Secondary Data Analysis
Where and when are children and young people most likely to be involved in crime, and which local contextual factors best predict these concentrations of crime?
It is well established that crime and violence is highly concentrated, both geographically and by time of day. Most offending is also concentrated in small geographic areas, and it is likely that local environmental settings (e.g., physical infrastructure) and their wider context (e.g., neighbourhood characteristics such as poverty, social cohesion) are predictive of crime in these areas. Knowing where and when violence is mostly likely to happen is vital in order to reduce the harm it causes and ultimately to prevent it.
Our understanding of the specific times and places where crime involving children and young people across England and Wales tends to concentrate remains limited. Similarly, little is known about the contextual factors that predict these concentrations of crime among children and young people, and how these crimes and their predictive factors may differ by age, crime type and ethnicity.
This project will aim to further our understanding across a number of UK police force areas and provide actionable insights to tackle crimes against children and young people.
The main research questions in this project are:
This study draws on small geographical area analysis of police-recorded crime data involving children and young people, and contextual data sources including the open-source Ordnance Survey datasets (to identify built and natural infrastructure that are predictive of crime), ONS Census 2021 variables (to establish the sociodemographic context of different geographical areas), UCL Space Syntax database (to measure pedestrian movement) and NTU Community Engagement Area Classification (to measure social cohesion and collective efficacy).
Spatial and temporal concentrations of crime involving children and young people will be analysed using Geographic Information System (GIS) software and advanced statistical models (Getis-Ord/GI*, generalised Gini coefficient and Seasonal Autoregressive Integrated Moving Average [SARIMA] models). This project will also examine whether these concentrations vary when using different micro-level geographical areas and different temporal scales (e.g. hours, days, months), whether there are concentrations in specific areas which have not been observed in other aggregated analyses, and whether concentrations vary by age, gender and ethnicity of the children and young people involved in crime.
Using binary logistic regression and multivariate analysis, this project will also examine the contextual factors that best explain these spatial and temporal concentrations, and whether these remain the same across different spatial and temporal scales.