Networks of Youth Violence: A Relational Approach to Prevention

Why we funded this 

Rising violence affecting young people, particularly knife crime, has become an urgent public concern. Between 2013/14 and 2023/24, homicides among 13–19-year-olds in England and Wales increased by 64%, while knife-related hospital admissions among 0–17-year-olds in England rose by 58%. Over the same period, knife-related homicides among 13–19-year-olds surged by 141%. 

A substantial body of research has identified individual trauma, adverse school environments, poor family relationships, and residence in high-crime neighbourhoods as key risk factors for youth involvement in serious violence. However, these individual and contextual factors provide only a partial explanation. Growing evidence suggests that social relationships and peer networks play a critical role in shaping patterns of violent offending and victimisation, yet these relational dynamics remain underexplored. Existing research is largely based on data from in highly segregated US cities with high levels of firearm violence, focuses predominantly on adult offenders and co-offending networks, or is confined to organised crime settings. In addition, most studies treat social relationships as static, failing to capture how networks form, change, and influence behaviour over time. This project addresses these important gaps by examining the role of evolving social networks in youth violence within the UK context.

What are the main research questions 

The main research questions this project aims to address are:  

• How is violence structured within networks involving vulnerability young people? What is the network potential for violence contagion, clustering, and cycles of violence? 
• What individual, relational, and area-level factors increase the likelihood that a young person will (a) commit violence or (b) become a victim of violence? 
• What factors make a young person’s behaviour more likely to escalate to violence, or increase their risk of becoming a future victim of violence? 

What the analysis involves 

The project utilises a social network analysis approach to investigate the role of relational factors in the emergence and spread of violence among young people who have already come to police attention. Social network analysis is based on the idea that relationships matter in explaining social phenomena, as the structure of such relationships and an individual’s position within that structure impact outcomes, such as which young people are likely to commit violence or be a victim of it.  

The project will first employ descriptive network analysis including degree centrality, betweenness centrality, and community detection algorithms. Then, the dataset will be split into time periods to identify factors that impact future behaviours and introduce latent space models to identify positions within criminal networks. Finally, it will adopt longitudinal models to understand temporal evolution of violent behaviour.