Did Twitter help spread campaign video?
Coverage of ‘manipulated media’ failed to convince viewers it was misleading
OTTAWA—News coverage of Twitter’s labelling a Liberal campaign video as “manipulated media” served to amplify that video’s reach and failed to convince viewers the information was misleading, according to a new report.
A research team working with McGill University and the University of Toronto found that only 10 per cent of respondents saw the video on Twitter, where it was originally released by Liberal candidate Chrystia Freeland last Sunday.
But 62 per cent of respondents in their survey said they encountered the misleading video through the resulting news coverage. And despite most of that news coverage describing the video as misleading, most respondents still believed the video to be truthful (52 per cent) or were not sure what to believe (13 per cent).
The findings present a challenge for news media covering the federal campaign, according to researcher Aengus Bridgman.
“I think the media is in an incredibly difficult place on misinformation issues in general, and there have been numerous instances when that amplification has occurred,” Bridgman, research director with the Canadian Election Misinformation Project, said in an interview Friday.
“The most dangerous takeaway of this piece for me is that huge difference between whether Liberal (respondents) believe Erin O’Toole wants to privatize health care and whether Conservatives believe that,” he said. “You have these very polar opposite reactions to this information.”
The video posted by the Liberal campaign on Twitter featured comments O’Toole made in 2020 about provinces experimenting with some privately provided health care within Canada’s system of universal coverage.
It showed O’Toole saying he would permit private service providers in health care, and that he believes private sector investment can drive innovation. Edited out of the video, however, was O’Toole’s statement that he nonetheless would “make sure universal access remains paramount.”
The tweet was shared by many other Liberals, including Liberal Leader Justin Trudeau.
But the video was also quickly labelled as “manipulated media” by Twitter, a social platform that has been under intense pressure — including from the Trudeau government — to address the issue of misinformation and disinformation being shared during election campaigns.
The Liberals said they are appealing Twitter’s labelling, and the edited video remains posted on Freeland’s account.
On the campaign trail Friday, Trudeau again attacked O’Toole on the issue, accusing the Conservative leader of wanting to “hollow out” universal access to health care by permitting more private care.
“What Erin O’Toole is saying is that we should have ‘choice’ in health care — and what that means is that he believes people who are wealthy should be able to pay and cut ahead of the line and get private services,” Trudeau told reporters.
O’Toole’s platform says a Conservative government would increase federal transfer payments for health care by six per cent annually — a policy it estimates would increase provincial health care funding by $60 billion over a decade. The Conservative leader said this week that he supports a “public and universal system, full stop.”
But public health care advocates warn any increase in the number of for-profit health care providers in the country could lead to a “two-tier system.”
Bridgman noted that parties playing fast and loose with the facts is not exactly a new phenomenon in Canadian politics or elections. But ironically, flagging misleading content in political advertising might actually serve to draw voters’ attention to those ads.
“For sure this is going to be a part of politics going forward, and I think the media and the parties are going to be aware of it,” Bridgman said.
“If you have an arbiter like Twitter, if you can get the right label, it’s going to boost your signal? Then that’s a new tool.”
The Canadian Election Misinformation Project’s survey included responses from 806 Canadians between Aug. 23 and 26. Bridgman said that because the sample is not yet fully nationally representative, it could not provide a margin of error.