Understanding the adoption of military equipment by local law enforcement agencies in the United States

Abstract Objective This article examines the diffusion of paramilitary police units across the United States to consider the computational turn in social science. Methods The process of police diffusion is modeled by comparing traditional predictive models with agent‐based modeling (ABM) under constant and flexible binding conditions. The data are drawn from the Census of State and Local Law Enforcement Agencies 2000, 2004 and the United States Cities Database. Results ABM was not superior to traditional predictive models when it came to predicting policy diffusion and that flexible binding provides superior predictive capabilities over constant binding. Conclusions Public policy scholars should exercise caution in adopting computational methods as they may not offer advantages over traditional predicative analysis. Policy diffusion of police militarization is not driven by geographical proximity but may reflect hierarchical influences in the form of federal policy or the influence of far‐flung organizations perceived to be similar. This may cause inappropriate policies and practices to be adopted.