Nicky Morgan is right. According to the Sunday Times she plans to scrap plans for tough new tests for primary school children. So she should, but not for the reasons that some teachers would give. She should do so because the testing system in state schools is a mess. Having decided not to make a bad system worse, she should now take the next step and make a bad system better.
Like all parents who have seen their children (in my case five of them) progress through state primary and comprehensive schools, I reckon I am an expert. As a pollster, I also know that most parents trust their local teachers’ judgements more than those of ministers, civil servants and the folk who run Local Education Authorities.
However, my argument is rooted not just in personal and polling experience but also in numbers and logic—and the realisation that our testing culture takes up too much time that should be spent helping kids to learn new things and engage their curiosity about the world around them.
School tests have two functions: to find out how pupils are doing—and to find out how schools are doing. They sound much the same, but they are not. In truth, by combining the two, our system does neither well. By its very nature, teachers have a huge incentive to teach to the test, by narrowing the focus of their classes, administering mock exams and so on. Most of us tend to accept all this, even if we grumble a bit. Parents (generally) want their daughters and sons to do well; teachers (generally) want their schools to do well. As a parent I was no exception, even though I knew the way to find out properly how my kids were doing was to talk to their teachers, not look at their SATs scores.
Were we designing a new system from scratch, I would restrict the purpose of primary school tests to measuring the performance of the schools, not the children. Instead of giving children long literacy and numeracy tests, with all the attendant coaching and mock exams, give them a short paper with (say) ten multiple choice questions on both subjects. This would generate 20 bits of information on each child. This is not nearly enough to grade an individual pupil; but in a class of 30, this would generate 600 answers—plenty to iron out any anomalies in the performance of particular kids, and tell us whether the teacher was any good. (Of course the calculation would need to measure value-added, taking account of each pupil’s past record, not crude scores.)
The key to this is that no data for individual children would be released. The tests would not be used to determine which class or set they would move to. Parents would then be in a position to insist—either directly or via parent-governors and the head teacher—that precious school time should be used for proper education, not pre-test training.
I hope Ms Morgan is thinking along these lines. According to the Sunday Times, she “will signal that measurement of the basics in literacy and numeracy at 11 is non-negotiable—but it may be by teacher assessment rather than by an external exam.”
The logic of that position is inescapable. Use external exams to measure schools, and keep teachers up to scratch—and then trust teachers to make the right judgements about our children. Precisely because the residual tests could be so short to administer, and so non-threatening to the pupils themselves, we would end up with more robust schools statistics, a better education, happier children and healthier teacher-parent relations. What’s not to like?
