Howard T. Welser from Cornell University, Eric Gleave from University of Washington and Microsoft Research, and Danyel Fisher and Marc Smith from Microsoft Research are bringing to or attention a study of major importance on social roles in online discussion groups. Based on a structural signature approach, with a thorough theoretical background, a complex methodology, and a rich bibliography, the study has the goal of researching and establishing strategies and techniques of deducing such social roles from structural and behavioral signatures.
Given the multitude of possible social roles to be found within an online discussion group, the researchers limited their area of study and focused on only one role, that of the ‘answer person’ in contrast with other kinds of roles, like the discussion person. The contributing behavior of a person who offers valuable information and advice in threaded online conversations can be explained by a desire to obtain social goods like status. The authors consider the patterns of contribution as structural and behavioral signatures, thus indicators of social roles and they develop visualization strategies for observing, recording and analyzing these signatures. Also, the authors consider that they can leverage this structural and behavioral data in order to identify other types of social roles in online discussion spaces and to construct a taxonomy of such roles.
Figure 1: Authorline for an answer person
Figure 2: Authorline for a discussion person
This line of research has many more good implications, as it can serve to the elaboration of predictive models for online social role behavior adopted by participants and for variations in online discussion environments. Visualization is a method still to be refined. Authors of the study gathered data from three newsgroups, but they already started to record data from 45 other newsgroups, so they can enlarge and better test their initial conclusions. The method is not limited to newsgroups, but extensible to mailing lists, other discussion forums, blogs, and to wikis. Also, they believe that the best results in studying online social roles are to be found by both intersecting and integrating multiple methods. Multiple methods can provide different sets of metrics that can be submitted to systematic comparison.
Finally, this kind of sociological research into the depths and hidden structures of the social interaction mediated by computer systems can provide new grounds for social network software, ways of predicting the evolution of an online discussion group and also methods of improving the dynamics between its members by working on the social roles performance in the online arena.