Grammar schools & stereotype threat

In the debate about grammar schools, everybody assumes that it is those who pass the 11+ who should be admitted to them. But why should this be? Mightn���t it be more efficient instead to admit those who do badly?

I ask because of some recent experiments by Victor Gonzales at Tilburg University. He got people to solve some Raven���s matrices. Then he randomly gave out medals and asked people to redo the tests. He found that low-ability people who got a medal improved substantially, whilst high-ability people who didn���t get one did just as well.

The reason for this lies in self-serving biases. Low ability people who got a prize thought they were good and this increased their ambition and self-belief, thus motivating them to do better. As Arsene Wenger said, ���If you do not believe you can do it then you have no chance at all���. However, high ability people disregarded the signal sent by the lack of a prize, believing they were good anyway.

I know, I know. At this stage many of you are shouting: this looks like the sort of priming result that is notoriously hard (pdf) to replicate.

However, I suspect it does have real-world relevance. Stereotype threat ��� our tendency to live up or down to stereotypes ��� is a real thing. As Rachel Kranton has shown, our (self)-identity shapes our behaviour, and this identity isn���t wholly exogenous but is at least in part shaped by our environment and by how others perceive us.

For example, girls in mixed schools are less likely to choose to study ���masculine��� subjects such as maths and science because being surrounding by boys reminds them more of their femininity and so prompts them to act girly. Jeffrey Butler has found that inequality leads to unwarranted perceptions of one���s ability. And Robert Oxoby has shown that low expectations ��� such as those formed by growing up in poverty ��� can reduce people���s ambitions and make them satisfied with less. People really can end up like a dog that���s been beat too much.

You might think this is a case for admitting 11+ failures into grammar schools. Low ability kids who get the prize of a place will raise their performance, whilst high-ability kids who fail won���t do worse: think of Will MacKenzie or Adrian Mole at their sink schools believing they were better than the rabble. Overall, academic performance will improve.

However, I don���t think this inference is warranted. This will only work if the low-ability kids are deceived into thinking they are good. This is probably not a trick that can be pulled off for long.

Perhaps a stronger inference would be that grammar schools are a bad idea because there���s danger that stigmatizing some people as being of low ability might prove to be self-fulfilling.