Exploring the Latest Findings in the SEBP New Research Bulletin
- Matt Bland
- 2 hours ago
- 5 min read
Our new series of updates on new studies relevant to policing practice have been published in the last few weeks. In this bulletin the research spans recruitment and organisational fairness, hotspot detection, violence against women and girls, investigative practice, policing technology, officer wellbeing, and the use of artificial intelligence in policing.
Below is a selection of recent empirical studies and evaluations with potential implications for policing policy and practice.
Have we missed anything? Let us know - coo@sebp.police.uk
Police Vetting Decisions and Ethnicity
Watson, Katus, Shah, Barnes, Melia & Sutherland
Journal of Experimental Criminology (2026)
Method: Randomised vignette experiment with 1,778 vetting decisions by police vetting professionals
This study examined whether ethnicity influences police vetting outcomes. Vetting professionals reviewed fictional applicant profiles where ethnicity was randomly varied. Across five of seven scenarios, ethnicity had no statistically significant effect on vetting decisions. However, in two scenarios involving more complex applicant backgrounds, minority ethnic applicants were less likely to be approved.
The results suggest that vetting processes are largely consistent, but that bias may emerge when decision-making becomes more ambiguous. Monitoring outcomes and reviewing decision guidance may help ensure consistency in these situations.
Tags: recruitment, organisational fairness, legitimacy
A Globally Optimal Algorithm for Hotspot Detection
Martin Boldt
Crime Science (2026)
Method: Algorithm development and empirical evaluation of 1.75 million crimes across Malmö, Boston and New York
This paper introduces a new algorithm designed to identify crime hotspots more effectively than commonly used spatial techniques. Using crime data from three cities, the algorithm captured a larger proportion of crimes within hotspot areas than standard kernel density approaches.
The study also presents a faster “greedy” version that approximates the optimal solution while significantly reducing computation time. The approach may offer crime analysts improved tools for identifying priority patrol locations.
Tags: hotspot policing, crime analysis, resource allocation
Invisible Harms: The Hidden Health Impact of Fraud
Skidmore, Halkon, O’Connell, Meenaghan & Button
Police Foundation / NIHR (2026)
Method: Survey (n=311) and qualitative interviews
Fraud is the most common crime experienced in the UK, yet its health impacts remain poorly understood. This study surveyed fraud victims and conducted interviews across two police force areas.
Most victims reported negative health impacts following victimisation, including stress, anxiety and sleep disruption. A substantial proportion reported behavioural changes such as withdrawal from activities, and a smaller group reported severe distress including thoughts of self-harm.
The report argues that fraud should be treated as a significant public health issue and calls for policing and support services to recognise the wider harms experienced by victims.
Tags: fraud, victim support, wellbeing
Link:
Policing Violence Against Women and Girls: Trust and Engagement
Aisha K. Gill & Sundari Anitha
Policing & Society (2026)
Method: Mixed methods study: 52 police officers in focus groups, 55 case files, body-worn video review
This study examines police engagement with racially minoritised victims of violence against women and girls. Researchers found that victims’ needs and expectations were not always fully met, and that inconsistent communication and investigative practices could undermine trust.
However, the study also identified examples of promising practice, particularly where officers demonstrated empathy, proactive safeguarding and strong partnership working. The authors argue that improving trust requires both organisational learning and stronger engagement with affected communities.
Tags: VAWG, trust and confidence
Investigating Intimate Partner Sexual Violence
Cassandra Wiener, Andy Myhill & Merili PulleritsPolicing & Society (2026)
Method: Qualitative study with practitioner interviews and case analysis
This research examines how police investigate sexual violence within intimate relationships. The study argues that these cases are often poorly understood because they sit between domestic abuse and sexual offence frameworks.
Forces that approached these investigations through a domestic abuse lens were better able to recognise coercive control dynamics and support victims effectively. The authors suggest that clearer investigative frameworks and training could improve outcomes for victims.
Tags: domestic abuse, sexual violence investigations
Dynamic Guardianship and Residential Burglary
van Sintemaartensdijk, Frerichs, Friehs & de Vries
Journal of Experimental Criminology (2026)
Method: Behavioural experiments using virtual-reality burglary scenarios
This research explores whether “dynamic guardianship” signals such as automated lighting, smart cameras and self-closing blinds, influence offender decision-making.
Participants placed in simulated burglary scenarios responded differently to these signals. Cameras appeared to increase perceived risk, while other cues had weaker effects. The findings suggest that smart-home technologies may contribute to deterrence but should be considered complementary to traditional situational prevention measures.
Tags: burglary prevention, situational crime prevention
Measuring Police Burnout Using Brain Monitoring
Chen et al.
Scientific Reports (2026)
Method: Functional near-infrared spectroscopy (fNIRS) with machine learningSample: 33 active police officers performing simulated operational tasks.
Researchers used brain-imaging technology to measure stress responses in police officers during simulated tasks. A machine-learning model trained on the brain activity data was able to classify burnout levels with high accuracy.
Although still experimental, the research suggests physiological indicators may eventually complement traditional self-report measures when assessing officer wellbeing.
Tags: wellbeing, burnout
Social Media Campaigns and Adolescent Relationship Violence
Seddig, Bartz, Bliesener, Rühs, Schauten & ThomsenJournal of Experimental Criminology (2026)
Method: Randomised survey experiment of 1,973 adolescents, 772 follow-up
Researchers evaluated a social-media campaign (in Germany) promoting healthy relationships among young people. Participants exposed to campaign videos and audio content demonstrated improvements in knowledge and intentions to seek help.
Attitude changes were smaller, but some effects emerged over time. The study suggests digital campaigns may increase awareness but may need to be combined with broader prevention efforts to achieve sustained behavioural change.
Tags: youth violence prevention, digital interventions
Link:
Natural Language Processing and Police Reports
Lukmanjaya, Halmich, Butler, Cook & Karystianis
Crime Science (2026)
Method: Scoping review of 61 studies
This review examined how natural language processing techniques have been used to analyse police narrative reports. Across the studies reviewed, machine-learning models showed promising performance in tasks such as classifying domestic violence cases and extracting information from incident narratives.
However, the authors highlight inconsistent reporting standards and limited discussion of ethical issues. They call for improved governance frameworks and shared benchmarks for applying NLP techniques to policing data.
Tags: AI, policing data, intelligence analysis
Generative AI and Financial Crime
Tiwari, Zhou & Lee
Crime Science (2026)
Method: Quantitative systematic literature review (94 studies)
This review examines how generative AI is influencing financial crime. The literature shows that offenders are increasingly using AI tools for fraud automation, voice cloning and synthetic identity creation.
At the same time, law-enforcement agencies are beginning to apply AI techniques to detect suspicious transactions and identify emerging fraud patterns. The authors emphasise the importance of developing governance frameworks and analytical capabilities to respond to AI-enabled crime.
Tags: fraud, cybercrime, AI
