Homophily and Influence in Dynamic Networks
Homophily describes the empirical finding that individuals select friends and acquaintances with similar characteristics, i.e. similar individuals share the same personal networks. This results in homogenous networks regarding societal, attitudinal, and behavioral characteristics. In return, the homogeneity of someone's personal network limits her/his perspectives on attitudes and behavior. Therefore, the individuals are directly influenced by their proximate environment. Although the interaction between homophily and influence seems obvious in various societal processes, the concepts are mainly popular in sociology. The dissertation project aims to bring the concepts homophily and influence into political science. An example of application in the arena of political science could be the opinion formation processes on political issues in parliaments, in small regional entities, or in groups of individuals.
Steglich et al. have developed a set of statistical models which enables researchers to quantify the effects of homophily and influence. In a first step, the effects of homophily and influence are estimated simultaneously. In a second step, simulations based upon the estimated parameters allow identifying the causes for the similarity of the network members. Thus, determinants both on the individual and on the network level can be incorporated into the explanatory model. Steglich et al. combine actor-based models with empirical data. The application of this model with data which will be newly gathered in the dissertation project also means a test of robustness of the model itself.