Surrey junior school head Betty Kerr had to break some bad news to her Year 6 classes last week.
"They had just come in from the playground all hot and sweaty after the lunch break and I had to explain that we'd like them to do a maths test," said the head of Crawley Ridge School, Camberley. "They were a bit taken aback but I told them: ' When something isn't right you have to fight, and you are part of that battle'."
Mrs Kerr felt obliged to ask the children to take the test because she suspected that the key stage 2 maths test they had been given earlier in the term had prevented them from showing their true capability. She believed that too little time had been allocated, and that the questions were unduly complex.
Having seen the key stage 3 paper for levels 3-5, she was convinced that her children would score higher marks if they took the test that 14-year-olds had sat eight weeks earlier. And so it proved.
Most of the 67 children achieved the same level, but six pupils who had been awarded level 4 in the key stage 2 test moved up to level 5, and seven level 3 pupils achieved level 4. Only three moved down from 4 to 3, and one from 3 to 2.
Mrs Kerr accepts that her unofficial test results are open to criticism, too - there was no external moderation of the marking, for example - but feels that the exercise was worthwhile none the less.
"The children had been demoralised by the key stage 2 test but they felt very positive about the key stage 3 test," she said. "All of them felt that they had achieved. It was very similar to last year's key stage 2 pilot in format and presentation - the children were given an hour to answer 14 questions, compared with 35 minutes for 20 questions in the key stage 2 test."
Mrs Kerr has also returned English papers that she believes were wrongly marked, and is sending letters of complaint to the Department for Education and Employment and the School Curriculum and Assessment Authority.
"In fact, these results won't matter to the secondary schools in our area because they trust our assessments of children," she said. "But I am worried that these results will be used for a value-added exercise in the future. "
"How will such calculations be possible when we have tests like these and when SCAA keeps changing the content and marking criteria?"