Estimation of the Network Effects Model in a Large Data Set

Effects of interpersonal context are processes best represented by the network effects model. However, the size of the data matrices involved has until now prevented application of the model to large data sets containing information on individuals. The proposed adaptation of Doreian's maximum likelihood estimator substantially reduces the amount of computer core memory required for estimating the model, so that estimation can be carried out on a typical mainframe computer. An empirical illustration of the network effects model using the Adolescent Society data is presented.