Successful and Unsuccessful Pirate Attacks Worldwide: A Situational Analysis

Abstract

Using concepts drawn from situational crime prevention theory, this study compares successful and unsuccessful pirate attacks (n = 4,638) against ships worldwide and the situational factors that help prevent such attacks. The results show that when a ship’s crew takes proactive self-protective measures that increase the perceived effort (increasing speed, employing evasive maneuvers) and increase the perceived risk (embarking private security, having watchman present, raising alarm, increasing lighting, anti-piracy) of perpetrating an attack, unsuccessful attacks are significantly more likely after controlling for environmental influences. Despite a few common data limitations, the study contributes to the crime prevention literature by analyzing piracy from a micro level instead of a macro level. Future research should examine how the piracy “ecosystem” contributes to the problem and the costs and benefits counter-piracy activities.

Keywords:

• piracy
• opportunity
• shipping industry
• situational crime prevention
• logistic regression
• vessel security

Acknowledgment

We wish to acknowledge Dr Bridget Coggins, Professor of Government, Dartmouth College (USA), for collating and making available part of the original data-set from which we worked, Dr Eric Piza, John Jay College of Criminal Justice, for his comments on an early draft of the paper, and the anonymous review members who made this a stronger contribution to the literature.

Notes

¹ A “Citadel” is a fortified safe-room aboard a ship designed to protect the crew if pirates board the vessel.

² See http://www.icc-ccs.org/piracy-reporting-center/live-piracy-report. Retrieved on June 15 2014.

³ A Tolerance value of .01 or less generally indicates the presence of multicollinearity. Tolerance scores ranged from .712 to .996. The variance inflation factor (VIF) score is the reciprocal of Tolerance, and VIF scores that exceed 10 indicate multicollinearity. VIF scores range from 1.006 to 1.405. A Condition Index score should not exceed 30, nor should two or more variance proportions for a given predictor variable be equal to or greater than 50. The Condition Index score is 9.227 and the variance proportions across all predictors ranges from .00 to .74 (see Meyers, Gamst & Guarino, 2006, p. 212). Cook’s Distance score should be lower than 1, the values ranged between .000 and .738; DFBeta for the Constant score should be lower than 1, the values ranged between −.272 and .660 (see Field, 2009, p.44).

Additional information

Funding

Funding. This work was supported by the City University of New York, Research Foundation (grant number 66804-00-44).

Notes on contributors

Jon M. Shane

Jon M. Shane, PhD, is an associate professor in the Department of Law and Police Science at John Jay College of Criminal Justice in New York City. He received his PhD in criminal justice from Rutgers University, School of Criminal Justice. His research interests include violent crime and police policy and practice. His theoretical interests include situational crime prevention, routine activities, and social disorganization.

Shannon Magnuson

Shannon Magnuson, BA, is a master’s degree candidate in the Department of Criminal Justice at John Jay College of Criminal Justice in New York City.  She received her BA in criminology from University of Florida.  Her specialization is Corrections Administration.