NewsWhip Spike was a crucial tool in fighting misinformation during France’s election campaign, by allowing editors to predict how big problematic stories were set to go.
Last year’s US election campaign highlighted the effects of misinformation created specifically for distribution on social media.
Now publishers and platforms are looking for strategies to fight back against the wave of misleading memes, fake stories, one-sided analysis and other problematic content online.
One of the groups that has been to the forefront of this effort is First Draft, a non profit organisation made up of leading verification and UGC experts looking for journalism based solutions to the problem.
Taking a proactive approach to combating the spread of fake news, the First Draft team looks to share best practices and new approaches to getting journalists and newsrooms up to speed on how to recognise, verify, and debunk misleading stories online.
So far this year, the group’s flagship project has been CrossCheck, an initiative aimed at combating the spread of fake news on social media during the French presidential election campaign, involving newsrooms and technology partners, including NewsWhip.
As First Draft Research Director Claire Wardle told us on the blog earlier this year, the aim of the collaborative project was to cut down on the spread of ‘misinformation’ about the candidates and campaigns.
The idea is that some cooperation between newsrooms can save a lot of time and effort in quickly debunking rumours, and gives the media generally a stronger platform to stand on together.
“I would argue there is a need for journalists to help audiences navigate the information ecosystem. People want to know what it true and what isn’t. But it makes no sense for 50 journalists in 50 different newsrooms to debunk the same rumour or to verify the same video that emerges on YouTube. There are better ways for journalists to be spending their time. So we believe the answer is collaborative verification.”
CrossCheck kicked off in February, when 37 participating newsrooms, including leading French national titles like Le Monde and Libération, digital and regional newsrooms like Rue 89, La Voix du Nord, and international newsrooms and agencies such as Bloomberg, Agence France Presse and France 24.
NewsWhip’s role in the project was to provide free access to our Spike dashboard, along with training, to participating newsrooms. Spike is a powerful real time content discovery platform, tracking the spread of content across social media networks, from sources around the world. Everything tracked in Spike is categorised by language, country and topic, making it easy for journalists to drill down into what their audience is sharing and talking about each day.
In the context of fake news and misinformation, this is a powerful asset.
“Spike was probably the most used tool to help us out in what we were doing (with CrossCheck)”, explains Sam Dubberley, project manager for CrossCheck.
“Basically because throughout the campaign, we always had this tension between ‘putting fuel on the fire’ or the question of ‘are we too late’, and the story has travelled halfway around the world already. That was really where Spike came into its own.”
It’s a fundamental question that journalists and other disinfo debunkers always pose to themselves and their colleagues when weighing up how to deal with a misleading story they’ve found online.
By addressing something that obscure and irrelevant to the larger political debate, the newsroom risks putting fuel on the fire by giving the perpetrators more exposure than the original piece warranted. But by waiting a day, the story could have gone super viral, already misleading readers, and leaving the newsroom playing catch up to produce a debunk that would only reach a fraction of the viewership of the problematic story, meme or image itself.
Spike has a predictive feature, which allows users to accurately forecast how much engagement content will achieve in the near future. In the context of debunking misinformation, this feature is extremely useful to newsrooms, as CrossCheck’s participants discovered.
“We’d find a URL or a tweet, or a Facebook post, and we could see how fast it was spreading. The predictive feature was so useful in deciding whether or not to write about a story. That was the shining light for Spike,” Sam explained.
And the predictive feature in Spike managed to pinpoint where early stage memes and misleading stories would end up with a large degree of accuracy. Spike’s predictive feature allows news organisations to get a data-informed view of what stories will be big on social media later in the day.
The model looks at the early volume of social engagement to find the ‘social velocity’ of different stories – or how fast they are spreading on social platforms. Based on that information, the system can then predict how much engagement the story will achieve over the course of the day. The more the system knows about the story, the further it predicts. For instance, if a story is one hour old, Spike can predict what its social engagement will look like in six hours time.
“The accuracy of the predictive feature was terrifying,” said Sam. “When something was predicted to get a lot of engagement, it actually did. Those metrics were really valuable in deciding whether to follow a story or not.”
Every day, the Crosscheck team came across egregious examples of hoaxes and purposefully misleading memes and other content about the campaign themes, and especially the candidates themselves.
The work was shared by the journalists from the newsrooms themselves, and the collaboration aspect was managed by a group of 12 journalism students, with Slack acting as a critical communication tool. Sam explains the daily work process.
“We’d check out Spike every morning and look at the stories from the known viral French sites that were gaining traction. Then we’d also have a list of stories from other sources like the newsrooms, or Twitter or Facebook.”
Spike was then used to see which stories should be covered.
“(Spike) allowed us to say ‘ok, we should cover this, because we can see in Spike that it’s going to get 8,000 shares in the next 24 hours, even though it only has 100 shares now.’ To have that early warning mechanism was hugely useful,” says Sam.
CrossCheck partners covered all sorts of stories on the campaign trail, from a report claiming that Emmanuel Macron washed his hands after shaking hands with voters, to a claim that the French government prevented overseas polling stations from displaying Marine Le Pen’s election posters.
“We saw a lot of stories that were slightly right, but took a fact, and pushed it to the edge of truth. I think the one thing we see about well done misinformation is that there’s a little bit of truth in it, which gives it a bit of reality. Then we have to investigate from there on in.”
Sam says it’s still too early to say whether the CrossCheck project had a demonstrable impact on the election’s coverage, and First Draft will be conducting analysis into the results to help guide future projects. “To give those answers, we need to talk to voters in France, audiences, the main media players, to get an idea of what people made of CrossCheck,” explains Sam.
For now, awareness of the problem has risen, meaning that newsrooms are getting closer to addressing the problematic stories authoritatively when they first appear.