[month] [year]

IEEE-ISTAS20

Dr. Kavita Vemuri and her student Vivek Kaushal  presented a paper virtually on A large scale study on the impact of clickbait headlines on credibility of news at IEEE Symposium on Technology and Society (IEEE-ISTAS2020) from 12 – 15 November 2020 at London. This paper has been selected for publication in the prestigious journal IEEE Transactions on Technology and Society. Dr. Kavita Vemuri and Vivek Kaushal explain their work:

Clickbait headlines are compelling and designed to lure readers to click on the linked article. In online news media, a click on a news headline commonly translates to revenue for the publisher through advertisements on the linked page, or a probable conversion to a subscription. But, with an information overload in online news media and lower reader attention span, clickbait can result in readers either forming misleading knowledge maps based solely on headlines, or having a dissatisfactory user experience when information in a linked article does not match the curiosity induced by its headline. To understand the impact of clickbait headlines on the credibility of news items when controlling for articles’ content, we designed and conducted an experiment on real world news articles from the Webis-Clickbait-17 dataset. Additionally, the role of readers’ age and epistemic curiosity on clickbait propensity was also studied. Two datasets of 6 articles each of news relevant to North America and India were selected and each article was assigned a clickbait and a non-clickbait headline based on a pre-test for clickbait perception. The selected headlines (3 clickbait and 3 non-clickbait) were arranged in no particular order on a screen.  200 participants (100 each of English-speaking American and English-speaking Indian) were recruited on Amazon’s M-Turk platform. They were instructed to click and read through any two articles of their choice.  A credibility questionnaire was filled by the participants after reading each article while epistemic curiosity survey data and demographic data were collected at the end of the experiment. We observed that clickbait headlines significantly reduced the credibility of news items. A correlation between age and propensity for clickbait was observed, while no significant difference was found in the number of ‘clicks’ on clickbait and non-clickbait headlines. No correlation was observed between diversive epistemic curiosity and propensity for clickbait, while a weakly negative correlation was observed between specific epistemic curiosity and propensity for clickbait. We hope that our results on the impact of clickbait on credibility are useful for media companies in drafting policy decisions on the usage of clickbait and that insights from our findings may lead to a better end-user experience in terms of access to credible information for news readers.