Encouraging pupils to question their teachers on what they do and do not understand about a subject is the single most effective way of improving education.
This finding comes from what is believed to be the largest ever overview of education research, involving a synthesis of more than 50,000 studies covering all aspects of schooling across the English-speaking world.
The research suggests that raising the quality of pupil-teacher interactions is key to improving education systems, which makes debates about school type, school-by-school performance and class size, among others, look irrelevant.
These findings are being taken seriously by the Department for Children, Schools and Families, whose officials were presented with them over the summer.
The research also supports backers of Assessment for Learning, the feedback-based teaching system - although the academic behind the study has joined other observers in arguing that the UK government has "hijacked" the term.
The research, published this week, was carried out by Professor John Hattie, of Auckland University, who has spent the past 15 years analysing education research.
As a result of his research, he has developed a new assessment system, called Visible Learning, which is designed to improve teacher-pupil interactions.
Professor Hattie's work has involved synthesising some 815 meta-analyses of education research, each mainly relating to developed English-speaking countries such as the US, the UK and Australia. These cover at least 83 million pupils over the period 1976 to 2007.
His research has enabled him to create a league table of the most effective ways to raise achievement, out of 138 possible approaches.
The upper rankings of the table are dominated by programmes designed to improve the quality of pupil-teacher interactions, suggesting that transforming this aspects of education is the key.
The table's top-rated approach is pupils assessing themselves: getting children to reach a view on their levels of understanding and feeding this back to their teachers.
This suggests that there is a very high correlation between the levels of progress pupils believe they have made and their actual performance in tests.
Professor Hattie said this undermined the belief that pupils needed constant external assessment.
The second most effective approach is using "Piagetian programs" to raise achievement. This method is founded on the work of Jean Piaget, the Swiss psychologist who believed that children go through successive phases of development. Piagetian programs aim to set pupils work that is one step ahead of their current level.
Other approaches that feature in the top 10 include: formative assessment schemes - using assessment to decide pupils' next steps in learning; "teacher clarity" - teachers being explicit in what they want their charges to do; "reciprocal teaching" - pupils taking turns to teach the class; and "feedback" from pupils to teachers and vice versa.
By contrast, some aspects of education over which there is much debate appear to be much less relevant to children's achievement than many might think.
Of the top 30 most effective interventions, only two - providing accelerated education for pupils who might benefit from it, and managing behaviour - could be attributed to school management, despite much policymaking energy being directed at struggling institutions, Professor Hattie said.
"In most western countries, take two students of the same ability, and it matters not which school they attend," he said.
Reducing class sizes, the research found, was towards the bottom of the rankings in terms of effectiveness, at 106th.
Professor Hattie's research found that while in theory cutting class sizes could be beneficial, this would only be the case if teachers changed their teaching strategies, which tended not to happen.
Other relatively unimportant influences on performance included setting pupils by ability; faith schools; US charter schools, which are similar to England's academies; frequent testing; and pupils' gender.
Transferring children between schools and policies that hold low- performers back a year were found to be damaging.
Studies of the effect of television on learning also suggested it had a small negative effect.
Professor Hattie told The TES that teachers needed to use information they gained on pupils' understanding to set them more challenging work.
They also needed to allow pupils the freedom to make mistakes, as errors were powerful learning tools.
"A teachers' job is not to make work easy. It is to make it difficult. If you are not challenged, you do not make mistakes. If you do not make mistakes, feedback is useless."
- `Visible Learning: A Synthesis of Over 800 Meta-Analyses Relating to Achievement' is available at Pounds 24.99 (paperback) and Pounds 75 (hardback) from Routledge.