Chair: Matthew Ashby

WG-PLACE: Data innovations: measurement, sources and comparisons

Building: A
Room: 01

Author: Lee Won Do , Manchester Metropolitan University Crime and Well-being Big Data Centre

Mark Ellison, Manchester Metropolitan University Crime and Well-being Big Data Centre
Title: Police Risk Assessment of Domestic Abuse: the Mediating Role of Space and Time
In England and Wales, police forces deploy an evidence-based Domestic Abuse, Stalking and Harassment, and Honour-Based Violence (DASH) assessment tool when responding to reports of domestic abuse. The DASH assessment tool is primarily utilised to undertake risk prediction / estimation at the incident, though it can also be used to undertake harm and needs identification as well as demand management (Medina Ariza et al. 2016). This paper explores whether risk prediction (High, Medium and Standard) derived from 27 DASH questions differs across space and according to the time of year. Moreover, and over a period when the number of domestic abuse associated crime have seen significant growth, whilst the level of police funding has seen significant decline, the paper questions the consistency of risk prediction through time. The paper draws on a dataset of 360,000 DASH assessments, risk assessment outcomes and victim characteristics for the period 2011-2017. It deploys probabilistic and heuristic machine learning-based algorithms to evaluate the existence and degree of spatially and temporally weighted decision-making and to offer guidance as to how this might be overcome. The importance of this research rests in supporting equitable service delivery in an era of fiscal strain.
Keywords: domestic abuse; risk assessment; DASH; police; demand
Author: Ashby Matthew , Nottingham Trent University

Title: Looking for Neighbourhood-Level Variations in Crime Seasonality in Multiple Cities
The frequency of crime varies throughout the year, with many crime types being more common in the summer. Recent research has begun to consider whether these seasonal patterns of crime might be different in different neighbourhoods. However, existing studies in this area have been largely descriptive, based on data from a single city and based on monthly counts of crime. This limits the generalisability of this research and leaves open the possibility that important patterns have been obscured by using aggregate counts. The present study attempts to further research in this area by searching for local variations in seasonality a) using data from multiple cities, b) using daily counts of crime and c) using inferential tests that are suitable for seasonal time-series data.
Keywords: crime seasonality; crime patterns; crime analysis; spatio-temporal patterns of crime
Author: Wallace Stephanie , Manchester Metropolitan University Crime and Well-being Big Data Centre

Karolina Krzemieniewska-Nandwani, Manchester Metropolitan University Crime and Well-being Big Data Centre
Title: Hidden Spatial Inequalities in the Exposure to Crime?
It is well known that exposure to ‘recorded’ crime varies markedly across urban space (Hope et al, 2001; Rey et al, 2012). It is less clear, however, whether ‘recorded’ crime reflects a true representation spatial crime patterns. The ‘reporting’ of crime types is known to vary (Tarling & Morris, 2010), but does the ‘reporting’ of crime also exhibit spatial variance and does this manifest in ‘recorded’ crime? This paper seeks to explore hidden spatial inequalities in the exposure to crime through interrogating the relation between calls for service (reporting crime) and crime (recorded), paying attention to the factors that mediate this relation. The research draws on data from two large metropolitan cites in the UK and employs time-series structural equation modelling and cluster analysis techniques to examine the existence of causal relationships at different spatial (and temporal) scales. The results of the research indicate stark distinction in the ‘reporting’ of crime across urban space and in the relation between ‘reported’ and ‘recorded’ crime, influenced by the level of deprivation. ‘Recorded’ crime masks significant spatial inequalities in the exposure to crime.
Keywords: calls for assistance; recorded crime; spatial inequalities; deprivation
Author: Cook Will , Manchester Metropolitan University Crime and Well-being Big Data Centre

Jon Bannister, Manchester Metropolitan University Crime and Well-being Big Data Centre
Title: A Longitudinal Study of the Relationship Between the Decentralisation of Urban Poverty and Crime
The suburbanisation of poverty thesis posits that urban change in developed economies is increasingly described by the decentralising of poor neighbourhoods from the city core to the city periphery. Such a dynamic has been observed in US and UK cities (Bailey and Minton, 2017). These changes may have important implications for access to employment and the delivery of public services. However whilst the suburbanisation of poverty has gained empirical support, there has been little consideration of its effects on the pattern of crime and disorder with cities. Using detailed neighbourhood level data on offences reported in two British cities, we examine the spatial distribution of crime and its relationship to the decentralisation of neighbourhood deprivation over the period 2001-2016. This analysis comprises: i) a descriptive analysis of the changing distribution of crime within a city; and ii) a longitudinal model of neighbourhood level crime based on neighbourhood deprivation. The findings contribute to understanding how the spatial distribution of poverty affects crime and provide some of the first empirical evidence of the social effects of the suburbanisation of poverty.
Keywords: poverty; deprivation; neighbourhoods; crime concentrations
Marked presentations to PDF Generate ics file