In an age of virality, it appears that the only thing that travels faster among humans than a novel coronavirus is information. Now, with the exponential spread of COVID-19 cases in recent weeks, stories and data about the virus are everywhere – on social media, in the media, and filling up pre-print databases.
Reactions to news of outbreaks and associated containment measures in Switzerland have ranged from fear and panic-buying, to concern over the economic impacts of such measures. Meanwhile, others seem to be wondering: “what's the big deal? It's just the flu!”
With the digital data deluge evolving every day – and even by the hour – it’s no wonder public perceptions of the pandemic vary so greatly. Indeed, Robert West, who leads the Data Science Lab (DLAB) in the School of Computer and Communication Sciences, believes that it’s going to take more than statements from scientists to communicate the facts of the COVID-19 situation, and to encourage curve-flattening behaviors like frequent hand-washing and social distancing.
“There is a very human aspect to facts; it’s not just about the facts themselves, but about how people understand them,” says West, who is also a member of the UNIL-EPFL Center for Digital Humanities (dhCenter).
“When you go from scientific papers to social media, audiences don’t have the same background, and we need to understand how they read these kinds of information. That’s something many scientists don’t care much about, and it’s where digital humanities research can really play a role.”
Choosing a messenger
West describes an ongoing research project, funded through the Collaborative Research on Science and Society (CROSS) program, which underscores the importance of understanding how people understand data, especially when it comes to controversial or risk-laden issues. He and his colleagues set out to understand how people’s opinions on four hot-button topics – climate change, abortion, vaccination and immigration – were swayed based on the statements of celebrity spokespeople.
The researchers randomly mixed and matched real statements with real celebrities, to test their hypothesis that statements attributed to known and respected spokespeople had more influence than those from unknown or disliked sources. Study subjects also read statements attributed to an “expert”, who unbeknownst to them was made up by the researchers.
“We expected experts to have a bigger effect on changing opinions than disliked celebrities, but our preliminary results actually show that experts have the least impact, and that their statements can even backfire. I think this is one example of how digital humanities can help in this crisis, because it can help shape the strategies we use to educate the public.”
Three tips from a data scientist
EPFL professor Robert West offers some advice for how to navigate COVID-19 data without falling victim to false information, whether from fake news or simple misinterpretation.
Check the sources of data interpretations…and of the data itself. “I think it’s obvious advice, but even I have to remind myself of it, because as humans we have an innate hunger for sensational news,” West says. In addition to looking for reputable sources (universities and respected news outlets top social media, for example), he advises always reading graph and infographic captions carefully. Since different countries have different approaches to collecting epidemiological data and testing for COVID-19, their numbers aren’t always comparable.
Read the axis labels. Check to see if graphs are linear or logarithmic, as this makes a big difference when it comes to understanding exponential growth.
Keep perspective. Bear in mind that data on the COVID-19 pandemic is constantly changing. “What we’re seeing online reflects events that are unfolding in real time, and what we read today may be different tomorrow,” West says.