The association between bullying and violent crime 

Why we funded this 

Bullying is common. An estimated 35% of 10–15-year-olds in England experience in-person bullying each year. Evidence shows that bullying is linked to later violence, and that children involved in bullying – whether they are bullied, bully others, or both – are more likely to carry a weapon.  

But there are limits to what the evidence shows. Most studies measure bullying at a single moment in time, so they cannot tell us whether bullying that keeps happening, or bullying at a particular age, matters more. Most rely on children reporting their own offending rather than on police records. Most look only at children who bully others, and not at children who are bullied. And few studies have been carried out in the UK, which matters because the picture of violence here – particularly around weapons – is very different from the United States, where most of the evidence comes from. 

What are the main research questions?  

The aim of this research is to examine the association between being bullied and bullying others, and later violent behaviour and weapon carrying. 

The main questions this project sets out to answer are: 

• What is the association between bullying others or being bullied and both self-reported violent behaviour and official sanctions for violent offending? 
• Does this association change depending on how many times in childhood a child is involved in bullying, or on the age at which it happens? 
• What is the association between involvement in bullying and self-reported weapon carrying in adolescence? 
• What is the association between physical bullying specifically and both self-reported violent behaviour and official sanctions for serious violence? 

What did the analysis involve? 

This study uses linked data between the Avon Longitudinal Study of Parents and Children (ALSPAC), the National Pupil Database (NPD) and Avon & Somerset Police Data. ALSPAC is a cohort study that recruited pregnant women with a due date between April 1991 and December 1992, living in and around Bristol. It resulted in an initial sample of 14,901 children. These children and their parents have been tracked continuously into adulthood. The analysis included 9,944 of them. 

Children reported on bullying at ages 8, 10 and 12, covering behaviour that happened repeatedly at school or on the journey to and from school – from name-calling and spreading lies to being hit or beaten up – both done to them and by them. Children were counted as involved if this happened several times a month or more. The outcomes were self-reported violent behaviour at ages 17 and 18, police-recorded serious violence between ages 13 and 24, and self-reported weapon carrying at age 15. The research uses logistic regression, controlling for a wide range of individual, behavioural, family, school and neighbourhood factors, with a further method used to test whether the number of times a child was involved in bullying, or the age at which it happened, made a difference. 

Please refer to the analytical protocol for more details. 

Key findings 

Interpretation and implications 

The results of this study show that the consequences of bullying may extend beyond the immediate harms experienced during childhood, contributing to behaviours with serious criminal justice and safety implications into adulthood. The report found strong evidence that bullying perpetration (including physical bullying) was associated with both self-reported and police-recorded violence, even after accounting for child-, family-, and area-level factors. This association was strongest for children who engaged in bullying at multiple ages. This suggests that supporting children who demonstrate bullying behaviours should be a focus of early intervention and support. While this study found no evidence that being a victim of bullying was associated with later violence perpetration, once accounting for contextual factors, all forms of bullying involvement, including victimisation, were associated with a higher likelihood of weapon carrying. This is the first time evidence from the UK has suggested a potential link between bullying victimisation and precautionary violent behaviours, even where these behaviours may put someone at greater risk of harm.  

Several considerations should be taken into account when interpreting these findings. First, the dataset comprises children living in and around Bristol and captures bullying experiences during the mid-to-late 2000s. Consequently, more contemporary forms of bullying, such as cyberbullying, were not captured. In addition, only 5% of the ALSPAC sample consisted of participants from ethnic minority backgrounds. Although this reflected the demographic composition of the study area at the time of recruitment, the small number of participants from these groups meant that analyses by ethnicity could not be conducted. Together, these factors limit the generalisability of the findings to the contemporary UK population. Finally, although the results provide strong evidence that bullying was associated with later involvement in violence and weapon carrying, even after adjustment for a wide range of contextual and individual-level factors, the nature of the study design means we cannot conclude that bullying had a causal impact on later involvement in violence and weapon carrying. 

Please refer to the final report for more detail on these findings, how to interpret them, and their implications.