Exploring Pathways to Children Missing from Education in Wales: A Longitudinal Linked Data Approach

ObjectivesChildren missing from state education is a significant policy concern, particularly since COVID-19. This study extends previous work by considering repeated cross sections of national child populations to show change over time and a cohort-based approach to better understand the longitudinal determinants of missingness from the state education system. MethodsWe link routine health and education administrative datasets in the SAIL Databank, including GP- and school registered children living in Wales along with those educated outside mainstream schooling (Educated Other Than At School, EOTAS). Using both cross-sectional and longitudinal approaches we assess for which years children were resident in Wales, and for which of those they were in state education. We use recorded low-level geography codes to determine whether that young person lived at a residential property or a communal establishment/institution (e.g. boarding school) and whether they are still resident in Wales. ResultsInitial analysis indicated that approximately 24,000 (6.0%) of children were estimated to be missing from state education in 2021, describing the variation by age (4.5% to 8.1%), and local authority. That work has now been updated with the most recent data to show the change in the pre-COVID to post COVID era. Furthermore, we have extended that cross sectional approach to establish a cohort for those expected to complete schooling in 2021 and to examine educational missingness longitudinally. Findings provide evidence on the importance of absence and exclusions as pathways to becoming children missing from state education. ConclusionLinking education to health records provides a robust estimate of children missing from education. Triangulating the cross-sectional approach with the updated cross section and cohort approaches enables more robust evidence of the determinants of children missing from education at the national and individual levels.