Empirical Advances for the Study of Weblogs: Relevance and Testing of Random Effects Models

The global blogging phenomenon has caught researchers by surprise. While past research has focused on the motivations behind blogging, since the number of weblogs is growing at such an unprecedented rate, the roles played by their visitors should not be overlooked. This chapter presents an empirical analysis of weblogs from a methodological perspective that demonstrates empirical advances in this research area. Owing to the large number of weblogs in our sample and the persistent use of certain technological features in some of these weblogs, accounting for individual blog-specific effects through traditional dummy-variable regression is infeasible. We introduce the use of random effects in empirical weblog research. A random effects model allows researchers to encapsulate blog-specific effects in a regression without significant degree of freedom losses, as in typical dummy variable regressions. The modeling approach also allows the analyst to confront issues and problems with data singularity, which happens because of the linear dependency between blog-specific dummy variables and the persistent use of technological features. By developing and testing a random effects model that accounts for mutual influences between blog visits and comments, we show that blogger-visitor interactions play an important role in affecting blog popularity. Further, the use of technological features may not always raise the number of blog visits, and hence may not add value to bloggers in terms of increasing their blogs’ popularity . We discuss some implications of these findings for research and practice.