ABSTRACT
Abstract: As information systems (IS) and technologies proliferate, an unprecedented amount of business and individual activities have become mediated and captured by computers. Consequently, Internet technology has enabled academic researchers in information systems to access rich data sets (e.g., Weblog files, interpersonal communication history) that heretofore have been unavailable. These newly available data provide IS and e-commerce scholars with an excellent opportunity to observe individual activities online in an unobtrusive manner. Existing empirical methods, however, may be insufficient to analyze the Internet-enabled data. In this chapter we propose the application of choice-based sampling and weighted exogenous sampling maximum likelihood estimation to properly handle large-scale dyadic data sets representing infrequently occurring events. We then illustrate this empirical method using an example drawn from an open source software development context. The empirical advance associated with this approach is that it allows the researcher to analyze the influences of relational factors between the two entities involved in an event without excessive sampling cost and computation time. Through this research, we demonstrate that Internet technology not only brings along efficiency gains in research data collection, but also has the potential to help support the development of innovative empirical methods to analyze the data and eventually develop more in-depth explanatory theories.
