As the Toronto mayoral election approaches, more and more polls will be published in an attempt to gauge the state of the race. Can anything be learned from the last election's experience with the polls?
In the early days of the 2010 Toronto mayoral campaign, the race was close and the field was crowded. George Smitherman narrowly had the advantage, averaging about 31 per cent in polls conducted between April and June of that year. Rob Ford was not far behind at 27 per cent, and one poll even put him in the lead.
But support for other candidates was also relatively robust. Joe Pantalone and Rocco Rossi each boasted about 13 per cent support, with Sarah Thomson not far behind. These three candidates alone were garnering about two-fifths of the vote, limiting the ability of either Mr. Smitherman or Mr. Ford to push ahead.
The race in 2014 has some similarities, but not many. Rather than two front runners there are three, with Olivia Chow, John Tory, and Mr. Ford each registering above 20 per cent. In the only polling done in the first half of the year, Ms. Chow was well ahead with about 37 per cent support. Her numbers were holding steady. Mr. Ford's support dropped from 31 per cent in March-April to 25 per cent in May-June. This drop in support boosted Mr. Tory, who went from 24 per cent to 28 per cent support.
But unlike 2010, there is no second tier of candidates polling in the double digits. Instead, Karen Stintz and David Soknacki each poll at around 4 per cent support.
There are also some important differences in the campaigns themselves. There was no incumbent candidate in 2010. And the sordid history of the incumbent in 2014 is unlike any in Canadian political history.
The current state of the race is somewhat confused. The two most recent polls, conducted at the beginning of July, both agreed that Mr. Ford stood in third place. But Ms. Chow led in one poll, the survey conducted by Forum Research, the firm that has consistently placed Ms. Chow in front, while the other survey, conducted by Nanos Research, gave Mr. Tory the edge with a level of support not seen since Ms. Chow officially threw her hat in the race.
If we use the poll numbers of these two surveys as a range, and remove the undecideds from the Forum poll, that would put Mr. Tory at between 28 and 39 per cent support, Ms. Chow between 33 and 38 per cent, and Mr. Ford between 22 and 27 per cent. If we average them out instead, Ms. Chow is placed narrowly ahead with 36 per cent to 34 per cent for Mr. Tory and just 25 per cent for Mr. Ford.
A newer poll, conducted by Forum Research and published by the Toronto Star on Wednesday, suggests the race has tightened between all three candidates. Ms. Chow was awarded 29 per cent, Mr. Tory 28 per cent, and Mr. Ford 27 per cent support among all voters (among decided voters, that would increase to around 30, 29, and 28 per cent support, respectively). This new poll has done little to ease the confusion over the state of the race, though it suggests that Ms. Chow's support may have taken a slide.
At this stage of the race in 2010, the picture was somewhat clearer. After the close race in June, Mr. Ford moved ahead in August with an average of 39 per cent to 28 per cent for Mr. Smitherman. The three other candidates each were polling around 11 per cent, suggesting that Mr. Ford had taken support away from all of his rivals. Mr. Ford would also lead in all polls conducted in September.
That is when the campaign was transformed by the thinning out of the field. By mid-October, the number of competitive candidates dropped from five to three, as Mr. Rossi and Ms. Thomson bowed out. This gave Mr. Smitherman a boost. Mr. Ford had increased his support to around 41 per cent in September, and that only inched up to 42 per cent in the first half of October. But in the narrower field, Mr. Smitherman went from 25 per cent in September to 40 per cent in early October.
In the end, the final polls of the campaign showed Mr. Ford making some more gains as Mr. Smitherman's numbers faltered. But the departure of two of the three second-tier candidates had made the race a much closer affair.
Could something similar happen in 2014? If the three front runners stay in the race, there are few other votes available. Nanos and Forum awarded 93.5 per cent of the ballots to the three of them. Polls have suggested that all three candidates, but primarily Ms. Chow and Mr. Tory, could benefit if Mr. Soknacki and Ms. Stintz withdraw, but it is unlikely that this could be an election-deciding event.
In terms of the reliability of the polls, the record is mixed for 2010. Polls conducted up to October 18, a week before the election, showed a very close contest, rather than the 11.5-point margin that Mr. Ford actually won by. But the two final polls of the campaign by EKOS Research, conducted up to October 21 and 22, were far closer to the mark: 44 to 48 per cent for Mr. Ford and 33 to 36 per cent for Mr. Smitherman. The actual result was 47.1 per cent to 35.6 per cent.
A lesson from this is the need to continue polling as close as possible to election day in order to capture any late swing. Another lesson might be that every poll over-estimated the support of Mr. Pantalone. Support for him ranged between 13 and 17 per cent in the final polls of the campaign, when he actually took 11.7 per cent. If the race turns into a close two-way contest by October, we might expect the third-place finisher to under-perform expectations as voters cast a ballot instead for a potential winner.
At this stage, however, third place looks likely to go to Mr. Ford. When forced to choose, his supporters mostly go to Mr. Tory. But at the same time, after all that has happened, his remaining supporters must be rather committed. If their man looks set to lose, will they only then turn their backs on him?
Éric Grenier writes about politics and polls at ThreeHundredEight.com.