How the media twisted a healthy discussion about "crowd size" into a story about election result denial

We see you

There’s no accounting for taste, as our Mothers might say, but there’s no excuse for the recent nonsense on the federal campaign trail.

Last Wednesday morning, I wrote a Star column that tried to square the circle around relative crowd sizes (see prior post “Does crowd size matter in an election campaign? Mark Carney had better hope not” Apr 9-25). I drew upon the firsthand experience of several former staffers and volunteers from campaigns dating back to 1965, and was lucky to be able to quote a “legend” (Freddy Watson) — someone who’d driven a Tory campaign bus in four different decades.

In that piece, I reflected on three things:

1. Pierre Poilievre is drawing far larger crowds than Mark Carney, yet - “Perhaps it means nothing more than that Poilievre’s voters are ‘motivated,’ while Carney’s polling lead indicates a broader base of support from folks who don’t have time for speeches.”

2. Pros are perplexed - Watson says he “could usually tell by day 14 if we were going to win or lose. I don’t know what to make of these large Poilievre crowds.”

3. Poilievre’s unusually large crowds may be meaningless, but - If pollsters are missing a bunch of voters, as they did during the 2016 United States presidential election, isn’t this what it would look like?

What that piece didn’t do was suggest, even obliquely, that the polls were wrong. I know enough about the people behind several of these firms to be confident that their work is both consistent and professional. Some of the leaders behind these firms are more popular with the Liberal government than others, but if six or eight different pollsters produce similar data over a two week period, that’s not because they’re colluding against a particular political party.

During a process event on Thursday in Milton, Ontario about housing, a journalist asked Mr. Poilievre “if he believes the polls and if he will accept the results of the election no matter the outcome.”

Although he passed on the first part of the question, it shouldn’t have surprised anyone that Mr. Poilievre said “Yes” when it came to accepting the outcome. Not that he had to be asked, of course.

The Globe and Mail, Friday

Suddenly, the media corps travelling with Mark Carney felt it newsworthy to ask the Liberal Leader “if he believes the polls and commits to recognizing the election results.”

“Of course,” Mr. Carney said, after making a public-safety announcement in Brampton. “This is Canada.”

The Liberal Leader said there are “certain parties that just import all their slogans and their policies from America, but let’s not import that nonsense into Canada.”

“We will have, and we are in the midst of, a fair and open and free election in Canada,” he said.

There isn’t a sane person in the country who has suggested that Canada isn’t running a fair election, but the Press Gallery appears to have turned a healthy discussion about relative crowd sizes (“What makes the 2025 campaign weird is that neither Watson nor I can recall a time when the guy with the largest crowds wasn’t also winning”) into a specious debate about whether or not we are on the verge of Donald Trump-style election denial games.

Back at Liberal Campaign HQ, any day is a good day when journos are talking about Donald Trump and not Mr. Carney’s defence of Liberal candidate Peter Yuen’s “connections to Beijing-friendly groups that have endorsed annexation of Taiwan and played down China’s repression of its Muslim Uyghur minority in Xinjiang province.”

The Toronto Star, Friday

For decades, politicians from around the world have relied on some form of the saying “The only poll that counts is the one on election day” to brush off media questions about their party’s relative standing during an election campaign. There are references to the very phrase in the archives of both President Gerald Ford (R) and Jimmy Carter (D) from the 1980 U.S. Presidential Campaign, as well as in the UK and Australia. There’s even a Wikipedia page, it’s that common.

During the 1957 election race, Progressive Conservative Leader John Diefenbaker was asked how he felt about polls showing his party trailing Uncle Louis. His memorable line lives on to this day: “I was always fond of dogs as they are the one animal that knows the proper treatment to give to poles.” The shorthand is now simply “Polls are only good for dogs,” although I’m not sure Dief uttered those exact words. Regardless, Dief turned out to be right — he beat Mr. St. Laurent in 1957.

Over the last 50 years, federal turnout has fallen from 71% to a range of 59-68% (let’s ignore 2021 given Covid), although the change elections of 1979 & 1984 saw 75% of eligible voters deposit something in a ballot box. Whether due to polling failings or higher-than-normal turnout to come, there’s nothing “tinfoil hat” about asking this question during the 2025 campaign given these relative crowds:

“if pollsters are missing a bunch of voters, as they did during the 2016 United States presidential election, isn’t this what it would look like?”

A Star colleague covered Mr. Poilievre’s rally in Brampton on Wednesday night, and filed a column under the headline I attended a Pierre Poilievre rally to figure out the disconnect between his large crowds but shrinking support in the polls. According to the daily polls from Nik Nanos CM, for example, Conservative support has been rock solid since March 28th, but headline writers don’t always track what we actually write. More importantly, are we conflating a tried-and-true tradition of debating polls with something more sinister?:

So how do his supporters square what they’ve just seen with all the polls that tell them Mark Carney and the Liberals have a solid lead? On the way out of the rally I ran into a group who had unfurled a big blue and white banner blaring the question: “DO YOU BELIEVE THE POLLS?”

In the context, the expected answer was clearly “no.” “Screw the polls, man,” offered one Poilievre supporter passing by. One of the banner’s organizers, Paul (no last name please), maintained the pollsters can’t be trusted. “They’re biased. They admit it.” Another organizer, Mike, argued the barrage of polls showing a Liberal lead are aimed at softening up voters for a Carney win: “It’s just a step toward making that inevitable.”

There’s a lot of stuff like this all over social media, as you’d expect, and it’s being validated by elements of the so-called mainstream media. A columnist in the Sun wrote after the Brampton rally that a Liberal victory is something “which the election industry is continually pushing.” I’d never heard that expression. It implies the pollsters aren’t just wrong but are part of a cabal conspiring to “push” a message they presumably know is wrong.

The next step along that path might well be to discredit the election result in advance, on the grounds that voters were fed false information. Both Poilievre and Carney, in fact, were asked on Thursday whether they believe the polls and will accept the result on April 28. Poilievre simply said “yes” to the last question but ducked the first. Carney seemed appropriately appalled by both questions. “Of course,” he said. “This is Canada. We always accept the results of an election.”

I doubt it will go that far but we shouldn’t dismiss the importance of the disconnect between rallies and polls. The mistrust on display in Brampton isn’t going away.

Speaking of my colleagues, in 2015, The Star’s Donovan Vincent arrayed a series of successes and failures by the polling community in a piece titled Canadians have had a steady diet of public opinion polls. Are they for the dogs? Errors in sampling were highlighted, as was the reality that “we can weight for the known knowns, we can adjust for the known unknowns, but we can’t adjust for the unknown unknowns. That’s part of the problem.” This piece proves that one can engage in a straight-up analysis of poll veracity without devolving into election denial territory.

I’ll given the final word to Nate Silver, who knows more about polling than almost anyone on the planet. In the wake of the 2024 Presidential campaign, he had this to say:

Where things look worse is statistical bias: the extent to which polls miss in the same direction and consistently underestimate Democrats or Republicans. Historically, the direction of bias has been unpredictable cycle-to-cycle. For example, you’d see polls underestimate Democrats one year (say, 2012) and Republicans the next (2014 and 2016), with no clear year-to-year pattern.

But the polls have underestimated Republicans across the past three presidential elections.

The polls underestimating Trump again is concerning, especially because of how much work pollsters put into avoiding this exact result.

By consciously introducing election denial to a discussion about relative crowd sizes, serving those who want to unfairly paint Mr. Poilievre as Trump Light, the Press Gallery again contributes to a bias narrative that reminds us all why “overall trust towards news in Canada” has dropped “by almost 20 percentage points since 2018,” according to the Reuters News Institute.

MRM

(this post, like all blogs, is an Opinion Piece)

(Ed note: thx to an eagle-eyed Subscriber, this post has been updated to reflect that it was St. Laurent and not Pearson in 1957; Doh!)