Generative Explanations of Crime: Using Simulation to Test Criminological Theory*

Abstract

This study demonstrates that computational modeling and, in particular, agent-based modeling (ABM) offers a viable compatriot to traditional experimental methodologies for criminology scholars. ABM can be used as a means to operationalize and test hypothetical mechanisms that offer a potential explanation for commonly observed criminological phenomena. This study tests whether the hypothesized mechanisms of environmental criminology are sufficient to produce several commonly observed characteristics of crime. We present an ABM of residential burglary, simulating a world inhabited by potential targets and offenders who behave according to the theoretical propositions of environmental criminology. A series of simulated experiments examining the impact of these mechanisms on patterns of offending are performed. The outputs of these simulations then are compared with several well-established findings derived from empirical studies of residential burglary, including the spatial concentration of crime, repeat victimization, and the journey to crime curve. The results from this research demonstrate that the propositions of the routine activity approach, rational choice perspective, and crime pattern theory provide a viable generative explanation for several independent characteristics of crime. * The authors would like to thank the following people for their continued assistance, advice, and above all else, patience in the preparation of this submission: Susan Donkin, Peter Cassematis, Joe Clare, and Danielle Reynald; the anonymous reviewers for their perceptive analysis and insightful comments; and Charlotte S. Navarro and Rex Osborne for their continued inspiration and support of these ideas. Direct correspondence to

Crime and Justice in Digital Society: Towards a ‘Digital Criminology’?

“…Criminological research has also sought to engage with these increasingly automated, algorithmic and computational capacities as they relate to crime data analytics, law enforcement and justice system practices (Berk 2008;Birks, Townsley and Stewart 2012;Brantingham 2011). There is to date, however, a comparative dearth of criminological research that has begun to empirically and critically explore the range of challenges and opportunities presented by 'big data' analytics.…”

An analysis of crime hot spots using GPS tracking data of children and agent-based simulation modeling

“…When micro-level studies were conducted, these studies tended to rely on surveys (Miethe et al 1987, 1990, Osgood et al 1996 and they could not examine the fully dynamic nature of spatio-temporal interactions among the three core elements of crime. Several recent studies (Liu et al 2005, Groff 2007a, 2007b, Wang et al 2008, Birks et al 2012) that innovatively examined the routine activities theory through simulation modeling are exceptions.…”

“…Simulations have recently become a new and attractive medium for criminologists to test theories and evaluate intervention strategies (Groff 2007a, 2007b, Dray et al 2008a, Liu and Eck 2008a, Birks et al 2012. Increasing interests in this approach may be illustrated by the dedication of a special issue on simulations by the Journal of Experimental Criminology edited by Groff and Mazerolle (2008) and by the publication of an extensive collection of simulation studies in crime analysis edited by Liu and Eck (2008b).…”

“…While traditional statistical research is often described as a top-down approach, where aggregate data of crime, offending, and victimization are examined through a statistical model (e.g. a regression model) to make inferences about underlying causal mechanisms, simulations are described as a bottom-up approach where micro-level interactions are directly examined and are used to generate macro-level patterns (Birks et al 2012).…”

Modelling patterns of burglary on street networks

“…Indeed, this potential, along with the increased availability of geographical data and the development of a wide array of tools for the analysis of complex social systems, has inspired significant recent interest within the modelling community. Within this domain, attempts to model burglary range from mathematical approaches (Berestycki and Nadal 2010;Pitcher 2010;Short et al 2008) to those employing agent-based simulation (Birks et al 2005;Groff 2007a;Johnson 2008;Malleson et al 2009), with these playing complementary roles. This divergence is common to crime modelling as a whole: although the possibility of generalised analysis means that mathematical models might offer greater insight, they have so far failed to match agent-based approaches in terms of their scope for the http://www.crimesciencejournal.com/content/2/1/10 incorporation of detailed individual-level behaviour, with only some examples occupying a middle ground (Davies et al 2013;Short et al 2008).…”