Research update #2: Countering misinformation, and demographic differences in news consumption
The best way to tackle misinformation is to understand as much as possible about where and how it spreads.
In the second in a series, our Research Manager Amy Sippitt takes a look at some of the latest findings and updates about the spread of misinformation and the culture of factchecking.
Honesty in public debate matters
You can help us take action – and get our regular free email
More evidence that backfire effects are rare…
An article by Andrew Guess and Alexander Coppock reporting the results of two studies with workers on Amazon’s MTurk platform and one nationally representative sample in the US failed to find backfire effects when presenting people with attitude-inconsistent information on gun control, minimum wage and capital punishment.
They conclude: "We generally find that subjects update, if at all, in the direction of the information. ... our results suggest that while backlash may occur under some conditions with some individuals, it is the exception, not the rule".
The results are particularly interesting because they fail to find backfire even for controversial issues like gun control. But the finding of “if at all” shows there’s still a lot of room for improvement.
On a similar theme… a separate, smaller study of 700 MTurkers in the US by Mirya Holman and J. Celeste Lay set out to see how people react when presented with a news article on election fraud.
The articles varied whether PolitiFact, Mother Jones or Breitbart were referenced (albeit, briefly) as having found no evidence of election fraud.
They mostly found little change in beliefs at traditional statistical confidence levels (5%), suggesting in this case neither non-partisan nor partisan sources made a big difference.
Only Democrats seeing the PolitiFact or Mother Jones references showed some reduction in belief in election fraud. (The authors also report further results found outside of a 5% confidence level).
Tackling vaccine misinformation requires listening and engagement
Targeted social media can combat misinformation, says the Vaccine Confidence Project – which is doing research to detect early signals of rumours and scares about vaccines.
They find: "No single strategy works for all types of misinformation, particularly among those who are already sceptical. Educational materials and resources are important, but limited; health officials and educational campaigns often fall short because they craft messages based on what they want to promote, without addressing existing perceptions.
“Dialogue matters. Strategies must include listening and engagement”.
People may pay more attention to commentary on Facebook posts featuring news articles than to the text in the article previews
Research by Nicholas Anspach and Taylor Carlson randomly assigned different Facebook posts featuring a Yahoo! News article on a Trump approval rating poll to 954 MTurkers.
They found that comments in the post questioning the poll’s accuracy appeared to have a greater influence on readers than the text in the preview of the news article.
Comments contradicting the article preview also led audiences to question the accuracy of the article’s content – and regard both the news outlet and the person posting the comment as less trustworthy.
Distributed forms of news online (like social media and Google) are particularly important for people from lower socio-economic backgrounds
Research by the Reuters Institute based on a survey of about 2,000 people in the UK found:
- There’s little difference on average in the number of sources people use offline, but lower social grade individuals use significantly fewer online sources on average.
- "Lower social grade individuals are significantly less likely to go direct to news providers, whereas lower and higher social grade individuals are equally likely to rely on distributed forms of discovery (relying on social media, search engines, and the like)."
So they conclude that distributed forms of discovery are particularly important for lower social grade individuals as news consumption moves more online.
They suggest that social inequality in news consumption may be greater in countries with less of a tradition of popular tabloid newspapers and less widely used public service media.
Caveats:
Their sample excludes 8% who said they had not consumed any news online or offline in the past month.
News avoiders tend to be from lower socio-economic backgrounds, so inequalities are likely to be larger if these individuals are taken into account.
Age differences in news consumption
The Pew Research Center has released a report comparing news consumption and media trust between younger and older adults across eight Western European countries.
Across the eight countries, adults aged 18 to 29 were twice as likely to get news online than from TV.
The UK stands out as the only country where younger adults are still using public news organisations (the BBC) almost as much as older adults. In other countries, younger Europeans were less likely to use public news organisations and had more varied media diets.
Younger Europeans were also more likely to rate the media poorly for coverage of key issues. For example, in the UK, only 35% of 18-29 year olds said the news media did a very/somewhat good job at investigating the government, compared to 57% of those aged 50+.
Why is there a gender gap in news avoidance?
One study by Benjamin Toff and Ruth Palmer set out to explore why individuals avoid news, interviewing 43 lower- and middle-income people in the UK (36 of which were women) who said they followed the news less frequently than once a month.
They found that caregiving responsibilities (which tend to fall more on women) – especially in homes with children and working parents – tended to leave people feeling they had little time and emotional energy left for news.
Some wanted to shield their children from what they perceived to be negative content, and some tended to outsource news consumption to partners or family members (male or female).
Some female respondents also appeared to see interest in the news as predominantly a male characteristic.
Caveats:
There aren't that many people who actively identify as news avoiders (this year’s Reuters Digital News Report found about 7% of adults online in the US and UK say they've accessed no news in the past month) – so participants are at the extreme end of news avoidance.
Women also made up the majority of the participants, so the research can’t offer much of a comparison to male news avoiders.
(We’ve written here about the apparent gender gap in factchecking audiences too).
If you’d like more to read…
Here’s a study of factchecking initiatives in Bangladesh, India and Nepal, and the challenges they face. Read more
Here’s a quick overview by Rasmus Kleis Nielsen on disinformation research needs. Read more