“One of our concerns is that we hope schools don’t change things overnight as a result of this,” says Jeremy Hodgen, professor of mathematics education at University College London (UCL) and at the Observatory for Mathematical Education at the University of Nottingham.

Hodgen was one of the lead authors of a recent study commissioned by the Education Endowment Foundation (EEF) that investigated the impact of different grouping practices in key stage 3 maths teaching.

The research compared the attainment and self-confidence of Year 7 and Year 8 students taught in mixed-attainment classes with those taught in classes set by prior attainment.

There were two key findings: that students with high prior attainment make slower progress in mixed-attainment classes; and that setting does not significantly harm the attainment of students with low prior attainment.

Taken together, these findings might seem to be a strong vote of confidence for setting in maths - and a sign that secondary schools using mixed-ability grouping in the subject should reconsider the approach.

Setting vs mixed-ability grouping

But, as Hodgen and his co-author Becky Taylor, professorial research fellow at UCL, explain, the picture is not really that simple.

“Attainment grouping is a complex thing. It’s not something you can change overnight,” says Hodgen. “Changing things would be disruptive for students. It would be disruptive for teachers. And it is important to remember that you can do either [setting or mixed-attainment grouping] well or badly, and they both fit into wider practice of teaching and learning.”

This isn’t to say that schools shouldn’t take note of the findings, the researchers add, rather that they shouldn’t make any rash decisions that could have unintended consequences.

“We are concerned about the high attainers in mixed-attainment classes, and we think that there are some things that schools can do in order to address that,” Hodgen says.

To find out what those things are, and to learn more about how this new research adds to our understanding of ability grouping, we sat down with Hodgen and Taylor to discuss their latest study and what schools should and shouldn’t take from its findings.

There is already a large body of research looking at attainment grouping in schools. What does this new study add?

Hodgen: This is the culmination of more than a decade of work. We had a study funded by the EEF called the Best Practice in Grouping Students study, which ran between 2014 and 2018. It was a pilot study looking at best practice in mixed attainment and a larger study looking at best practice in setting.

One of the things that came out of that was that it’s really difficult to do a randomised controlled trial in this area. The main reason is it’s quite hard to change your attainment grouping, once it is in place.

So we proposed to the EEF a naturalistic study, which would compare schools that were already doing setting with schools that were already doing mixed attainment.

Taylor: I think one of the things that’s really novel about this is that we were able to directly compare the effect of setting and mixed-attainment grouping. Best Practice in Grouping Students didn’t do that. It looked at the impact of our particular model of setting, and the impact of our model of mixed attainment. This new study was, I think it’s fair to say, the most robust study, directly comparing the two practices.

Hodgen: I’d add that most of what we know about ability grouping is from studies conducted in the 1960s and 1970s in the United States, where the way they group students is really different. So this is the most robust study that’s been done on the kind of grouping that’s done in England.

How did the study work?

Hodgen: We started in 2019, but then had to stop and restart because of the Covid-19 pandemic. So this is our seventh year of this project.

We recruited two matched groups of secondary schools: one group using mixed attainment in maths and one using setting.

We had some criteria by which we defined “setting”. For example, there had to be at least three groups.

Some schools told us that they were doing mixed attainment when they were really doing a kind of setting: they have a top set and a bottom set and loads of middle sets, so in effect, most of the students are in a wide middle set. But that’s not quite mixed attainment.

In the end, we looked at 97 schools, so about 15,000 students.

And what did you find?

Hodgen: The first finding about low attainers not being significantly harmed by setting was kind of what we expected. There was a small difference, but it wasn’t significant.

The finding about high attainers was a wee bit of a surprise. We expected to see a gap between how much progress they made in mixed groups, compared to the progress they made when grouped with other high attainers. But the difference is quite large, so that was a concern.

Taylor: High prior attainers’ scores did not go down in schools using mixed-attainment grouping, but they did make less progress, compared with middle prior attainers, than the equivalent group of kids did in schools with sets.

Hodgen: We also looked at self-confidence both in maths and in general. We found that these were lower in mixed attainment schools for all groups of students. We were a bit surprised here about the low attainers, because the previous research has suggested that low attainers have a poor experience in bottom sets, though that prior research hasn’t done a lot of direct comparison between setting and mixed ability, of course.

Alongside the big impact analysis, we had an implementation and process analysis. We had a number of case study schools, in which we observed lessons and interviewed teachers, and there we found the mixed-attainment classes to be more similar to the lower sets than the top sets. We saw a range of practice, but, in general, the top sets were faster paced.

I’d be careful about how I say this, but I don’t think that the experience in many of those top sets was always that good. You might say that it was a bit performance-related?

Taylor: Yes, we saw a focus on GCSEs right from Year 7.

Hodgen: Right, there was a focus on examinations in all the groups, on picking up method marks, things like that. But it was in some of the top sets where GCSEs were mentioned specifically. Bear in mind that this is Year 7. They have five years to go, so you’d hope they’d be learning something before they talk about exams, really.

What other differences in teaching practice did you see?

Hodgen: All the mixed-attainment schools were interested in challenging their high attainers. But one of the things we saw was challenges being used that weren’t so related to the activity that had gone on in the lesson. They were often challenges from programmes like the NRICH [University of Cambridge resources], which are designed for collaborative rather than individual work, and there was little discussion about that challenge when the class came together to talk about things at the end of the lesson.

Overall, in comparison to our previous work, we saw more teacher-led activities and less differentiation and less paired and collaborative work. We think that might have something to do with the current Teaching for Mastery focus in the subject.

Taylor: Some of the teachers and heads of maths we spoke to talked explicitly about how they were working with mastery in their classes, but I think it would be fair to say that the overwhelming majority of the lessons that we observed had the same shape: some kind of review questions, the teacher providing some new input from the front, children doing some individual practice and then maybe a little bit of feedback at the end.

I think out of 47 lessons, only two of them contained any kind of meaningful small group work. So that was really different to what we saw in our Best Practice in Grouping Students observations, back in 2015.

What are the implications of these shifts in practice?

Taylor: It was very striking that we saw almost no differentiation. We saw teachers spending time with different children within the class, so that was differentiation by support from the teacher. Off the top of my head, I don’t think we saw very many teaching assistants. So these are typically teachers working on their own in classrooms.

I’ll add that one of the things we were interested in was the extent to which children had ownership of the maths they were learning, and the extent to which they were developing as mathematicians. And I think it would be fair to say that what we saw in a lot of the lessons we observed was that the teacher was in control of the maths in the classroom. So there was less opportunity, even in high sets, for the young people to take ownership.

What I mean by that is young people seeing themselves as mathematicians and having mathematical agency, being able to come up with their own ideas and participate actively in maths lessons. In the lessons that we saw, this was the case only some of the time. Most of the time students only responded to the teacher’s questions and weren’t given the opportunity to come up with their own mathematical ideas.

Hodgen: One of the advantages of a mixed-attainment class is making use of the students themselves. There’s a lot of knowledge in the class and students who can help other students. And sometimes the way that a student can explain something is closer to another student than the way a teacher can explain it. So if what you’re doing is just teaching to the middle, you’re often not making use of that.

What you want, both in mixed-attainment classes and in sets, is not so much to have everybody doing the same thing, but for the students to have a similar experience, whilst being challenged in different ways, through questioning and through differentiation by outcome.

It’s important to emphasise that schools can do mixed attainment well or less well - and similarly, that they can do setting well or less well.

The EEF study focused specifically on maths teaching. Do you think these findings have implications for teachers of other subjects?

Taylor: It is important to note that this was a study looking at one subject. Having said that, we did choose maths because maths teachers are particularly keen on setting. For Best Practice in Grouping Students, we looked at English and maths, and found many more schools that were using mixed-attainment grouping in English than in maths.

I think probably teachers should be asking themselves the same questions about stretching high attainers, whatever the subject.

Hodgen: Yes, we did choose maths specifically because we thought it would say something about other subjects. Subjects are different, of course, but we suspected some of these things might be happening in other subjects, too. Although we don’t know for sure, as we haven’t looked specifically at those other subjects.

You mentioned that schools can do mixed-attainment grouping well. What does good teaching practice look like here?

Taylor: In terms of mixed attainment, the key thing, which is really clear from this research, is to be thinking about high attainers: how they’re included in the lesson and how they’re being stretched, but stretched in a way that is integrated into the overall teaching of the class.

So that means looking for ways to extend and stretch them within the same topic, rather than doing an add-on activity or something that’s separate that they do on their own when they finish the other work.

Hodgen: Teachers should also focus on the kind of questioning that goes on in the discussion or plenary you’re using as a wrap-up activity, because that means you can make use of what students actually know in a way that benefits all the students. It benefits high attainers to explain and articulate their mathematical ideas, and it benefits middle and lower attainers to hear and respond to those articulations.

What about best practice for teaching in sets? What do we know about what this looks like?

Taylor: For schools that are using setting, the advice is around looking at who’s teaching each of your groups, and making sure that everybody’s got fair access to the most qualified teachers or the teachers who are considered to be the most skilled. You also need to think about moving pupils around between groups when it’s appropriate to do so, so that pupils don’t get stuck in a low group, for example. If they make some progress, make sure they can move up.

Hodgen: It’s important to provide challenges for all students, whatever their prior attainment, and give all chances to reason about maths. But that goes for mixed attainment, too, of course.

One of the arguments often levelled at mixed-ability grouping is that this approach makes it more difficult for teachers to manage student behaviour. Is that something that was reflected in your findings?

Hodgen: Teachers didn’t raise this with us. Mostly, when asked about different forms of grouping, they referred to issues around pace and depth of content. In lesson observations, we largely observed no behavioural issues and students were mostly engaged and on-task.

There were a couple of lessons with behavioural issues, a low set and a mixed-attainment class. But I wouldn’t read too much into this. In each case, there were clearly other things going on that explained it.

If there was one thing that you hope schools would take away from the findings of this latest study, what would it be?

Hodgen: Don’t rush into making changes. What we’re really concerned about is that schools and teachers and headteachers look carefully at the research and think about how they can support all kids, whatever attainment grouping method they’re using.

Taylor: I’d add that one of the things we asked teachers in the study about was how much professional development they were getting around teaching maths, and it was quite minimal. There was a lot of variation between schools and between teachers, but, on average, it was quite a small amount. And so that might be another thing to think about.

Some of the points that this study raises would be brilliant areas for maths departments to look at together and to spend some time learning about together, and developing their practice together.