Our eyes are wonderful at detecting associations in data because our brains are wired to simplify complicated patterns and relationships. Within a few seconds, you should be able to figure out that 25 is the next number in the sequence 1, 4, 9, 16. A computer program may need to make millions of calculations just to detect a pattern in the chess pieces that is readily visible to an accomplished player. The problem is that the patterns we perceive in the data may not meaningfully characterize its actual behavior. In the same way that our brains are able to “see” elaborate paintings in the clouds, we are sometimes able to “recognize” what are spurious relationships in economic data.

This chapter will present the time plots of various types of terrorist incidents, including bombings, assassinations, kidnappings, and skyjackings. Some features of the data will be clear even to the casual observer; for example, there is no decidedly upward trend in any of the incident series. However, many other features of the data may not be readily apparent, so statistical analysis may be required to draw inferences about the data. Toward this end, the chapter uses the basic tools of spectral analysis and intervention analysis to formally analyze data of terrorist incidents. Spectral analysis enables us to estimate the cyclical patterns in the various terrorist incident series, while intervention analysis allows us to measure the effects of important structural changes on the incident series. For example, using intervention analysis we can examine the effects on skyjackings of the introduction of metal detectors in airports, and the behavior of terrorists since the attacks of 9/11.