Your editorial on February 29 and the views of Ronnie Smith, to which you refer, are based on a misunderstanding of the research on both formative assessment and class-size reduction.

The research evidence on class-size reduction - from both the UK and the United States - shows that if you reduced class size from 30 to 20, you would get an extra one or two students passing a test. But if your classes are already down to 23, then you have already had most of that benefit. Reducing class-size further, to 20, will increase your costs by about 25 per cent once you’ve paid for all the new classrooms you’ll need - for a negligible increase in student achievement.

And formative assessment does not require detailed monitoring of every student as an individual - indeed, the research suggests that this is not particularly useful, even if it were practicable. Marking the work of students is hugely expensive of teachers’ time, and has little impact on student achievement. Far better for the teacher to spend time outside lessons planning good questions to use, so that the students’ thinking can be corrected before they leave the classroom, and with the whole class in one go, rather than one at a time, after they have gone away.

Put bluntly, time spent planning good teaching has a bigger impact on student achievement than the same amount of time spent marking.

At the moment, teachers don’t have enough time to work collaboratively to plan good lessons, which is why I suggested reducing teacher contact time to 15 hours, as a quid pro quo for larger classes. Classes of 23 (the average, I believe, in Scotland), with teachers teaching 22 hours per week cost the same as classes of 34 with teachers teaching 15 hours per week and require the same number of teachers. It’s just that the latter will produce higher student achievement, which is what I thought everyone wanted.

Dylan Wiliam, deputy director, Institute of Education, University of London.