Clickbait, the intentional use of exaggerated and misleading content to entice people to click on a link to a particular web page, is a phenomenon that has grown rapidly in recent years. Clickbait has become problematic in the post-truth era as it can de-value digital content and erode people’s trust. The practice is especially common in social media where wider audiences can be reached more rapidly. Despite studies of clickbait being conducted on various social media sites, there has been little investigation of WeChat, the most popular social networking site in China. In this paper, we investigate clickbait behaviour in WeChat by analysing two samples (17,898 and 18,316 articles) for clickbait using supervised clickbait classifiers where an F1-measure of 0.834 is obtained using Naïve Bayes. We train and test the classifier by manually annotating a sample of 3000 examples for clickbait. Results show that approximately 70% of our WeChat examples are likely to be clickbait. We find that articles from publishers categorised as Funny, Anime, Entertainment and Culture exhibit the most clickbait, as well as posts from publishers in specific regions, such as Guangdong, Beijing and Jiangsu. We discuss the implications of results and provide recommendations. As far as we are aware, this is the first large-scale study of clickbait activity in WeChat.
CORE (COnnecting REpositories)
CORE (COnnecting REpositories)