Methodology

National pupil projections

Published

Introduction

This document provides background information on the ‘National pupil projections – future trends in pupil numbers’ release. It explains the concepts and methods used to calculate the figures and provides an overview of the data sources and other relevant information.

For 2021 it also explains differences in the figures being released this year, as a result of the covid pandemic, and the consequent limitations of the dataset. 

Projection methods

The pupil projection model creates national projections for the number of pupils in England by age and by school phases and types. It uses ONS population estimates and projections, information on birth registrations in England and data derived from the School Census. 

Each model uses the most recent ONS population projections, based on the estimated population at the appropriate mid-year, as well as ONS’s monthly birth figures. The population estimates and projections are modified from their original mid-year basis to simulate the definitions used in the School Census (a January count with pupils’ ages as at the previous 31 August).

Census data, normally up to and including the model year is used to create participation rates, calculated as a series of historic census totals measured against population estimates. These provide a starting point for the model, with the school census data informing assumptions about future proportions of the school population by type of school. 

Separate figures, both historically and into the future, are calculated for part time and full time pupils.

ONS’s population projections are based on their most recent estimate of the resident population. The population includes all usually resident persons, whatever their nationality. Long-term migrants are defined as individuals who change their country of usual residence for a period of at least a year, and in such cases their country of destination becomes the country of usual residence. Pupils who are short-term migrants, and are staying in England for less than a year are included in the School Census and are therefore reflected in the pupil projections. They are not separately identifiable.

Model frequency

These pupil projections are produced annually. However, they are based on ONS’s national population projections, which are produced once every two years. 

In the years where there are no new ONS projections then the pupil projections are an update of the previous year’s projection, with an added year of actual data, rather than a new projection in their own right. 

2021 reporting year

Pupil counts for the historic years are derived from data collected via the School Census. 

This year 2021 school census data would normally be added into the model. However, the dataset shows unusual drops in the pupil numbers in certain school types, most significantly nursery & primary and alternative provision schools. The 2021 nursery & primary school population is 34,000 lower than was forecast, a notably higher difference than usual. The alternative provision population for 2021 is 24% lower than had been forecast.

These drops are believed to be temporary, as a result of the pandemic, rather than long-term changes. If they were used in the projection model they would result in population decreases across future years which are not considered to be realistic estimates of the pupil population over the next ten years.

Therefore it has been decided not to include the 2021 school census data in this update of the pupil projection model. Thus 2021 is still a projected year and the model, having not been scrolled forwards, is still a 2020 model, albeit with additional actual 2019 data incorporated.     

Note these census figures will need to be included, alongside those from 2022, in the new pupil projection model produced next year. However, the 2022 school census results will show whether the drops seen this year were indeed temporary. The methodologies for projecting into the future can then, if required, be adjusted accordingly.

Key drivers to change in the projections

The school census data is measured against the ONS population figures to create participation rates (both overall and for the different school types) for each historic year. The participation rates from the most recent years are then used to calculate the pupil numbers going forward. Thus if the school census figures are inaccurate they can affect not only the year in question but the whole of the projection model.

As these pupil projections are based on ONS’s national population projections, alongside participation rates, they are directly affected by the assumptions made by ONS. Changes in the population who are of school age is largely driven by changes in the birth rate rather than direct immigration. However, that birth rate is in turn affected by any increase in the number of children born to non-UK born women (compared to those born to UK-born women). The number of children born to non-UK born women increased by around 75 per cent between 2002 and 2013 (the years in which many children currently in schools were born), although this was a period of increased births generally.

After a decade of increases the total number of births in England dropped in 2013, remained stable for the next three years, then have dropped again. ONS’s short-term birth assumptions in their national population projections will have been reduced as a result of these changes. 

Alternative scenarios

The ONS national population projections, which are the base for this projection of future pupil numbers, are based on assumptions which best reflect demographic patterns at the time they were selected.

However, there are inherent uncertainties in projecting the future size of the pupil population. As such ONS also produce a number of alternative – known as variant – projections based on different assumptions of future fertility, migration and life expectancy. 

As there is no new information added to the model that affects the variants, these have not been rerun for this year.

Note: the ONS are currently consulting on the content of their mid-2020 national population projections, which are the last projections being produced before the results of the 2021 national census are made available. They are currently not intending to produce any variant projections. If that proves to be the case then the department’s 2022 pupil projections, due to be produced using these new national population projections, will again not include variant projection results. 

Relationship with school capacity publication

This release presents the Department’s main national-level pupil projections. The department’s data on school capacity is produced by a different methodology and includes local authorities’ own forecasts of future pupil numbers, based on local-level information, such as inter-authority movement of pupils. It provides forecasts by national curriculum year group whereas these national pupil projections forecast by actual age.

The school capacity data is published annually as an official statistical release, which can be found here.

The national pupil projections presented in the current release are produced within the Department, at national level only. They cover a longer time period than in the school capacity release. Differences in the methodologies used by individual local authorities for pupil number forecasts mean that the aggregated totals in the school capacity release will differ from the national-level projections presented here, but the trends in both are broadly similar.

Accuracy

In any year, the projected figures are subject to increasing margins of error the further into the future they run. There is also a greater level of uncertainty for smaller and more detailed categories, such as particular age bands or smaller categories of school. Factors which may contribute to differences between projections and actual data include: 

  • the underlying population trend, which depends on assumptions about births, mortality and migration;
  • differences between short-term assumptions in the ONS population projections, used in one model, and subsequent population estimate data incorporated into the next year’s model;
  • participation for all pupils, but particularly amongst pupils aged under 5 and over 15, which may differ from the rates assumed;
  • school reorganisation by local authorities, which may affect the balance between primary and secondary schools;
  • variations in the proportions attending different school types, such as independent schools;
  • the introduction of new Government policies affecting schools or population.

Pupil projections rely heavily on the population projections produced by ONS which are widely used across government for planning purposes. National population projections are National Statistics, which means they undergo regular quality assurance reviews and are produced free from political influence. The projections are produced using demographic assumptions about future levels of fertility, life-expectancy and net-migration, based on analysis of trends and expert advice. The assumptions are agreed in liaison with the devolved administrations, following consultation with key users of projections in each country and advice from an expert academic advisory panel.

Change from early projections

As for the last few years, this release does not include projections for pupils aged 16 and over. The projections prior to 2013 were aligned to overall 16-18 participation rate projections produced for the Education Funding Agency 16-19 Funding Statement. This has been replaced with a much simpler funding letter to 16-19 education and training providers which does not include projections that are comparable with the pre-16 projections in this publication. This is a permanent change.

Consistent with the July 2014 onwards publications, this release takes out dual-registered pupils from schools classified as alternative provision settings. This removes double counting in pupil numbers from this school type, which was present in previous statistical releases. For this reason, the alternative provision data from 2002 to 2010 has been removed.

Definitions

Date of count

  • Figures relate to January of the year shown.

School types

  • State-funded schools – a total of the mainained nursery, primary, secondary, special schools and alternative provision settings.
  • State-funded primary schools – primary and middle deemed primary schools
  • State-funded secondary schools – secondary, middle deemed secondary and all through schools, including city technology colleges, studio schools and and university technical colleges.
  • State-funded special schools – includes all special schools apart from non-maintained special schools and general hospital schools
  • Non-maintained special schools – constitute a separate category of school which are not included in the state funded total
  • Alternative provision settings – includes pupil preferral units and alternative provision academies and free schools, but not other types of alternative provision
  • Independent schools – includes all schools that are not state funded, ,except non-maintained special schools.

Age

  • Figures calculated in January, age as at previous 31st August.
  • Under 5 – ages up to and including age 4
  • Primary ages – 5 to 10
  • Secondary ages – 10 to 15

Part-time and full-time equivalent

  • Part-time pupils assumed to represent half of a full-time pupil to give full-time equivalent (FTE) totals.
  • From 2003 to 2011 state-funded schools could record part-time pupils at any age.  From 2012 pupils aged under 5 or over 15 can be recorded by state-funded schools as part-time, but all pupils aged 5 – 15 are recorded as full-time.
  • Independent schools can record pupils of any age as part-time.