Formal testing is on the increase in secondary schools, it emerged this week, as one of the country's leading education academics warned of the "dangerous" unreliability of such assessment.
The Association of School and College Leaders (ASCL) told TES there had been a "big shift" towards the use of more internal formal exams for all year groups in secondaries.
The organisation said the move had been sparked by the government's GCSE and A-level reforms, which reduced or removed coursework and placed renewed emphasis on exams.
The ASCL said some schools used results from internal tests to decide which ability groups or sets pupils should be placed in. But respected education academic Dylan Wiliam warned that the practice could lead to half of pupils being put in the wrong groups.
A `big shift'
Cherry Ridgway, curriculum and assessment specialist at the ASCL, said: "A few years ago, there wouldn't be many schools where children did formal exams in the hall in every subject in every year. Schools are moving towards that, and that's quite a big shift.
"Heads are saying there will definitely be more formal exams in exam halls, and they're struggling to use it formatively. An increase in testing reduces time to learn from results."
Benfield School in Newcastle introduced formal end-of-year tests in core subjects for all pupils last summer. "We started doing it to prepare students for linear exams," headteacher Neil Walker said. "It's to help them develop the study skills required."
Ms Ridgway said that some school leaders planned to make greater use of pupils' test results to decide on ability sets in core subjects, believing this to be "logical" because of the fresh emphasis on exams in the reformed GCSEs.
But Professor Wiliam said that even if a school used a reliable test to place children in sets for maths, only 50 per cent of those children would be in the "right set".
"That's just a logical consequence of the less-than-perfect reliability and the less-than-perfect validity of assessments," he said. "That's why it's very dangerous to depend on a single measure."
Ms Ridgway said most schools had previously used a combination of tests and teachers' observations of classroom performance. She added that she shared Professor Wiliam's concerns about the potential implications of using tests to put children into sets.
"A lot of research evidence shows that students perform to teacher expectations, and a teacher teaches to the level of the set," she said. "That means it's even more important to make sure that if you're going to use sets, you get it right."
Professor Wiliam (pictured, right), of the UCL Institute of Education in London, cautioned against reliance on test results, saying that errors of measurement were "quite substantial" in GCSE papers.
He said that a GCSE psychology foundation tier paper's error measures led to two-thirds of entrants receiving a score within five marks of "what they should have got".
However, Professor Wiliam added, this meant that in a class of 30, one student would get a score either 10 marks higher or 10 marks lower than they should have received. "You won't know which kid, and you won't know whether the score they got was 10 higher or 10 lower," he said.
He added that the validity of a test could be limited by a range of factors, such as if a maths test were worded in a way that required a high level of literacy to be understood, so it effectively tested students' literacy rather than their mathematical ability.
Professor Wiliam said exam results would become more reliable if schools spent more time testing students, but that this could have a negative effect on teaching.
Ofqual is responsible for ensuring that qualifications are reliable. But the regulator declined to comment.
`You'd be foolish not to give more tests'
Fullbrook School in Surrey is planning to introduce extra practice exams for Year 10 students and is considering extending these to Year 9 pupils to help them prepare for the reformed GCSEs.
"If pupils are going to have to sit more externally set tests, we have to give them more practice at that," says principal Anne Turner, pictured. "You'd be foolish not to."
She says the results will provide more information about performance and progress, likely to be taken into account when setting pupils.
"We might have a different set of data on which to decide sets," Ms Turner adds.