Is it better to group students based on their prior ability, in mixed ability groupings, or something in between? This questions besets schools, departments and teachers on an annual basis. There are unique conditions, supports and pressures that mean it is a difficult issue with no silver-bullet answer. The best we can do is survey the best research, undertake some disciplined inquiry and apply it to our unique context.

Here is just a sample of the evidence:

The Education Endowment Foundation Toolkit: The Toolkit, now commonly used in UK schools, is much cited in terms of defining the view of grouping according to the research. It asserts that setting by ability has a small positive impact for the high attainers, but a potentially damaging impact on low attainers:

“Overall, setting or streaming appears to benefit higher attaining pupils and be detrimental to the learning of mid-range and lower attaining learners. On average, setting or streaming does not appear to be an effective strategy for raising the attainment of disadvantaged pupils, who are more likely to be assigned to lower groups.

On average, studies show that higher attaining learners make between one and two additional months’ progress when set or streamed compared to when taught in mixed ability groups…

Evidence suggests that the impact of setting is more detrimental to low attaining pupils in mathematics, who do better in mixed attainment groups, and that setting or streaming particularly affects upper primary and lower secondary education. The effects appear to be less clear-cut in other subjects, though negative effects are reported for low attaining pupils across the curriculum.”

Despite such evidence, setting by ability is still predominant and a highly popular model for schools. The reasons are varied and complex. Setting is perceived as easier to manage, both in terms of the teaching (the many-headed beast of differentiation stirs) and in terms of behaviour management. Many teachers and school leaders argue that you can better tailor instruction to more homogenous ability groupings.

Parents can often put a great deal of pressure on schools to have setting by ability. Of course, most parents suffer from the ‘Dunning Kruger effect’ – that is to say, they think their child is above average and that they would of course benefit by setting and be sitting proudly in the ‘top set’. I have yet to have a parent request setting and be happy to have their child languish in the doldrums of the ‘bottom set’.

The Toolkit raises some great questions and points to consider:

– Flexible within-class grouping is preferable to tracking or streaming for low attaining pupils.

– It is important to recognise that a measure of current attainment, such as a test, is not the same as a measure of potential.
Have you considered how the differences in grouping will enable more effective teaching for all pupils, including lower attaining pupils?

– How will you monitor the impact of ability grouping on pupils’ attitudes to learning and their engagement?

As Chris Husbands shows, politicians are in support of setting by ability as it appears to be a vote winner, regardless of the evidence base. He cites interesting research from Peter Mortimore about how student sense of themselves as learners (or their self-concept) can be damaged by being excluded from the top groups in a setting model – see here. This has implications for a school like ours that professes to focus teaching and learning founded on the principles of all aiming to possess a ‘growth mindset’. Does setting by ability imply a fixed notion of ability in the minds of our students?

Conversely, a large-scale Metlife study from America has shown that 43% of High School teachers in the USA found that groups have become so mixed that they found it hard to teach them successfully. Whether this self-reported issue actually translated into student outcomes is unknown, but how we feel as teachers about our groups is crucial. The expectations (and the effect on learning that can have) we have for our students is similarly crucial. As the adage goes, whether we think we can, or think we cannot, we are probably right.

In the student grouping debate we focus on the differences between whole class groupings, whereas we miss the focus that is perhaps even more crucial: within-class groupings. Wed that to the quality of teaching being experienced by those groups of students and perhaps we should be asking a series of different questions:

– What teachers are securing the best student outcomes and what are they doing in our specific context and subject domain?

– What are the best methods and approaches for within-class groupings? What does the evidence on good ‘cooperative learning‘ (group work) say and can I leverage ‘peer tutoring‘ effectively?

– Given the spread of ability in groupings, either by setting or mixed, how do I best use oral and written feedback to improve learning?

– Are the issues I am trying to tackle with our grouping questions about learning difficulties/challenges in our subject/school, or are they about behavioural problems? How do these two issues interrelate in my context?

– How will I help students better understand what learning looks like in my subject domain (metacognition)? 

– How will I ensure my students are highly motivated to learn in our context? How crucial is student motivation in my context?

– In a mixed ability model, how are our ‘high starters’ still challenged in class and beyond?

– How will I ensure all students have mastered the content and what will I do if small groups haven’t done so, without holding back those who have (mastery learning)?

I do not pretend that these are easy questions to answer, but this specific focus on the quality of teaching and learning should trump the more prevalent debate about class grouping. If too many of the questions are problematic because our student grouping choice is hampering learning, then we need to study the problem with forensic intent and change the grouping. This may be a year-on-year decision and it will likely be subject specific and cohort specific. It certainly won’t prove as simple as just deciding to group by ‘setting or mixed ability?’