For days leading up to the Alberta election, every public opinion poll suggested the upstart Wildrose Party was poised to end the Progressive Conservatives' 41-year rule in the province.
But less than an hour after polls closed Monday night, it became clear that the PCs would be celebrating a surprising, historic victory. They won 61 seats, clinching a 12th consecutive majority government.
Their Wildrose rivals got 17 seats, far less than the polls had projected.
"I never felt that the polls were reflecting what I was hearing during the campaign," Premier-elect Alison Redford told CTV's Power Play on Tuesday.
So did the pollsters botch their surveys or was there a major shift in voters' attitudes over the weekend? Did a swath of Liberal and NDP supporters decide to vote strategically at the last minute, fearing the consequences of a far-right Wildrose government?
Those questions are at the centre of an ongoing debate about the validity of election polls in Canada and the way political surveys are conducted.
"Polls are, at best, guesses about what's going to happen and they are inherently limited," Western University political science professor Cristine de Clercy told CTV News Channel on Tuesday.
Polls reflect voter attitudes that can suddenly – and drastically – shift during the course of an election campaign, de Clercy said.
She pointed out that in the last federal election, "there was some distance between some of the major polling houses and the actual election results."
But analysts and pundits say a variety of factors could have influenced the outcome of Monday's vote.
In the last two weeks of the campaign, two Wildrose candidates came under fire for their controversial remarks about race and sexual orientation.
An anti-gay blog post written last year by Edmonton candidate Allan Hunsperger surfaced as the campaign intensified, while Calgary candidate Ron Leech suggested in a radio interview he had an electoral advantage because he is white.
Hunsperger said he was entitled to his religious beliefs and said he wasn't intolerant. Leech apologized.
Party leader Danielle Smith stood by them, despite the backlash. Both men lost the election.
Coupled with existing fears about Smith's perceived ideologies, the controversy may have prompted many Albertans to opt for a strategic vote to stop Wildrose in its tracks, Dave Rutherford, a Conservative radio talk show host in Calgary, told Power Play.
Journalist and blogger David Climenhaga, on the other hand, argued that "pollsters, as a group, really blew it."
He told Power Play that Albertans were inundated with pollsters' robocalls during the campaign, and many voters were averse to answering the phone.
"The only people who bothered to answer those questionnaires were committed people and I believe that skewed the result in favour of ideological parties," Climenhaga said.
A University of Lethbridge professor said he believes "a combination of all of the above" resulted in Monday's stunning election results.
Geoffrey Hale, who teaches political science, told Power Play that at least one last-minute election poll picked up the shift in favour of the PCs.
Whatever led the Tories to an unexpected victory, de Clercy said it's important to remember that polls' sample sizes and methodology can have serious limitations.
For example, if pollsters are only calling landlines, they won't be catching younger voters who are more likely to only use cellphones, she said. Depending on the time of the day and the geographic area, pollsters may be capturing a segment of voters that doesn't represent a wider public opinion or reflect the nuances of voting habits.
"Pollsters are trying to keep up with the trends," de Clercy said, but they don't always succeed.
A re-examination of the use of election polls and the reporting of their results may be needed, she said.
"It may well be time, in the wake of the Alberta election, to go back and think a little bit more carefully about how to regulate polling information."
