Alberta’s election is a horse-race – or not, depending on who you ask.
A series of polls have rolled out since the provincial election campaign began. Those for CTV, Global, the Calgary Herald and Edmonton Journal show a statistical tie between the top two parties, with Alison Redford’s PCs and Danielle Smith’s Wildrose party within the margin of error of one another. The most recent had the PCs ahead, 37 per cent to 34 per cent.
Then there’s the Sun.
Its poll, by Forum Research and based on automated responses on Monday evening, put the right-wing Wildrose in majority territory, with 41 per cent support to 31 per cent of PCs.
Some celebrated the results, with conservative talk radio host Charles Adler saying on Twitter that Ms. Smith was “putting the boots to Dinosaur Dynasty.” But the results left some wondering why one poll stuck out.
“Aren’t we all,” said Ian Large, vice-president, Alberta, of Leger Marketing, which used live interviews over the four days before the election to put together a poll, also released Wednesday, for the Journal and Herald. “Notwithstanding the Forum poll, all of the polls have put it all roughly in the same ballpark – neck and neck, slight lead for the Tories.”
Both companies have been around for a while – Leger 26 years, Forum 19 – but use substantially different methods.
Forum’s approach is known as “IVR,” or Interactive Voice Response, that uses a recording and not live interviews. It polled landlines Monday evening and proportionally broke down responses by four regions – Edmonton, Calgary, rural north and rural south.
Forum president Lorne Bozinoff attributed much of the difference in results to the fact they polled on the day the writ dropped (though the campaign had been expected for weeks.) The other polls were done in the days before the writ dropped.
“We’re not asking ‘if an election is called, if this, if that,” Mr. Bozinoff said, adding his polls have consistently shown Wildrose gaining (as most have).
The Leger poll, meanwhile, is based on live interviews, and is weighted for demographics like age. Mr. Large said his company uses responses, also from landlines, to build a population breakdown of the province (Forum wouldn’t reveal its “secret recipe”) and claimed robo-polls like the one Forum used are unproven.
“For some reason, the push polls seem to overstate the right-wing vote,” Mr. Large said. Forum dismissed that claim.
Leger also spreads the calls out over several days, and the weekend, to reach a broad sample of people when they’re not at work. Forum did it in one night. “You can’t compare polls on a weekday compared to a week,” Mr. Bozinoff said. “I’m not going to comment on one poll or another.”
He said Forum’s track record speaks for itself – it correctly forecast the seat total in Saskatchewan’s election last November, when Premier Brad Wall’s Saskatchewan Party won 49 of 58 seats.
The Leger Marketing poll surveyed 1,215 people in live phone interviews in the four days before the election. It has a margin of error of 2.8 percentage points, 19 times out of 20.
The Forum Research poll was done on March 26, drew 1,069 automated phone responses, and is considered accurate within three per cent, 19 times out of 20 – though Mr. Large also disputes that, saying it’s harder to forecast accuracy of an IVR poll. Mr. Bozinoff dismissed that, saying the method is tried and tested in the United States.
Similar discrepancies were found in Ontario this month, where a Forum poll and Nanos Research poll were 14 per cent apart - Nanos pegging Liberal support at 42 per cent, and Forum with it at 28, said Michael Ornstein, a sociologist at York University who is studying the issue. In an e-mail, Prof. Ornstein wrote that he suspects Forum's methodology is “biased towards older people who are more likely to be home and are more conservative,” saying the 14-point gap in Ontario is “way too large to reflect sampling error.”
The difference may also be because people are less willing to tell a live interviewer, as compared to an automated system, they plan to vote Wildrose, said Tim Woolstencroft, managing partner of The Strategic Counsel, a Toronto-based polling firm.
"The Alberta electorate is going through fundamental re-structuring and sometimes it is challenging for polls to pinpoint this type of fundamental change because it is occurring quickly," Mr. Woolstencroft said in an e-mail. His agency hasn't yet done polling in the election but said "it is pretty clear from these two polls that that the PCs are in deep trouble."
The polls do indeed share certain themes – Wildrose is stronger in the rural south and Calgary, while the PCs are stronger in Edmonton, traditionally a less-conservative community. All show the PCs and Wildrose well ahead of the New Democrats and Liberals.
And it’s clear that the huge lead the PCs had only weeks ago has evaporated.
“Everyone agrees this is not going to be a big coronation,” Mr. Bozinoff said.
Campaigning in Edmonton Wednesday morning, Ms. Smith shrugged off the poll numbers.
“Four weeks ago, I think the PCs were bragging about how they were going to get 82 seats and that just shows you how much polls can change in a four week period of time,” she said. Ms. Smith said her party is campaigning in all 87 ridings as if the Wildrose is 10 points behind.
“We’re certainly are not looking at the numbers of today and thinking it’s going to reflect anything like we’re going to see on election night,” she said, “We just simply don’t know.”
Ms. Redford said she's not concerned about polls in the tight race, and that voters face a clear choice.
"I think it's a very volatile election and a lot of people are saying that," she said. "With respect to polls, I think we’ll know on election day what the ultimate poll is."
Her party has been overhauled since she won its leadership on Oct. 1, she said. "It’s a Progressive Conservative party with a new leader that started on Oct. 1, that’s developing a new agenda."
The Leger poll showed Ms. Redford (an international human rights lawyer) had wide leads on Ms. Smith in her competency and ability to represent the province on the national and international stage, while Ms. Smith (who has a more prominent folksy streak) had a lead in likability.
In short, voters like Ms. Redford’s resume, but prefer the cut of Ms. Smith’s jib.
Both agencies anticipate scores of public polls will be released as the campaign drags on – only then will people know if Forum was prescient or off-base.
“We’ll know in a week, for sure,” Mr. Bozinoff said.
Ms. Smith is campaigning in central Alberta Wednesday morning, while Ms. Redford toured Suncor Energy’s oil sands mine in Fort McMurray, Alta.