Put a child of cleaner from Shanghai or Singapore up against the scion of the western elite in a standardised test and guess who will come out top? According to the latest research, the western kids will trail their Asian counterparts by the equivilant of a whole school year. This prompted another bout of anxiety of a kind that has become increasingly common since 2001, when the global Pisa survey of educational attainment was first published. Parents once drew comfort from steady improvements in school-leaving grades in places such as the UK. Confronted with evidence of how their children’s accomplishments compared to those of students in faraway places, many westerners have taken fright.

Next week Elisabeth Truss, a British education minister, will lead a fact-finding mission to Shanghai to try to find out what the schools there are doing right. Yet in their rush to copy the winning formula of high-performing countries in east Asia, politicians risk drawing the wrong conclusions. Schools in Shanghai are very different from those in Ms Truss’s constituency in southwest Norfolk. But not all of those differences play a role in Shanghai’s superior performance. Some are irrelevant. Some may even be harmful. And some will be idiosyncratic features of the school she happens to visit, rather than representative of the system. It is easy to point out how a good school differs from a bad one, and conclude that you have found the secret to high achievement – but it is also lazy, unscientific and wrong.

East Asian countries have the highest average test scores, but also some of the most uneven results. The gap between their best and worst pupils is larger than anywhere else. Should British teachers try to increase the gap between the best- and worst-performing students, in the hope of dragging up the average? No. There is no evidence that the inequality of outcomes in east Asian schools contributes to raising standards. The west should not copy its rivals methods before it knows which ones are responsible for their success.