Peer-reviewed article
Hidden Harm-Exploring the Utility of Geostatistical Analysis to Identify Child Criminal Exploitation (CCE).
Keaney-Bell A, Walsh C ·
Summary
Written by AI, not yet AICS's view · claude-haiku-4-5-20251001This study uses geostatistical analysis to predict where child criminal exploitation is likely to occur by analyzing spatial patterns of risk factors such as deprivation, violent crime, and educational attainment. Researchers integrated seven open-source datasets and applied spatial analysis techniques to identify high-risk areas. The method reduced the geographic area requiring investigation from 14.45% to 0.84% of land area, enabling more targeted resource allocation. The approach combines criminological theory with geospatial methods to support safeguarding strategies.
Key points
- Contextual risk factors for child criminal exploitation are spatially clustered rather than randomly distributed
- Using four search parameters reduced ward locations by 85-99% and affected population from 17.91% to 1.41%
- The study integrated seven open-source datasets on education, age demographics, violent crime, deprivation, and paramilitary violence
- Geospatial analysis enables more efficient targeting of limited public resources for child protection
- The model is limited by absence of multi-agency datasets and requires ground-truthing for validation
Limitations
The study did not include multi-agency datasets, which limits validation of the predictive model. The analysis is methodological rather than evaluating an implemented intervention, so it cannot directly demonstrate whether targeted resource allocation actually reduces child criminal exploitation.
Abstract
This interdisciplinary study integrates criminological theory with geospatial methods to analyse large, multi-format datasets using geostatistical techniques. The aim is to predict where Child Criminal Exploitation (CCE) is likely to cluster, based on the spatial convergence of contextual risk factors. Drawing on insights from General Strain Theory (GST) and prior research on CCE, this study integrated seven open-source datasets capturing educational attainment, age demographics, violent crime, deprivation, and paramilitary-related violence. These variables were operationalised to construct a proxy measure for strain. Spatial analysis was conducted using ArcGIS Pro, including the Data Interoperability extension, to enable efficient integration and interrogation of multi-format geospatial data. Geospatial analysis demonstrated that contextual risk factors for CCE are spatially clustered. Using four search parameters, a small subset of wards with elevated risk were identified. This resulted in a reduction in ward locations by 85-99%, land area under investigation from 14.45% to 0.84%, and affected population from 17.91% to 1.41%, enabling more targeted and efficient resource allocation. As understanding of the contextual factors contributing to CCE improves, this methodological approach offers scalable and data-driven means of identifying high-risk areas. By integrating geospatial analysis with criminological theory, the model supports more effective safeguarding strategies and prioritisation of limited public resources. This study is limited by the absence of multi-agency datasets, which were beyond its scope. Future research aims to incorporate cross-sector data to validate and refine the model through ground-truthing, enhancing its predictive accuracy and practical applicability.