COLLEGE PARK, Md. – Reading – or making – comments is a big part of consuming online news. But as anyone who has ever read comments knows, the quality can range wildly from good to bad to ugly. How can a comment moderator of a news website provide readers the best material when there are hundreds – sometimes thousands of comments?
The answer may be a new journalism tool called CommentIQ. Developed by an interdisciplinary team of researchers at the University of Maryland, College Park Campus, CommentIQ is an advanced computer tool designed to help comment moderators find and sort the best material for their readers.
Assistant Professor Nick Diakopoulos is a computer scientist who heads up the Philip Merrill College of Journalism’s Computational Journalism Lab. He is also a member of the UMD Human Computer Interaction Lab (HCIL). Diakopoulos says CommentIQ is designed to create the tools a moderator needs to visually find and sort high quality comments that readers have contributed to news articles. “This type of tool is far more advanced than the tools that are currently available to industry, enabling a new way of moderating comments that doesn’t focus on removing the low quality stuff as much as highlighting the high quality stuff.”
The interdisciplinary research project includes faculty from Merrill College and the University of Maryland’s iSchool (College of Information), a student from the iSchool, as well as a students from Computer Science (College of Computer, Mathematical and Natural Sciences). The CommentIQ research is made possible thanks to a Knight Foundation-funded prototype grant. That support has opened doors to allow the team to begin working with the Coral Project that aims to re-invent commenting for news websites. “We evaluated our prototype in terms of how it would change work practices with real comment moderators from the Washington Post, Baltimore Sun, Wall Street Journal, and New York Times, giving the work much more ‘real-world’ relevance,” said Diakopoulos.
New Paper Focuses on Comment Moderation
The University of Maryland research team will present a new paper on their findings today, May 9 during an ACM (Association for Computing Machinery) Conference on Human Factors in Computing Systems in San Jose, California. It is considered the flagship conference on human computer interaction research.
“Supporting Comment Moderators in Identifying High Quality Online News Comments” was selected as a best paper honorable mention by ACM (which is just 5% of the papers submitted).
The research paper was co-authored by Deokgun Park, Simranjit Singh Sachar, and Niklas Elmqvist.
A Q&A With Professor Nick Diakopoulos
1) Why is there a need for a tool to help website comment moderators sift through the good – and the bad – comments readers make?
National news outlets routinely publish articles that attract hundreds and even thousands of user comments. These comments often provide valuable feedback and critique, personal perspectives, new information and expertise, and opportunities for discussion, not to mention an occasional dose of profanity and vitriol. The varying quality of comments demands a high level of moderation and curatorial attention in order to cultivate a successful online community around news. CommentIQ helps moderators sort the great from the not-so-great comments at scale by applying visual analytics in a way that helps them work effectively and efficiently.
2) Some news stories bring thousands of comments – would your tools help the best ones “rise to the top” while pushing the worst to the bottom?
CommentIQ is explicitly designed to help identify the top quality comments. This is done interactively and through visualization so that professional moderators can define the criteria that are important in a given community or discussion context. We do not currently provide support for removing the worst of comments, however this could be an area for future work.
3) You’re already working on ways to reinvent commenting. Ideally – what do you think this will look like in five or ten years?
In the future online commenting will inform journalism in a host of new ways. For instance, reporters will have advanced tools for identifying new information sources and experts from the community based on past comments. End-users will have new tools for sorting and filtering commentary based on point of view, topic, or sentiment. And business-people will make use of comment analytics to drive revenue decisions about likely subscribers.
4) What does collaborative research like this (working with Computer Science and the iSchool) say about research going forward here at Merrill College and the University of Maryland? Is this a model for the future?
Most of the world’s problems don’t fit in a tidy box that can be solved by a single discipline. CommentIQ draws on disciplinary expertise from journalism, visual analytics, data analysis, and human computer interaction in order to solve a real problem and advance the state of the art for comment moderators. Interdisciplinarity is in the DNA of the University of Maryland. The CommentIQ project is a model for the future of research in which experts in different fields come together to focus on applying their knowledge to solve specific problems.
Professor Diakopoulos and members of his CommentIQ team are available for interviews. Please contact Dave Ottalini for more information:
Senior Communications Manager
Philip Merrill College of Journalism