Fears of election meddling on social media were overblown, say researchers

Hype versus the reality (perhaps Canada not important enough…). The hype was in both mainstream and ethnic media:

Now that the election is over and researchers have combed through the data collected, their conclusion is clear: there was more talk about foreign trolls during the campaign than there was evidence of their activities.

Although there were a few confirmed cases of attempts to deceive Canadians online, three large research teams devoted to detecting co-ordinated influence campaigns on social media report they found little to worry about.

In fact, there were more news reports about malicious activity during the campaign than traces of it.

“We didn’t see high levels of effective disinformation campaigns. We didn’t see evidence of effective bot networks in any of the major platforms. Yet, we saw a lot of coverage of these things,” said Derek Ruths, a professor of computer science at McGill University in Montreal.

He monitored social media for foreign meddling during the campaign and, as part of the Digital Democracy Project, scoured the web for signs of disinformation campaigns.

Threat of foreign influence was hyped

“The vast majority of news stories about disinformation overstated the results and represented them as far more conclusive than they were. It was the case everywhere, with all media,” he said.

It’s a view mirrored by the Ryerson Social Media Lab, which also monitored social media during the campaign.

“Fears of foreign and domestic interference were overblown,” Philip Mai, co-director of the Social Media Lab, told CBC News.

A major focus of monitoring efforts during the campaign was Twitter, a platform favoured by politicians, journalists and partisans of all stripes. It’s where a lot of political exchanges take place, and it’s an easy target for automated influence campaigns.

“Our preliminary analysis of the [Twitter hashtag] #cdnpoli suggests that only about one per cent of accounts that used that hashtag earlier in the election cycle can be classified as likely to be bots,” said Mai.

The word “likely” is key. Any social media analyst will tell you that detecting bonafide automated accounts that exist solely to spread a message far and wide is incredibly difficult.

#TrudeauMustGo and other frenzies

A few times during the campaign, independent researchers found signs that certain conversations on Twitter were being amplified by accounts that appeared to be foreign. For example, the popular hashtag #TrudeauMustGo was tweeted and retweeted in large numbers by users who had the word “MAGA” in their user descriptions.

But this doesn’t mean those users were part of a foreign campaign, Ruths said.

“It’s very hard to prove that those MAGA accounts aren’t Canadian,” he said. “How can you prove who’s Canadian online? What does a Canadian look like on Twitter?”

Few Canadians use Twitter for news. According to the Digital News Report from the Reuters Institute for the Study of Journalism, only 11 per cent of Canadians got their news on Twitter in 2019, down slightly from 12 per cent last year.

Twitter’s most avid users tend to be politicians, journalists and highly engaged partisans.

Fenwick McKelvey, an assistant professor at Montreal’s Concordia University who researches social media platforms, said he feels journalists overestimate Twitter’s ability to take the pulse of the voting public.

“Twitter is an elite medium used by journalists and politicians more than everyday Canadians,” McKelvey told CBC News. “Twitter is a very specific public. Not a proxy for public opinion.”

In fact, most Canadians — 57 per cent — told a 2018 survey by the Social Media Lab that they have never shared political opinions on any social media platform.

Tweets for elites

For an idea of just how elitist Twitter can be, take a look at who is driving its political conversations. For some of the major hashtags during the election — like #cdnpoli, #defundCBC and the recently popular #wexit — just a fraction of users post original content. The rest just retweet.

And the users who get the most retweets, the biggest influencers, represent an even tinier sliver of Twitter users, according to data from the University of Toronto’s Citizen Lab, another outfit that monitored disinformation during the campaign.

“What we thought was a horizontal democratic space is dominated by less than two per cent of accounts,” said Gabrielle Lim, a fellow at the Citizen Lab.

“We need to take everything with a grain of salt when looking at Twitter. Doing data analysis is easy, but we’re bad at contextualizing what it means,” Lim said.

So why this focus on Twitter if it’s such a small and unrepresentative medium for Canadians? Because it’s easy to study. Unless a user sets an account to private, everything posted on Twitter is public and fairly easy to access.

On the other hand, more popular social networks like Facebook make it much harder to harvest user content at scale. A lot of misinformation may also be shared in closed channels like private Facebook groups and WhatsApp groups, which are nearly impossible for outsiders to access.

But even taking into account those larger social media audiences, the evidence shows that Canadians are getting their news from a variety of sources, Lim noted.

Although the threat posed by online disinformation to Canadian democracy was overblown in the context of the 2019 campaign, Ruths said he still believes it was important to be alert, just as it’s important to go to the dentist even if no cavities are found.

And he suggests that journalists looking for evidence of bot activity apply the same level of rigour as the people doing the research.

“We saw a lot of well-intentioned reporting,” he said. “But finding suspected accounts is not the same as finding bots. Saying that MAGA accounts don’t look like Canadians’ doesn’t mean they’re not.”

Source: Fears of election meddling on social media were overblown, say researchers

Conservatives, Liberals rank economy high in immigration file, but diverge on integration, cultural values, says survey

What one would expect given other polling data but still of interest:

Conservatives and Liberals tend to agree that jobs and the economy should rank high when it comes to the immigration file, but concerns for the plight of refugees and integration of immigrants depends on where one falls on the political spectrum, suggests a new study released today.

Whereas many Conservatives prioritize on cultural values, national security, and jobs, the Liberals and NDP place less importance on those concerns, according to a survey from the Digital Democracy Project, a months-long effort that the Public Policy Forum and McGill University’s Max Bell School of Public Policy are leading.

“Partisans differ in terms of what they’re talking about when they talk about immigration, what dimensions … they think about,” said Peter Loewen, a political science professor at the University of Toronto’s Munk School of Global Affairs, one of the report’s authors.

Respondents were asked to rank eight dimensions related to immigration, including social services and welfare; diversity and multiculturalism; and illegal immigration, were of high concern.

For example, Conservative partisans expressed more concern over illegal immigration than other partisans, with 42 per cent saying it’s a concern, compared to 28 and 27 per cent of respondents who identified as Liberal and NDP supporters, respectively, the survey suggested.

The data, based on an online panel survey of 1,559 Canadians, was conducted from Sept. 11 to 16. Online polls are not considered to be truly random and cannot be assigned a margin of error.

There are marked differences between the Liberals and NDP on the immigration file, too. NDP supporters rank Canada’s responsibility towards welcoming refugees as higher on the list over jobs and the economy, while Liberal supporters indicated it as less of a priority, with 29 per cent choosing it as a top concern, compared to 44 per cent who identify with the NDP.

Researchers also found that most Canadians are misinformed about Canada’s immigration levels and refugee intake. Asked how many refugees Canada admitted in 2018, only 12 per cent answered correctly, 61 per cent were unsure, and 24 per cent said it was higher than the actual figure of 28,000.

“The worrying takeaway is that the more people are exposed to traditional news, to social media, the more likely they are to give incorrect answers about immigration levels, refugee intake levels,” Prof. Loewen said. “People are taking misinformation from somewhere in the ecosystem.”

Nativism could also explain differences in views on immigration policy, the report noted. In seeking to measure the level of nativist sentiments with a series of questions, researchers found that while Canadians “exhibit modest levels of nativism,” Liberal and NDP supporters have lower scores than Conservative supporters. (To measure respondents’ openness to nativist sentiments, they were asked to rate six statements, including whether they agree “immigrants take jobs from real Canadians” and if Canada “would be stronger if we stopped immigration,” on a five-point scale.

Attempts to provide information on the economic benefits of immigration had an influence on respondents’ perception of immigration, according to the study. Half of respondents were given an excerpt from a 2018 Conference Board of Canada report that said immigrants are key to economic growth. Among those who weren’t given the report, 23 per cent said immigration was bad for the economy and 57 per cent said it was good. Those figures changed slightly to 19 per cent and 63 per cent among those who viewed the report.

“While theories of motivated reasoning suggest that partisan respondents will reject information that doesn’t conform to their existing values or beliefs, the effect of this intervention was stronger for right-leaning partisans than for left-leaning partisans,” the report noted. “…This suggests that providing the public with relevant information could also influence their opinions on public policy, and that nativism is not as much of an immutable sentiment as commonly believed.”

Though Canada isn’t immune from nativist and populist sentiments, the report noted that such expressions don’t mimic the trends in the U.S. and “far-right parties in Europe.” The report suggested that the embrace of populist sentiments is “most common” among NDP supporters than Liberals, while the Conservatives are in between.

Previous studies from the Digital Democracy Project have looked at how Canadians consume and share media and its effects on their support for policies in the lead-up to the federal election.

Source: Conservatives, Liberals rank economy high in immigration file, but diverge on integration, cultural values, says survey

Canadians who hold strong links to political parties more likely to be misinformed about politics, study finds

Depressing but not surprising. Strong beliefs lead to more “automatic thinking,” to use Kahneman’s terminology rather than questioning a belief or being more open to different beliefs:

A new study found Canadians who hold strong partisan beliefs are more likely to be misinformed about key political issues than more politically neutral voters.

Data released Wednesday by the Digital Democracy Project found “strong partisan” Canadians were more often incorrect when answering a set of 10 basic questions about current political issues. Those who had no partisan affiliation, or weaker ties to a political party, were less likely to give an incorrect answer.

The study asked 10 questions that had relatively clear answers — like whether or not Canada is currently on track to meet climate change commitments under the Paris Accord (no), or whether the deficit was greater in 2018 than it was in 2015 (yes).

The results suggest the more partisan a voter is, the more likely they are to give an incorrect answer. But they also suggest — perhaps counterintuitively — that the more traditional news Canadians consume, and the more time they spend on social media, the more likely they were to give an incorrect answer.

“Media consumption is as often equipping partisans with arguments to support their position as it is correcting them on facts, because the facts on these things are actually kind of hard to pin down,” said University of Toronto Professor Peter Loewen, one of the academics behind the study, on Wednesday.

“Over the course of the (federal election) campaign I think you’re going to find that people are going to have different sets of facts depending on what their views are, and they’re going to find those informed by what they read in traditional media outlets.”

The Digital Democracy Project is a partnership between the Public Policy Forum and the Max Bell School of Public Policy at McGill University. In the weeks leading up to the federal election, the project is tracking Canadians’ media consumption, social media usage, and the digital discussion around Canadian politics to put together a picture of how political information flows through the electorate.

The initial report, released publicly Thursday, found most Canadians trust traditional media organizations to accurately report political news. The survey asked respondents to rank their trust on a scale of zero to 10.

“Canadians trust mainstream news organizations (5.8) at similar levels as their friends and family (6.0). Canadians are comparatively much less trusting of the information provided by major political parties (4.8), and in what they read on social media (3.3 for all respondents, 4.2 for respondents who indicated they used social media for political news in the past week),” the report stated.

And unlike American voters, Canadians don’t seem to choose where they get their news from based on their partisan leanings. Liberal, Conservative and NDP voters reported they get their news from roughly the same outlets, with CTV Online and CBC Online leading across partisan lines.

But the report also found a disconnect between what voters are concerned about and what political issues take up the most oxygen in the press.

Respondents to the survey listed the environment, health care and the economy as the three most pressing issues. Looking at the Twitter conversations of a sample of 300 journalists, the report found that while a lot of discussion was devoted to environmental issues, the journalists paid little attention to health care and almost no attention at all to the economy.

On the other hand, the journalists spent a lot of time talking about ethical issues (think: the SNC-Lavalin affair) and foreign affairs, the public listed both categories as much less important.

Source: Canadians who hold strong links to political parties more likely to be misinformed about politics, study finds