This project will develop a design-oriented science of online volunteer communities (OLVCs) such as Wikipedia, SourceForge, and TripAdvisor. These web sites attract hundreds of thousands of people - not to play games or socialize - but to produce content. The content produced by OLVCs is valuable. Wikipedia is among the ten most popular sites on the Web. SourceForge is the hub of the open source software movement, used by thousands of projects to organize their work. TripAdvisor brings together travelers whose reviews help each other find and vet hotels and services. Despite these successes, OLVCs face real challenges. Most fail quickly, and even the most successful have problems. Over two-thirds of SourceForge projects become inactive fairly quickly. Even Wikipedia struggles: vandalism is a constant problem, two thirds of the articles are stubs - the lowest quality article, which are too short to provide encyclopedic coverage - and recent data suggests that Wikipedia's constant growth may have halted and even reversed.
A design-oriented science of OLVCs could help solve these problems. Four specific research activities will be carried out to produce this science. First, the project will mine results from prior research to create models, formulate hypotheses, and define key success metrics. Second, it will develop algorithms and interfaces to help OLVCs function better. Third, it will evaluate the mechanisms in real OLVCs, for example, with experiments that measure the extent to which different social comparison interfaces motivate volunteers to step up their activity. Finally, the project will abstract the results to develop generalizable design principles for online volunteer communities.
This project will develop a broad set of software mechanisms to solve important and general problems of online volunteer communities. The algorithms and interfaces will be evaluated systematically in real OLVCs, thus providing an empirically grounded body of knowledge about their effectiveness. In addition, the results of this project can feed back to advance social science theories originally developed for offline volunteering.
Wikipedia. SourceForge. TripAdvisor. These web sites attract hundreds of thousands of people – not to play games or socialize – but to produce content. The content produced by these online volunteer communities is valuable: Wikipedia is among the five most popular sites on the Web. SourceForge is the hub of the open source software movement, used by thousands of projects to organize their work. TripAdvisor brings together travelers whose reviews help each other find and vet hotels and services. Despite these successes, online volunteer communities face real challenges. Most fail quickly, and even the most successful have problems. Over two-thirds of SourceForge projects become inactive fairly quickly. Even Wikipedia struggles: vandalism is a constant problem, there are sharp arguments about the quality of work, it is increasingly difficult for new editors to contribute, and its rapid growth has slowed and may even be reversing. This project addresses these problems by developing a design-oriented science of online volunteer communities. It draws on an extensive body of social science research on what motivates volunteers in offline communities. The project consists of three major types of research. The first type was the development of a computational theory of participation in online communities. This is an agent-based model that synthesized several social psychological theories and led to novel recommendations for community design based on the integration of these theories. For example, the model highlighted the tension between focused online discussion around the topics on which a community is organized and the off-topic discussion that support personal relationships. The model showed that communities produce more benefits to their members and are more successful in terms of growth, retention and activity when they filter content on a personalized basis. The second type of research consists of empirical studies of the factors associated with the success of online communities, at multiple levels of analysis – for example, the success of the community as a whole, success at socializing, retaining and supporting group members and success at group production. The goal of these empirical studies was to develop micro-theories of factors responsible for group success. We conducted empirical research that led to theories of shared leadership, diversity or contribution, and used the conclusions from the empirical research and theorizing as the basis for interventions to improve the functioning of online communities. An example is our research on shared leadership. One challenge online volunteer communities face is getting the volunteers to work on tasks that are important to the organization, but not necessarily to the individual volunteers. Unlike employment organizations, they are less able to use hierarchical control in which managers direct subordinates’ work activities. Some of our research identified how online communities, which are similar to off-line, volunteer organizations in lacking strong hierarchical management, are still able to organize coordinated action on the part of community members. Many of the empirical studies were conducted in the context of the online encyclopedia Wikipedia. For example, many communities rely upon peer leadership, in which fellow community members directly influence each other through communication. For example, in Wikipedia formal leaders, such as administrators, communicate to regular Wikipedians more on a per capita basis than do regular Wikipedians, and a message from them is more influential than is a message from a regular Wikipedian. However, because there are very few formal leaders and tens of thousands of regular Wikipedians, over two-thirds of influence exchanges in Wikipedia occur among peers. Online communities also use the confluence of group-level goals and members’ social identities as community members to align the activities of different individuals. Thus, for example, Wikipedia uses the group goal setting mechanism called Collaborations of the Week to put members of a Wikipedia project literally on the same page. The dungeons and fierce monsters that participants in multi-player games need to overcome serve the same group goal-setting purpose in game guilds. Although most of our empirical research examined the influences on the success of volunteer communities, we believe many of the same principles apply in conventional organizations. For example, our research has shown that the environment of other communities surrounding a particular community can influence its success. Sharing members with other communities allows ideas and best practices to flow among them, even as they compete for members. These principles apply both to the online communities of volunteers in Wikia and to paid employees at IBM. The third type of research applies the theory and empirical findings generated in the project to produce interventions designed to improve the function of particular online volunteer communities. For example, we used our research on the socialization of newcomers to online communities to develop a regime to select, train and integrate volunteer tax experts into the community providing support to customers for Quicken’s TurboTax product and students into the community writing psychology articles in Wikipedia.
