Dr Angelina Nazarova has been awarded one of 17 new ADR UK research fellowships, designed to provide new insights into pressing policy questions using flagship administrative data sets. The evidence generated will have the potential to help policy makers, local authorities, schools and public services to understand how to target support and improve outcomes.
Read more about the ADR UK fellowships using ECHILD data here
Dr Nazarova’s project, SENLOSS: The effects of losing special educational needs support at educational transitions in England will investigate how often children lose special educational needs support at key school transitions and what happens afterwards. It will help identify which children are most likely to support.
Project details
This project aims to explore the following research questions:
- How often do children lose SEN support when they move between stages of school, and which children, schools and areas are most affected?
- What effect does losing support have on children’s learning, engagement with school and health?
- How do these effects differ according to children’s type of need and other characteristics, and which groups are most affected when support is lost?
The methodology used in this study:
- The project will follow children who were receiving school-based SEN support before moving to a new stage of school. It will focus particularly on the move from Year 6 to Year 7, as children start secondary school, as well as the earlier transition from Reception to Year 1. Using ECHILD, the research will compare children whose support continued with those whose support stopped.
- Children who lose support may already differ from children who keep it, so the project will use several statistical approaches to make the comparisons as fair as possible:
- Machine learning (a method called gradient boosting) will identify which characteristics of children, schools and local areas are most strongly linked to losing support.
- Target trial emulation will be used, which uses existing records to imitate the fair comparison you would get from a randomised trial.
- An event study will follow the same children in the years before and after they lose support, comparing how their outcomes change with how outcomes change for similar children who kept it. This is known as a ‘difference-in-differences comparison’: each child is measured against their own earlier record as well as against other children, which cancels out pre-existing differences between the two groups.
- A causal forest searches across many characteristics at once to find groups of children most affected by losing support.
Findings will be expressed in practical terms, such as differences in exam results, additional days of school missed and additional use of health services. This will help policymakers assess the potential consequences of withdrawing support alongside the resources required to provide it.