A nationwide survey of 107 secondary school teachers has found that just 10 per cent this last school year found the levels achieved by their year seven pupil levels to be consistent with their scores in their key stage two SATs tests which they take in their final year in primary school, year six.

Nearly 80 per cent of those English, maths and science teachers surveyed found that up to a third of their pupils' true abilities were lower than their SATs results suggested.

The main source of this "inflation" of the key stage two results was reported by teachers as being "teaching to the test" – essentially, cramming. This conclusion is supported by previous independent and international research showing that rises in key stage two SATs test results have not necessarily indicated a rise in pupil standards. Rather, it indicates better preparation for passing tests.

Aside from the money wasted on testing which doesn't provide an accurate picture of abilities, the flaws in key stage two testing are having highly problematic repercussions for pupils and teachers alike. Primary school teachers are finding themselves compelled to narrow the curriculum to boost results, lowering not just pupil morale, but also their own.

Secondary school teachers are having to patch up the holes left in pupils' learning, as well as facing accusations of not having made enough progress with pupils. As one of the science teachers surveyed commented: "We do baseline testing so that we can show what we have done with them – by using the key stage two results it would look like we hadn't made any progress."

However, pupils are the biggest losers in the equation, not only bombarded with test pressure, but most worryingly, losing out on a rounded education – not just subject-wise, but actually within English, maths and science.

The question is what is leading to this teaching to the test, or "coaching", and most importantly, how can it be stopped?

The root problem is the way in which testing has become "appropriated" by the Government for its own purpose: demonstrating a rise in standards often by generating one artificially through the testing process, as their policies have not succeeded. This is where cramming creeps in, both through official guidance and unofficial pressure on teachers to do whatever they can to ratchet up results. The testing process has resulted in a scenario in which the outcome is more important than how it is achieved. Subsequently, higher primary school test scores have, particularly over the last decade alongside ambitious targets, often represented less learning and worse educated pupils.

Something which the Civitas survey of secondary school teachers highlights is that coaching is contributing to a "pile-up"of inflated pupil levelling – a cumulative effect which will frequently have started early on in the key stage one testing of seven-year-olds and continuously grown through key stage two and into secondary school.

Aside from the difficulties which this presents for value-added based evaluations, the danger is that pupils complete their school careers with considerable gaps in their learning. This has contributed significantly to a concurrent rise in national test and exam grades and a need for universities and employers to re-test school-leavers.

The answer to the SATs problem, however, is not scrapping primary testing. Discussion around coaching for the tests in primary schools often leads to the conclusion that the root of the problem is testing per se. The evidence on what has gone wrong in testing strongly suggests that this is an erroneous position. Testing itself is not the problem. Testing can be stimulating for pupils and useful in terms of measuring how effective teaching and school policies are; if testing is used effectively, it can indeed be a valuable accountability tool.

The solution, instead, is testing which gauges a truly randomised snapshot of learning, rather than the testing happening today, whereby the sum of learning all too often becomes only that snapshot.

Annual unseen testing carried out at any point in upper primary school (between years three and six), without forewarning on content and timing, would achieve this.

This would provide a more accurate picture of learning levels and progress in a school. In short, the solution is testing which tests learning, rather than test preparation.


Anastasia de Waal is head of family and education at the Civitas think tank.