It’s Easier Than You Think: Modeling Dynamics in Time-Series-Cross-Sectional Count Data

• In Event: Models for Time-Series Cross-Sectional Data

Brief Overview

Direct inclusion of a lagged dependent variable is problematic in count-outcomed GLMs. Current solutions are not easily extended to a panel setting. We introduce a panel based alternative for modelling persistent count data and apply it to terrorism.

Authors

• John Niehaus, Texas A&M University

• Scott Jeffrey Cook, Texas A & M University-College Station