Uncovering the mechanisms of racial disparity and cumulative disadvantage in the sentencing of children for drug-related offences in England and Wales: an examination of direct and indirect effects

This thesis examines whether race is associated with harsher outcomes in the sentencing of children for drug-related offences in England and Wales, and whether disparities arise through direct or indirect (mediated) race effects. Using the administrative Data First dataset, it focuses primarily on custodial sentencing while also analysing discharges, Youth Rehabilitation Orders, Referral Orders, and non-convictions. Extending prior research, the thesis employs mediation analysis (in the form of path analysis) to test indirect race effects via remand and guilty plea, interpreted through a micro–meso–macro racialisation framework and informed by qualitative interviews with defence lawyers. The findings show that Black, Mixed Heritage, and Other ethnicity children were significantly more likely than White children to receive a custodial sentence, and all racially minoritised children were also more likely to be remanded in custody. Crucially, where no direct race effect on custody is detected, an indirect effect via remand is present – demonstrating that models which control for remand but ignore mediation underestimate the total effect of race on sentencing outcomes. Modelling both direct and indirect pathways suggests that remand is likely the key mediator through which race effects are transmitted and cumulative disadvantage is reproduced, connecting pre-sentencing inequalities to harsher sentencing outcomes. Analysis of non-convicted cases suggests that weaker cases involving racially minoritised children might be progressing to court rather than being diverted, indicating upstream inequalities. Overall, unequal diversion and remand decisions create a cumulative disadvantage that increases the likelihood of harsher sentences for racially minoritised children. Methodologically, the thesis demonstrates the value of mediation for sentencing research in England and Wales: only by modelling indirect pathways can the full extent of racial disparities be revealed, with the true scale larger than suggested by traditional models that treat racially influenced variables as neutral-controls.