While every public poll conducted during the municipal election campaign predicted victory for Mayor-elect Brian Bigger, one stood out from the rest, for a couple of reasons.
It was the Prime Contact poll, released Oct. 10 – more than two weeks before the vote. That poll stood out because it was an Interactive Voice Response (IVR) survey, better known as a robocall.
It also stood out because it most accurately predicted the outcome of Sudbury's mayoral race. When excluding undecided voters, the Prime Contact poll pegged Bigger's support at 45 per cent, and he won the race with 46.32 per cent support. It gave second-place finisher Dan Melanson 16 per cent support, and he finished at 19.25 per cent, while John Rodriguez was pegged at 18 per cent, and he actually finished with 17.38 per cent.
That contrasts with three polls commissioned by Northern Life, which also gave Bigger the lead, but by a much smaller margin. The Oraclepoll survey conducted Sept. 24-25 had Bigger ahead with 33.2 per cent of decided voters, compared with 23.7 per cent for Melanson and 22.3 per cent for Rodriguez.
The second, conducted Oct. 14 and Oct. 16, had Bigger at 34.9 per cent, Melanson at 21.7 per cent, Rodriguez at 19.8 per cent The final poll (conducted Oct. 21-22) found Bigger had 31 per cent support, with 23 per cent for Melanson, and 21 per cent for Rodriguez.
Rival campaigns quickly dismissed the poll as skewered, and questioned the validity of IVR polling methods.
“When they hear the recorded messages, it turns people off,” said Oraclepoll Research CEO Paul Seccaspina.
It's the sort of criticisms Prime Contact CEO Josh Justice has been hearing for years. His firm was the first to put Rob Ford in the lead in 2010's race to be mayor of Toronto, he said, and it opened them up to derision from other candidates and some media.
“So a bunch of (media outlets) got together and spent $25,000 or something on a poll, and found out the same thing,” said Justice, whose Canadian operations are based in Hamilton. “And we had him not only leading, but way ahead. And we nailed it.”
Their poll of voter intentions in Sudbury were also strikingly accurate. The difference, Justice said, is the IVRs, which offer several advantages over traditional live polling. For one, they are designed to exclude people who don't intend to vote. That helps provide a more accurate picture of voter intentions.
“Even with a turnout of 50 per cent, that still means half the population didn't vote,” he said. “So if someone (who doesn't vote) says they support a candidate, it doesn't mean very much.”
IVRs can also be done quickly, and reach a larger number of voters in a shorter period of time, he said. Prime Contact's Sudbury poll interviewed 1,085 people, compared to 500 for the two Oraclepoll surveys, and 400 for the third.
“We also employ actual statisticians who examine the data,” he said.
They do a lot of work between elections for governments and organizations, he said, and don't “seek to get their name in the paper,” the way traditional polling firms have done.
Services include focus group surveys and gauging public opinion for governments. Clients have in Canada have included Ontario, British Colombia and various Liberal Party offices across the country.
“This technology is becoming the world standard,” he said.