Friday, October 06, 2006

# Posted 9:27 AM by Taylor Owen  

HOW TO LOSE AN ESTABLISHMENT POPULARITY CONTEST: I am sure that there are American and British equivalents to Warren Kinsella, but I can’t think of any off the top of my head. He was a Chrétien strategist, is in a punk band, and now has an enormously popular Canadian political blog and column in the National Post. His role in the Canadian political debate, however, hinges on a wonderful irony; he is the ultimate insider who ruthlessly mocks the establishment. (Imagine if John Stewart had spent ten years as Clinton’s chief of staff prior to doing the Daily Show. Or, if Carville actually turned on the establishment that still looks to him for salvation.)

In any case, his column this week is a truly superb example of this irony, and is just as applicable internationally as it is in Canada. In it he mocks the use of polling data by the major media outlets, and the associated punditry. All of which predicted a very close race between Rae and Ignatieff in the delegate selection process last week. The results were roughly Ignatieff 30% (will be closer to 35% once said and done), Rae 19%, Kennedy 17%. Dion 16.5%, and, Dryden/Volpe/Brison all under 5%. Here, however, is his skewering of what the polling/punditry predicted:

However, if you relied upon a mid-September poll of the Gandalf Group -- as did the Parliamentary tabloid called the Hill Times -- you can be forgiven for being gobsmacked. There, the newspaper and Gandalf reported that Dryden was supported by approximately 20% of Liberals (and Canadians) nationally -- with Michael Ignatieff running a distant third. Um, wrong, by a factor of 400%, 19 times out of 20.

Next up for a trip to the woodshed: the Globe, with another mid-September poll, this one by Allan Gregg's Strategic Counsel. In front-page story accompanied by a large headline (to wit, "Ignatieff clings to slight lead"), Gregg said: "If you had money to put on it, you'd bet Rae right now." Uh-huh. Sure. And then you'd lose your money, Allan.

By a unanimous decision of our panel of judges (me, myself and I), the Chicago Daily Tribune Foul-up Finalist is the Toronto Star, for the paper's Sept. 25 page one headline fumble: "Rae now poised to become Liberal leader." In Linda Diebel's accompanying news story, Ekos Research asserts that Bob Rae has "emerged as the leading candidate in the Liberal leadership race." Said Ekos president Frank Graves: Liberal poll respondents underwent a "very careful screening process." Apparently not careful enough.

There were a smattering of other wince-inducing boners, such as the Toronto Sun's Peter Worthington ("conventional wisdom" is that Rae has the "momentum," wrote the veteran columnist), or the Vancouver Sun's Barbara Yaffe (Ignatieff is "a long shot ... he should get himself a good set of worry beads, pronto").

So who is to blame? The media organizations (with some justification) will blame the pollsters. The pollsters (again, with some grounds) will pin it on the folks responding to their polls, who are increasingly unenthusiastic about confessing their innermost thoughts to complete strangers on the telephone.

The loser, naturally, is the reader. The reader deserves better. And if polling mistakes keep getting made, then the media needs to re-examine its enthusiasm for polls.

Ask Harry S. Truman. He knows.

Wonderful. Seriously though, the Canadian media is going to have to take a real look at how they use the crack that political polling data has become leading up to the next federal election.

(2) opinions -- Add your opinion

To be slightly more relevant than the poster above and to be slightly more fair to the pollsters. The issue in the race is that the voters in the leadership campaign are a very select group, and not anywhere representative of the population at large, nor even of liberals in general. So unless you had a representative sample of only people elegible to register in support of one of the candidate, polling is going to be wildly skewed. An inability to properly do one of these polls probably has little to you with ability to do a poll representative of voters.

However, that does leave the question of why bother doing and reporting these polls.
Post a Comment