Will the polls be right?

British Columbians are voting tomorrow, meaning many are paying attention to political polls in an effort to glean who the next leader of the province may be.

Pollsters aren’t united in their view of what’s to come. Most have the B.C. NDP walking away with a mandate to run the province for another four years. At 338Canada.com, the B.C. NDP is predicted to have a 69 per cent chance of forming government.

Mainstreet Research sees a rise of a party that has never been in power – the B.C. Conservatives.

Mario Canseco, president of Research Co, says it's a close race, though he's leaning to the NDP in a race he said started unusually, but turned out to be pretty standard.

Once BC United bowed out, he said the BC Conservatives had about 40 per cent of support and the NDP had about the same.

“What we saw happen, especially after the debate, was a little bit of a drop in support for the B.C. Conservatives,” Canseco said.

"I think there was an opportunity to really connect with the population on the debate and have a couple of moments where people could really look at John Rustad as a serious way. We didn't really see that. So now it's a slightly larger lead for the NDP at the start of the week.”

Canseco said their polling shows the party is doing well in Metro Vancouver, which is where most of the seats are.

“They managed essentially to claw their way back in Southern B.C. and in the Fraser Valley, so areas that look like blowouts for the B.C. Conservatives are no longer safe,” he said.

“It's essentially a situation that would suggest that the NDP would form the government if the numbers hold.”

Odds are less favourable for the NDP in the Okanagan, though there are a variety of factors that remain to be seen.

Lingering questions

“In places like Southern BC and the Fraser Valley, where the NDP did better in 2020 than many other elections this century, are we going to see people who used to be BC liberals flocking back and voting for the for the BC Conservative because they are now the non-NDP option?” Canseco said.

Part of what is going to make this race interesting, Canseco said, is that there has been a consistently high level of support for the NDP with the over 55 age bracket across the province.

“This is a very different scenario from the one that we had before, where it was young people gravititating towards the NDP and the over 55 saying, 'no, no, no, we got a good thing going with Gordon Campbell and Christy Clark’.”

And there’s also the youth vote that’s hard to nail down.

"Right now they're looking at the B.C. Conservatives as an option, people who are dissatisfied with the state of affairs— particularly on the economy and jobs and housing— are seriously considering the B.C. Conservatives,” Canseco said.

“We know from previous elections, not just in B.C., but across the world, that it's a group that doesn't vote at the same rate as their older counterparts.”

So, what they like won’t necessarily show up on a ballot.

“Especially if the weather is bad,” he said.

A recipe for success?

In a place like B.C., with three and a half million voters, Canseco you find a sample of 800 gives you a margin of error of 3.5% but it's not a matter of simply finding 800 people.

Polling isn’t an exact science, but each organization deploys its own distinct method to get a prognostication with some degree of accuracy, and nobody is likely to spill the details on their secret sauce.

“I try not to look at what anybody else is doing because it can distract you from what happens,” he said.

“After the election, that's when you do the comparisons, we were closer than this, or we were closer than that on that party? But you know, part of, part of what happens if everybody's going to have their own motivations for doing this and their own ways of structuring the questionnaire.”

How that structure comes to be, however, isn't something any pollster is going to talk about freely.

“It's not a question of lacking transparency,” Canseco said.

“There are certain things that that you wouldn't share with people who are competing in the same field. There’s a system for it, but not one I’m going to share.”

Size doesn’t matter

The only way to create a perfect survey of what every British Columbian thinks is to poll every British Columbian.

That, however, isn’t going to happen so Canseco explained pollsters create a sample that is reflective of the population as a whole.

“If you doing something national, usually we have a sample size of 1,000, but it's not just 1,000 Canadians from wherever,” he said.

"You are trying to replicate the Census. So if the census shows that 51 per cent of Canadians are women, then your sample has to be 51 per cent women.”

Then geography and a variety of other issues come into play.

"You could be talking to 25,000 people, but if that hasn't been properly balanced, and it's not representative of the region that you're polling, then you're not going to get anything that is close to being accurate,” he said.

Another change in recent in the last 15 to 20 years is the use of online panels for surveys.

"These are groups of people who have agreed to take surveys, and they might get a survey about which detergent they use or how often they go out for dinner and the next survey might be something more political,” Canseco said.

"As long as you apply the same rationale, as far as having a sample that is representative of the population, then you can certainly use online panels.”

That’s a common way to go now, he said, but in the early stages of polling the telephone was king.

“As we started to get closer to the end of the century, people started to get tired because they were getting phone calls all the time… and the conversations went on for 20 to 25 minutes,” Canseco said.

“Some industry figures believe that the golden standard should continue to be the telephone, and you know, I don't think you can make that argument anymore when people hang up on you, 97 per cent of the time.”

The industry in turn moved toward the online model, and Canseco has been following how that’s changed since the first online poll was launched in 2007.

“It’s easier to find willing respondents in online panels that want to participate,” he said.

“They'll get an email from us, they'll log on to our our platform and tell us how they feel about things… once we hit the desired numbers down the tables it’s done.”

And the winner is

Canseco is proud of his record, but admits he’s fumbled a couple of times.

“We all have a tendency to fixate on the ones that go poorly,” he said.

"You know, nobody is going to write a headline that says, ‘Pilot lands plane safely, all the passengers are out’.”

“Pilot crashes,” however, that’s a different story.

For those who are paying attention, however, a win matters.

“It helps you talk to clients, and say to the clients ‘we know what we're doing, because in the elections we called this outcome, this is our record,’” he said.