Federal election polls say little about how Northerners plan to vote - Action News
Home WebMail Saturday, November 23, 2024, 07:19 AM | Calgary | -12.2°C | Regions Advertise Login | Our platform is in maintenance mode. Some URLs may not be available. |
NorthAnalysis

Federal election polls say little about how Northerners plan to vote

Between now and the all-but-called federal election scheduled for Oct. 19, Canadians will be in for a steady stream of opinion polls measuring the horse race between the three major parties vying for power.

With just 3 seats and a tiny population, the territories are mostly left out of national surveys

From left, Prime Minister Stephen Harper, Green Party Leader Elizabeth May, NDP Leader Thomas Mulcair, and Liberal Leader Justin Trudeau. There won't be more than a handful of Northerners included in national election polls of 1,000 or so people, meaning results for the territories are totally unreliable on their own. (Canadian Press/Reuters)

Between now and the all-but-called federal election scheduled for Oct. 19, Canadians will be in for a steady stream of opinion polls measuring the horse race between the three major parties vying for power.

Accompanying many of those polls are seat projections, which incorporate measured levels of voter support to predict how many, and which, seats each party will win.

Those polls and projectionscontain lots of data about how each party is faring in each province. But what do they tell us about who is likely to win in the three Northern ridings?

Not much, it turns out. With just 0.3 per cent of the Canadian population combined, there won't be more than a handful of Northerners included in nationalpolls of 1,000 or so people, meaning results from the territories are totally unreliable on their own.

It says absolutely nothing about what people in the North are actually thinking.- ricGrenier

"There's going to be a lot of polling done and the poor territories are going to get left out of it mostly," saysric Grenier,founder of the polling website threehundredeight.com, who also runs CBC's Poll Tracker.

Grenier recalls seeing Northern data from a recent poll that had the Liberals at 66 per cent support and the NDP at 33 per cent. But before any partisans get excited, consider the number of Northerners surveyed by that poll: three.

You're projecting

Still, the Poll Tracker has seat projections for the North (as of writing, the Conservatives, NDP and Liberals are each forecast to win a seat North of60). So how does that work?

Grenier says in southern Canadahe creates seat projections by taking the last election's results and adding in regional breakdowns from current polls. In the North, where there's almost no data, he adds national trends to riding results from the last vote.

But Grenier admits this approach has its limitations.

"It says absolutely nothing about what people in the North are actually thinking," he says.

"It's just a way to make some sort of estimation for the North instead of just saying it will probably be the same as the last election."

Lorne Bozinoff, thepresident of Toronto-based polling firm Forum Research, says it's common practice to make Northern seat projections by rolling the territories in with neighbouring provinces. Beyond that, he says the territories are represented according to their share of the population, like anywhere else. But he adds that Forum won't release regional breakdowns for the North because on their own, the figures aren't statistically reliable.

"There's nothing special about calling the North," Bozinoff says. "It's just an issue of sample size."

Donna Larsen, owner of Datapath Systems, a market research company based in Marsh Lake, Yukon, says making seat projections for the territoriesbased on national polls is of limited value. She points to the Green Party's relative popularity in Yukon as an example.

"Applying federal numbers to Yukon numbers won't translate exactly," she says.

A poll done in Yukon during the last election showed Larry Bagnell's Liberal campaign with a 20-point lead. Bagnell lost to Conservative Ryan Leef, some say because Liberal voters thought it was safe to stay home on voting day.

"They can't possibly be calling a national election based on a national survey. It's not going to work."

Snapshots in time

Datapath Systems did polls during the last federal election campaign in Yukon with sample sizes of around 350 people. They drew the ire of Larry Bagnell's Liberal campaign when one pre-election poll showed the Grits with a 20-point lead. Both the Liberal and winning Conservative campaigns pointed to the poll as contributingto Ryan Leef's victory in 2011, because Liberal voters thought it was safe to stay home on voting day.

That's still a sore point for Larsen four years later.Polls, she says, are justsnapshots in time, so blaming a poll for defeat is akin to shooting the messenger.

"A poll is just a random sample of a small group of people on a particular point, day and time," Larsen says. "That's all it is."

The other question is, with just three seats out of 338, how much doresults in the territories matter? The Conservatives have long wanted to take out N.W.T. MP Dennis Bevington, a New Democrat, so they could claim all three Northern seats for symbolic reasons.

But with current (southern) polls predicting an extremely close three-way race, the North could matter more than usual.

Grenier's poll averages have the NDP and Conservatives tied nationally at 31 per cent (as of July 24), with the Liberals close behind at 26 per cent. That would translate into a Conservative minority government, but give the NDP and Liberals enough combined seats to form a coalition government (a whole other issue entirely). Meanwhile, a recent Forum seat projection had the NDP winning 120 seats to the Conservative Party's 112.

All this might mean the politician's old campaign chestnut is true: the only poll that really counts is election day.