news
A Look Back, At Advance Mayoral Polls Past
Some of the mayoral contenders—and all of the front-runners—before the Better Ballots debate on June 1. Photo by Shaun Merritt from the Torontoist Flickr Pool.
Rob Ford is now polling alongside George Smitherman in the run-up to this year’s mayoral election (recent drug-related scandals notwithstanding). But how reliable are these polls? Torontoist has compiled a selection of pre-election polling data from every mayoral race since amalgamation, and you might want to take a look at it before you start planning your Ford Escape™. (Or, if you’re planning on voting for him, maybe your Ford Excursion?)
Pre-election polls have their uses, but no poll is able to gauge public sentiment with perfect accuracy. Most news outlets, when they report on polls, will include a margin of error, which usually falls a few percentage points to either side of the poll’s findings.
Polls can fall prey to a wide variety of systematic methodological errors. For example, if a poll were conducted over the phone, and the pollster neglected to canvas any cell phone users, he or she might inadvertently under-report the opinions of younger voters, who are generally less likely to have a landline. Trained pollsters have techniques for minimizing mistakes and bias, but there is always a small chance that some oversight or statistical fluke could skew results.
Municipal elections generally are not as intensely polled as federal or provincial races.
Here are some charts we prepared, with information from all four post-amalgamation elections (excluding the current one). All the data were drawn from archived newspaper clippings and press releases. This is a representative sample, but not a comprehensive one. We’ve undoubtedly missed a few polls along the way.
* = Of decided voters.
† = Of those likely to vote.
The 2006 and 2000 races, which involved strong incumbents, contrast sharply with the other two, which were competitive throughout. In both the 1997 race to determine the first “megacity” mayor, and the 2003 race to replace Mel Lastman, in which David Miller upset Barbara Hall, advance polls were able to predict outcomes accurately, but there were some aberrations along the way.
In 1997, for example, polls from as early as April accurately predict Mel Lastman’s eventual win, but they drastically underestimate public support for Barbara Hall, who ultimately lost to Lastman by a margin of just 6%.
In 2003, a poll from May shows Barbara Hall with a large lead over all her competitors. A poll from October still shows her ahead of David Miller by 7%. It isn’t until late October, about two weeks before the election, that Miller, who ended up defeating Hall by a margin of 34%, finally polls decisively in first place.
“When you don’t have an incumbent running,” said Derek Leebosh, a senior associate with Environics Research Group, which has been conducting polls dealing with Toronto municipal politics since the early eighties, “a lot of what you’re measuring is basically how many people have heard of these people.”
“The polls are important more as a kind of self-fulfilling prophecy than in terms of how people actually vote once there starts to be a major focus on the campaign,” Leebosh told us.
“They don’t just describe the campaign. They also become a factor in the campaign.” As when fundraisers and volunteers decide to throw their support behind a candidate whose polling looks promising, rather than divide their resources between two candidates with similar politics, one of whom doesn’t seem to stand a chance.
Leebosh believes polls were a factor in David Miller’s success in 2003. Left-leaning voters who were backing Barbara Hall might have switched allegiances after early surveys showed Miller, initially perceived as a long-shot, to be a viable candidate.
But how seriously should we take this latest round of advance polls, which appear to point to a tense runoff between Smitherman and Ford? Well, it’s worth noting, at least, that in the springtime 2003 poll showing Miller so far behind Hall, only 20% of respondents said they were undecided. In the polls that show Ford next to Smitherman, the percentage of uncertain voters is nearly 40%.
Literally anything could happen before election day. Hell, plenty has happened already.
Hat tip to @mambamayor for suggesting we examine past polls, and to Environics Research Group for providing us with some of their archived poll information.