I wonder whether one of the reasons that school leaders pursue highly prescriptive marking policies is that they misunderstand some of the research around feedback. Yes, feedback has powerful effects. However, some of these can be negative and, as John Hattie suggests in “Visible Learning”:
“It was only when I discovered that feedback was most powerful when it is from the students to the teachers that I started to understand it better.”
Writing a comment on a piece of work is not the same thing as feedback. Instead, it is one potential way of providing feedback to students. Comments in exercise books don’t even act as feedback if they are not received and the message certainly won’t be received if we try to convey too many points at once or if a grade or score is also present. I suppose this is why the surreal practice has evolved whereby students then have to write comments about the comments in order to demonstrate that they have read them.
Given these widespread beliefs, I decided that I’d tell you about what I have have been involved with to improve feedback at my school. It is part of a wider whole-school initiative to get smarter about our use of data (although it wouldn’t look the same in all areas – a critical point). I have invented the statistics that I am going to show you, both for purposes of clarity – there has been a lot of tinkering as we have gone along – and also because I have not sought permission to use real data.
Imagine two classes sit a maths test. In this case, I teach one of the classes and a different teacher has the other class. He is more experienced than me at this level and has a strong track record. Let’s have a look at a crude measure; the average scores on each question:
The test was on probability and question 2 was the only question on the binomial theorem. It was a non-calculator test and students had to calculate some relatively simple combinations by hand. This might not mean anything to you but the key point is this; it looks like my group have underperformed on question 2 given that they get similar scores to the other group on the rest of the questions.
Our curriculum is tight; we are both using the same materials in class and are ostensibly following the same lesson plan. This evidence is not research-standard – there are too many variables – but it does strongly imply that there is some meaningful difference in the way that we taught this concept. Given that 90%+ of what we did must have been the same, it should be pretty easy to figure out what that was.
And it is easy. Following a discussion, I learn that my experienced colleague had expanded on a particular example and started a discussion about the use of Pascal’s triangle to find simple combinations. So I go away and teach that strategy to my own students in time for the exam and we write it in to the curriculum so that everyone will teach it next year.
I’ve been through a cycle like this a few times now. It’s not always straightforward and it can be hard to tease out the differences between teachers who think they’ve all done the same thing. It usually turns on the use of a particular example. In one instance, we discovered that a teacher had used an additional step which made a procedure more explicit. She hadn’t even realised this until we asked her to show us what she did and we noticed the difference. Now, everyone does it this way.
We also sometimes find ourselves side-tracked into discussions about things that are not the cause of the difference. For instance, in the example above it is tempting to discuss whether my class has less able students. Yet if this were true, why doesn’t it show up on the other questions? So there is a discipline in focusing on what is important. Teachers also need to have a certain level of expertise; a necessary requirement for effective forms of inquiry.
You don’t even have to have groups that are operating at the same level to use an approach like this – if a generally lower performing group scores more highly than a higher performing group on a particular question then that tells you something interesting.
But please don’t get the wrong idea. We do not analyse everything in this way. A mania for doing that would probably be as unhelpful as a prescriptive marking policy.