SHARENEWS: Predicting the Shareworthiness of ‘Real’ and ‘Fake’ News in Europe

In this project based, we analyze parts of the Facebook URL Shares dataset. It is funded by the Social Science Research Council.

From the abstract:

Given the central role that information plays in democratic societies, one of the pressing challenges for the social sciences is to better understand which information (news, but also disinformation, misinformation, and malinformation) citizens are exposed to. While traditional media are still the main source of information for some groups of citizens, in other groups, this exposure is increasingly shaped by what others share on social media. However, our understanding of what actually gets shared, or not, is limited. The proposed project aims at explaining and predicting the sharing of information in the domain of news and politics in four European multiparty systems—Germany, Italy, the Netherlands, and Poland. We know from decades of research on news values that some characteristics inherent to or ascribed to news events (such as the relevance, importance, unique character, or involvement of elite actors) influence what is considered newsworthy; in addition, the specific framing of a news story (for instance, putting an emphasis on conflict or human interest) can lead to increased attention. We have comparatively little evidence, though, about the role of these features when it comes to the shareworthiness of news. We then compare whether these features have a different role when comparing legitimate news with so-called fake news, when comparing mobile devices with desktops/laptops, and when comparing countries.

Damian Trilling is the principal investigator. More info on the whole team here.