June 18, 2004 - Voters beware! If you believe the polls in the media, then you must think Canada has a democracy based on proportional representation.

Take the Pollara poll on the National Post's front page yesterday. The poll is of 1,267 Canadian voters, 36% of whom said they would vote for the Conservatives, 31% for the Liberals, 16% for the NDP, and 12% for the Bloc Quebecois.

Contrary to popular delusion, receiving the endorsement of a sample group does not equal victory. In the 2000 election, the Progressive Conservative Party had 12% of the vote but won 4% or 12 seats in the House of Commons. Hardly a victory for The Right Honourable Joe Clark, but a much better showing than both the New Democratic Party (13 seats) and the Bloc Quebecois (38 seats). What matters in the elections (and not to the polls) is the concentration of votes in each riding, not the percentage of votes received overall.

Still, the impression given by the Conservative Party-leaked Pollara poll is that the Tories are kicking the Bloc Quebecois` butt. However, the Bloc is only running for Quebec's 75 seats and only has to worry about 5.5 million voters, a quarter of the Canadian voters. If you accept the premise of the poll, that 12 per cent Bloc vote has the party holding a comfortable 46% of the decided Quebec vote. Add the detail that Quebec voters elected only one Progressive Conservative to Parliament and the picture presented to Can West News readers contradicts the impression of a Tory victory. Instead, the Bloc looks set to kick the Tory and the Liberal butts out of Quebec.

The poll`s margin of error, plus or minus two percentage points, also raises questions about the so-called Conservative lead. With such a margin of error, the results are too close to say with confidence that the Conservatives are ahead of the Liberals. The truth is that the Pollara poll only tells us that the Liberals and the Conservatives are on the same playing field. We don't know whether this sample of 1,267 people might change their minds tomorrow. Answering a telephone survey is much easier than making a final decision to vote. They might not even go vote.

Canada`s election participation rate has dropped significantly since the 1988 "free trade" election that gave Brian Mulroney a second term. The chattering classes have fingered Ontario as the key battleground. However, Ontario is where voter participation has dropped the most in Canada. Only 58% of eligible Ontario voters cast their ballot in the 2000 election. Will those 3.2 million Ontario voters participate in this election? These disillusioned people are the political wild card in Ontario.

But this question of participation, that strikes at the core of Canadian democracy`s legitimacy, is not what is on the front page of the July 17 National Post and Ottawa Citizen. Nope. It is the Pollara poll.

The front page exposure shows the ideological dedication of the editors to overcome the standards of good journalism such as fact-checking and independent sources. Buried at the back of the story, the Post not only admits "senior party insiders provided the poll to CanWest News Service" but that the pollster himself, Pollara chairman Michael Marzdini, could not be reached for comment. At one point in the story, the writer claims "Conservatives now expect to pick up 65 ridings or more" but does not even bother to attribute it to a source.

The release of the poll was clearly a successful bid by the Conservative Party to get the National Post to be its cheerleader and demonstrate momentum. Based on the polling numbers, there is no explanation for why the Conservatives would win 61% or more of Ontario`s ridings with just 36% of the sample`s votes.

Polls are commissioned for their extraordinary propaganda value and the value placed on them by gullible assignment editors. There is no real connection between the numbers given by the polls and what will happen on election day. Polls are theoretical, the statistical equivalent to a paint-by-numbers picture. The broad sweep of these polls make them nearly meaningless. Of course, meaning and elections rarely intersect, especially when there is a candidate to promote.