National reference tests: an overview

The first reference tests in English and maths will take place in March 2017 and in future years will be used when awarding GCSEs.

We are introducing the new national reference tests to provide additional information to support the awarding of GCSEs. The purpose of the tests is to provide evidence on changes in performance standards over time in English language and maths in England at the end of year 11. The tests should allow us to see, over several years, if there is a change in how students perform at the national level, which we will consider carefully in our awarding.

The tests will provide additional information that we expect will significantly advance our ability to detect genuine changes in performance and so improve awarding. We will act cautiously as we build our understanding of the information that the tests will provide and how this is used in awarding GCSEs.

Test design

There will be separate tests for maths and English. They will be based on the new content for GCSEs in maths and English language, which students will take for the first time in summer 2017.

The questions may not be in exactly the same style as the new GCSE questions, but they will be suitable for students who are studying for the GCSE. The questions used in the tests will not reflect any particular exam board’s style of GCSE questions. We will publish some example questions later this year.

Each test is divided into several booklets. Each student takes just one booklet, so no student has to take the whole test.

The reference test for maths, unlike the GCSE, will not be tiered. Most of the questions will be based on the GCSE content that all students taking the qualification have to study. There will be a few questions at the end of each booklet that will be based on the content only studied by those preparing to take the higher tier GCSE, but these will be clearly marked.

The reference test for English will include questions that require students to write longer responses as well others that can be answered with short, one sentence responses. There is no testing of spoken language in the English test.

The questions will remain largely the same from year to year to enable performance standards to be compared. Therefore, the questions used in the tests have to remain confidential and will not be published, and we will not be able to provide individual feedback to students.

Timetable

The new tests will be developed over the next 18 months and the first national reference tests will be taken in March 2017. In future years the information from the tests will be taken into account when GCSEs are awarded.

In September and October 2015, we will trial questions that have been developed for the new tests. In March 2016, we will hold a preliminary reference test. This will involve the same number of schools and students as future tests but the purpose of the tests in 2016 is to confirm that all the operational arrangements are working well. The preliminary test is a very important step that will allow us to introduce the reference test smoothly in 2017.

Who takes the test?

Each year, a random sample of 30 students from around 300 schools will be asked to take a test booklet in maths or English. The students will be in year 11 and will take the test in early March, before they take their GCSEs (in late May and early June). The test will take each student around an hour to complete. The results will be analysed only at the national level; there will be no results for individual schools or students.

We would expect all randomly selected pupils to take the test. This will provide the most statistically sound sample. However, headteachers will have the option to exclude pupils from the testing - at their discretion and in exceptional circumstances. This option would be used for a specific reason such as accessibility or situations where taking the test may cause undue distress for the pupil. NFER, which is administering the tests on our behalf, will make the test as student-friendly as possible and allow all reasonable and appropriate access arrangements so it is hoped that such exclusions will be needed only in rare cases.

NFER, will select the schools to take part each year. We will use a different sample of schools each year. The selected schools will provide a list of all their students in year 11 that they expect will take the GCSE (in maths or English language). From this list NFER will randomly select which students will be asked to take the test.

How will it be administered in schools?

NFER is working with us to develop the tests. In order to minimise the burden on participating schools, a test administrator from NFER will come to the school on a pre-agreed day and carry out the tests. The school will need to provide a room where the students can take the test and NFER’s administrator will invigilate it. The test will last around an hour for each student.

What results will be published and when?

Towards the end of August each year we will publish the national measures of performance in the tests, around when exam boards publish GCSE results. We will also explain how performance in the national reference tests has been taken into account in awarding the GCSEs.

The results of the reference tests will set out the percentage of students in that year who are predicted, based on the tests, to achieve at least a grade 4, a grade 5 or a grade 7 in their GCSE. The tests will show if these percentages change from year to year. And it is these changes, if any, that will be taken into account when GCSEs are being awarded. We are developing the details of what we will publish and will confirm this later in the year. The first time that we will able to compare performance between years will be in 2018.

There will be no results for individual students or schools. So, for example, no result can be recorded in students’ school records. The only information will be for England as a whole.