> Backgrounder: The polls - according to Pickup

Backgrounder: The polls - according to Pickup

Document Tools

Print This Page

Email This Page

Font Size
S      M      L      XL

Mark Pickup, 778.782.8640; mark_pickup@sfu.ca
Marianne Meadahl, PAMR, 778.782.4323

October 1, 2008
Simon Fraser University political scientist Mark Pickup’s website, The Polling Observatory, compiles data from polls released by Canada’s major polling houses to produce “a better estimate of vote intention” for each party in the current federal election campaign – based on those estimates, here’s his latest summary:
  • The Liberals are struggling to put the brakes on a decline in support that has continued since the beginning of the campaign. There are early signs that the decline has subsided - at least temporarily. 
  • The Conservatives are struggling to break through what appears to be a ceiling of approximately 38 to 39 per cent of the decided vote and have even recently experienced a slight decline in support to about 36 to 37 per cent. This has left the Liberals about nine to 10 percentage points behind the Conservatives.
  • The NDP have been benefiting from the Liberals' decline and are currently polling around 16 to 19 per cent of the decided vote.
  • Since the beginning of the campaign, the Bloc and the Green party have been neck and neck in the polls - somewhere in the range of seven to ten per cent of the decided vote.

Given the movement in support seen so far in the campaign and the number of days until election day, there is still opportunity for things to change, he says.

“This is only the story for the popular vote,” Pickup notes. “How the popular vote will translate into seats isn't entirely clear. The Bloc with its regional focus will gain a substantially greater number of seats than the Green party with the same vote percentage.

“The distribution of votes for the Liberal party also tends to mean a greater efficiency in terms of turning votes into seats compared to the NDP.”

Pickup says the estimates of vote intention correct for the systematic bias of each polling house (as exhibited during and since the 2006 election campaign) and the measurement error of each poll.