Exclusions are the most severe punishments available in the English school system and involve removing students from school on a temporary basis (a suspension) or permanently (an expulsion). It is also one of the most persistent and well-established risk factors for future criminal behaviour. In our data, nearly 50% of the young people we observe in custody aged 15-17 experienced an exclusion while in Year 10, compared to 6% among pupils that do not end up in custody. While the relationship between exclusion and later involvement in the justice system is well-recognised (eg see the Timpson Review of School Exclusion), it is not known whether the relationship is causal – that is, whether being excluded actually causes more young people to commit crime and end up in custody.
Today we publish research that provides evidence suggesting that school exclusions do have a causal impact on youth custody. These results highlight an onward negative criminal justice outcome of education policy that, to the best of our knowledge, has not previously been quantified.
What we did
This research was carried out as part of a broad study funded by the Nuffield Foundation on the educational influences and labour market consequences of youth custody. Like the other studies in this project, we used population-level administrative data, including the National Pupil Database (NPD) to investigate school experiences and youth custody.
Specifically, we used data for four cohorts of students, focusing on whether students are permanently excluded or suspended in year 10, when pupils are 14-15 years old, and their experiences of custody age 15-17 (inclusive). Our analysis used academisation, a change in school status that moves schools out of direct local council control to greater self-governance, to understand the impact of exclusions on custody. Previous research has found that upon conversion academies enforce stricter discipline policies. Therefore, assuming that academisation itself does not increase the likelihood of custody, any increase in exclusion can be viewed as a random shock, and we could use this to assess the impact of exclusion on custody. To make this assumption more plausible, we look only at schools that eventually convert to academies but differ in when they convert.
What we found
We found that attending a school that converts to an academy in Year 10, the year when pupils are most likely to be excluded, increases the probability of receiving a suspension or permanent exclusion by 3 percentage points. This compares to a raw exclusion rate of 6% among all pupils in our sample, or 5.3% among pupils in the sample that attend schools that had not yet converted to academies. This implies that academising resulted in an extra pupil (4 instead of 3) being excluded in every two classes of pupils. This result aligns with previous research – that academies enforce stricter discipline policies and make greater use of suspensions – and demonstrates that academisation is a good predictor of pupil exclusions. .
Based on this increase in the probability of exclusion as a result of academisation, we found evidence that receiving an exclusion resulted in an increase in the probability of custody age 15-17, with impacts varying depending on the type of exclusion. Specifically, receiving a permanent exclusion increases the probability of custody by 33 percentage points. For suspension, the increase is 1.3 percentage points. These are sizable increases considering the custody rate of 0.1% among pupils as a whole.
What this means for policymakers and research
Our interpretation of these results is that exclusion presents a small but non-ignorable risk of increasing the likelihood that a pupil will end up in custody. Even acknowledging the potential limitations of this research – including critically whether it is plausible to assume that academisation does not impact on the likelihood of custody in any other way except through exclusion, and lack of statistically significant results on some robustness checks – this potential risk warrants, at the very least, further exploration.
Given the consequences (and cost) of both exclusion and custody, the possibility of a negative effect should be taken seriously by researchers and policy practitioners alike. This evidence does not necessarily mean that exclusion as a policy should cease – although some may advocate that. What it does mean is that headteachers, teachers and policymakers should take seriously the idea of exclusion as an intervention that can and should be better understood in terms of its short and long-run impacts, on both pupils excluded and their peers. We have provided evidence here only on the impacts on excluded pupils.
We recommend further research to better understand the relationship between exclusion and crime, building on this research. In particular, we think that further administrative data linkages would allow researchers to study contacts with the criminal justice system before custody, such as convictions and arrest, and episodes of custody later in life. Additionally, because exclusion and custody are fairly rare events, looking at experiences for more cohorts of students would increase the sample size and therefore confidence in the results.
Regardless, this research provides critical evidence on the impact of exclusions on pupil outcomes. If this can be combined with other basic facts about the costs and benefits of exclusion, this will support the development and use of evidence-based policy decisions for school behaviour policy to improve outcomes for all pupils.
This research has been funded by the Nuffield Foundation, as part of a larger project (with the Universtity of Wesminister, the Universtiy of Bath, and the London School of Economics) providing quantitative evidence on which young people are imprisoned, the school experiences that increase the risk of imprisonment and the impact prison has on subsequent labour market outcomes.