The folly of random pay trials

Yes, by all means let's do as Andrew Oswald suggests (TES, August 10) and conduct a proper randomised controlled trial into performance-related pay (PRP).

To ensure that the effects are detected, we would need about 200 schools - 100 to implement PRP and 100 to act as controls. To make it a proper randomised trial, we would need to find 100 schools that were prepared to implement PRP whether they believe in it or not (otherwise it would not be randomised).

To make it a fair trial, we would need to give the same amount of extra money to the control schools, but they would have to give the same amount to everyone.

Then we have to agree how we measure performance. No one believes that results are the only outcome of schooling, so we would also need to devise measures of the wider benefits, such as willingness to engage in further learning, reduced delinquency etc. etc.

To give PRP time to influence the behaviour of teachers, and the achievement of pupils, the enhancement would need to be at least 20 per cent and we would need to run the experiment for about five years to affect teachers' behaviour and pupils' achievement.

The cost of the experiment, just for secondaries, would be around pound;160 million, and then the Government would probably ignore the results if they didn't come out right. On second thoughts, maybe we shouldn't bother...

Professor Dylan Wiliam Assistant principal King's College London Stamford Street, London SE1

Log in or register for FREE to continue reading.

It only takes a moment and you'll get access to more news, plus courses, jobs and teaching resources tailored to you