Reflection 5 – Spencer Jenkins

The article for today was “Harnessing the Wisdom of Crowds in Wikipedia: Quality Through Coordination,” a work by Carnegie Mellon researchers Kittur and Kraut. This paper was concerned with the editing process of a Wikipedia article. One of the most appealing features of Wikipedia is that anyone can edit an article on the site. As the authors point out, this allows for crowdsourced wisdom with a very fast edit time for inaccuracies. This collaborative editing process is not without difficulties, however. Some level of coordination is required for a cohesive article with a meaningful structure and consistent style. Through an empirical study, the authors examine whether the number of editors working on an article is related to the article’s quality. 

What the researchers found is interesting. Essentially, an increased number of editors only benefits quality when there is an appropriate level of coordination. The paper makes an important distinction between two types of coordination: implicit and explicit. Explicit coordination occurs through talk pages, a place for meta-discussion on articles. Implicit coordination occurs through editor concentration. By having one or a few editors doing a majority of the work, coordinating without explicit communication, the overall structure and tone can be implicitly set for future editors. The authors found that communication in both forms helps in the quality of the article. Interestingly, the benefits of explicit coordination decrease as the number of editors increases. This contrasts with their findings for implicit coordination; when work is highly concentrated, the quality of the article improves with the number of editors. This is the opposite result of the case in which work is more evenly distributed. The authors also found that the benefits of coordination are greater earlier in the life cycle of the article.

While the results make intuitive sense, I found this to be a useful study. The paper was well-written and easy to follow. I’m very curious to how this research translates to an open source project of a different kind. The authors briefly mention the Linux operating system as an example of this. A software project such as this requires a different type of collaboration than a Wikipedia article, though there are still many similarities. I’m very interested to see whether the same types of benefits exist in such a project.