Wednesday was going to be "Omni Day" at Probe Research. That's the day every three months or so I spend entirely on the phone talking to dozens of clients and potential clients about our omnibus survey of Manitoba adults.
For nearly two decades our quarterly "omni" has provided accurate and affordable insights to our clients. It's probably our flagship research service and our firm is very proud to offer this strategic tool to a wide cross-section of private, public and not-for-profit organizations.
But this Wednesday was different. The previous evening, Canadian pollsters had taken a spectacular -- and very public -- belly flop when British Columbia's Liberal Party was re-elected despite the confident assurances of pollsters to the contrary.
And it was not just some pollsters who missed the mark -- they were all off... and by a lot.
I just couldn't muster the enthusiasm I normally devote to promoting our omnibus. Instead, I opted to give the shamed pollsters a couple of days to wipe the egg from their faces and try to explain their way out of this one. Our omni would have to wait.
As it turned out, these B.C. election polls were not worth the paper they were printed on despite the fact some of Canada's most prestigious media outlets (The Globe & Mail, Global News, Sun Media, Vancouver Sun, CTV and others) carried breathless and detailed analysis of this evolving numerical horse race.
It is true reliable polls can help the electorate form more informed choices but these same measurements, when imprecise, can cause considerable damage.
We will never really know, for example, the extent to which these wildly off-the-mark polls may have influenced the tone of media coverage of the 2013 B.C. election campaign. Nor can we really embrace the truthfulness of findings on the importance of campaign issues, party leadership or anything else purportedly "measured" in these same surveys.
While industry insiders surely will be engrossed in the forensic story behind their B.C. folly, citizens will and probably should care far less about the precise reasons for these miscalculations. The public had been misinformed, plain and simple.
At this point, British Columbians and other Canadians would be well within their prerogative to sweep into the dustbin the polls that had so conspicuously littered this election landscape and to ensure their grounds remain clean of this type of debris in the future.
Time away from my desk and phone afforded me the opportunity to reflect on our own record of accuracy and to consider looming future research challenges -- even from the glass house we gingerly inhabit.
Our most recent election survey, taken for the Winnipeg Free Press in the final stretch if the 2011 Manitoba election, was (humbly) a virtual bang-on projection of the actual voting outcome.
But a decade earlier in the 2003 provincial election, a similar election survey taken by our firm for the Winnipeg Free Press and Global News provided a less-accurate projection of the final outcome (but not off nearly to the degree of the recent B.C. polls).
In 2003, as was the case with the B.C. polls, we overestimated the level of voter support for the NDP. And the lessons we learned in 2003, I believe, offer critical insights as to what went wrong in B.C.
As local historians and political junkies may recall, in the lead up to the 2003 election Manitoba's governing New Democrats, fronted at that time by the charismatic Gary Doer, were at the top of their political game. A series of quarterly polls taken by our firm between 1999 and May of 2003 showed unwavering and seemingly insurmountable NDP dominance with voters.
When the NDP dropped the writ a full year and a half before an election was officially required, there was a strong sense, just as there was in B.C., the election was a "done deal."
Our late campaign poll pegged the NDP with a startling 29-percentage-point lead over the Progressive Conservatives.
Sometime after the 2003 election, Doer told me the release of our poll served as a brake on his party's campaign momentum. According to Doer, voters greeting him on the doorstep seemed to be reacting as if he had "already won" even though several days were left on the campaign calendar.
Another important similarity between the conditions in Manitoba in '03 and this B.C. election was the low turnout of voters. Manitoba turnout in 2003 had taken a dramatic nosedive -- plummeting to 54 per cent from nearly 70 per cent in the previous election. Only about half of B.C. voters bothered to vote on Tuesday.
Our 2003 post-poll analysis revealed turnout rates in historically "safe" NDP constituencies were down by nearly 20 per cent on average, while voter participation in Progressive Conservative strongholds had not eroded nearly to this degree. An obvious explanation for this differential turnout would be that many NDP-inclined voters were satisfied an NDP victory was a fait accompli and, as such, victory would surely occur with or without their support.
Further investigation will reveal whether or not a similar stay-at-home New Democrat phenomenon occurred in B.C. this week.
Scott Mackay is president of Probe Research Inc., a Winnipeg-based public opinion and marketing research organization.