MUCH AS it goes against the grain, I have to admit that very occasionally Chris Woodhead, the chief inspector, gets it right. He is absolutely right to insist that the national curriculum tests are unreliable as a basis for making year-on-year comparisons and for assessing progress against national targets.
At the same time Chris Whetton of the National Foundation for Education Research and Nick Tate of the Qualifications and Curriculum Authority, defending the tests (TES, January 8), are also right in many of their criticisms of the chief inspector's views. How is that possible?
All three of them fail to distinguish between using national tests to provide measures of pupils' performance across a wide range of curriculum-related criteria in order to make comparisons over time and using nationally standardised tests to measure specific aspects of pupils' performance at a particular time.
The only way to use national tests reliably to provide year-on-year comparisons is to do exactly what Whetton, Tate and Woodhead do not want to do - use exactly the same tests administered in exactly the same way successively over a period of years. As Whetton points out, to increase the reliability of the tests they could have more questions and cover more ground so as to encompass more elements of the programmes of study than do the current tests.
This need not lead to the overload on pupils which he implies. For the purpose of national year-on-year comparisons (but not of individual school accountability) the tests could be divided into sub-tests and each sub-test administered each year confidentially to a nationally-representative sample of pupils. In that way progress against targets nationally could be assessed far more reliably.
To complement such testing, short standardised tests focusing only on key elements (eg David Blunkett's "reading, writing and adding up" plus, presumably "taking away", "timesing" and "sharing") could be devised and administered each year to as many pupils as the Government, schools or parents wanted. These standardised scores could provide parents with information about individuals' performance in the so-called basics.
Yearly aggregated profiles of such scores could be used by central government or the Office for Standards in Education for accountability purposes for as long as the mutual distrust between officialdom and the teaching profession persists.
Professor Colin Richards, 1 Bobbin Mill, Spark Bridge, Cumbria