Despite a burgeoning literature on people’s acceptance of fake news, we find few studies have been devoted to exploring why and how people share news, particularly fake news (cf. Halpern et al., 2019). Using social network analysis across 4 social media forums (Twitter, Facebook, Instagrama and Whatsapp, we find that people tend to filter out real news and share fake news. Our two-step experiment suggests that people do not invoke their analytical capabilities when sharing news. Instead, partisan affinity as well as their social networking (discussion of politics, and internet usage) strongly predicts news items they share. However, when asked to assess news, analytical thinking matters and there are no significant differences between people from different parties. Since people normally do not have the time to cross-check and analyze the validity of the news items they encounter, they tend to make a lot of mistakes when it comes to sharing fact-checked items. Indeed, fake news infects networks more so than true news, and most worryingly of all, everybody shares fake news.