Modelling the school admissions system exposes educational inequalities but also highlights ways these can be addressed.


Project overview

Every September a new cohort of students move from primary to secondary school, however, in Bradford and across the country, many students are not offered a place at one of their preference schools. The aim of this project was to understand whether the school admissions system favours some groups more than others, producing inequality. It also stands as a case study in how data science techniques can be used by Local Authorities to inform policy and improve lives.

 

Data and methods

This project used data from Bradford Local Authority, detailing the school preferences and destinations of over 7,000 children allocated places at Bradford’s secondary schools in 2019.

Spatial Interaction Models (SIMs) were used to model flows of students from their homes to schools. This exposed the effects different selection criteria used by schools (including religious selection and the use of ‘fair banding’) had on student intakes, as well as the potential impacts of changing these.

Data visualisation and mapping were also key to communicate findings.

 

Key findings

Modelling and analysis of the school admission system over 2019 revealed significant inequalities in access to education across the Bradford District. One major issue identified was that the most advantaged students enjoyed almost certain access to a good school, but students from disadvantaged backgrounds did not.

As shown in the following graph, students from the most disadvantaged areas in Bradford gained places at schools with a range of Progress 8 scores (a key school performance measure), from exceptionally high to very low. Comparatively, almost all children from the most advantaged areas were offered a place at a school with an above average Progress 8 score.

Figure 1: Violin plot showing the distribution of school Progress 8 scores for the school a child attends compared with the Index of Multiple Deprivation given to the child’s home LSOA, Bradford District, 2019.

 

This can be linked to disparities in who gets their school preferences satisfied under the current admissions system as children in Bradford’s most deprived areas were around five times more likely to not receive any offers from their preferred schools than children from the least deprived areas. the following map shows that in some areas as high as 35% of children did not receive an offer from a preference school.

Figure 2: Percentage of students per LSOA not receiving an offer from any of their preference schools, Bradford District, 2019.

Figure 3: Percentage of students per LSOA receiving an offer from one of their top three preference schools, Bradford District, 2019.

 

The different selection criteria of schools can be linked, in part, to this inequality of uncertainty. Several schools in Bradford use a policy of ‘fair banding’ where all children applying to the school are asked to sit a test and a proportional number from each ability band (assigned by the test scores) are given places using random allocation. This leaves applicants with no assurance they will receive an offer at that school, and the areas surrounding these schools appear to have increased numbers of students not receiving their preference offers.

 

Value of the research

This research stands as example of how data modelling techniques can be used by local government to assess and inform policy. The findings of this project can be used to achieve greater equality within the school admissions system in Bradford and potentially across the UK.

 

“This was and is an excellent piece of work. It provides an immediate, practical benefit, by shedding light on an issue affecting disproportionately the most disadvantaged pupils in Bradford.

It also stands as an effective example of improving public understanding of science, showing how data science and different presentational tools can be used to improve understanding of a live issue, by policy makers, front line professionals and other non-scientists.”

Representative from Bradford Opportunity Area.

 

 

Insights

  • Inequalities exist in who gets their preferences satisfied by the school admissions system and who gets into ‘Good’ schools.
  • Some uncertainty can be addressed by changing schools’ selection criteria.
  • This project demonstrates how modelling and analytics can be used in the public sector to assess and inform fairer policies.

 

 

People

Holly Clarke, LIDA Intern

Ning Lu, Senior Research Fellow, Born in Bradford, University of Leeds

Mark Mon Williams, Professor of Cognitive Psychology, Leeds University, Lead for the Centre of Applied Education Research

Allison Heppenstall, Professor in Geocomputation, University of Leeds

Nick Malleson, Professor of Spatial Science University of Leeds

 

Partners & Funders

CAER and Bradford Opportunity Area