For nearly 15 years you have carried out close to 1,200 meta-analyses on which factors are very influential on learning and which are not. Is it possible to apply your findings to all the schools regardless of any context and country?

Most of the research studies have come from the more developed countries so some care is needed when translating beyond this. It seems one rule of thumb is that the findings may be more generalised when the within school variance is much larger than the between-school variance – and in some parts of the world the between is much larger – and hence the school factors would become more important. For example, the variance between schools, based on the 2009 PISA results for reading across all Organization for Economic Cooperation and Development (OECD) countries, is 36 percent, and variance within schools is 64 percent. For Australia, it is 18 and 72 percent; Canada, 20 and 80 percent; Finland, 8 and 92 percent; New Zealand, 16 and 84 percent; the UK, 24 and 76 percent; Sweden, 9 and 91 percent; and the USA, 30 and 70 percent” (John Hattie, personal communication, July 24, 2016).