You state that bilingual programmes have an effect size of 0.36, so not so significant. What do you define as bilingual progammes and where, around the world, was this data taken from? Why do you think the effect size is so low?

First, do not think of the average of .036 as ‘significant” or not – This 0.36 means students in bilingual programs outperform those in other programs 0.36 of a standard deviation – which I would have thought was most worthwhile. But the devil is in the details and happy to send any of these meta-analyses” (John Hattie, personal communication, December 22, 2018).