Before the next election: Important reading on the last one
When I cracked the spine of The Canadian Federal Election of 2006 I expected the kind of turgid phrasing and obtuse theorizing that is all to common in contemporary Canadian political science. Regardless of the quality of its content, by its very nature, this subject attracts a limited audience at the best of times. In this case, given the unimaginative presentation and title, I had reason to expect the worst.
Yet once past the blandness of the book's design and into the meat of its articles, I found a very readable collection of topical essays about the previous federal election - including one must-read standout, which I'll get to in a minute. As a whole, the book is a concise and rewarding package, providing the back-story and context to the actors and parties who played key roles in the recent 2006 federal election. The work could serve as a useful accompaniment to the recently published Right Side Up by Paul Wells, and provides particular value for government relations practitioners - most of us, after all, are confined in our electoral experiences to a single political party, whereas this book, with its ecumenical approach to all political parties, gives valuable insight from all other camps.
True, the book does seem to belie its academic roots in the poor quality and seemingly secondhand information used to assess the successful Conservative campaign. Simply put, the Conservatives, tight-lipped even with their friends at the best of times, saw no need to share anything with the writers of this book, and both sides suffer for it. But despite that weak point, this book tells how each party performed in the election, and identifies overall trends in crucial aspects such as candidate nominations and electioneering practices. An excellent example of this is the piece by Stephen Clarkson, How the Big Red Machine Became the Little Red Machine, an article strongly critical of the narrow-minded and vindictive mindset of the people around Paul Martin, and the consequences of their decade-long civil war with the Chretienites, which combined to cripple the Liberal campaign.
Another standout is the article by Alan Whitehorn, offering an insider's account of the NDP election campaign and how it dealt as directly as it could with the need to stem the flow of soft NDP voters casting strategic votes for the Liberals in an effort to stop a surging Stephen Harper and the Conservative Party - an effort that was largely successful, as the NDP picked up a number of seats from the sagging Liberals.
The gem in this collection of essays, however, is clearly the one by Michael Marzolini of Pollara, despite its weak title, Public Opinion and the 2006 Election. Marzolini has done an incredible service to the Canadian political community by publishing this frank, succinct article on how Canada's pollsters performed in the 2006 election, although a more descriptive title might have been Leading Canadian pollster shows how his fellow pollsters were mostly right in 2006 Election, but notes glaring errors, particularly by Allan Gregg and the Globe and Mail. Okay, perhaps that's a little long...
But the topic is timely, especially for PAAC members with a taste for polling and for pollsters. After being embarrassed by the glaring difference between the last pre-election polls and the actual ballot results in the 2004 federal election, Canadian news media and their hired guns, the pollsters, entered the 2006 campaign determined to do more and better polling in order to avoid being stung again. Consequently, 2006 saw more preparation for published polling from more firms than ever before. The daily tracking polls by SES/CPAC produced 26 published reports during the campaign, in addition to nightly broadcasts on the Cable Public Affairs Channel (CPAC). The Strategic Council/ Globe and Mail published 22 reports.
Marzolini's narrative shines as he uncovers other glaring errors committed during the campaign, the most interesting and consequential of which originate with the Strategic Council's work for the Globe and Mail. Marzolini also covers the other polling companies and their efforts during the campaign, but the Strategic Council/Globe miscues are the most interesting and important, not because they did a bad job overall, but due to their high prominence and wide readership.
The first noted error in the campaign comes from what Marzolini characterizes as a case of media outlets manufacturing news. The Globe and Mail, its editors apparently bored after days and days of no movement in public opinion, published a December 5 front page article under the headline "Liberals Surging in Ontario," reporting a 2 percent gain, which was well within the poll's 4.1 percent margin of error. This "surge" quickly dissipated since it was, in fact, a mere statistical blip.
Worse, though, was a second front page headline on December 12: "Liberals Snatching NDP Votes in Ontario," which reported NDP support dropping in Ontario from 17 percent to 9 percent. In fact, nothing of the sort had occurred. One of the Globe and Mail writers had simply mixed up figures on "voter intent" and "campaign momentum." After a full two days of damage to the NDP campaign, The Globe owed up to its error when it ran a terse correction but no apology.
On January 17, six days before the vote, the Globe's published polls showed Harper's Conservatives in majority territory, with an 18 point lead. This poll was far off from where other pollsters were tracking - SES had the Conservatives with only a 5 percent lead, and Ekos showed them with just an 8 percent lead. Marzolini says this was not just a rogue poll, but instead was due to bad methodology, including poor questionnaire preparation. "The first question placed in front of the vote-preference question skewed the results," Marzolini says. "The first question in the Strategic Counsel survey reads, 'Do you support a change in government?' If a respondent answers, "Yes," indicating they believe the Liberals should be defeated, it then becomes difficult to follow this with "I'm voting Liberal" when asked for a vote choice," Marzolini points out. Even though the January 17 poll was prominent, especially due to the Globe and Mail's reputation and readership, no apology or retraction was ever issued for this poll. The article also reviews the substantial differences between the relatively inexpensive media polls and the much more in-depth internal polling done for the respective parties.
Marzolini's article is so important that I strongly recommend it as essential reading for every active GR practitioner. It is easily one of the most substantial yet accessible critiques I've encountered on the subject of recent Canadian political polls and election polling practices. Hats off to Michael Marzolini and The Dundurn Group for publishing such great work.
The entire book: Recommended. The Marzolini article: Strongly recommended, in fact, a must-read.