Research on spatial crime patterns have grown greatly and an increasing number of police forces worldwide have invested in geographical information systems (GIS) and different tools to predict future criminal activity.
As society keeps getting more digitalized, Norwegian policymakers too have shown an increasing interest for new technological solutions that can enhance prevention initiatives, inform resource allocation, and make the police more efficient. Although it is an official goal to put knowledge-based approaches in the forefront of Norwegian crime policy, there is little to no empirical research on where and when crime happen in Norway. We also do not know what kind of software might be beneficial for strategic decision making within a Norwegian context.
Contemporary criminology has established crime is highly concentrated at microgeographical places, that most places show stable developmental trends and that a few specific places consistently account for the highest crime rate in a city. Surely there are some nuances in the literature, but if it is true that some hot spots are chronic and most patterns stable – we wouldn’t need any advanced software to predict crime. Simple calculation and visualization of historical crime patterns would be adequate for employing prevention strategies or allocate police resources. Moreover, if these spatial features of crime are highly stable, we can assume local police (or other relevant actors) know where hot spots are located already. Then, we might wonder if we even need spatial crime analysis at all. This would indeed explain the large research gap in this field in Norway. But one can wonder what spatial stability would mean in practice… If crimes patterns are that stable, does it mean the police should always employ their resources at chronic hot spots? And, on a more theoretical note, does it imply crime is unaffected by environmental changes and crime prevention strategies used in the past? This sounds utterly naïve and would in fact contradict the very foundation of the Crime and place-literature, which made us wonder how the claims of spatial stability has been established in previous research and how we can best make use of spatial crime analysis in Norway.
Since there is no consensus of how to measure stability, our current research investigate how temporal variation is taken into account and how stability is operationalized and established in different research designs. We use a k-means clustering algorithm for longitudinal data to investigate heterogeneity in areas’ trajectories over 20 years. We also assess the robustness of our results with Group-based trajectory modeling which is commonly used in the field. In contrast to previous research, we focus on within-group variation, and demonstrate that clusters of seemingly homogenous areas have important within-group differences. This means spatial stability might be overestimated when these models are used on annual data as previously done. Without added scrutiny of within-group variation, places might wrongfully be considered to have stable developmental trends or to be the most criminogenic ones over time, even if ‘chronic’ hotspots might not exist.
To be able to spot temporary flare-ups of criminal activity early on, future research should focus on finding important fluctuations for specific crime types in different areas and look for systematic variation in spatial crime patterns. This analysis might be hard to comprehend by experience-based knowledge or without more sophisticated tools that can account for spatio-temporal interactions. A software that can perform more than basic crime mapping might therefore be beneficial in Norway, but it is a need to empirically test different models to find the best-suited one. Thus, we do need (detailed) spatial crime analysis in Norway. Since these kind of analysis allows us to recognize changes in environmental features or preventive strategies that correlates with changes in criminal activity, it can also enhance our general understanding of crime and crime prevention.
Annica Allvin is a criminologist and doctoral researcher at the Norwegian Police University College (PHS) and the faculty of Social Science at the University of Oslo (UiO). Her research focus on spatial- and temporal crime patterns and the use of digital crime mapping and spatial forecasting techniques in (Norwegian) policing and crime prevention. Her areas of expertise is criminology, preventive- and predictive policing, quantitative methods and geographical analysis.