Melanie Butler looks at evidence-based language policies

Countries around the world have spent many years trying to improve the English language levels of their children. With the emergence of Pisa tests of English, the demand for effective language policies will increase. So, what might they do, and will it actually work?

Perhaps the most commonly adopted measure is starting English earlier, but children in many countries are already starting in year one.

“In Asia, the favourite solution is sticking a native speaker graduate in 
each state school. Japan has been doing this for 40 years with seemingly 
little impact.”

More importantly, there is evidence that when it comes to results on formal language tests, starting younger may not be better. In 2016, a large German cohort study followed two groups of children, one which started English at 8 years-old, the other at 11. While the early starters got higher test scores at age 12, by 16 the late starters were ahead.

The problem may lie with the tests. Early learners, like native speakers, tend to be better at recognising grammaticality and using collocations, but do less well in tests of overt knowledge. One reason, perhaps, that the Pisa ranking of the UK actually goes down when you add in our scores on EFL tests.

In Asia, the favourite solution is sticking a native- speaker graduate in each state school. Japan has been doing this for 40 years with seemingly little impact. The Korean government became so disillusioned with the lack of progress from its own scheme that it started to cut it, until angry parents demanded it back.

A more successful model was used in Uruguay’s plan Ceibal. Senior schoolteachers with a specialism in English, often non- natives, appeared in the class online – working with the classroom teacher. US company Eleutian has successfully used a similar approach in Korea and China.

In Europe, there is a long tradition of sending children on school trips to language schools in English-speaking countries. It’s not clear how much language they actually learn in one week, though language anxiety has been shown to drop dramatically, which in turn is shown to improve speaking scores.

If there is one thing that Anglosphere schools are really good at, however, it is test preparation – they routinely get Ielts results up half a band score in two weeks. For best results, though, they need to know the task type a test uses, so short, test-oriented school trips will work best once we know what the Pisa test actually looks like.

And finally, in desperation, countries will turn to the British Council and beg for mass teacher training in communicative methodology.

Again, the evidence suggests this will not always work. Two large research assessment exercises, one from the Welsh government, the other from the University of Oxford, found that the language level of the teacher had much more impact on student outcomes than any methodology. If the teacher does not speak the language to a good level, nobody learns anything.

That’s the reason why Xavier Gisbert, mastermind of Madrid’s successful Clil programme, insisted that all teachers must speak the target language to C1 and paid for them to take classes in local language schools and to attend summer schools in the Anglosphere. The level of the student will never exceed that of the teachers, he tells me.

So, if you want to improve the language level of your students, improve the language level of your teachers.