Property Crime at O’Hare International Airport: An Examination of the Routine Activities Approach

Abstract

The September 11, 2001 terrorist attacks permanently transformed aviation security, generating more intensive security-related practices. Although these enhanced security measures primarily sought to prevent future terrorist attacks, they also may have provided a secondary benefit of reducing property crimes at airports. The present case study examines changes in airport security at Chicago's O’Hare International Airport post-9/11 in the context of routine activities theory. The study first posits that increased security measures at O’Hare should have resulted in increased capable guardianship, thereby reducing the number of suitable targets and criminal opportunities for motivated offenders. After identifying various changes in airport security at O’Hare since 9/11 in the context of routine activities theory, the study then examines changes in larcenies from January 1999 to June 2006 using an interrupted time series analysis. The ARIMA model indicates that in the post 9/11 O’Hare Airport environment, there was a statistically significant and persistent reduction in reported larcenies.

KEYWORDS:

• Routine activities
• airport security
• 9/11
• airport crime
• terrorism

Notes

1. While research appears to support the use of deterrence-based models in counterterrorism efforts (see, e.g., Dugan, LaFree, & Piquero, 2005), at least one study has identified the backlash effects (increasing future criminal behavior) that may result from using a deterrence-based model (LaFree et al., 2009).

2. The generality of routine activities theory (Cohen & Felson, 1979) presented a methodological challenge. Following the procedure described by Kleck, Britt, and Bordua (2000), it would have been useful to compare the effects of the 9/11 attack on the larceny time series, relative to a type of airport crime that theoretically should not have been impacted by changes in guardianship (similar to the approach utilized by McDowall, Loftin, & Wiersma, 1992, among others). If an effect was observed on the “comparison time series” one might have been able to conclude that any apparent changes in the larceny time series were caused by an unmeasured variable. However, it is nearly impossible to envision a type of crime that should not have been impacted by 9/11, and the associated security response.

3. The models described herein also were estimated on a time series that included September 2001. The conclusions derived from those models were virtually identical to those presented here.

4. Some analysts advocate selecting the impact model empirically (i.e., statistically estimating a series of possible models and choosing the one that best fits the data). However, Kleck et al. (2000) caution against this practice. They suggest that a priori theorizing should be used to select an appropriate impact model. Should the theoretically appropriate model not fit one's data, the authors argue that analysts should accept the null hypothesis of “no effect” rather than searching for alternative specifications (which may, by random chance, fit the data). This article agrees with, and utilizes, the approach suggested by Kleck et al. (2000).

5. All ARIMA analyses were estimated using SAS 9.1.

6. The authors also estimated a second-order moving average (i.e., ARIMA 0,0,2), a first-order autoregressive (i.e., ARIMA 1,0,0), a second order autoregressive (i.e., ARIMA 2,0,0), and a mixed model.

7. The authors attempted to obtain comparable crime data from several other airports covering the time period under analysis and were refused on the basis of “national security concerns.” Attempts also were made to lengthen the pre-intervention portion of the time series, but no additional data beyond what is analyzed here could be obtained.