Is Ofqual's plan to end grade inflation a test too far?

Schools may struggle with extra exam to check validity of results

From 2017, amid a welter of changes to exams, thousands of GCSE pupils in England will also be faced with completely new national tests in English and maths.

The national reference test (NRT) should have no direct impact on the lives of the teenagers who take it, as individual results will not be published. But it promises to make a huge difference to teachers across the country.

If it works as planned, Ofqual's proposed new test will ensure that there is never a repeat of the 2012 GCSE grading controversy, where schools believed that thousands of pupils were denied the results they deserved because of a "statistical fix". Perhaps even more enticingly for teachers, the test should finally bring an end to the debate about "dumbing down" that takes place every time GCSE grades go up.

The NRT is meant to provide exam boards with an independent measure of the ability of a year group. If GCSE performance suggests that the proportion of pupils achieving good grades should increase, the results from the new test - to be sat by a representative sample of pupils - can be used to see whether that really is the case. This means that no one would be able to claim there had been a rise in the percentage of top grades because GCSEs had got easier. And Ofqual would be able to show that grading decisions were based on a current, independent measure of student ability rather than statistics drawn from primary school tests taken years before.

In theory it is a neat solution, which assessment expert and Ofqual external adviser Professor Robert Coe has described as "the way forward". But there are plenty of potential problems, not least that thousands of stressed GCSE candidates would be asked to take yet another test in the run-up to exams.

Glenys Stacey, Ofqual chief regulator, said the sample of pupils needed for the NRT would have to be sufficiently large to give reliable results, but would still only be a "small proportion".

"We are talking about thousands of students but not many thousands of students, and you have got a national cohort of 600,000," she said. "Most schools won't be bothered by this."

But some schools and their pupils will be. And the timing of the test is important, to ensure that it is a contemporaneous measure of the national cohort. "Those Year 11 [15- and 16-year-old] students that are selected will be taking that test shortly before their GCSEs so it is as close as possible to the real event," Ms Stacey conceded.

That is bound to raise fears about distracting pupils from crucial revision. Asked whether the NRT would be compulsory for those selected for the sample, Ms Stacey was non-committal. "I am not sure that in our culture we could insist on students taking it," she said. "But discussions are still to be had."

There is already some precedent for making sampling tests mandatory. The 2011 Education Act gave the secretary of state the power to force state schools to participate in international education surveys such as the Programme for International Student Assessment (Pisa) or the Trends in International Mathematics and Science Study (Timss).

Then there is the question of how to ensure that the NRT's content stays secret. The idea is that the assessment will remain the same each year so that it stands as a consistent measure over time. But with thousands taking the test annually, it might be difficult to stop details leaking even though there is no incentive for an individual student to do well. Ms Stacey did not go into detail but said that, unlike GCSEs, the test would not be sent out to schools to run. "There are tighter ways in which we would administer such tests," she added.

Ofqual has been able to look at the experience of other countries when confronting these issues, but this has uncovered more problems. "We know that in Hong Kong, for example, students sit the [reference] test but don't always put their hearts and minds into it," Ms Stacey said. "You know, why should they? You need a mechanism to identify when a student has been switched off rather than switched on. In Hong Kong we found they do that by referencing the reference test result with the individual student's actual result in the real qualification and if there is not a sufficient correlation then you take those out of your reference test."

But that solution might be seen as undermining the whole point of the NRT - to provide an independent reference point for GCSE grading. If results were then used to check the validity of student performance in the reference test, how independent a measure would the new test really be?

Reference tests around the world

  • England used "pre-tests" taken by children in representative samples of schools to help develop national curriculum tests.
  • Australia uses a common national assessment, the Australia Scaling Test, to enable comparisons between state-wide assessments.
  • In the Netherlands, reference examinations or questions have been used to set and maintain performance standards in national examinations.
  • Hong Kong uses reference tests to maintain standards in its diploma examinations.
  • Many states in the US also use "anchor" tests or questions to maintain the comparability of standards in their state-wide assessments.

    Log in or register for FREE to continue reading.

    It only takes a moment and you'll get access to more news, plus courses, jobs and teaching resources tailored to you