I am ambivalent about ability grouping. Not least, I am ambivalent about the name. Calling it ‘ability grouping’ implies that we can make some sort of prospective assessment of what students will be able to do. Yet my 21 years of teaching experience has been one where students constantly confound any predictions I might make. In the U.K., ability grouping is usually called ‘setting’, but nobody else knows what that means and it is hardly a description of anything. At my school, we call it ‘grouping for instruction’ and I prefer this term but I doubt it will catch on.
A lack of clarity bedevils the whole area of ability grouping. For instance, in some approaches, groups of students stay together for all of their lessons – known as ‘streaming’ in the U.K. – whereas the more common approach, in my experience, is to regroup students for different subjects.
Definitional issues such as these dog the research on ability grouping. So do many other confounding variables. For instance, I suspect it is common for the least advanced classes to be assigned disproportionately more new and inexperienced teachers, which is quite the reverse of what they need, with heads of department snaffling more of the advanced groups. Not only do new and inexperienced teachers have weaker pedagogical content knowledge, but they are likely to have more classroom management problems and this is compounded by the fact that students may be grouped indirectly on the basis of their behaviour and motivation.
Any department area that wishes to group students has to decide the basis on which to do this. Standardised tests are, well, standardised, and so they are more objective. Yet, they are few and far between. So, if used as a basis of grouping, students can stay in the same group for a long time. If we take class assessments and teacher judgements into account then there can be issues of unconscious bias.
Many years ago, as head of science, I went to observe a third set i.e. group three out of four. The students were calling out answers to the teacher’s question or, rather, one student of Nigerian descent was calling out answers to the teacher’s questions. And he was getting them all right. On investigation, I discovered that he had missed an assessment and that this had been recorded by the teacher as a score of zero. When this score was taken out of his average, he looked like he should be assigned to set two. So we made the move. Soon he was in set one. He took physics at A Level and went on to study engineering at Cambridge.
Despite the potential pitfalls, and with no great enthusiasm, I have continued to make use of ability grouping in the departments I have run. It allows a teacher to somewhat tailor the teaching to the needs of students without being placed in the absurd position of trying to teach 4 or 5 lessons simultaneously. Nevertheless, I remain ambivalent.
Which is why I was enthusiastic about a research project at the UCL Institute of Education in London that, as I understood it, sought to investigate best practice in ability grouping versus mixed ability teaching. By ensuring that they tested the best possible version of each model, researchers would be able to strip out the confounds that had muddied earlier research and answer the question once and for all.
However, the team behind the project have now released a paper that has given me pause for thought. It is called, “The symbolic violence of setting: A Bourdieusian analysis of mixed methods data on secondary students’ views about setting.” How will a group that thinks ability grouping is ‘symbolic violence’ be able to dispassionately evaluate best practice in ability grouping?
As is the fashion in education research, the paper draws on the ideas of a French philosopher. This time, it is Bourdieu. Interestingly, E. D. Hirsch has argued that Bourdieu’s influence over the French education system has led to a decline in academic outcomes.
The authors of the new paper conducted a survey of students and how they feel about being in upper, middle or lower sets. The particular lens the researchers used becomes apparent when students give the wrong answers:
“Yet, even among those in the lowest sets, there were relatively few explicit views challenging the fairness of setting (or recognising the cultural arbitrary on which it is based), which we interpreted as exemplifying how misrecognition helps ensure that such processes are seen as legitimate, and are thus perpetuated, often with the compliance of the dominated. “
Falsifiability is clearly an issue when you can interpret any response as supporting your theory.
And so I have to wonder, what will the best practice study find? I think I already know.