The recent explosive growth of online social networks presents the IT field with new research opportunities and challenges to understand their evolution over time and their impacts on individual behavior. Both theoretical research and empirical studies are beginning to appear. Surveys have been the preferred approach to collect data as a basis for effective testing of hypotheses. However, several are limitations associated with questionnaires. One of the major problems is that they reveal information that the user is willing to provide which might be biased or even inconsistent with her actual opinion or behavior. As social networking sites become more popular, researchers attempt to collect observable data directly from social networks. However, privacy concerns constantly emerge, making it increasingly difficult to collect detailed data efficiently and effectively. Facebook and other social networks opened their platforms to outside developers. By developing a Facebook application, we are able to collect a comprehensive set of panel data about user behavior by observing users activities inside the application. For example, postings on discussion boards, rating other people's works, inviting friends, and many other interactions can be obtained for analysis. We also gain access to valuable information about the user such as gender, geographical area, level of education, and in some cases information as specific as religion and political inclination. There are several advantages to these data sets. First, all of them are without the bias normally associated with surveys. Second, with observation of many users' behavior over long periods of time, we are able to use econometric methods such as fixed effect estimation to deal with unobservable individual effects which may cause endogeneity problems. Third, with long-term observation of users' behavior, we can potentially study the dynamics of users' behavior, which is rather difficult with survey data. The tutorial will illustrate these ideas by referring to two applications we developed: bRkthRgh and MyHateList, both currently operating on Facebook.
Whinston, Andrew B., "Opportunities and Challenges in Analysis of Social Netowrks" (2009). AMCIS 2009 Proceedings. 288.