On Ability Grouping and Gifted Education – Part 2

In Part 1, I sought to:

  • clear away the terminological difficulties that complicate international discussion of this complex and sensitive issue;
  • set ability grouping within the wider education policy-making context, specifically the relationship between excellence and equity and the choice between prescription and school autonomy; and
  • summarise the key points from an objective review of international research on the benefits of setting and tracking relative to mixed ability grouping.

In a nutshell, the research tells us that, while mixed ability settings tends to favour the lowest achievers and setting/tracking is typically better for the highest achievers, when we look at the impact on the attainment of all pupils there is little difference in overall outcomes, because the benefits and disbenefits broadly balance out.

But, compared with mixed ability scenarios, setting/tracking is more likely to increase attainment differences between low achievers and high achievers and, because disadvantaged learners are more prevalent in lower sets/tracks, setting/tracking is relatively more likely to increase the achievement gap between rich and poor.

I argued that, when we develop gifted education policy – and even when we advocate for gifted learners – it behoves us to understand and accept this reality, rather than taking a partial view that concentrates solely on the benefits for the group whose interests we exist to serve.

The fundamental point is this: it cannot be right to insist on improvements in G&T education which can only be achieved at the expense of other learners. Instead of a simplistic either/or debate, we need to look at more sophisticated hybrid solutions that retain the benefits of the two different models and mould them into a better ‘third way’.

The more promising elements of such a ‘third way’ will begin to emerge from deeper reflection on the research findings.

Setting versus tracking/streaming

If we look at the issue from a UK perspective, there is seemingly broad professional consensus that streaming is less effective than setting. Streaming typically depends on crude distinctions that fail to recognise the full profile of a learner’s strengths and weaknesses. So it is quite likely that learners – even many gifted learners – will find themselves relatively under-challenged in some subjects and relatively over-challenged in others.

Conversely, setting enables schools to tailor their initial selection criteria and pedagogical approach to the subject in question. It is quite conceivable that some subjects will use sets while others remain mixed ability – that is standard practice in many English schools. It may also be possible to run sets and mixed ability groups in parallel in the same subject – eg a set for high achievers and/or a set for low achievers with other learners taught in mixed ability groups.

Setting also provides greater flexibility for schools to move pupils when necessary, in recognition that children develop at different rates in different subject areas, with lags and spurts at different times. By offering a more flexible, dynamic model, they are better able to respond quickly when a learner needs extra challenge or greater support in one particular part of the curriculum.

This is not to say that setting inevitably delivers these advantages – rather that effective setting practice does so. As with the identification of gifted learners, it can be all too easy for poorer schools to undertake setting as a once-and-for-all judgement rather than an arrangement kept under regular review and informed by careful analysis of individual pupils’ progress.

I am not sure to what extent setting (maybe we should call it ‘subject-specific tracking’) has been adopted in the US, though it clearly exists in some areas as an alternative to the cruder standard tracking approach. Setting does appear to offer more inherent advantages than the cruder alternative, but there is no reason in principle why elements of both could not be combined into a workable hybrid.

And we should not discard entirely the possibility of a workable model based exclusively on tracking. Advocates of both tracking and of setting are right to assert that, given the right kind of implementation, it ought to be possible to utilise either technique to raise standards for all learners while simultaneously narrowing achievement gaps. The big – and seemingly unanswered – question is how to realise that in practice.

How Big are the Attainment Differences Between High- and Low-Achievers?

Within the research canon, the scale of the performance gap between different tracks/sets is disputed territory. This is probably because of the statistical difficulties involved in isolating the effect of the pedagogy from the effect of prior achievement attributable to gender, ethnic and socio-economic background (amongst other variables).

More recent research suggest that the gap is probably less significant than was earlier believed. Some argue that it is negligible, but there does not seem to be a reliable body of evidence to support that contention. And, even if there were, this does not impact on the ‘sorting hat’ effect that is potentially so detrimental to those from disadvantaged backgrounds (an effect that we know well because it also undermines the effective identification of G&T populations the world over).

Teacher Quality and Deployment in Tracks and Sets

One might reasonably hypothesise that performance gaps between upper and lower tracks, streams and sets are much less likely to appear in schools which ensure that their more experienced subject-specialists are more regularly deployed at the lower end of the spectrum.

Conversely, given the impact of teacher quality on pupil performance, it would be no surprise to find that schools which deployed their best teachers at the upper end were more likely to see performance gaps widen.

This is not to suggest that it would be appropriate to load all the worst teachers into the upper sets, but merely that the distribution might be weighted more towards the lower end than it is now in a typical school. For it is an inconvenient truth often ignored by gifted educators that, by and large, high-achieving gifted learners tend to benefit from the typical distribution of teaching talent across sets or tracks.

It may or may not be attributable to teacher quality, but there does seem to be some research evidence that students in upper sets and tracks typically encounter a more enlightened pedagogy, with more opportunities for student engagement and interaction, whereas lower sets more often experience a more didactic ‘remedial’ teaching style.

Teacher Quality and Deployment in Mixed Ability Settings

In schools that follow a mixed ability approach, there is a different kind of issue about the distribution of pedagogical skill. In this context, the most important consideration becomes the effectiveness with which teachers can cater for widely differing needs within a single classroom.

While the best teachers may be able to differentiate their challenge and support across a mixed-ability class of 30 or more students, the critical issue is whether this can be achieved consistently by all teachers. This is straying into yet another contentious research area – and not one that I have yet explored in any detail.

So I am not equipped to challenge the orthodox anecdotal view – often heard from the parents of gifted learners in particular – that this is a bridge too far for many teachers, particularly if the school cannot afford to reduce the class to a more manageable size. So, they argue, these teachers need the pedagogical ‘crutch’ of a narrower achievement range within which to operate.

In passing, we should not forget that effective differentiation remains an issue even within a smaller top set or upper track. There may be a tendency, particularly amongst inexperienced staff, to assume that it is more acceptable to teach to the middle in a top set, so underserving those who are struggling to stay with the set and those who need extra challenge.

We must also acknowledge potential subject differences. It seems to be a widespread view amongst professionals that setting is potentially more valuable in the linear subjects such as maths and modern languages where the acquisition of subject knowledge is critical to progress, whereas in English and arts subjects it is not quite so essential to divide pupils on the basis of past achievement. Whether this view is reliably supported by the research literature is more open to question.


The case for and against detracking seems to hinge on this issue of teachers’ capacity to differentiate effectively in mixed ability settings. If they are not up to mark, it is of course the pupils at either extreme who are most likely to suffer, including the high-achieving gifted learners.

And it is high-achieving learners from disadvantaged backgrounds that are most vulnerable. They are more likely to be found in schools in disadvantaged urban areas: precisely the schools that cannot attract a critical mass of high quality teachers able to differentiate effectively in a mixed ability setting. It is another inconvenient truth that the better teachers are found disproportionately in more advantaged areas and schools, despite the redistributive impact of initiatives like Teach for America/Teach First.

This means that – contrary to the conclusions above about the overall negative impact of sets on performance gaps – detracking may unintentionally widen the excellence gap we have discussed in previous posts on this blog.

In my view, detracking is unlikely to succeed as a crude stand-alone adjustment. It would be much more likely to work as part of a sophisticated suite of reforms that supplement the support available for all learners, but especially those at either end of the achievement spectrum, whether that be by means of in-class ability grouping, pull-out for catch-up and extension, targeted 1:1 tuition, additional enrichment, independent learning opportunities…and so on.

The critical risk that proponents of gifted education must guard against is that, when such a suite is designed and implemented, all the funding and support is targeted at the lower achievers, while comparable support for the high achievers is neglected on the mistaken assumption that they will succeed regardless. That is against the spirit of personalised education and is manifestly unfair and inefficient.

A more sophisticated approach

Too often, the debate about tracking/setting and mixed ability settings is over-simplified into a straightforward either/or choice between two mutually exclusive options. But in reality the real issue is how effectively a school deploys its professional and para-professional resources and a rich pedagogical armoury to personalise learning for all its pupils.

Schools need to think carefully about the needs of all their learners and to devise solutions that will enable them to deliver improved standards for all (excellence) and a narrowing of achievement gaps (equity). The latter must not be secured by holding high-level achievement steady while the poorer performers catch up.

In the English context, as we move towards greater institutional autonomy, this requires each and every school to secure the necessary expertise to make such decisions with confidence, to evaluate their reforms properly and to adjust them in the light of the evidence.

The trick, as always, is to provide a broad, flexible and non-bureaucratic review framework that supports the weaker schools without limiting the options of the stronger ones. In an ideal world this would be mediated by a School Improvement Partner or similar, but it should also be capable of direct use by the school’s staff and governors, possibly in partnership with another school, so providing some degree of objective external scrutiny.

On the surface, the politicians seem confident in schools’ ability to deliver and are freeing them from several constraints upon their autonomy; but it would appear that they also continue to hanker after prescription in certain areas – perhaps those where they hold strong personal beliefs, or those (like setting) where they know that opinion in schools is divided. For I would be wrong to give the impression that setting in English schools is a ‘done deal’.

Once again there seems to be a dearth of reliable national data. But a crude estimate can be derived from OFSTED 2007/08 observation data quoted in a 2009 PQ answer. This suggests that only around 15% of primary lessons are organised on the basis of ability, whereas the comparable figure for the secondary sector is 45%. So, even in the secondary sector, the small majority of lessons is still likely to be mixed-ability.

There is a case for including within a review framework an expectation that setting will be considered, but only as part of a much wider-ranging assessment of school organisation, teacher deployment and classroom pedagogy. It doesn’t make sense to consider setting in splendid isolation.

As part of this process, schools would need to satisfy themselves that a decision to introduce setting would not inhibit them from narrowing achievement gaps. They might be invited to consider setting as a default element of their strategy if they have no evidence to suggest that an alternative arrangement would result in better achievement outcomes for their high achievers and their low achievers, especially (but not exclusively) those from disadvantaged backgrounds.

Some Innovation Pointers

Such a framework should overtly encourage schools to experiment and innovate, to evaluate carefully and to learn from each other’s experience.

Ironically, while mainstream US opinion seems relatively wedded to tracking, US educators (and Australian educators too) have experimented much more thoroughly than England with hybrid solutions, innovative classroom grouping techniques and other promising approaches to organisation and differentiation.

Let us hope that the greater autonomy promised to schools here allows them to be equally innovative. They might begin by giving more serious attention to the potential of vertical grouping – where pupils are organised on the basis of achievement, but across year groups rather than within them. With one or two honourable exceptions vertical grouping is not deployed here as a means of improving differentiation, but rather as a pragmatic solution to small classes in rural schools.

And we could do with significantly more English school experience of cluster grouping, where the G&T pupils (or low achievers for that matter) are concentrated in a single mixed ability class with a specially trained teacher rather than being dispersed across a range of classes, so relatively isolated from opportunities to learn with their peers.

For the research demonstrates two things above all else:

  • There is no single right answer to the question how best to organise and differentiate within a school; and
  • We have very much more still to learn about what constitutes effective practice per se.


September 2010

On Ability Grouping and Gifted Education: Part 1

After a recent #gtchat on this subject and follow-up posts from fellow participants, I felt an urge to set out my own perspective on this vexed question.

I cannot pretend to have reviewed the voluminous research undertaken on both sides of the Atlantic, let alone worldwide, but what I have digested leads me to the comments below.

Before we summarise the research evidence, let us begin with three important pieces of context.



First – On Excellence and Equity

Regular readers will know my view that the balance between these two principles has a powerful impact on the nature of gifted education policy at all levels of our various education systems.

Like non-identical twins, excellence and equity are often found together. Dressed in various guises, they regularly find their way into politicians’ statements of their wider educational objectives. For example, England’s Secretary of State for Education maintains that his Government’s top educational priorities are ‘raising standards’ (aka excellence) and ‘narrowing gaps’ (aka equity). It wouldn’t take much effort to unearth similar statements from most of his opposite numbers around the developed world.

Politicians are often rather coy when asked to say whether one of these twins is more important to them than the other. They tend to be given equal billing regardless of party political viewpoint.

But, given a longer term perspective, they can be imagined more accurately as sitting at either end of a policy seesaw. At any particular time, one of the pair is usually in the ascendant, but it always seems as though that elevated status inevitably results in the other gaining ascendancy in its turn, though perhaps not until the next administration, or even the next but one.

Please excuse the dreadful mixing of metaphors but, as with G&T education – and other critical issues like school admission – ability grouping is sensitive territory. It forms part of the troubled borderlands of education policy where the twin principles of excellence and equity often clash and have to be reconciled (or, failing that, to coexist in a state of some tension).

Second – On Prescription and Autonomy

The squaring of excellence and equity within these educational badlands can be attempted within the declared education policy of the authority, state or country concerned, or it can be devolved to schools. This applies to ability grouping, gifted education or any similarly contentious issue where the twin issues are wrestling for ascendancy.

The second option – devolution – seemingly enables politicians to sidestep the issue, which is often attractive because the alternative centralised prescriptive approach risks leading them into hot water and/or requires a commitment of additional resources that they do not have to spare.

But it is not feasible to admit such reasons openly, so devolution tends to be justified on the grounds that the administration should not be fettering the discretion of the professionals. In the case of ability grouping, they will argue that it cannot be right for them to trespass in the secret garden of pedagogy (to use a phrase that will resonate with English educationalists). That is rightly the province of the school, which must be free to reach decisions without meddling governmental interference, taking full account of its unique circumstances and the particular needs of its pupils.

Both prescription and autonomy have their weaknesses. In the case of autonomy, the problem is that a government is typically accountable to an electorate for its education policies and will need at some point to demonstrate success in achieving its stated priorities. If too many schools act in a way that runs counter to those priorities, the government must either admit failure or – much more likely – use evidence selectively to put the best possible spin on affairs.

A sensible administration will not trust solely to their powers of communication however. If there is a chance that the collective decisions taken in schools will not deliver the outcome they stand for, they will also want to deploy the various ‘policy levers’ available to surreptitiously encourage schools towards the ‘right’ decision. Examples of such levers are the distribution of funding, or the inclusion of relevant issues within inspection and quality assurance processes, or a set of identified priorities for teachers’ professional development.

Sometimes there are inherent contradictions in a government’s policy that need to be reconciled. This is a problem regardless of whether prescription or autonomy is in the ascendant. In the case of autonomy, there is a strong risk of mixed messages, schools are divided over the best approach and policy levers are introduced that pull in different directions.

On occasions there can be policy contradictions and a conflict between prescription and autonomy. Before the recent General Election, the senior Conservative partners in what has become the Coalition Government committed to freeing teachers and schools from government interference (autonomy) and narrow gaps in performance between advantaged and disadvantaged learners (prescription, equity) and ensure that setting became much more prevalent in schools (prescription, excellence).

It will take some sleight of hand to pull off that particular trick, as we shall see. But maybe the commitment to setting has fallen by the wayside – it did not feature in the Coalition Government’s priorities for education reform and we wait to see whether it will reappear in the forthcoming Schools White Paper.

Third – On Terminology

A further similarity between ability grouping and G&T education is that different terminology is used in different parts of the world. This makes it much more likely that discussants will become disputants, because they do not share a common pedagogical language.

In the US context, one most often encounters the term ‘tracking’ and its rather inelegant antithesis ‘detracking’.

As I understand it, ‘tracking’ is typically used to mean dividing pupils into different educational programmes – either on the basis of ability or past achievement – which apply across all or most of the school experience. The equivalent term in the UK is ‘streaming’.

I cannot find recent and reliable data to illustrate how prevalent tracking is in the US. Some of the research studies imply that it has been almost universal in US high schools, at least until ‘detracking’ began in earnest around a decade ago. Conversely, while streaming was once popular in England, it now seems comparatively rare. It seems to survive in a few comprehensive schools that are competing with selective schools and so opt to establish a ‘grammar stream’ but is otherwise discarded as poor practice.

It is also unclear to what extent US tracking practice has morphed into what we in the UK call ‘setting’ – the grouping of pupils according to ability or achievement in specific subjects. As far as I can see, the stereotypical generic and tripartite tracking model in the US still seems common, but I stand to be corrected. Setting is much more prevalent in UK secondary schools, especially in the so-called ‘linear subjects’, and has also increased in primary schools, predominantly in the core subjects of English and maths.

‘Detracking’ is the process of dispensing ‘tracking’ in favour of exclusively mixed ability classes. The online literature implies that this is happening in many schools and states across the US but, yes you’ve guessed it, reliable data seems rather thin on the ground.

Tracking, setting and streaming all take place across the school, or at least across an entire year group within a school. They are therefore clearly distinct from ‘selection’ – which is the shorthand term we in the UK apply to approaches that sort pupils into different schools – and to ability grouping at the classroom level.

So, with our terminological differences out of the way, we can move on to consider what the vast body of research actually tells us about the effectiveness of tracking, detracking, streaming and setting.

Top Line Conclusions from the Research

The first and most obvious point is that the research conclusions are contested and – to be brutally honest – not always entirely objective. Put very crudely, those with an equity focus tend to marshal the arguments in such a way that helps them conclude that tracking/setting/streaming is detrimental to learners from disadvantaged backgrounds.

Conversely, those with an excellence perspective (including most gifted education researchers) tend to work towards the conclusion that it is markedly beneficial to high achievers, if not to G&T learners per se.

Both groups tend to downplay the evidence that does not fully support their position.

For a more balanced view I turned to an extensive and objective 2005 literature review for England’s education ministry. This concludes that the research worldwide can be interpreted as follows:

  • no single form of grouping benefits all pupils and there is little attainment advantage associated with setting – ie no significant difference between setting and mixed ability classes in overall attainment outcomes across all pupils
  • ‘at the extremes of attainment’ low-achieving pupils show more progress in mixed ability classes and high-achieving pupils show more progress in sets
  • lower sets tend to contain a disproportionate number of boys, pupils from those ethnic groups that tend to underachieve and pupils with SEN
  • there are aspirational and behavioural disadvantages to setting, predominantly amongst lower attainers and there is a correlation between disaffection and setting, particular for pupils in the lowest sets
  • higher sets are more likely to have experienced and highly-qualified teachers whereas lower sets experience more changes of teacher and are less likely to be taught by a specialist in the subject.

The terminology is tailored for a UK audience but these conclusions apply in equal measure to tracking versus mixed ability settings. Since the gains of high achievers are offset by the losses to low achievers – and middle achievers tend to do broadly the same in either scenario – neither setting nor tracking has little overall impact on the raising of standards (excellence).

However, compared with mixed ability teaching, setting and tracking tend to increase performance gaps between high achievers and low achievers per se. Moreover, because pupils from disadvantaged backgrounds are found disproportionately in the lower sets or tracks, this serves to widen socio-economic achievement gaps (equity), as well as some ethnic minority achievement gaps and the gender gap between girls and boys.

Next Time

In Part 2 we will look more closely at some of the issues beneath these headlines – playing back into the mix the wider contextual issues above. But, for now, my point is this: we, as an international community of gifted educators, ought to be prepared to accept this as the broad baseline consensual position, rather than holding on to an alternative and partial version of reality that is focused too narrowly and exclusively on the needs of gifted learners.


September 2010