Social Feedback: Social Learning from Interaction History to Support Information Seeking on the Web

Abstract

Information seeking on the Web has become a central part of many daily activities. Even though information seeking is extremely common, there are many times when these tasks are unsuccessful, because the information found is less than ideal or the task could have been completed more efficiently. In unsuccessful information-seeking tasks, there are often other people who have knowledge or experience that could help improve task success. However, information seekers do not typically look for help from others, because tasks can often be completed alone (even if inefficiently). One of the problems is that web tools provide people with few opportunities to learn from one another’s experiences in ways that would allow them to improve their success.

This dissertation presents the idea of social feedback. Social feedback is based on the theory of social learning, which describes how people learn from observing others. In social feedback, observational learning is enabled through the mechanism of interaction history – the traces of activity people create as they interact with the Web. Social feedback systems collect and display interaction history to allow information seekers to learn how to complete their tasks more successfully by observing how other people have behaved in similar situations.

The dissertation outlines the design of two social-feedback systems, and describes two studies that demonstrate the real world applicability and feasibility of the idea. The first system supports global learning, by allowing people to learn new search skills and techniques that improve information seeking success in many different tasks. The second system supports local learning, in which people learn how to accomplish specific tasks more effectively and more efficiently. Two further studies are conducted to explore potential real-world challenges to the successful deployment of social feedback systems, such as the privacy concerns associated with the collection and sharing of interaction history. These studies show that social feedback systems can be deployed successfully for supporting real world information seeking tasks.

Overall, this research shows that social feedback is a valuable new idea for the social use of information systems, an idea that allows people to learn from one another’s experiences and improve their success in many common real-world tasks.