Most education initiatives do not impress Julius Lang - but he said assessment for learning had taken him aback. "I'm still amazed by it," he said. "I have seen so many things happen in education, but I really believe in this one."
Mr Lang is among the first of a new breed: a chartered educational assessor. For four months he has been part of a pilot project working with schools to improve their day-to-day assessment practice. Teachers he has worked with appear to share his enthusiasm.
Mr Lang, a former secondary head of English who is now a senior examiner for the AQA board, was one of 34 testing specialists who applied and were granted the chartered status by a new professional body, the Chartered Institute of Educational Assessors. The idea is for them to start spreading their knowledge and enthusiasm. Mr Lang is working with Maundene School, a successful primary in Chatham, Kent.
On arriving in the school in December, he began doing an audit of the assessment information held on each child. He said: "There was shedloads of it. They had a programme, for example, which could colour-code each child into green, amber and red, according to whether they were reaching their target for the year." Yet he wanted to know whether the data was always accurate, and if this high-performing school was getting as much progress out of each of its pupils as it could. Some of the test information, he said, was clearly wrong.
He asked if one boy, who had achieved a level 3 in a maths test in Year 4, was truly that good, and his teachers said: "Of course not. He can barely add up".
Mr Lang then worked with the teachers to help them come up with assessment tools which could be designed to provide more reliable information on pupil progress and then be used immediately to inform teachers.
The school introduced them first for writing in Year 1 and maths in Year 4. In both subjects, pupils were typically set tests at the end of six week units of work.
Now, in Year 4 maths, teachers Sarah Dean and Claire Procter conduct daily assessments in class.
These involve sitting down with groups of children, asking them individual questions and then recording what each pupil does and does not know. The teachers worked with Mr Lang to design a checklist to record each child's progress.
Ms Proctor said: "We used to do a lot of summative assessment - end-of- unit tests, which put a lot of pressure on the children. The pupils did not show their true potential. They would have worked on most of the content of the test, but remembering it all at once is tough. This way, we assess them more often through single questions, such as, `can you tell me what the graph means?'
"The children like it, because they do not think about their levels and the stress of taking a test."
In writing, Helen Williams and Elizabeth Wintle give their pupils an assessment every two to three weeks. Mr Lang said: "The teachers had proof that they had improved writing in those few weeks. Some children had gained dramatically. What came out was that the way you assess improves teaching."
The negative side of the pilot appears to have been teacher workload, a factor in the project's seemingly high attrition rate. Of the 28 schools that started working with the chartered assessors last year, six have dropped out. Yet teachers at Maundene have not been put off. Nigel Jones, the deputy head, said the trial had not been as hard to set up as he had expected, and Ms Wintle said assessments in class were straightforward.
A final verdict on the trial, including its effect on pupil progress, will have to wait at least until June, when an independent evaluation is published.
But Graham Herbert, who is leading the project for the institute, said that report was likely to be positive. A similar trial will also begin at colleges in June, with another one to follow at secondary schools in the autumn.
The Chartered Institute of Educational Assessors offers copious support materials on assessment, including advice on how to write questions to test your pupils.