Few mentions of polling in the provincial election campaign in British Columbia go without a question of whether the polls should be believed. The B.C. Liberals are banking some of their re-election hopes on this doubt, and reminders of how the polls missed the call in Alberta's election last year are never far from their lips. But if they are hoping for a repeat of Alberta, they are likely to be disappointed.
The ingredients that led to the monumental error in calling the provincial election in Alberta in April 2012 do not exist in British Columbia, making it very unlikely that a failure of the same magnitude will occur on May 14.
That is, if a polling failure did indeed take place in Alberta's election. There is no question that some of the polling that was done during that campaign was sloppy, perhaps in large part due to inexperience. Of the seven polling firms that were active throughout the election campaign, only two of them had any prior experience polling at the provincial level in Alberta. And one of those experienced firms did its last polling a full week before the vote was held.
Of the remaining five firms, two or three of them were publicly taking part in their first election campaign anywhere, and the other two were participating in an Albertan provincial election for the first time. In addition to this inexperience, the pressure to make a correct call in a first foray into Alberta would have been quite high, and the polling phenomenon known as 'herding' – when polls tend to group together as the election date approaches, minimizing the likelihood of being the odd-man out when the ballots are counted – might have been a factor.
With the campaign so lopsided, there are not nearly as many polling firms active this year in British Columbia. But the two main firms, Angus-Reid and Ipsos-Reid, both have long histories in the province and offices in B.C. They are unlikely to make the same mistakes.
But one of the reasons that the polls missed out on the election call in Alberta was due to a late swing in voting intentions that occurred in the final days of the campaign. A poll by Angus-Reid taken just after the election showed that as many as 39 per cent of Albertans made up their minds on how to vote in the last three days before casting a ballot, and almost one in four on election day itself. The final public poll of the campaign showed that a swing was taking place, and internal party polling showed the same thing. Simply put, most of the polls in Alberta missed the call because they stopped polling too far from election day.
Could the same thing happen in British Columbia? There are a lot of arguments against it. Unlike in Alberta, where the Wildrose party was relatively new, untested and fielding a slate of unknown candidates, the New Democrats in British Columbia have formed government before and Adrian Dix is well known to the public. The comments from a few of Wildrose's candidates that seemed to confirm the fears some Albertans had of the party (compounded by Danielle Smith's initial defense of them) gave voters second thoughts, whereas with the NDP most British Columbians know more or less what they are going to get.
Add to that the fact that the polls have not shifted to any great degree in British Columbia for the past year, whereas support for the Alberta Tories swung heavily to Wildrose in the first week of the campaign, and the more extensive organization and get-out-the-vote capability of the B.C. New Democrats compared to Alberta's Wildrose, and there is little reason to expect a late swing to occur.
But the most persuasive argument against an Alberta surprise in British Columbia's election may be the gaping difference between the numbers in this campaign and those of 2012. Throughout the Alberta campaign, the Progressive Conservatives trailed Wildrose by an average of 10 points, and in only two of 22 polls was the margin more than 14 points. In British Columbia, the B.C. Liberals have trailed the New Democrats by an average of 18 points over the last 22 polls, and only six of those have showed a narrower gap than the widest margin of any poll during the Alberta campaign.
A more decisive factor may be the comparative approval ratings of the incumbent premiers. In a late campaign poll by Forum, Alberta Premier Alison Redford was shown to have an approval rating of 41 per cent. A survey by Angus-Reid a few days before the vote gave her an approval rating of 46 per cent, little different from the 49 per cent approval rating the firm gave her on the eve of the campaign's start.
Christy Clark, however, does not have comparably good approval ratings. She has done no better than 31 per cent approval in polls over the last two months, with net ratings (approval minus disapproval) anywhere between –27 and –42. Ms. Redford's approval ratings made it easier for voters to switch over to the Progressive Conservatives (in retrospect, more should have been made of these numbers), but Ms. Clark's abysmal ratings make it more difficult for undecided voters to go over to the B.C. Liberals.
If Ms. Clark is able to boost her party's support in the next few weeks and vastly improve her own approval ratings, she may be able to put her party in a position where an Alberta surprise becomes plausible. But even that would require a major change in the race, in addition to the long-shot hope that the experienced pollsters in British Columbia will miss the call. The fine print in most polls explains how their findings apply in 19 out of 20 cases, meaning there is a 5 per cent chance they could be wrong. Those odds might be as good as it gets for Christy Clark.
Éric Grenier writes about politics and polls at ThreeHundredEight.com .