A poll by Ipsos-Reid for the provincial election in British Columbia had some comparatively good news for Adrian Dix and the B.C. New Democrats, as it put his party 10 points up on the B.C. Liberals - a more comfortable margin than the three other polls that have been released since last week's debate. As a result, Dix has a more comfortable lead over the Liberals in the projection, and the likely ranges now put him entirely in majority territory once again.
The Greens and Conservatives have dropped a little to 9.6% and 7.1%, respectively.
The odds that the NDP would win the most seats if an election were held today have improved to 78.6%, while he has a 94% chance of winning the popular vote on May 14. The large difference between the two are primarily due to two things: it is possible to win the most seats and lose the popular vote, and the probabilities are calculated differently. The probability of the NDP winning the popular vote is based on how the polls have moved, and have been wrong, in the past. The probability of the NDP winning the most seats is based on how the seat projection model has performed in other elections. It would only need to get six seats wrong (an accuracy rating of 93%) to put the two parties in a tie.
Ipsos-Reid's poll suggests that the gap between the two parties has closed, but sets off a few fewer alarm bells than the polls from Forum, Insights West, and Angus-Reid. But a narrowing gap was expected.
The chart below updates an article I wrote for The Globe and Mail before the B.C. campaign had kicked off. It originally showed that the margin the NDP had over the Liberals was greater than the one the Liberals had in the last two elections, and that it was holding relatively steady. But as you can see, the margin has closed very quickly and in a more dramatic fashion than in either 2005 or 2009.
And it needs to be noted that the last, steep closing of the gap in 2001, 2005, and 2009 were primarily due to errors and/or late movement between the last polls of the campaign and election day. And in every case, those errors and late swings were to the advantage of the B.C. New Democrats. We will find out whether this phenomenon is exclusive to the B.C. Liberal Party or just the party that is leading the polls next week.
The Greens were up one point to 10% and the Conservatives were down four points to 7%.
That gain by the Liberals is significant, but Ipsos is also judging Conservative support to be lower than in other polls - and probably more realistically.
Unlike the Angus-Reid survey, Ipsos-Reid finds the gender gap to still be in place: the two parties were tied among men while the NDP had a 20-point advantage among women (!). Also of note is that, according to the poll, roughly 1-in-4 supporters of the Green and Conservatives parties say they could change their mind by election day. But unless Christy Clark starts picking in the pockets of the NDP, they would need about 60% of Green and Conservative supporters to change their mind and vote Liberal in order for her to close the gap entirely.
Regionally, Ipsos-Reid shows the same sort of numbers that the other firms recorded in Vancouver and on Vancouver Island. It appears that on the island the Greens have stopped making gains. But in the Interior and north, Ipsos-Reid still gives the NDP the edge. Angus-Reid, Insights West, and Forum all gave the nod instead to the Liberals. Much of the election will be decided here.
The next round of polls could be very revealing. When you look at the chart showing how the gap has closed in other elections, you have to believe that, if the polls continue to show gains for the Liberals, Clark would have a very good chance of pulling off a big upset next week (and if she does, we'll never hear the end of the "polls having been wrong", rather than the reality of minds having changed over the course of a campaign). But if the polls instead show little change in the numbers, one might conclude that Dix will be able to hold on and win. There isn't much time left for another lurch in the polls.