Cyber-Human Systems

• Social Media Analytics

| March 15, 2018 – February 29, 2020

Empirical and Design Investigations to Address Misleading Online News in Social Media

This research employs social computing and human-centered approaches to understand the relationship between people and technology in the context of online news. It specifically investigates online news sources that propagate fabricated stories, explores how users engage with these sources on social media, and examines ways to nudge users to be more conscious consumers of online news. This investigation aims to provide new perspectives that address digital misinformation by focusing on the following questions: How can we establish differences between mainstream sources and misleading sources of online news? How can we nudge people to be more careful and conscious consumers of online news?

The research is being conducted along two symbiotic lines of inquiry. The first involves empirical investigation of misleading online news sources to develop a deep understanding of their behavior along the following threads: topical and writing style differences from mainstream sources, user engagement distinctions, and the corresponding temporal changes. This inquiry is based on the following data: a professionally curated list of online news sources along with their credibility labels from expert fact-checkers, and tweets sent out by these news sources over a period of at least a year. The second thrust of this research focuses on exploring design interventions to increase people’s awareness while they read news on social media sites. Specifically, it will investigate two classes of design nudges on Twitter. The first intervention, emphasize, nudges users to reflect on the ambiguity and uncertainty present in certain news posts. Emphasize automatically detects whether a social media news post from a mainstream source has been questioned and highlights those question tweets for the news reader. The second intervention, de-emphasize, will be triggered whenever news posts originate from misleading sources, to make that post less visible in an attempt to minimize exposure to misleading online news. The human-centered evaluations that accompany the deployment of these interventions will provide qualitative and quantitative evidence about user experiences, as well as measurements about the efficacy of these interventions.