2002) -> YYC (2002-2022) -> YVR (2022-present)
Exp:
Default

Quote:
Originally Posted by killer_carlson View Post
well, it called the last AB provincial election almost completely bang on.
Not bang on, but still pretty close:

Quote:
Vote projection: 44.5% New Democrats, 25.9% Wildrose, 23.7% Progressive Conservatives, 3.2% Liberals
Vote result: 40.6% New Democrats, 27.8% Progressive Conservatives, 24.2% Wildrose, 4.2% Liberals
Error: 10.7 points, or 2.7% per party

The New Democrats and Wildrose were over-estimated, and the Tories under-estimated, but all of their results fell within the 95% confidence interval. Part of the reason for this was an adjustment made to the Liberals and Alberta Party, dropping their support to reflect their lack of candidates. This made the projection less accurate.

Seat projection: 55 New Democrats, 25 Wildrose, 6 Progressive Conservatives, 1 Alberta Party
Seat result: 54 New Democrats, 21 Wildrose, 10 Progressive Conservatives, 1 Liberal, 1 Alberta Party
Error: 10 seats, or 2.0 per party.
Riding projection accuracy: 86.2%
Projected seat range: 48-61 New Democrats, 17-31 Wildrose, 4-17 Progressive Conservatives, 0-1 Alberta Party
Seat range error: 1 seat, or 0.2 per party
Winners called in projected ranges: 93.1%

The seat projection model performed very well, identifying the correct winner in all but six ridings. Every party finished close to the average projection, while all but the Liberals fell within the likely ranges (and the Liberal result was within the 95% confidence interval).
http://www.threehundredeight.com/201...ck-record.html
MarchHare is offline   Reply With Quote