Wednesday, March 2, 2011

Ipsos: Tories Lead by 16

The latest Ipsos poll (via ThreeHundredEight.com) puts the Conservatives squarely in majority territory, with 43% of the vote. This is a very strong survey across the board for the Tories, except in Québec where their 19% is still average. For the Grits, this was actually not bad in terms of their support level, and in fact the 27% in Québec is quite encouraging. However, the Tory-Grit gaps, at 12% in Ontario and 19% in Atlantic Canada, would spell electoral disaster for the Liberals.

So where did all the Conservative votes come from? Well, this was a very bad survey for the NDP (13%) and for the Greens (5%), though the latter may be due to a methodology change: as explained by Éric at ThreeHundredEight.com, Ipsos has opted to stop prompting for the Green Party. My averaging method for the national popular vote has been modified to account for this change.

Not surprisingly, this poll brings the Tories closer to that magic 155-seat mark in the aggregate projection, though they are not there yet:

CON - 152
LIB - 70
BQ - 53
NDP - 33

This is the absolute worst projection for the Liberals since this blog was created in July 2009; it is also just one short of the highest point for the Tories. The average Conservative national lead is at 12.1%.

Also, readers of Éric's blog have probably noticed that he has a new projection model in place. The change in the model gave 4 extra seats to the Conservatives and 9 to the NDP, at the expense of 13 seats for the Liberals. Over the past 1.5 years, Éric's old method has, compared to my projections, consistently disadvantaged the Tories and especially Dippers to the benefit of the Grits. His methodology change makes the seat projections from our two blogs much closer. I feel a little vindicated, but mostly reassured that I was not way off in the left field!

However, it remains true that my methodology depreciates old polls faster than Éric's. This would explain why I peg the Tory lead at 12.1% now compared to 10.4% for ThreeHundredEight.com.

No comments: