The long-term impact of school exclusion on risk behaviour: a retrospective cohort study

This study is being led by Prof Catia Nicodemo (University of Oxford, Brunel University of London). The study started in 2025 and aims to finish in 2027.

‍This project was reviewed and approved through ECHILD's data access process. Learn how ECHILD data is kept safe.‍

  • We want to investigate how school exclusions affect risky behaviour later in life.

  • School exclusions are linked to antisocial behaviour and mental health issues. However, evidence up to date are limited to correlational studies focused on short-term outcomes. This study provides more causal evidence of the impacts of exclusions on risky behaviours in early adulthood.

  • Using a retrospective cohort design and event study around a policy reform in 2012 which reduced the number of exclusions, we evaluate how exclusion affects risky health behaviours such as substance use, alcohol misuse, traffic injuries and early pregnancies.

Find out more

  • Why does this matter?

    Exclusion from school is associated with many negative outcomes, but we do not have much robust causal evidence on its exact impacts.

    What do we want to find out?

    How does school exclusion affect antisocial behaviour such as substance use, alcohol misuse, traffic accidents or early pregnancies.

    What we will do

    Using a retrospective cohort design and event study around a policy change in 2012, we analyse the impact of school exclusions on these health outcomes between cohorts and over time.

  • Background and aims

    School exclusion, through suspension or expulsion, is linked to negative outcomes like delinquency and mental health issues. Existing research is limited by short-term focus and correlational methods. This study addresses these gaps by investigating the causal impact of exclusion on health risk behaviours in adolescence and early adulthood. The findings aim to inform policies to reduce exclusionary practices and support affected students.

    Methods and analysis

    Using an event study and retrospective cohort design with linked longitudinal data, we examine how exclusions affect risky health behaviours such as substance use, alcohol misuse, experience injuries related to traffic accidents and early pregnancies.

    We focus on the 2012 policy reform in England as a natural experiment where parental rights to request reviews of exclusions were introduced, reducing the rates of exclusions. We compare cohorts of children born between 1991-2001, some of whom reached the end of compulsory education before (control) and some after (treated) the reform. We adjust for confounders such as socioeconomic status, prior academic achievement, school attendance and regional differences.

    Imputation techniques will be adopted where appropriate to overcome issues with incomplete data. To address potential bias from unmeasured confounders, we include a wide range of covariates in regression models and run robustness. Event study design will also include control cohorts to account for secular trends, and placebo tests will be conducted to validate the findings.

    ECHILD data used

    • HES: Outpatients (2005-2023), A&E (2007-2020), Emergency Care (2020-2023), Admitted Patient Care (2005-2013), Critical Care (2007-2023), Mental Health Minimum Dataset (2006-2016) + mother-baby link (2005-2022)

    • NPD: School Census (school + pupil level, 2005-2022), Early Years Census (2007-2022), Pupil Referral Unit Census (2009-2013), Alternative Provision Census (2007-2022), EYFSP (2005-2019), Key Stages 1-5 (2005-2022), Absences (2005-2022), Exclusions (2005-2021), National Client Caseload Information (2005-2021), Children Looked After Return (2005-2021) and Children in Need Census (2008-2022)