Julita Vassileva

Adaptive Rewards Mechanism for Sustainable Online Learning Communities

Biography
Julita Vassileva is an associate professor in Computer Science at the University of Saskatchewan, Canada. With her students at the MADMUC Lab (http://bistrica.usask.ca/madmuc/) she investigates mechanisms for encouraging participation in online communities, user modeling and user persuasion techniques, advanced learning technologies, trust and reputation in multi-agent systems. More information about her projects and publications is available at: http://julita.usask.ca

Position Paper
The proliferation of online communities may lead designers and researchers to the conclusion that the development of custom-made communities for particular purpose, for example, to support a class, is straightforward. Unfortunately, this is not the case. Although software providing basic community infrastructure is readily available, it is not enough to ensure that the community will “take off” and become self-sustainable. Ensuring a critical level of participation is important problem of peer-to-peer and online communities.

We developed a motivational strategy based on theories from social psychology (social comparison, reciprocation) to encourage users to contribute resources. The motivational strategy is implemented as a set of hierarchical memberships that can be gained by users, if they perform actions that help the community: contributing new resources, rating resources, taking care of the quality of resources and ratings that they contribute. Each membership rewards the user with particular interface appearance, visibility / status in the community shown in visualization based on a star-sky metaphor.

These strategies were tested last year for three months in a small-scale online community called Comtella for sharing web-links to class-related articles. They were shown to be successful in motivating participation and numerous contributions for a period of 3 weeks, after the motivational mechanism was introduced, followed by sharp decline in the last week. We weren’t able to pinpoint the reason for this—it could have been that the coursework load became too high, or that the theme of the last week was not so interesting for the students. Another reason could have been that the novelty effect of the motivational interface has worn off. A final reason, that seems most likely, could be an effect called “ageing of the community” (Q. Jones). It is characterized by a small number of users providing a large proportion of the contributions, of poor quality. Users feel swamped by a mass of unwanted information, i.e. experience information overload and as a result, withdraw. Our strategy could not ensure a sustainable level of contributions and it seems that exactly ensuring mechanisms for sustainable on-line communities is one of the hottest topics in social computing (J. Preece, P. Resnick).

To ensure self-moderation, we have augmented our motivational strategy with a user- and community- adaptive rewards mechanism that can be embedded in any on-line community software, to regulate the quantity of the contributions and encourage users to moderate the quality of contributions themselves. Users with high membership gain more power to rate contributions of other users. Rating contributions is rewarded with points that can be used to increase the visibility of ones’ own contributions. This mechanism has been applied again in the Comtella online community, supporting undergraduate students to share class-related web-resources and the results of the currently ongoing experiment are very encouraging.

The presentation will discuss some of the controversial issues in the design of social motivational and rewards mechanisms that we encountered.

Back to Social Computing Symposium 2005