Ipsos Reid shows BC Liberals up 9%
Posted November 18, 2008 by Sacha Peter - Link
Category: Polls
Comments Off
Just one day after Angus Reid showed the NDP up 5% over the BC Liberals, Ipsos Reid’s poll shows the BC Liberals still comfortably ahead, 44% to the NDP’s 35%. The Green party polled at 16%.
Here is a link to a Vancouver Sun article describing the poll – I could not locate the source document (Update, November 19 – Source documents found here). 800 British Columbians were surveyed between November 5 to November 12. I have updated the charts in the Polls page. I would be curious to see the regional splits in the Ipsos poll.
Also, Vaughn Palmer’s commentary is fairly insightful, so I will not repeat his comments here.
His two last paragraphs (said humorously) suggested “averaging the two” to lead to a BC Liberal 42%, NDP 40% result is statistically feasible in a meta-analysis, providing that the respective sample spaces of both surveys were randomized and there was no/little correlation between the two samples (e.g. duplicate people were not sampled). Of course, if BJ’s suspicions are correct, the Angus Reid survey’s sample space is biased and such a meta-analysis cannot yield accurate results.
For both political parties, they will likely be conducting their own private surveys to reconcile the huge differences between the two polls.
One small sample that can easily be proven is the spread in the difference between the 2005 elections and the 2008 by-elections; in the 2005 general election, the BC Liberals had an actual vote of 45.8%, while the NDP had 41.5%. The NDP performed better in the lightly-attended by-elections than the BC Liberals did. If 44/35/16 is the projected poll, one would surmise that the Green party’s performance would be more reflective in the by-election results, which didn’t occur. One inference is that the Green party vote is highly inflated in polling, but this is hardly news.