Mapping the bias of police records: Project final report

Angelo Moretti, David Buil-Gil

Research output: Book/ReportCommissioned report

282 Downloads (Pure)

Abstract

Police-recorded crimes are the main source of data used by police forces and researchers to analyse the distribution of crime in geographic areas. These data, however, are affected by measurement error arising from underreporting and recording inconsistencies across police jurisdictions. Not all persons are equally willing to report crimes to the police and cooperate with police services, and practices followed by the police to record crime vary across police jurisdictions. Both these issues affect the ‘dark figure of crime’ (i.e., all crimes unknown to the police), which vary across areas. The open question that this project addresses is whether micro-level maps of police-recorded crimes (i.e., maps produced from aggregating crimes at very detailed spatial scales) suffer from a higher risk of bias than maps of crime produced at larger scales, such as neighbourhoods and ward. While communities unwilling to cooperate with the police may concentrate in some micro-places more than others, and thus the ‘dark figure’ may vary across small areas, larger geographies aggregate more heterogeneous social and demographic groups, and thus the proportion of crimes unknown to the police may be more similar across areas. We utilise data from the UK Census and the Crime Survey for England and Wales to generate synthetic crime data in Manchester, UK, and analyse if micro-level crime maps are affected by a larger risk of bias than maps of crime aggregated in neighbourhoods. The main findings of this project are:
• The proportion of crime unknown to the police varies substantially across micro-places.
• The proportion of crimes unknown to the police is similar across neighbourhoods.
• Micro-level maps of crime are affected by a larger risk of bias than maps of crimes aggregated at larger spatial scales.
• The risk of bias in micro-level crime mapping is attributed to the fact that social groups unwilling to cooperate with the police concentrate in some areas more than others. This is less of a problem when aggregating crimes in neighbourhoods.
• Future work is needed to address measurement error in crime data.
Original languageEnglish
Place of PublicationManchester
PublisherManchester Statistical Society
Number of pages9
Publication statusPublished - May 2021

Research Beacons, Institutes and Platforms

  • Cathie Marsh Institute

Fingerprint

Dive into the research topics of 'Mapping the bias of police records: Project final report'. Together they form a unique fingerprint.

Cite this