CMU-CS-09-175
Computer Science Department
School of Computer Science, Carnegie Mellon University



CMU-CS-09-175

Collaborative Online Video Watching

Justin D. Weisz

December 2009

Ph.D. Thesis

Also appears as Human-Computer Interaction Institute
Technical Report CMU-HCII-09-106

CMU-CS-09-175.pdf


Keywords: Online video, collaboration, chat, social television, enjoyment, distraction, sociability, audiences, linguistic analysis, human computation, social proxies, MovieLens, YouTube, Facebook

With the rise of broadband Internet, watching videos online has become a popular activity for millions of people. Many web sites encourage people to contribute, rate, and comment on video content, and more recently, to share their experiences with each other in real time. This dissertation explores the user experience of simultaneously chatting with other viewers while watching videos online.

Watching a video and talking with others is a form of multitasking that reduces information processing quality. The first part of this dissertation examines distraction, enjoyment,and sociability in collaborative watching. In a series of field studies and laboratory experiments, I show that small groups of friends and strangers enjoy chatting while watching videos together, despite chat's distractive effects in both text and audio channels.

The second part of this dissertation examines challenges of interacting with other viewers in large-scale broadcasts. I argue that viewers can chat with their friends and monitor the activities of the rest of the audience without feeling overwhelmed by using visual chat summaries and a social proxy representation of the audience.

The final part of this dissertation shows how three types of information can be inferred about a video from the raw chat log data of groups that watched the video together: a set of tags that describe the video, ratings of the video, and a profile of peoples' enjoyment of each part of the video. This information can be used to improve the quality of video search and recommendation engines, and provide behavioral-based feedback on viewers' enjoyment to content creators.

This dissertation provides new insight on the distraction from multitasking in an entertainment context. For the videos studied, it shows that although distraction does degrade information processing, it does not significantly harm the entertainment of social experience. This dissertation also provides concrete designs and recommendations for user interfaces for large-scale online video broadcasts. Finally, this dissertation demonstrates that information about a video that would be difficult or impossible to infer through a computer algorithm can be learned from the social interactions that occur as viewers watch together.

337 pages


Return to: SCS Technical Report Collection
School of Computer Science

This page maintained by reports@cs.cmu.edu