National and local averages show only a small discrepancy between test results and teachers' own assessments, but this conceals a diverse picture for individual schools. Some show twice as many pupils achieving level 4 in the teacher assessment as in the test; others did far better in the tests. Nationally, teachers' own assessments were slightly higher than test results: 2.8 per cent more were rated level 4 in English, 5.8 per cent more in maths and 2.9 per cent more in science.
While there is no reason for these scores to be identical - since they do not test all the same things - big differences suggest something is amiss.
Junior teachers did not have as much training in assessment as their infant counterparts, but inspectors say teachers' accuracy in assessing their pupils is improving as they become used to the process. However, some were still simply "guessing". In some cases, tests may not have been marked correctly by external markers.
In one school's case, the Department for Education and Employment made a mistake in the table. The governing body of Sundon primary in Luton, Bedfordshire, which had decided to withhold its teacher assessment results, was very surprised to see them listed in the national table at 96, 97 and 99 per cent respectively for English, maths and science. "I really can't think where they got the data from," said deputy head Diane Beard. The figures were incorrect and diminished the impact of the school's high-in-the-league test results of 63 per cent reaching level 4 or above in English, 60 in maths and 74 in science.
In two Lewisham schools, teacher assessment showed up to twice as many children achieving level 4 as test results in some subjects. However, the table shows high levels of absence during the tests or of children being "disapplied" because of special needs.
Meanwhile, Leicestershire Year 6 teacher Calvin Davies of Westfield county primary in Hinckley, found himself explaining why his school's teacher assessment percentages were well below the test results. Essentially, he said, many of the children were near the boundary between levels 3 and 4. There were only a few points difference between the numerical scores, but this had made the difference. He said the school was amending the children's transfer documents for secondary school to show if they were early level 4, middle level 4 or top level 4. He said speaking and listening, not included in the tests, often lowered teacher assessment results by a level. But he had done them early in the year, and the children had later improved.
In future, including this year, he would assess oracy later in the year.
Overall, Sir Ron Dearing's famous pledge that national curriculum tests and teachers' own assessments would have equal value rings hollow this week. The TES is the only national newspaper to publish the teacher assessment results, which have otherwise formed no part of the national debate.
However, their dutiful publication by the DFEE alongside the test scores means that they will be available to parents and schools locally, and that they have at least a token importance.
The teacher assessment covers a broader range of attainments than the tests, including the hard to pin down areas like speaking and listening, exploration of science and using and applying maths. It represents a teacher's judgment of a pupil's work over time, rather than a snapshot of what they can do on one summer afternoon under test conditions.