The small world of Canadian polling is in a huge uproar.
"It's like a brawl has broken out in the chess club," Tim Powers of Summa Communications said on Thursday.
Ottawa-based Summa owns the upstart polling firm Abacus, one of the companies on the receiving end of an open letter published on Wednesday by Darrell Bricker and John Wright of the venerable Ipsos Reid.
The letter was addressed to the Ontario media, and chided them for "ignorance" in publishing what it called second-rate polling data on the current provincial election campaign that show the Tories and Liberals easily outdistancing one another or running neck-and-neck depending on which numbers you choose to believe.
But a number of Ipsos's rivals see the diatribe as an attack on their integrity and business practices – not surprisingly, given that Mr. Bricker and Mr. Wright describe unnamed competitors as "hucksters selling methodological snake oil."
Abacus pollster David Coletto thinks he prompted the outburst by releasing a survey on Tuesday that used a non-traditional online method – a sliding-scale ranking of the kind often used to gauge consumer brand preference – to show a surprising nine-point lead for the Ontario Tories.
"We're the new guys in town, we're trying new things, and yes, we did have different numbers," the 29-year-old researcher said. "And I understand that creates confusion. But I take offence at being described as a huckster....I'm not going to be pressured just because established companies don't like what I'm doing or feel threatened."
Mr. Bricker took issue with the Abacus survey because it didn't ask the standard question that allows polls to be compared: if you had to vote tomorrow, who would you vote for?
"It should come with a flashing light that says hey, this is different, this has never been tried in a provincial election campaign before, so take it for what it's worth," Mr. Bricker said. "It's really not a measure of a vote so much as it's an indication of comfort or affinity with a particular political brand."
Pollsters like to see themselves as political interpreters, the all-knowing intermediary between the candidates' promises and the voters' innermost desires. "They're a breed apart," Mr. Powers said. "They're opinion savants and Svengalis of the public mood. And with that position comes a particular perception of the world. They're very sensitive about the trade they practice."
The traditional pollster's instinct is to be cautious about polls that deviate too far from conventional wisdom: But the impatience of a round-the-clock news cycle coupled with a period of extreme electoral volatility – witness the NDP surge in Quebec and the near-disappearance of the Bloc Québécois in the spring federal election – has increased the media acceptance of odd poll results.
"The polls that are counterintuitive get attention," Mr. Bricker said. "... Instead of someone sitting back and saying, 'Hold on a second, I don't know if this can be right,' the attitude is, we've got to get this out fast because it's a new trend."
But the open letter from the Ipsos pollsters did more than just counsel caution in uncertain times, when new technology has made traditional telephone surveys less dependable, and young people's reluctance to vote gives disproportionate power to an older, more conservative demographic.
By deriding competitors' methods, Ipsos left the impression it cared as much about its own competitive advantage as it did about the state of the industry.
"I'm not sure where Darrell gets the right to have a bully pulpit to hector the media and other pollsters' relationship with the media," said Frank Graves of Ekos Research, which uses an automated technique called Interactive Voice Response Polling that the Ipsos letter called "tremendously biased."
Mr. Bricker disagrees. It's really a general statement about how we're not going in the right direction, how we have to be more careful. Too many people think polling is fun-with-numbers, that it's just a publicity game."