The provincial election in British Columbia is now less than two months away. And that means the polls are back.
After the failure of the polls in the 2013 vote, British Columbians might be forgiven for looking at them with skepticism. But the performance of the polls since 2013 suggests there is no reason to assume they are unreliable or that British Columbians are somehow unpollable.
Nevertheless, there is justification for looking at the latest numbers with a fair bit of caution.
- Christy Clark to announce panel to investigate fundraising
- With spend-happy budget, Liberals look to neutralize NDP
Pollsters have gotten skittish in British Columbia. This time four years ago, 27 polls had been conducted in the province in the previous 12 months. But only eight polls have been published in the last year in B.C. and three of them over the last month.
Those polls, all conducted by Mainstreet Research for Postmedia, suggest the margin between Christy Clark's governing B.C. Liberals and John Horgan's New Democrats is narrow — the latest survey conducted at the beginning of March put the NDP lead at just four points among decided voters.
The NDP has been polling at between 37 and 39 per cent over the last three polls, compared to a range of 33 to 37 per cent for the Liberals, 13 to 17 per cent for the Greens and 10 to 13 per cent for the B.C. Conservatives.
B.C. NDP ahead, but uncertainty abounds
Even with leads in Greater Vancouver and on Vancouver Island, which should put Horgan's NDP in a good position to secure a majority government, the margins are close enough that an NDP win is far from a sure thing even if the vote was held today.
But it won't be. The election is scheduled for May 9 and the latest Mainstreet poll pegged undecideds at 25 per cent. These undecideds would not need to break towards the Liberals by huge numbers to overturn the NDP's lead.
And campaigns matter — Adrian Dix's NDP had a 15-point edge at this point in 2013. That lead in the polls was cut in half by the end of the campaign. When the votes were finally counted, Clark's Liberals ended up winning by just over four points.
Whither the Conservative voter?
Another wildcard in the 2017 vote is the B.C. Conservative Party. Currently without a leader, the party is unlikely to field anything close to a full slate of candidates. That will leave a significant amount of its 10 to 13 per cent support among decided voters up for grabs.
Mainstreet's polling on who Conservative voters say is their second choice is muddled — the Liberals stood to be the main beneficiaries in one poll, the NDP the next. The party's position on the right of the political spectrum, however, suggests the Liberals probably have the better chance of scooping up more of that orphaned support in ridings where the Conservatives do not put up a candidate.
Even without the context of the 2013 polling miss, these factors would give British Columbians good reason to consume polls with some caution — the election looks little better than a toss-up at this stage. But is skepticism based on that 2013 failure warranted?
The polls in 2013 and since
No public poll published before votes were cast four years ago placed the B.C. Liberals ahead of the NDP. A number of explanations have been floated for that error, including the demographics of those that turned out and online pollsters' problems with assembling a representative sample.
Online polling in 2013 was certainly more bullish on the NDP's chances than the polls conducted over the telephone or via automated interactive voice response technology (IVR), where participants punch in their responses to recorded questions. The final online polls of the campaign gave the NDP an average lead of just over nine points. Telephone and IVR polls put the gap at just 3.5 points, with some of those polls out of the field almost a week before election day.
Mainstreet's B.C. polling is being conducted by IVR. But online polls did well in the 2009 B.C. provincial vote and elsewhere since 2013.
The last federal election demonstrated that British Columbians aren't unpollable. Surveys conducted in the province, including online polls, were very close to the final result in 2015. The polling average was within two points or less of the final result for every major party.
Polls done in every other province since 2013 have also been close to the mark. The last polls of the federal campaign were on target nationwide.
The polls were very accurate in the most recent elections held in Manitoba and Saskatchewan last year — including the surveys conducted by Mainstreet Research, which will be one of the most active pollsters in the coming B.C. campaign.
The Alberta lesson
The temptation to dismiss the polls after the scars of 2013 might be powerful. Taking polls as gospel is never a good idea — they should be read with a critical eye. But ignoring them entirely can make for an uninformed voter.
That is what happened in Alberta in 2015. Albertans had been burned by the polls in the 2012 provincial election. Three years later, the NDP had soared to first place but Albertans didn't believe the numbers.
- The Pollcast: Listen to Éric Grenier's podcast
Even in the last week of that 2015 provincial campaign, Albertans were still telling pollsters they expected the Progressive Conservatives to hold on to the power they first captured in 1971. But those same polls, surveying the very same people, showed that Albertans intended to vote the New Democrats into office instead.
On election day, despite their expectations, that is just what they did.