<$BlogRSDUrl$>
Spreadsheet Scandal 
Posted to Jim Elve's Group Election Blog.

The AG has reported that there are serious flaws with my predictions. After looking at my treatment of undecided voters (brought on by Colby Cosh's post yesterday) I realized I wasn't doing what I had intended.

Basically, I was punishing those parties that had less than their popular vote in the 2000 election (Cons in the West, Liberals in Ontario) and inflating the parties that had improved on their numbers from the election (NDP in Ontario notably).

Here is the revised seat count based on the Feb 16th poll - a slim Liberal majority - helped mainly by an Atlantic Canada rout.


QuebecOntarioWestAtlanticTotal
Lib.23871727154
Con.01361579
N.D.P.0313016
B.Q.5200052


Also, in a post yesterday, I had presented poll numbers that would result in a Conservative minority government - it changes slightly and is shown below:


Date:Feb 16, 2004May 16, 2004
Firm:Ipsos-ReidCons Minority
Quebec
Lib.3138
Con.1010
N.D.P.88
B.Q.3938
Ontario
Lib.4136
Con.2635
N.D.P.2224
West
Lib.2530
Con.3949
N.D.P.2020
Atlantic
Lib.4742
Con.3233
N.D.P.1219



QuebecOntarioWestAtlanticTotal
Lib. 32 52 16 16 116
Con. 0 46 64 12 122
N.D.P. 0 5 11 4 20
B.Q. 43 0 0 0 43


I'll be sending out the corrected spreadsheet.

This page is powered by Blogger. Isn't yours?