Imagine what a difference it would make if children knew what they were good at and what they had to do to improve.’ With these stirring words, our trainer extolled the benefits of feedback. Teachers should make sure students were absolutely clear about where their work needed further crafting and provide them with time to do this. It was, and still is, a compelling vision. So what went wrong?

What went wrong was that feedback was interpreted as meaning marking. And not just any old marking – dialogic marking. Fast forward 10 years and schools were marking within an inch of their lives. The toll on teachers was terrible; there just weren’t enough hours in the day to get the marking done. Yet this toil had dubious actual impact on learning (Elliott et al., 2016).

So it was a surprise when I discovered that Dylan Wiliam, prime advocate of ‘feedback’, has little to say about marking per se, let alone labour-intensive dialogic marking. He describes feedback as anything that lets the student or the teacher know how well the learning is going. It might be something as fleeting as student facial expressions (2016). The term feedback, Wiliam explains, is borrowed from the engineering expression ‘feedback loop’ (2011). For example, a thermostat regulates temperature by measuring the current temperature, comparing this with the desired temperature and then doing something (activating a heating or cooling mechanism) to bring the current state in line with the desired state. The important part in all this is the response to the alert of a discrepancy. Without a mechanism to close the gap between the current and the desired state, feedback about the discrepancy is useless. In an educational context, the same applies. Feedback needs to be not only accurate in its diagnosis of what is wrong, but also helpful in enabling the learner to put it right.

Closing the feedback loop

Wiliam’s work on feedback has been interpreted as involving written comments intended to give the learner information on how to improve their work. Yet often the gap between where the learner is now and where the teacher wants the learner to be is too big to be bridged by a single comment. Wiliam uses the example of a student commenting on his science assignment. The teacher had written: ‘You need to be more systematic in planning your scientific inquiries.’ The student retorted: ‘If I knew how to be more systematic, I would have been more systematic the first time’ (Wiliam, 2011). The written comment, intended to close that feedback loop, is nothing more than a diagnosis of the problem. It’s like a red light on the central heating; it tells us something is wrong, it might even identify what the fault is, but it does not miraculously give us the knowledge of how to fix the problem. For that we need more teaching. The next step for this science student was another lesson, focused on what he couldn’t yet do. Feedback needs to effect change and cannot be reduced to a simple formula.

Another problem is that sometimes marking is so easy to action that the learner doesn’t have to think at all. As long as they mindlessly follow the teacher’s instruction, then the work will ‘improve’. Never mind that the student hasn’t learnt anything new so the ‘fault’ is likely to recur. It’s not just about closing the feedback loop, it is about closing it so that it stays closed. If we identify every missing full stop, every place value error for students, then we shouldn’t be surprised if they carry on omitting full stops and putting digits in the wrong columns.

Marking policy and effective feedback

Feedback is a powerful way of improving learning, but it has to be used thoughtfully. Our previous marking policy conflated feedback with marking and outlined a one-size-fits-all procedure intended to fit every subject. It assumed that a 10-minute ‘pupil response’ session at the start of every lesson was always the best way to ensure lasting improvements. Feedback was reduced to a simplistic formula. Sometimes marking was so specific it spoon-fed students, removing any actual learning. For example, our marking code identified all missing full stops with a triangle. Students did not have to reread their work to find where the missing full stops should go. They merely found a triangle and inserted a full stop. This was lots of work for the teacher and hardly any for the student, whereas Wiliam states that ‘feedback should be more work for the recipient than the donor’ (2011).

Extract from St Matthias School Feedback and Marking Policy: Start out with the assumption that all children can work independently given prior input and only increase the amount of intervention if the pupil really can’t get on without it. Give children take-up time; let them struggle for a bit, but above all, make sure they are the ones doing the hard work, not you.

Our new policy, based on Wiliam’s work, puts learners to work. Instead of marking, the teacher plans the next lesson around feedback they have gleaned from reading the students’ work. Strengths and weaknesses are highlighted with the whole class, with new teaching addressing any gaps that need more explanation. Students then have a substantial amount of time – often most of the lesson – to go back and hone their work in the light of this input. This requires effort on their part. They have to locate their own errors and think of their own improvements. Now, ‘the next step is the next lesson’. We therefore cover less, but students learn more. And teachers have their lives back.



Elliott V, Baird J, Hopfenbeck, T N, Ingram J, Thompson I, Usher N, Zantout, M, Richardson J and Coleman R (2016) A Marked Improvement? A Review of the Evidence on Written Marking. Oxford: Education Endowment Foundation.

Wiliam D (2011) Embedded Formative Assessment. Bloomington, IN: Solution Tree Press.

Wiliam D (2016) Formative assessment. In: Leadership for Teacher Learning: Creating a Culture Where All Teachers Improve so that Students Succeed. West Palm Beach, FL: Learning Science International.