Large, distributed, user-contributed content sites like blogs, web forums, Slashdot and Wikipedia, have significantly changed the character of mass social interaction online. As individuals participate, their actions are visible to other participants. Examining the work of both preparing content and maintaining the community is key to understanding the nature of these systems. Social translucence is a conceptual framework for thinking about how social actions in an online community facilitate future social interaction.
Reputation systems, systems that represent the reputation of participants in an online community, have been well studied in the domain of e-commerce. However, theoretical principles or guidelines to designing and developing generalized reputation systems have not emerged. The proposed research will answer four critical questions about the relationship between reputation systems and the existing framework of social translucence: How can social translucence frame the development of reflective social systems? What methods unpack the components of "reputation" to reflect the full range of valued activities in an online community? What technical methods effectively collect, mine or elicit data related to specific dimensions of an individual's reputation? How can systems be built that enable individuals to understand their reputation as it is collected and interpreted by other members of the community? Until recently developing a reputation system to elaborate the framework of social translucence was prohibitively difficult. But the accessibility of large community datasets, like that of Wikipedia, has mitigated one major stumbling block, the availability of data.
Intellectual merit: The proposed research will extend social translucence as a framework for the design and development of social and collaborative systems. The ability of developers of distributed contributor communities to build reputation systems that can be appropriated by the community will be illustrated through the use of social translucence. Lastly, these results will develop a situated model of reputation in distributed contributor communities - one that sees a majority of online work activity as part of a community member's reputation.
Broader impacts: The primary contribution will be an improvement in the interaction, contribution and sense-making of the ever growing diversity of distributed contributors to online communities. As more people participate in online communities, the level of interaction and trust is limited by the amount of information that participants have about one another. Effectively supporting these communities is an important challenge of social computing. The proposed approach puts socially translucent tools in the hands of the community - those who are in the best position to create, reflect, debate and refine socially translucent representations of behavior.
With the proliferation of popular social media systems—both in the world at large and within organizational contexts—it becomes increasingly valuable to understand how people organize themselves to achieve collective goals online. When such systems are used by large numbers of people and those people are engaged in massive numbers of disparate activities, it is impossible for an average participant to reasonably understand the nature of the community or to develop a meaningful understanding of individuals within the system. This lack of understanding presents a problem for people who are motivated to use such systems to collaborate on worthwhile societal challenges such as contributing to science, helping build knowledge in unfolding crisis events, or creating shared public resources. Our project explored the possibilities of using social translucence as a design approach for making humans and their actions in such systems sensible to people who want to understand and participate more effectively therein. Our exploration focused specifically on discovering techniques and end-user mechanisms for realizing social translucence in Wikipedia, though there is notable potential for extending this work into other social media systems, including programming code development environments, online public deliberation forums, and other forms of commons-based peer production. Through the project our team developed new techniques for data mining, aggregating user trace data into socially meaningful tokens, and representing system users to others in rich and end-user configurable visualizations. Through the project we worked with the broad community of Wikipedians who potentially benefit from such a system design. Using techniques of user-centered design, project work was strongly biased toward addressing public needs related to this new design approach. Project work was reported in a wide variety of public venues including research/discovery meetings with community stakeholders, publications in journals and conference proceedings, public demonstrations of the systems developed, and presentations at national and international research conferences (e.g., ACM’s Computer Supported Cooperative Work [CSCW], and the international OpenSym/WikiSym conferences) . Many students from diverse backgrounds working on degrees in varied fields at the University of Washington (Human Centered Design & Engineering, Information Science, Computer Science, etc.) were involved in the project. These students’ preparation in conducting scientific investigations, technical development work, and social analysis helped prepare them for productive careers in STEM-related fields. Upon graduation, these students went on to take professional positions, including at Microsoft and the WikiMedia Foundation, as well as in smaller technology companies, technology startups, and higher education. The computational equipment purchased through the grant provided students with resources to expand their learning in hands-on contexts, and it continues to support student learning even as the formal research project ends.
