A multidisciplinary team of Purdue researchers has been awarded a National Science Foundation grant to investigate how social media, especially Wikipedia articles and editors, shape public knowledge. The project “KredibleNet” aims to build a research community and to propose a research agenda for the study of reputation and authority in informal knowledge markets, such as Wikipedia. The team is lead by Sorin Adam Matei (PI) of Brian Lamb School of Communication, Elisa Bertino (Cyber Center), Michael Zhu and Chuanhai Liu (Department of Statistics) , and Luo Si (Department of Computer Science).
The KredibleNet grant will fund research and workshops that will reveal how social media leaders gain authority to shape public knowledge. The project will explore a series of large social media datasets, including one that includes over 250 million editorial changes made to all 7 million Wikipedia articles created between 2001 and 2010. The ultimate goal is to show how the most prolific editors are related by editorial work to other editors and how their networks of collaboration and the reputations these networks support shape the process of creating and disseminating knowledge. The project brings together a unique multi-disciplinary community of experts from the social sciences, computer sciences, and statistics to provide explanations for how expertise and reputation emerge in social media. A core project hypothesis is that Wikipedia in particular, and social media more generally, have become “knowledge markets” in which what is known is shaped by the collective work of a relatively limited group of individuals who engage in a gigantic social experiment of assuming and enacting “functional roles.” Assumed voluntarily and not bestowed by a higher authority, such roles are structured in a social order the KredibleNet researchers call an “ad-hocracy.” Although informal and fleeting, an adhocracy is no less structured and constraining than regular, more customary bureaucracies. “95% of content on Wikipedia is produced by 5% of Wikipedia editors, many of whom have been on the project for many years and know each other. Wikipedia has become a one-of-a-kind social order, whose leaders and missions we need to understand much better, if we want to understand why and how when we search for any common noun on Google, 95 out of 100 searches provide among the top five results a Wikipedia page” says Dr. Matei, the project principal investigator. “Because of its complex organization and huge size the Wikipedia dataset poses several interesting challenges with respect to its management. The Cyber Center will closely work with the researchers and students involved in the project to identify and address these challenges” says Elisa Bertino, co-PI of the project and director of the Cyber Center.
The KredibleNet project is supported by an international team of researchers lead by Marc Smith, a former Microsoft Research Scientist and currently the head of the Social Media Research Foundation and NodeXL project and by Dr. Luca De Alfaro, from University of California, Santa Cruz, the head of the WikiTrust project. The team also includes researchers Jure Leskovec (Stanford), Bernie Hogan (Oxford Internet Institute), and Ben Shneiderman (University of Maryland).
Through a series of workshops and use of an online platform, the KredibleNet nascent community will refine the conceptualization, operationalization and measurement of reputation as a component of online knowledge and create a roadmap for future research. The team will also develop and disseminate tools for data management, analysis and visualization of online reputation data and collaboratively design a prototype reputation measurement cyberinfrastructure platform that can be used for experimentation. This platform will include new data management and analysis tools to enable access to live social media data from sources such as Wikipedia, Twitter and YouTube. The resulting tools and dataset will equip the community with the capability to experiment with data-intensive analytic strategies and to better understand the nature of the reputation measurement problem.
The ultimate goal of KredibleNet is to shape the next generation of theoretical and analytic strategies needed for understanding how knowledge markets are influenced by social interactions and reputations. The community discussions will ensure that the infrastructure developed to fill the current conceptualization and measurement gaps, data management capabilities and analytic tools will provide as much benefit as possible to all the related fields in industry and the sciences; they will also likely give rise to new research synergies. The tools that will be developed will render the existing large databases amenable to analysis allowing scholars and practitioners to address a broad set of questions and gain valuable insights. The potential users of these tools, data, and ideas are quite widespread extending to multiple scholarly domains, policy communities, and industry partners. They will be made available publicly, so others may benefit from the results of this project to develop the next generation of “information gauges” that can help tomorrow’s information consumers make smarter choices.
The international and multidisciplinary nature of the team highlights Purdue’s successful strategy of combining varied talents and approaches under the umbrella of the innovative Discovery Park research unit, and especially of the Cyber Center. The Cyber Center mission is to Enable Discovery through Data by focusing on basic and applied research on topics such as data integration techniques, large scale data management and analytics, workflow systems, privacy preserving computing.