Re-evaluating Reactive Policing: A Minority Report

Chapter 1 (published in Crime Science) presents a scoping review on predictors of police response time (Verlaan & Ruiter, 2023b). The review synthesises 39 empirical studies from academic and grey literature, identifying 122 unique predictor variables categorised as kinematics, exogenous factors, human involvement, incident characteristics, innovations, and police strategies. Relationships are summarised in an extensive table and a preliminary causal model. The review highlights a lack of research into reporting time and the effects of rapid response, leading to the investigations in chapters 2 and 3.

Chapter 2 provides a historical analysis of how the lack of empirical research into reactive policing—and the dominance of proactive policing—arose. Tracing landmark studies from 1970s Kansas City to the Maryland Report, it argues that the dismissal of reactive policing was based on a limited and selectively interpreted evidence base. It concludes that reactive policing should be re-evaluated, beginning with the evidentiary foundation of its dismissal.

Chapter 3 examines this foundation in depth, focusing on the 1977 Kansas City Response Time Analysis (Bieck, 1977a) and its 1981 replication (Spelman & Brown, 1981). It argues that the rejection of rapid response rests on weak evidence: flawed designs, small samples, unreliable measurements, and outdated empirical contexts. Consequently, it contends that the standard view of rapid response should be reconsidered.

Chapter 4 (forthcoming in Journal of Quantitative Criminology) presents an agent-based model investigating how deployment strategies affect response times—specifically average time, percentage of fast responses, and percentage of late responses. Using empirical data, it compares random patrol, strategic positioning, and idling at the most recent call. Strategic positioning clearly outperforms the others, especially in achieving fast responses, while increasing police capacity has relatively minor effects.

Chapter 5 (in The Crime Data Handbook) discusses inferential statistics on police administrative data, emphasising their nature as population data rather than samples (Verlaan & Langton, 2024). A short review shows that statistical inference is frequently applied to such data without justification. The chapter calls for greater transparency in assumptions, more emphasis on descriptive analysis, and reconsideration of inferential approaches.

Chapter 6 (in Crime Science) examines measurement of reactive policing through administrative data, focusing on operationalisations of dispatched deployment time (Langton, Verlaan & Ruiter, 2023). It finds these measures fairly robust and shows that incident-level timestamps approximate unit-level data well in estimating the composition of dispatched deployment time.

Chapter 7 (in Police Practice and Research) replicates work by Ratcliffe (2021) and Lum et al. (2022) on reactive supply (Langton, Ruiter & Verlaan, 2023). Using Detroit 2019 data, it finds police spend about 50% of their time on crime-related calls, 46% on disorder, and 4% proactively.

Chapter 8 (in Tijdschrift voor de Politie) describes the reactive supply of policing across the Netherlands based on detailed unit-level timestamps (Verlaan et al., 2023). It reconstructs the composition of reactive supply for most frontline teams, showing strong variation across and within police regions and reflecting on the context-dependent nature of policing.

Chapter 9 (in Journal of Experimental Criminology) evaluates a place-based intervention in Amsterdam’s Haarlemmerbuurt co-designed with the local frontline team. The intervention increased police presence four- to fivefold and reduced crime-related calls by 23% in the treated area but showed displacement and worsening public perceptions. Officers also reported frustration with the self-directed nature of the proactive strategy.