Investigating Crime Pattern Stability at Micro-Temporal Intervals: Implications for Crime Analysis and Hotspot Policing Strategies

• DOI: 10.1177/0734016821996785
• Date: 01 June 2021
• Published date: 01 June 2021
• Subject Matter: Articles

Article

Investigating Crime Pattern

Stability at Micro-Temporal

Intervals: Implications for

Crime Analysis and Hotspot

Policing Strategies

Timothy C. Hart

1

Abstract

Studies of crime hotspot forecasts use various metrics to describe different characteristics of

prediction patterns. However, few investigations consider how the stability of crime hotspot,

estimated at relatively short temporal intervals, can impact hotspot policing efforts. In response,

using address-level incident location data that were collected from six law enforcement agencies in

the United States, the current study examines the daily stability of crime hotspots that were esti-

mated over a 1-year period. Results suggest that micro-temporal stability patterns in crime hotspot

forecasts are dependent on crime type, jurisdiction, and the interaction between these two factors.

Implications for crime analysis and future research are discussed.

Keywords

crime and place, micro-temporal analysis, crime forecasting

Hotspot policin g is an effective crime red uction and prevention strat egy that relies on analytic meth ods

designed to identify nonrandom patterns within crimelocation information. Many of these techniques

involve forecasting spatiotemporal crime patterns, based on patterns that have been observed in the

past. The effectiveness of hotspot policing is largely dependent on the results of these estimation and

prediction methods, and an increasing number of metrics designed to evaluate the performance and

characteristics of hotspot mapping results have emerged from the literature as a result.

One way in which the characteristics of crimehotspot patterns can be described is by measuring the

stability of crimeforecasts over time. For example, Adepejuand colleagues (2016) recently examined

multiple prediction methods using a Dynamic Variability Index (DVI) and found that kernel density

estimation(KDE) produceshighly stable forecastsacross consecutiveprediction intervals.In describing

1

Department of Criminology and Criminal Justice, University of Tampa, FL, USA

Corresponding Author:

Timothy C. Hart, Department of Criminology and Criminal Justice, University of Tampa, 401 W Kennedy Blvd, Tampa,

FL 33606, USA.

Email: thart@ut.edu

Criminal Justice Review

2021, Vol. 46(2) 173-189

ª2021 Georgia State University

Article reuse guidelines:

sagepub.com/journals-permissions

DOI: 10.1177/0734016821996785

journals.sagepub.com/home/cjr

their findings, researchers noted that evaluating hotspot stability is an important aspect of crime pre-

diction because subtle changes in hotspot patterns across estimation intervals may adversely impact

hotspot policingefforts. However, extanthotspot research focuses primarilyon the predictive accuracy

of various analytic techniques, rather thancrime pattern stability.

Nevertheless, research into crime pattern stability has produced empirical knowledge that has

been used to improve hotspot policing strategies. For example, Johnson et al. (2008) analyzed the

stability of crime patterns produced from KDE maps over multiple interpolation periods. They found

that hotspots with the highest concentrations of crime were not always the most stable. Based on

these results, they offered recommendations for how simple KDE methods could be modified so that

they could produce crime forecasts that could better inform strategic crime-reduction efforts.

Although actionable hotspot policing information was derived from this investigation, crime loca-

tion data were aggregated to a 6-month

1

interval, which could have masked important variation in

crime pattern stability that could have occurred within the interval.

Finally, many scholars have studied micro-temporal patterns in crime data, but few have exam-

ined the micro-temporal stability of crime hotspots (Adepeju et al., 2016, is a noteworthy exception).

For example, researchers have investigated changes in intraday crime risk (Lemieux & Felson,

2012) and the ebb and flow of the ambient population throughout the day and its correlation to

when and where crime occurs (Andresen, 2011). Analytic techniques used to detect micro-temporal

patterns in crime data have also been developed, such as Ratcliffe’s (2002) aoristic analysis.

2

However, additional research into crime pattern stability across micro-temporal intervals is needed

because understanding crime hotspot stability at a micro-temporal level can help improve our ability

to proactively combat crime and make our communities safer. The current study was undertaken in

response to this need for additional research.

The remainder of this article is organized in the following manner. First, a review of the relevant

empirical literature is provided, focusing on crime hotspot mapping techniques and the metrics

commonly used to assess them, with particular attention paid to the time intervals commonly used

in crime hotspot mapping. Next, the data and methods employed in the current study are described,

followed by a presentation of the results. The article concludes with a discussion of the current

findings and their implications for hotspot analysis. Limitations of the current investigation are also

presented, and recommendations for future research are offered.

Literature Review

The law of crime concentration represents decades of empirically based knowledge about crime and

place, summarized in a single statement: “[F]or a defined measure of crime at a specific microgeo-

graphic unit, the concentration of crime will fall within a narrow bandwidth of percentages for a

defined cumulative proportion of crime” (Weisburd, 2015, p. 133). Criminological theories that

attempt to explain these crime patterns argue that they are linked to the criminal opportunit ies

created by the built environment (Clarke, 1995; Jeffery, 1971; Newman, 1972), to the daily routine

activity patterns of potential offenders and victims as they intersect in space and time (Cohen &

Felson, 1979; Felson & Eckert, 2019), and to places within the environmental backcloth that gen-

erate and attract crime (Brantingham & Brantingham, 1984, 1991; Brantingham et al., 2016). These

insights play an important role in supporting law enforcement efforts to reduce and prevent crime

through hotspot policing.

Identifying Crime Patterns

Crime analysis is a fundamental component of most hotspot policing strategies, with identifying

higher-than-normal concentrations of crime being one of the analysts’ most important objectives

174 Criminal Justice Review 46(2)