Interrupted Time Series Analysis in the Study of Terrorism

Empirically informed counter-terrorism interventions are essential to effectively reducing terrorist violence and implementing sound policy. Interrupted time series analysis provides a rigorous methodology to carry out this important undertaking. In this chapter, we detail the standard approach to interrupted time series analysis in both univariate and multivariate settings, specifically autoregressive integrated moving average (ARIMA) and vector autoregression (VAR). An assessment of the impact of U.S. embassy fortifications provides an illustration of the statistical procedures, and sheds light on the issue of the intended and unintended consequences of security interventions. Our preliminary analysis suggests that enhancing U.S. embassy security reduced attacks on U.S. diplomatic targets without increasing attacks elsewhere. Rather, the hardening of U.S embassies resulted in a reduction of attacks against other U.S. interests.