Longer federal election campaign forces pollsters to adjust

The lengthy 78-day federal election campaign poses many challenges — to party finances, the stamina of politicians and the attention of voters. But the expanded duration of the campaign will present unique challenges to pollsters as well.

In the Canadian context, the extraordinary length of the campaign is an "unprecedented situation", according to Curtis Brown, vice-president of Probe Research. The ability for pollsters to gauge the state of the race from now until Oct. 19 "comes down to resources," he says.

• Listen to Election Pollcast, Eric Grenier's weekly podcast
• Interactive: Follow the polling trends with CBC's Poll Tracker

Media budgets that were only required to last some 37 days now have to be stretched for a period more than twice as long. That puts the squeeze on pollsters fortunate enough to have a paying relationship with a media outlet. For those pollsters that do not, it is their budgets that will be stretched.

Some of these companies add political questions to commercial omnibus surveys that are conducted on a monthly basis. Tracking a campaign requires a more frequent cadence of polling, which means conducting polls specifically for that purpose becomes a costly proposition if a pollster is not being paid for the work.

The effect will likely be, at least in the early weeks of the campaign, a reduced frequency in polling. With budgets tight, a similar number of polls may be published as in the 2011 campaign but over a longer period of time.

Changing strategies for a longer campaign

"The worst criticism we can get as pollsters," Christian Bourque, executive vice-president of Léger, said on the latest episode of CBC's Election Pollcast, "is if Canadians feel they are being over-polled during elections and that polling becomes the news story instead of the campaign itself."

Léger and some of Canada's other polling firms will instead be spacing their polls out more than they would have in a shorter campaign, but also increasing the sample size as well as the time spent in the field. That's the upside: polling more people allows pollsters to gauge with greater precision where things stand in different parts of the country.

Combined with a longer time spent in the field, this reduces the chance of a poll producing an outlier result, which might be harder to notice if a pollster does not have the data to put an isolated result into context.

But the length of the campaign also means pollsters may need to change their strategies.

"The ideal would be to conduct weekly tracking," says David Coletto, CEO of Abacus Data, adding that focusing on the later periods of the campaign, or around important events like debates, may be a better use of limited resources.

This is particularly the case considering the first part of the campaign is taking place during the summer, when it is harder to conduct polls in the first place. Canadians are on vacation, or just simply spending more time outdoors where telephoning pollsters cannot reach them.

Brown, of Probe Research, agrees, arguing the campaign will have three phases: the first being in August, as Canadians' engagement slowly builds; the second after Labour Day, when the pace of the leaders' campaigns will increase; and the third the final sprint to voting day itself, when many Canadians will make up their minds. 

An unfamiliar three-way race

The 2015 campaign is unique not only for its length, but also in that three parties are starting out with a legitimate shot at forming government. A party's chances of taking power may ebb and flow multiple times over the course of the next 10 weeks.

"Usually you will have one or two trends," says Bourque. "But this time around will we see three, four, five different trends? We don't know. We've never seen [a longer campaign] before."

The commitment of voters to their preferred party in this shifting landscape is also at question. The lead the New Democrats hold in the polls is relatively new, with the party having taken off in national voting intentions after the victory of the provincial NDP in Alberta.

"We don't know yet how soft [the NDP's] support actually is, or if it is firming up fairly quickly," says Bourque. If that commitment is still soft, it would not take much movement to turn the NDP's lead in the polls into a three-way tie.

This means pollsters will have to try harder to get at the motivations behind Canadians' voting intentions. Knowing how they will vote today is one thing, but having an idea of how they might vote on election day is an entirely different matter.

"The horse-race is one part of it," says Coletto, "but if we want to tell the story of this election responsibly, we need to look at where there is the potential for shifts, figure out who is going to vote, and ask more questions about people's openness to changing their minds."