Stalking is a patterned, escalating behaviour that remains chronically misunderstood and inconsistently managed within UK policing. Despite major legislative and policy developments over the past decade, the frontline response still hinges on tools that are outdated, poorly aligned with contemporary stalking behaviours, and unevenly applied across forces. This thesis critically examines how stalking is currently identified, assessed, and managed within UK policing and partner-agency practice, drawing on a mixed-methods approach including Domestic Homicide Review (DHR) analysis, comparative evaluation of stalking risk assessment frameworks, practitioner consultation, content analysis, correlation analysis, cross-tabulation, Cohen’s Kappa and Bland–Altman agreement analysis. These approaches were selected to explore classification reliability, agreement between frameworks, practitioner perspectives, and the operational feasibility of newly developed and proposed models. Across these sources, a clear pattern emerges: current approaches remain broken, typology-blind, overly reliant on static checklists, and heavily shaped by a postcode lottery rather than national consistency. In response to these gaps, this thesis develops and evaluates two interconnected models and proposes a third conceptual framework. The Stalking Perpetrator Spectrum (SPS) was developed as a typology-based classification model and explored through comparative analysis against an existing framework to assess consistency and potential utility in early risk identification. The Stalking Offender Risk Management model (SORM) was developed through practitioner engagement to support dynamic risk management and operational feasibility; it seeks to inform safeguarding decisions while targeting recidivism and supporting rehabilitation. Finally, the Violence Likelihood Analyser (VIOLA) is proposed as a conceptual framework, a theoretical “reverse VICAP” (Violent Criminal Apprehension Program), proposing a system that uses behavioural science, machine learning, and rolling intelligence updates to detect escalation patterns that static risk assessment tools routinely miss. Collectively, SPS, SORM and VIOLA illustrate how layered, typological and behaviourally informed approaches may improve consistency, transparency and responsiveness in stalking classification, risk assessment and management. The thesis contributes to knowledge by reframing stalking as a patterned, typological and dynamic form of harm requiring approaches that integrate behavioural threat assessment, victimology and contextual risk factors. It is proposed that the future of stalking risk assessment and risk management lies not in creating ever-longer checklists, but in building systems capable of recognising patterns, adapting regularly to new information, and working with, rather than against the realities of frontline policing. Finally, the findings further offer evidence-informed recommendations for future policy, operational practice and the continued development of stalking-specific risk assessment frameworks.
| Date of Award | 24 Jun 2026 |
|---|
| Original language | English |
|---|
| Supervisor | Jason Roach (Main Supervisor) & Rosie Campbell (Co-Supervisor) |
|---|