An anatomical comparison of fake-news and trusted-news sharing pattern on Twitter

Abstract

Online social networks allow users to share a variety of multi-media content on the World Wide Web. The rising popularity of such social networking platforms coupled with limitations in verifying the veracity of shared content has contributed to increase in misinformation on these media. Misinformation content such as fake-news and hoaxes, though often considered innocuous, may have high social cost such as influencing elections decision, and thus should be investigated carefully. Many researchers have studied various aspects of fake-news including automated ways to recognize it. However, a large-scale study comparing the sharing patterns of fake-news and trusted-news is missing. In this research, we take Ukraine, a country where fake news is common, as a case study. Using datasets generated by three different Tweets collection strategies, we present an anatomical comparison of fake-news and trusted-news sharing pattern on Twitter. Such a comparison enables to identify the characteristics of tweets sharing fake-news, and allows to find the users who are more inclined to share misinformation. Besides, we also study possible bot activities in the dataset. The top conclusions derived from this study are (a) Users sharing fake-news stories are more likely to include hashtags, and the hashtags used in Tweets sharing fake-news stories are similar to hashtags used in Tweets sharing trusted news. (b) Users sharing fake-news are also more likely to include mentions, but mentions used in tweets sharing fake-news and trusted-news are often different. (c) Tweets sharing fake-news have more negative sentiment. In contrast, tweets sharing trusted-news have more positive sentiment.