Adding Richer Credibility to Wikipedia

A central complaint about Wikiepedia is its credibility. (See Wikipedia, Truth and Citations) This past week, Robert McGrath, NCSA, University of Illinois, suggests ideas about how credibility can be added to collaborative community knowledge spaces at 2008 Microsoft eScience Workshop.

McGrath begins with an acceptance that shared knowledge communities like Wikipedia are a given. He suggests that what needs to be added is a more extended way to embed the knowledge, data and sources that form the foundation of the information or knowledge claim. He writes,

"We can imagine a 'better Wikipedia'; with a broader and deeper account of the purported knowledge presented. Rather than a single snapshot of knowledge, the artifact can be a complex web of knowledge including data, computation, and visualizations, and the history of the current artifact. Drilling down from the “article” leads to representations of the history, sources, and processes underlying the claims, including the data and software used, as well as citations and who did what. This enables evaluation of the knowledge (is it credible?), and comparison of
alternative accounts (e.g., using different data or assumptions)."

Unlike many other critics of Wikipedia, McGrath is not suggesting that we abandon the Wikipedia-type of shared knowledge communities, he is trying to add information that fits the Web 2.0 capabilities that makes these collaborative knowledge communities more credible.

He prvides a few suggestions about the types of information that might be available through a "who says so?" button on a Wikipedia article:

1. Notes by the writer that would comment on the sources of information.
2. Supporting or related documents (this is already a standard for Wikipedia).
3. Data, procedures, other information about how the writer arrived at his or her conclusions
4. Information about affiliations of the author or using Web 2.0 language, he refers to this as "social network information, reputation and trust relations."
   

He suggests that this accountability cannot be done through a centralized authority. What is needed he says, is

simple reliable mechanisms that enable users to “mash up” the required accountability. ... The general principle is to design flexible and reusable middleware that provides the “right” set of services, without “wiring in” a specific set of assumptions about how the systems must be used.

McGrath also acknowledges that the same level of accountability is not necessary within all types of collaborative communities, he notes, "Infrastructure should provide services that enable communities to implement their own culture of accountability." In short, the credibility of celebrity and/or sports information may evolve a different pattern of credibility than scientific reports.

These ideas are very promising in regards to helping us think about structures for developing a range of educational and scientific web-based collaborative knowledge efforts.