Using catchment areas to estimate school effectiveness

This project aims to assess secondary school effectiveness using an alternative approach to existing value-added measures, such as Progress 8. It is based on the use of catchment areas or distance cut-offs by oversubscribed schools. The project is motivated by research from the U.S. suggesting that value-added measures may be biased.

A ‘fuzzy’ regression discontinuity design (RDD) is being used to estimate the causal impact of the school attended on pupil attainment at ages 16 and 18, comparing pupils living just inside with those living just outside the catchment areas of the oversubscribed schools. The National Pupil Database contains information on pupils’ locations and attainment, while information on school catchment areas is provided by LEAs.

The estimates of school effectiveness generated by this approach will be compared with estimates from a range of value-added approaches to assess bias in the whole sample and in relation to school characteristics such as type, region and intake profile.

The main aim of the project is to inform policymakers and others interested in education performance about the accuracy of the measures currently used to estimate school effectiveness. If appropriate, it will make recommendations as to how these estimates could be improved. There is also the potential to use the novel identification approach in future research aimed at estimating longer-term school impacts on, for example, university and labour market participation, by linking with other administrative datasets.

This project is part of our broader programme of work on schools, including projects exploring school effectiveness and school choice.