August, the month when children aren't at school, is when the English spend most time talking about education, as results of GCSEs, A-levels and Key Stage 2 Sats (tests for 11-year-olds) are published. Alongside the stories about the nerdy geniuses who get 15 A-levels, newspapers argue about whether standards are rising or falling. This year, the bungled administration of Sats by a US-based company has added an extra twist.
Otherwise, it is all as predictable, and as depressing, as bank holiday rain. What almost nobody mentions is that the whole school assessment system, the most elaborate in the known universe, is a colossal waste of time and money. It is, if you like, a turkey. Or a parrot that should be dead, but isn't.
The annual cost of Key Stage 2 Sats - I shall leave aside GCSEs and A-levels for now - is around £20m. That covers production, transportation and marking of papers. You can add the teacher's time spent preparing for and administering the tests, estimated at 165 hours, or more than one-eighth of the contracted working year. No proper assessment of the tests' value for money has ever been carried out. But research suggests they have failed to raise standards or adequately monitor progress.
The latest evidence comes from the centre-right think tank Civitas. A survey of secondary school teachers found the vast majority believe that Sats give an inaccurate picture (usually an overgenerous one) of the abilities of a third or more of the children, their results artificially inflated by coaching.
You may want to treat a Civitas report sceptically, but alternative sources support its claims. The independent Primary Review, led by Cambridge University academics, has looked at all the available evidence (see www.primaryreview.org.uk). After Sats for 11-year-olds started in 1995, standards rose sharply, though nothing like as dramatically as official figures initially suggested. Since 2000, the results have scarcely improved at all. A research survey for the review (which publishes its final report next year) concludes that "teachers learned very quickly how to coach for the tests . . . but any benefit to be squeezed from the system by such coaching has long since been exhausted".
So, for dubious outcomes - other work for the Primary Review suggests reading standards are no higher than 50 years ago - children, parents and teachers endure high levels of stress and subjects other than the "basics" are squeezed into smaller corners of the school day.
Moreover, teaching styles have become more didactic and narrowly focused, which, research on how children learn suggests, is exactly the opposite of what pupils need. The same happened in the US when "No child left behind" legislation imposed a similar regime on schools. The Primary Review survey concludes that "as test scores have risen, educational standards, broadly conceived, may actually have declined".
Attempts to broaden the curriculum and inject more creativity have achieved little success. Now Ed Balls, the Children's Secretary, has asked a former HMI, Jim Rose, to review the primary curriculum with a view to "reducing prescription where possible". Rose will consider "the development of the whole child" and "widening . . . opportunities for child-initiated and play-based activity". Such notions would have been thought soppy a few years ago, so it's progress of a sort. Yet Rose is not allowed to consider "changes to the current . . . testing regime", which, it is widely agreed, is the source of the problem.
What is the answer? The flaw in Sats (and in most other exams) is that they try to measure standards generally while also assessing the progress of individual children. A single test cannot do both. The first job can be done by annual sampling, using papers that are much the same each year. The second is best done by internal teacher assessment, which, according to popular wisdom, must be more unreliable than external tests.
Popular wisdom is wrong. As Bristol University's Wynne Harlen explains, in another report for the Primary Review, an external test is necessarily limited to a small number of items and is, therefore, a poor measure of any individual child. In Sats, Harlen estimates, one in three pupils (a proportion similar to that estimated by secondary teachers in the Civitas survey) is given the wrong "level". Teacher assessment may be unreliable for different reasons but, because it can cover a wider range of items over a longer period, it has the potential for improvement.
Most importantly, ministers should abolish school league tables. As I have explained here previously, they make good newspaper copy but convey almost no useful information. As evidence to the Primary Review shows, England's testing regime is unique. Even Wales, Scotland and Northern Ireland have nothing quite like it. If Balls dares to act, it would be a rare act of political courage, worthy of a future Labour leader.