Modeling Collaboration Using Point Processes

In this chapter, we outline the uses of point processes and related methods for modeling temporal dependence in human interactions. We begin by describing our example, which was drawn from teamwork in sports. We then discuss three interrelated steps in analyzing the data: (a) the problem of defining and detecting temporal dependence among the activities of team members, (b) characterization of the dependence in terms of temporal clustering, and (c) the use of the Hawkes process to model the clustering. The third step provides a parametric model for describing and comparing statistical regularities of the interactions among individual team members or subsets of team members. We conclude by considering how this approach can capture aspects of team interaction that might be relevant for developing performance-based assessments involving collaborative problem solving.