Multivariate point process models in social research

Abstract

This paper advocates the use of multivariate point processes for modeling dynamic processes where several types of discrete outcomes occur repeatedly over time. The first section provides an overview of point process models with emphasis on the intensity function specification. The second section discusses complete and incomplete likelihood techniques for estimating parameters and briefly reviews the advantages and disadvantages of the two techniques. The paper ends with an empirical example from organizational sociology that illustrates the application of a multivariate point process model and complete likelihood estimation.