The Diffusion of Mainstream and Disinformation News on Twitter: The Case of Italy and France

In this work we provide preliminary results from an ongoing investigation on the Twitter diffusion of news pertaining to two classes of sources, namely websites which notably produce disinformation, i.e. misleading and harmful information, opposed to more traditional and mainstream websites which instead publish credible information. We used the Twitter Streaming API to collect a large-scale dataset of thousands of tweets containing links to news articles in two different countries, Italy and France. We show that mainstream news outlets generate a much larger engagement in both settings, with a larger discrepancy between the two news domains in France. We also show that only a handful of Italian outlets actively engage with Twitter users, whereas in France there is a larger number of outlets sharing misleading information which exhibit a non-negligible volume of shares. We observe a strong tendency towards sharing mainstream news in those users who also share non-credible information in both countries. Analyzing the diffusion networks of distinct news domains and countries, we observed that disinformation networks are more clustered and connected, but much smaller than the mainstream ones (with the largest discrepancy in the French scenario).