Alberta’s pollster puzzle sparks search for answers
Expert cites sea change in public opinion
Things were changing so quickly on the final days.
JANET BROWN, PUBLIC OPINION RESEARCHER
And now the navel-gazing begins.
As Alison Redford cruised to another Progressive Conservative victory, politicos and pollsters are trying to figure out why they predicted a vastly different story in the lead up to Monday’s vote.
The majority of polls heading into Monday’s election pegged Danielle Smith’s Wildrose party with at least a six-point lead.
The narrative — among pundits, pollsters and political observers — leading up to the ballot box showed the upstart party toppling Alberta’s 41-year-old Tory dynasty.
Instead, Redford’s Progressive Conservatives handily won 61 seats.
The problem wasn’t that there wasn’t enough data: 31 polls had been conducted by various companies between January and the Sunday before the election, including two dozen since the writ dropped, said pollster Ian Large of Leger Marketing.
The surveys were conducted with different methodologies: live calls, robocalls, online, and combination methods.
Yet all hinted at the same wrong story.
“There were very few outliers from any of the firms,” said Marc Henry of Thinkhq Public Affairs.
“I was actually surprised at how consistent the polls were. That’s sort of what makes the outcome so surprising.”
That all were so dramatically wrong seems unlikely, he suggested.
Another theory on the misleading poll numbers has emerged — that the “snapshot” nature of polls didn’t (and couldn’t) capture a lastminute voter stampede back to the ruling Tories.
“Public opinion changed dramatically, quickly and right at the end of the campaign,” said Janet Brown, a public opinion research consultant who accurately predicted the 72-seat Tory win in 2008.
This election, the model she used to do seat projections showed the Wildrose winning a convincing majority government, with between 50 and 60 seats.
The sea change came quickly, she said.
“Polls don’t have the right kind of calibration to track things on a daily or hourly basis,” Brown said.
“And I really think things were changing so quickly on the final days of this campaign.”
A series of controversies dogged the Wildrose in the last week of the campaign, including comments on race relations and homosexuals, by candidates Ron Leech and Allan Hunsperger.
On Tuesd a y, To r y strategist Stephen Carter suggested some polling firms owe Albertans an explanation.
“I think there’s some companies that need to explain to Albertans what they were doing. I think they were completely wrong,” he said. “I need to know if any of these companies were using their polling as a leader instead of a measurement.”
Carter suggested media should rely less on polling — or at least focus more on issues and trend lines.
The pollsters agree the volatile nature of the campaign made prognosticating tricky. Henry points out the Tories went from a wide lead in January, to a dead heat when the writ was dropped in March, to being down by double digits within the first week of the campaign.
Potential game-changers — such as the Wildrose missteps in the last week on the hustings — take a while to filter through to the public, Henry noted.
The only way to track such a dramatic shift as played out at the ballot box is to do overnight tracking, blended together to show a rolling average, Henry suggested.
In a pitched electoral battle like Monday’s votes, many Albertans may still have been mulling their choices right up to when they voted, said Large. That kind of decision-making is not easy to capture.
“What we’ve learned from this is we’ll probably be less reliant on the traditional interviewer administered poll, and more reliant on our online polling,” Large said.
“It’s quick, it’s just as accurate, we can get in-and-out in matter of a few hours or a night, and get reliable results quickly, so we can be more responsive to the shifts in the campaign,” he said.