A causal analysis of school exclusions and youth custody using administrative data and a bit of luck
Alex Sutherland, Giulia Tagliaferri
Abstract
Peer Reviews
No public reviews on file for this paper yet. If you reviewed it on a platform where reviews are public (OpenReview, ICLR, NeurIPS, ICML), you can paste yours below so the community can read it here.
Videos
No videos yet. Explain this paper in a talk, walkthrough, or lecture? Add one.
Taxonomy
TopicsHomelessness and Social Issues · Criminal Justice and Corrections Analysis · Youth Substance Use and School Attendance
Objectives
There is a strong relationship between being excluded from school and experiences of criminality and custody in later life. But we don’t know if this relationship is causal: does being excluded cause more people to commit crime and end up in custody? This study tries to answer this question directly.
Methods
We linked data from the National Pupil Database and the National Client Caseload Information System for four cohorts of Yr10 pupils attending state-funded schools in England.
We analysed the effect of being permanently excluded or suspended in Year 10 on the probability of custody at age 15-17 (inclusive) for more than a million pupils. We used academisation, when schools move out of local council control to greater self-governance, to understand the impact of exclusions on custody. We treat academisation as a ‘shock’ that increases the likelihood of exclusion then use a statistical approach called instrumental variables that enables us to leverage this change to understand the impact on later custody.
Results
Attending a school that converts to an academy in Year 10 increases the probability of a pupil receiving a suspension or permanent exclusion by 3 percentage points. A Year 10 pupil attending a school that academised resulted in a statistically significant increase in the probability of custody age 15-17, with impacts varying depending on the type of exclusion:
a permanent exclusion increases the probability of custody by 33 percentage points (statistically significant at 5%); for suspension, the increase is 1.3 percentage points (statistically significant at 5%)
Conclusion
Employing a novel dataset combined with an approach that allows us to distinguish causal relationships, we can make stronger claims about the impact of exclusion on custody (and thus crime). Our results highlight an onward negative criminal justice outcome of education policy that, to the best of our knowledge, has not previously been quantified. Exclusion presents a small but non-ignorable risk of increases in custody, warranting greater scrutiny.