Seat Projections accurate - polls were the problem. 
I have a crude spreadsheet that I used to make seat projections. I base the calculations on the 2000 results, use the poll numbers in the West, Ontario, Quebec, and the Atlantic regions to determine voter movement, and then project that movement onto each riding in the region.

If I use the actual popular vote numbers from the election and put them into my spreadsheet I get the following result - note it is based on a 301 House of Commons.


Some regions are a bit off but it is clear that the seat projection technique isn't the crap shoot that some pollsters (the ones that didn't use seat projections) want you to believe - they are looking for a scape goat.

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