The Politics of Selection: Grammar Schools and Disadvantage

This post considers how England’s selective schools are addressing socio-economic disadvantage.

Another irrelevant Norwegian vista by Gifted Phoenix

Another irrelevant Norwegian vista by Gifted Phoenix

It is intended as an evidence base against which to judge various political statements about the potential value of selective education as an engine of social mobility.

It does not deal with recent research reports about the historical record of grammar schools in this respect. These show that – contrary to received wisdom – selective education has had a very limited impact on social mobility.

Politicians of all parties would do well to acknowledge this, rather than attempting (as some do) to perpetuate the myth in defiance of the evidence.

This post concentrates instead on the current record of these schools, recent efforts to strengthen their capacity to support the Government’s gap closing strategy and prospects for the future.

It encourages advocates of increased selection to consider the wider question of how best to support high attainers from disadvantaged backgrounds.

The post is organised into four main sections:

  • A summary of how the main political parties view selection at this point, some six months ahead of a General Election.
  • A detailed profile of the socio-economic inclusiveness of grammar schools today, which draws heavily on published data but also includes findings from recent research.
  • An evaluation of national efforts over the last year to reform selective schools’ admissions, testing and outreach in support of high-attaining disadvantaged learners.
  • Comparison of the various policy options for closing excellence gaps between such learners and their more advantaged peers – and consideration of the role that reformed and/or increased selection might play in a more comprehensive strategy.

Since I know many readers prefer to read my lengthy posts selectively I have included page jumps from each of the bullet points above to the relevant sections below.

One more preliminary point.

This is the second time I have explored selection on this Blog, though my previous post, on fair access to grammar schools, appeared as far back as January 2011. This post updates some of the data in the earlier one.

One purpose of that earlier post was to draw attention to the parallels in the debates about fair access to grammar schools and to selective higher education.

I do not repeat those arguments here, although writing this has confirmed my opinion that they are closely related issues and that many of the strategies deployed at one level could be applied equally at the other.

So there remains scope to explore how appropriate equivalents of Offa, access agreements, bursaries and contexualised admissions might be applied to selective secondary admissions arrangements, alongside the reforms that are already on the table. I leave that thought hanging.


The Political Context

My last post on ‘The Politics of Setting’ explored how political debate surrounding within-school and between-school selection is becoming increasingly febrile as we approach the 2015 General Election.

The two have become inextricably linked because Prime Minister Cameron, in deciding not to accommodate calls on the right of his party to increase the number of selective schools, has called instead for ‘a grammar stream in every school’ and, latterly, for a wider – perhaps universal – commitment to setting.

In May 2007, Cameron wrote:

‘That’s what the grammar school row was about: moving the Conservative Party on from slogans such as ‘Bring back grammar schools’ so that we can offer serious policies for improving state education for everyone…

…Most critics seem to accept, when pressed, that as I have said, the prospect of more grammars is not practical politics.

Conservative governments in the past – and Conservative councils in the present – have both failed to carry out this policy because, ultimately, it is not what parents want….

…When I say I oppose nationwide selection by 11 between schools, that does not mean I oppose selection by academic ability altogether.

Quite the reverse. I am passionate about the importance of setting by ability within schools, so that we stretch the brightest kids and help those in danger of being left behind.’

With a Conservative Government this would be a motor of aspiration for the brightest kids from the poorest homes – effectively a ‘grammar stream’ in every subject in every school.

Setting would be a focus for Ofsted and a priority for all new academies.’

As ‘The Politics of Setting’ explained, this alternative aspiration to strengthen within-school selection has not yet materialised, although there are strong signs that it is still Cameron’s preferred way forward.

The Coalition has been clear that:

‘It is not the policy of the Government to establish new grammar schools in England’ (Hansard, 10 February 2014, Col. 427W).

but it has also:

  • Removed barriers to the expansion of existing grammar schools through increases to planned admission numbers (PANs) within the Admissions Code.
  • Introduced several new selective post-16 institutions through the free schools policy (though not as many as originally envisaged since the maths free schools project has made relatively little progress).
  • Made efforts to reform the admissions procedures of existing selective secondary schools and
  • Accepted in principle that these existing schools might also expand through annexes, or satellite schools. This is now a live issue since one decision is pending and a second proposal may be in the pipeline.

The Liberal Democrats have enthusiastically pursued at least the third of these policies, with Lib Dem education minister David Laws leading the Government’s efforts to push the grammar schools further and faster down this route.

In his June 2014 speech (of which much more below) Laws describes grammar schools as ‘a significant feature of the landscape in many local areas’ and ‘an established fact of our education system’.

But, as the Election approaches, the Lib Dems are increasingly distancing themselves from a pro-selective stance.

Clegg is reported to have said recently that he did not believe selective schools were the way forward:

‘The Conservatives have got this odd tendency to constantly want to turn the clock back.

Some of them seem to be hankering towards a kind of selective approach to education, which I don’t think works.

Non-selective schools stream and a lot of them stream quite forcefully, that’s all fine, but I think a segregated school system is not what this country needs.’

Leaving aside the odd endorsement of ‘forceful streaming’, this could even be interpreted as hostile to existing grammar schools.

Meanwhile, both frontrunners to replace Cameron as Tory leader have recently restated their pro-grammar school credentials:

  • Constituency MP Teresa May has welcomed consideration of the satellite option in Maidenhead.

The right wing of the Tory party has long supported increased selection and will become increasingly vociferous as the Election approaches.

Conservative Voice – which describes itself as on the ‘center-Right of the party’ [sic] – will imminently launch a campaign calling for removal of the ban on new grammar schools to be included in the Conservative Election Manifesto.

They have already conducted a survey to inform the campaign, from which it is clear that they will be playing the social mobility card.

The Conservative right is acutely aware of the election threat posed by UKIP, which has already stated its policy that:

‘Existing schools will be allowed to apply to become grammar schools and select according to ability and aptitude. Selection ages will be flexible and determined by the school in consultation with the local authority.’

Its leader has spoken of ‘a grammar school in every town’ and media commentators have begun to suggest that the Tories will lose votes to UKIP on this issue.

Labour’s previous shadow education minister, Stephen Twigg, opposed admissions code reforms that made it easier for existing grammar schools to expand.

But the present incumbent has said very little on the subject.

A newspaper interview in January 2014 hints at a reforming policy:

‘Labour would not shut surviving grammar schools but Mr Hunt said their social mix should be questioned.

“If they are simply about merit why do we see the kind of demographics and class make-up within them?”’

But it seems that this has dropped off Labour’s agenda now that the Coalition has adopted it.

I could find no formal commitment from Labour to address the issue in government, even though that might provide some sort of palliative for those within the party who oppose selection in all its forms and have suggested that funding should be withdrawn from selective academies.

So the overall picture suggests that Labour and the Lib Dems are deliberately distancing themselves from any active policy on selection, presumably regarding it as a poisoned chalice. The Tories are conspicuously riven on the issue, while UKIP has stolen a march by occupying the ground which the Tory right would like to occupy.

As the Election approaches, the Conservatives face four broad choices. They can:

  • Endorse the status quo under the Coalition, making any change of policy conditional on the outcome of a future leadership contest.
  • Advocate more between-school selection. This might or might not stop short of permitting new selective 11-18 secondary schools. Any such policy needs to be distinct from UKIP’s.
  • Advocate more within-school selection, as preferred by Cameron. This might adopt any position between encouragement and compulsion.
  • Develop a more comprehensive support strategy for high attaining learners from disadvantaged backgrounds. This might include any or all of the above, but should also consider support targeted directly at disadvantaged students.

These options are discussed in the final part of the post.

The next section provides an assessment of the current state of selective school engagement with disadvantaged learners, as a precursor to describing how the reform programme is shaping up.


How well do grammar schools serve disadvantaged students?


The Grammar School Stock and the Size of the Selective Pupil Population

Government statistics show that, as of January 2014, there are 163 selective state-funded secondary schools in England.

This is one less than previously, following the merger of Chatham House Grammar School for Boys and Clarendon House Grammar School. These two Kent schools formed the Chatham and Clarendon Grammar School with effect from 1 September 2013.

At January 2014:

  • 135 of these 163 schools (83%) are academy converters, leaving just 28 in local authority control. Twenty of the schools (12%) have a religious character.
  • Some 5.1% of pupils in state-funded schools attend selective schools. (The percentage fluctuated between 4% and 5% over the last 20 years.) The percentage of learners under 16 attending selective schools is lower. Between 2007 and 2011 it was 3.9% to 4.0%.
  • There are 162,630 pupils of all ages attending state-funded selective secondary schools, of which 135,365 (83.2%) attend academies and 27,265 (16.8%) attend LA maintained schools. This represents an increase of 1,000 compared with 2013. The annual intake is around 22,000.

The distribution of selective schools between regions and local authority areas is shown in Table 1 below.

The percentage of selective school pupils by region varies from 12.0% in the South East to zero in the North East, a grammar-free zone. The percentage of pupils attending selective schools by local authority area (counting only those with at least one selective school) varies from 45.1% in Trafford to 2.1% in Devon.

Some of the percentages at the upper end of this range seem to have increased significantly since May 2011, although the two sets of figures may not be exactly comparable.

For example, the proportion of Trafford pupils attending selective schools has increased by almost 5% (from 40.2% in 2011). In Torbay there has been an increase of over 4% (34.8% compared with 30.5%) and in Kent an increase of almost 4% (33.3% compared with 29.6%).


Table 1: The distribution of selective schools by region and local authority area and the percentage of pupils within each authority attending them (January 2014)

Region Schools Pupils Percentage of all pupils
North East 0 0 0
North West 19 20,240 4.9
Cumbria 1 833 2.8
Lancashire 4 4,424 6.6
Liverpool 1 988 3.3
Trafford 7 7,450 45.1
Wirral 6 6,547 30.5
Yorkshire and Humberside 6 6,055 1.9
Calderdale 2 2,217 14.2
Kirklees 1 1,383 5.5
North Yorkshire 3 2,454 6.5
East Midlands 15 12,700 4.5
Lincolnshire 15 12,699 26.9
West Midlands 19 15,865 4.5
Birmingham 8 7,350 10.4
Stoke-on-Trent 1 1,078 8.7
Telford and Wrekin 2 1,283 11.7
Walsall 2 1,423 7.0
Warwickshire 5 3,980 12.0
Wolverhampton 1 753 5.0
East of England 8 7,715 2.1
Essex 4 3,398 4.0
Southend-on-Sea 4 4,319 32.8
London 19 20,770 4.4
Barnet 3 2,643 11.6
Bexley 4 5,466 26.6
Bromley 2 1,997 9.0
Enfield 1 1,378 6.1
Kingston upon Thames 2 2,021 20.5
Redbridge 2 1,822 7.9
Sutton 5 5,445 30.7
South East 57 59,910 12.0
Buckinghamshire 13 15,288 42.2
Kent 32 33,059 33.3
Medway 6 6,031 32.2
Reading 2 1,632 24.1
Slough 4 3,899 37.4
South West 20 19,370 6.2
Bournemouth 2 2,245 23.3
Devon 1 822 2.1
Gloucestershire 7 6,196 16.2
Plymouth 3 2,780 16.3
Poole 2 2,442 26.8
Torbay 3 2,976 34.8
Wiltshire 2 1,928 6.6
TOTAL 163 162,630 5.1


Some authorities are deemed wholly selective but different definitions have been adopted.

One PQ reply suggests that 10 of the 36 local authority areas – Bexley, Buckinghamshire, Kent, Lincolnshire, Medway, Slough, Southend, Sutton, Torbay and Trafford – are deemed wholly selective because they feature in the Education (Grammar School Ballots) Regulations 1998.

Another authoritative source – the House of Commons Library – omits Bexley, Lincolnshire and Sutton from this list, presumably because they also contain comprehensive schools.

Of course many learners who attend grammar schools live in local authority areas other than those in which their schools are located. Many travel significant distances to attend.

A PQ reply from March 2012 states that some 76.6% of all those attending grammar schools live in the same local authority as their school, while 23.2% live outside. (The remainder are ‘unknowns’.)

These figures mask substantial variation between authorities. A recent study, for the Sutton Trust  ‘Entry into Grammar Schools in England’ (Cribb et al, 2013) provides equivalent figures for each local authority from 2009-10 to 2011-12.

The percentage of within authority admissions reaches 38.5% in Trafford and 36% in Buckinghamshire but, at the other extreme, it can be as low as 1.7% in Devon and 2.2% in Cumbria.

The percentage of admissions from outside the authority can be as much as 75% (Reading) and 68% (Kingston) or, alternatively, as low as 4.5% in Gloucestershire and 6.8% in Kent.


Recent Trends in the Size and Distribution of the Disadvantaged Grammar School Pupil Population

Although this section of the post is intended to describe the ‘present state’, I wanted to illustrate how that compares with the relatively recent past.

I attached to my 2011 post a table showing how the proportion of FSM students attending grammar schools had changed annually since 1995. This is reproduced below, updated to reflect more recent data where it is available

A health warning is attached since the figures were derived from several different PQ replies and I cannot be sure that the assumptions underpinning each were identical. Where there are known methodological differences I have described these in the footnotes.


Table 2: Annual percentage FSM in all grammar schools and gap between that and percentage FSM in all secondary schools, 1995-2013

Year PercentageFSM in GS Percentage FSMall schools Percentagepoint Gap
1995 3.9 18.0 14.1
1996 3.8 18.3 14.5
1997 3.7 18.2 14.5
1998 3.4 17.5 14.1
1999 3.1 16.9 13.8
2000 2.8 16.5 13.7
2001 2.4 15.8 13.4
2002 2.2 14.9 12.7
2003 2.1 14.5 12.4
2004 2.2 14.3 12.1
2005 2.1 14.0 11.9
2006 2.2 14.6 12.4
2007 2.0 13.1 11.1
2008 1.9 12.8 10.9
2009 2.0 13.4 11.4
2010 15.4
2011 2.4 14.6 12.2
2012 14.8
2013 15.1
2014 14.6

(1) Prior to 2003 includes dually registered pupils and excludes boarding pupils; from 2003 onwards includes dually registered and boarding pupils.

(2) Before 2002 numbers of pupils eligible for free school meals were collected at school level. From 2002 onwards numbers have been derived from pupil level returns.

(3) 2008 and 2009 figures for all schools exclude academies


Between 1996 and 2005 the FSM rate in all schools fell annually, dropping by 4.3 percentage points over that period. The FSM rate in grammar schools also fell, by 1.7 percentage points. The percentage point gap between all schools and selective schools fell by 2.6 percentage points.

Both FSM rates reached their lowest point in 2008. At that point the FSM rate in grammar schools was half what it had been in 1996. Thereafter, the rate across all schools increased, but has been rather more volatile, with small swings in either direction.

One might expect the 2014 FSM rate across all grammar schools to be at or around its 2011 level of 2.4%.

A more recent PQ reply revealed the total number of pupil premium recipients attending selective schools over the last three financial years:

  • FY2011-12 – 3,013
  • FY2012-13 – 6,184 (on extension to ‘ever 6’)
  • FY2013-14 – 7,353

(Hansard 20 January 2014, Col. WA88)

This suggests a trend of increasing participation in the sector, though total numbers are still very low, averaging around 45 per school and slightly over six per year group.


Comparison with FSM rates in selective authorities

In 2012, a table deposited in the Commons Library (Dep 2012-0432) in response to a PQ provided the January 2011 FSM rates for selective schools and all state-funded secondary schools in each authority containing selective schools.

In this case, the FSM rates provided relate only to pupils aged 15 or under. The comparable national average rates are 2.7% for selective schools and 15.9% for all state-funded schools.

  • Selective school FSM rates per authority vary between 6.0% in Birmingham and 0.6% in Wiltshire.
  • Other authorities with particularly low FSM rates include Bromley (0.7%), Reading (0.8%) and Essex (0.9%).
  • Authorities with relatively high FSM rates include Wirral (5.2%), Walsall (4.9) and Redbridge (4.8%).
  • The authorities with the biggest gaps between FSM rates for selective schools and all schools are Birmingham, at 28.0 percentage points, Liverpool, at 23.8 percentage points, Enfield at 21.8 percentage points and Wolverhampton, at 21.7 percentage points.
  • Conversely, Buckinghamshire has a gap of only 4.7 percentage points, since its FSM rate for all state-funded secondary schools is only 6.0%.
  • Buckinghamshire’s overall FSM rate is more than four times the rate in its grammar schools, while in Birmingham the overall rate is almost six times the grammar school rate. On this measure, the disparity is greatest in metropolitan boroughs with significant areas of disadvantage.


Proportion of disadvantaged learners in each selective school

I attached to my 2011 post a table setting out the FSM rates (all pupils, regardless of age) for each selective school in January 2009.

This updated version sets out the January 2013 FSM and disadvantaged (ie ‘ever 6 FSM’) rates by school, drawn from the latest School Performance Tables. (Click on the screenshot below to download the Excel file.)


GS excel Capture


Key points include:

  • The size of grammar schools varies considerably, with NORs ranging from 437 (Newport Girls’) to 1518 (Townley Girls’). The average NOR is slightly below 1000.
  • 24 of the 163 schools (14.7%) have suppressed FSM percentages. Since the lowest published percentage is 1.1%, the impact of suppression is that all schools at or below 1.0% are affected. Since no school returns 0, we must assume that all contain a handful of FSM learners. It is notable that six of these schools are in Buckinghamshire, three in Gloucestershire and three in Essex. Both Bromley grammar schools also fall into this category.
  • 67 selective schools (41.1%) have FSM rates of 2% or lower. The average FSM rate across all these schools is 3.25%.
  • The highest recorded FSM rates are at Handsworth Grammar School (14.4%), King Edward VI Aston School (12.9%) and Stretford Grammar School (12%). These three are significant outliers – the next highest rate is 7.8%.
  • As one would expect, there is a strong correlation between FSM rates and ‘ever 6’ rates. Most of the schools with the lowest ‘ever 6’ rates are those with SUPP FSM rates. Of the 26 schools returning ‘ever 6’ rates of 3.0% or lower, all but 7 fall into this category.
  • The lowest ‘ever 6’ rate is the 0.6% returned by Sir William Borlase’s Grammar School in Buckinghamshire. On this evidence it is probably the most socio-economically selective grammar school in the country. Five of the ten schools with the lowest ‘ever 6’ rates are located in Buckinghamshire.
  • A few schools have FSM and ‘ever 6’ rates that do not correlate strongly. The most pronounced is Ribston Hall in Gloucestershire which is SUPP for FSM yet has an ‘ever 6’ rate of 5.5%, not far short of the grammar school average which is some 6.6%. Clitheroe Royal Grammar School is another outlier, returning an ‘ever 6’ rate of 4.8%.
  • The highest ‘ever 6’ rates are in Handsworth Grammar School (27.2%), Stretford Grammar School (24.3%) and King Edward VI Aston School (20.3%). These are the only three above 20%.
  • In London there is a fairly broad range of socio-economic selectivity, from St Olave’s and St Saviour’s (Bromley) – which records an ‘ever 6’ rate of 2.5% – to Woodford County High School, Redbridge, where the ‘ever 6’ rate is 11%. As noted above, the FSM rates at the two Bromley schools are SUPP. The London school with the highest FSM rate is again Woodford County High, at 5%.

Another source throws further light on the schools with the lowest FSM rates. In October 2013, a PQ reply provided a table of the 50 state secondary schools in England with the lowest entitlement to FSM, alongside a second table of the 50 schools with the highest entitlement.

These are again January 2013 figures but on this occasion the rates are for pupils aged 15 or under and the only figures suppressed (denoted by ‘x’) are where no more than two pupils are FSM.

Sir William Borlase’s tops the list, being the only school in the country with a nil return (so the one or two FSM pupils who attend must be aged over 15 and may have been admitted directly to the sixth form).

The remainder of the ‘top ten’ includes eight selective schools and one comprehensive (Old Swinford Hospital School in Dudley). The eight grammar schools are:

  • Cranbrook, Kent – x
  • Adams’, Telford and Wrekin – x
  • St Olave’s and St Saviour’s, Bromley – 0.5%
  • Dr Challoner’s High Buckinghamshire – 0.5%
  • Dr Challoner’s Grammar, Buckinghamshire – 0.6%
  • Aylesbury Grammar, Buckinghamshire – 0.6%
  • Newstead Wood, Bromley – 0.6%
  • Pate’s, Gloucestershire – 0.6%

Comparing the data in my tables for 2009 and 2013 also throws up some interesting facts:

  • Some schools have increased significantly in size – Burnham Grammar School (Buckinghamshire), Sir Thomas Rich’s (Gloucestershire), Highworth Grammar School for Girls (Kent), Simon Langton Grammar School for Boys (Kent), Kesteven and Grantham Girls’ School (Lincolnshire), Carre’s Grammar School (Lincolnshire) and St Joseph’s College (Stoke) have all increased their NORs by 100 or more.
  • However, some other schools have shrunk significantly, notably The Skegness Grammar School in Lincolnshire (down 129), The Boston Grammar School in Lincolnshire (down 110), Fort Pitt Grammar School in Medway (down 132) and Slough Grammar School (down 175).
  • While recognising that the figures may not be fully comparable, there have also been some significant changes in the proportions of FSM pupils on roll. Significant increases are evident at King Edward VI Aston (up 5.9 percentage points), Fort Pitt (up 5.1 percentage points) and Handsworth Grammar (up 4.7 percentage points).
  • The only equally pronounced mover in the opposite direction is St Anselm’s College on The Wirral, where the FSM rate has more than halved, falling by 5.2 percentage points, from 9.8% to 4.6%.

Additional statistics were peppered throughout David Laws’ June 2014 speech.

He refers to a paper by DfE analysts which unfortunately has not been published:

  • In 2013, 21 grammar schools had fewer than 1% of pupils eligible for FSM. Ninety-eight had fewer than 3% eligible and 161 had fewer than 10% eligible. This compares to a national average of 16.3% across England. (The basis for these figures is not supplied but they more or less agree with those above.)
  • In Buckinghamshire in 2011, 14% of the year 7 cohort were eligible for the pupil premium, but only 4% of the cohort in Buckinghamshire grammar schools were eligible. In Lincolnshire the comparable percentages were 21% and 7% respectively.



Most commentary tends to regard the cadre of selective schools as very similar in character, leaving aside any religious affiliation and the fact that many are single sex establishments.

Although the fact is rarely discussed, some grammar schools are significantly more selective than others.

The 2013 Secondary Performance Tables show that only 10 grammar schools can claim that 100% of the cohort comprises high attainers. (These are defined on the basis of performance in statutory end of KS2 tests, in which they must record an APS of 30 or more across English, maths and science.)

At several schools – Clarendon House (Kent, now merged), Fort Pitt (Medway), Skegness (Lincolnshire), Dover Boys’ and Girls’ (Kent), Folkestone Girls’ (Kent), St Joseph’s (Stoke), Boston High (Lincolnshire) and the Harvey School (Kent) – the proportion of high attainers stands at 70% or below.

Many comprehensive schools comfortably exceed this, hence – when it comes to KS2 attainment – some comprehensives are more selective than some grammar schools.

Key variables determining a grammar school’s selectivity will include:

  • The overall number of pupils in the area served by the school and/or the maximum geographical distance that pupils may travel to it.
  • The number of pupils who take the entrance tests, including the proportion of pupils attending independent schools competing for admission.
  • The number of competing selective schools and high-performing comprehensive schools, plus the proportion of learners who remain in or are ‘siphoned off’ into the independent sector.
  • The number of places available at the school and the pass mark in the entrance tests.

I have been unable to locate any meaningful measure of the relative selectivity of grammar schools, yet this is bound to impact on the admission of disadvantaged learners.

An index of selectivity would improve efforts to compare more fairly the outcomes achieved by different grammar schools, including their records on access for disadvantaged learners.


Prior attainment data

In his June 2014 speech, Laws acknowledges that:

  • ‘A key barrier is the low level of free school meal pupils achieving level 5, typically a proxy for pupils you admit’.
  • However, in wholly selective areas fewer than 50% of FSM learners achieving Level 5 enter selective schools compared with two-thirds of non-FSM pupils:

‘We calculated it would require a shift of just 200 level 5 FSM pupils to go into grammar schools in wholly selective areas to remove this particular bias ‘

Alternative versions of this statement appear elsewhere, as we shall see below.

Using data from 2009/10 and 2011/12, the Sutton Trust study by Cribb et al explored whether advantaged and disadvantaged pupils with KS2 level 5 in both English and maths were equally likely to attend grammar schools.

They found that those not eligible for FSM are still more likely to attend. This applies regardless of whether the grammar school is located in a selective local authority, although the percentages and the gaps vary considerably.

  • In selective authorities, some 66% of these high attaining non-FSM pupils went on to grammar schools compared with under 40% of FSM pupils, giving a gap of over 26 percentage points. (Note that the percentage for FSM is ten percentage points lower than the one quoted by Laws. I can find no reason for this disparity, unless the percentage has changed dramatically since 2012.)
  • In isolated grammar schools outside London the gap is much smaller, at roughly 11 percentage points (18% non-FSM against 7% FSM).
  • In London there is a similar 12 percentage point gap (15% non-FSM versus 3% FSM)


Cribb Capture 1

A similar pattern is detected on the basis of KS2 maths test fine points scores:

‘Two points are evident. First, for any given level of maths attainment, pupils who are eligible for FSM have a noticeably lower probability of attending a grammar school. Indeed, a non-FSM student with an average maths score has the same probability of entering a grammar school as an FSM pupil with a score 0.7 standard deviations above average. Second, the gap in probability of attendance between FSM and non-FSM pupils actually widens substantially: non-FSM pupils with test scores one standard deviation above average have a 55% likelihood of attending a grammar school in selective local authorities, whereas similar pupils who are eligible for FSM have only a 30% chance of attending a grammar school. This is suggestive that bright pupils from deprived families are not attending grammar schools as much as their attainment would suggest they might.’

This rather calls into question Laws’ initial statement that level 5 performance among FSM pupils is ‘a key barrier’ to admission.

The study also confirms that pupils attending primary schools with relatively high levels of deprivation are much less likely to progress to grammar schools.

On the other hand, some 13% of pupils nationally transfer into selective schools from non-state schools and schools outside England. The researchers are unable to distinguish clearly those from abroad and those from the independent sector, but note that they are typically wealthier than state school transfers.

This masks significant variation between local authority areas.

Almost 34% of such pupils transfer in to grammar schools in Essex, as do 24% in Bromley, 23% in Wiltshire and 22% in Bournemouth and Southend. At the other extreme, only 6% are incomers in Kirklees.


Headteacher perceptions

The Sutton Trust released a parallel research report from NATCEN reporting the outcomes of interviews with a small sample of three primary school and eight grammar school headteachers.

The researchers found that:

  • Rightly or wrongly, many heads felt disadvantaged learners had relatively lower educational aspirations.
  • Disadvantaged parents were sometimes perceived to know less about grammar schools and place less value on the benefits they might confer.
  • Heads felt disadvantaged parents ‘often associated grammar schools with tradition, middle class values and elitism’. Parents felt their children ‘might struggle interacting with children from more affluent backgrounds’.
  • Grammar school heads highlighted the role of primary schools but ‘this was difficult when primary schools disagreed with assessment based entry processes and selective education in general’.
  • Heads felt grammar schools should provide more outreach and demonstrate their openness to everyone. It was suggested that, as grammar schools increasingly take in pupils from further away and/or from independent schools, this might further distance schools from their local communities.
  • It was widely acknowledged that learners from more advantaged backgrounds were coached to pass the entrance exams. Some grammar heads regarded tutoring as ‘good examination preparation’; others recognised it as a barrier for disadvantaged learners.
  • Although there are financial barriers to accessing grammar schools, including the cost of uniforms and school trips, grammar school heads claimed to deploy a variety of support strategies.


The preceding analysis is complex and difficult to synthesise into a few key messages, but here is my best effort.

The national figures show that, taken as a whole, the 163 grammar schools contain extremely low proportions of FSM-eligible and ‘ever 6’ learners.

National FSM rates across all grammar schools have fallen significantly over the past 20 years and, although the FSM gap between selective schools and all schools has narrowed a little, it is still very pronounced.

There is certainly a strong case for concerted action to reduce significantly the size of this gap and to strive towards parity.

The disparity is no doubt partly attributable to lower rates of high attainment at KS2 amongst disadvantaged learners, but high attaining disadvantaged learners are themselves significantly under-represented. This is particularly true of wholly selective authorities but also applies nationally.

Although the sample is small, the evidence suggests that grammar school and primary head teachers share the perception that disadvantaged learners are further disadvantaged by the selective admissions process.

However, the cadre of grammar schools is a very broad church. The schools are very different and operate in markedly different contexts. Some are super-selective while others are less selective than some comprehensive schools.

A handful have relatively high levels of FSM and ‘ever-6’ admissions but a significant minority have almost negligible numbers of disadvantaged learners. Although contextual factors influence FSM and ‘ever 6’ rates significantly, there are still marked disparities which cannot be explained by such factors.

Each school faces a slightly different challenge.

Transparency and public understanding would be considerably improved by the publication of statistical information showing how grammar schools differ when assessed against a set of key indicators – and identifying clear improvement targets for each school. 

There seem to me to be strong grounds for incorporating schools’ performance against such targets into Ofsted’s inspection regime.


Progress Towards Reform


The Sutton Trust Research

Although the Grammar School Heads’ Association (GSHA) argues that it has pursued reform internally for some years, a much wider-ranging initiative has developed over the last twelve months, kicked off by the publication of a tranche of research by the Sutton Trust in November 2013.

This included the two publications, by Cribb et al and NATCEN cited above, plus a third piece by Jesson.

There was also an overarching summary report ‘Poor Grammar: Entry into Grammar Schools for disadvantaged pupils in England’.

This made six recommendations which, taken together, cover the full spectrum of action required to strengthen the schools’ capacity to admit more disadvantaged learners:

  • Review selection tests to ensure they are not a barrier to the admission of learners from disadvantaged backgrounds. The text remarks that:

‘Some grammar schools and local authorities are already trying to develop tests which are regularly changed, less susceptible to coaching, intelligence-based and not culturally biased.’

  • Reduce the advantage obtained by those who can pay for private tuition by making available a minimum of ten hours of test preparation to all applicants on a free or subsidised basis.
  • Improve grammar school outreach support, targeting learners from low and middle income backgrounds. This should include: assurances on access to transport and support with other costs; active encouragement for suitable Pupil Premium recipients to apply; using the media to dispel notions that grammar schools are exclusive and elitist; and deploying existing disadvantaged students as ambassadors.
  • Using the flexibility within the Admissions Code (at this point available only to academies) to prioritise the admission of high achieving students who are entitled to the pupil premium. There is also a suggestion that schools might: 

‘…consider giving preference to students from low or middle income households who reach a minimum threshold in the admission test’.

though it is not clear how this would comply with the Code.

  • Develop primary-grammar school partnerships to provide transition support for disadvantaged students, enabling primary schools to provide stronger encouragement for applications and reassure parents.
  • Develop partnerships with non-selective secondary schools:

‘…to ensure that high achieving students from low and middle income backgrounds have access to good local teachers in their areas.’

The Sutton Trust also made its own commitment to:

‘…look at ways that we can support innovation in improved testing, test preparation, outreach, admissions and collaboration.

We will also commission independent analysis of the impact of any such programmes to create an evidence base to enhance fair access to grammar schools.’



Immediate reaction was predictably polarised. The GSHA was unhappy with the presentation of the report.

Its November 2013 Newsletter grumbles:

‘It is the way in which the research is presented by the Sutton Trust rather than any of research findings that give rise to concerns. Through a process of statistical machination the press release chose to lead on the claim that 6% of prep school pupils provide four times more grammar school pupils than the 16% of FSM eligible children. Inevitably, this led to headlines that the independent sector dominates admissions. The reality, of course is that 88% of all grammar school students come from state primary schools….

….Grammars select on ability and only 10% of FSM children reach level 5 at KS2 compared with a national average of 25%. The report, quite reasonably, uses level 5 as the indicator of grammar school potential. On the basis of this data the proportions of eligible FSM children in grammar schools is significantly greater than the overall FSM proportion in the top 500 comprehensives….

In 2012 just over 500 FSM children entered grammar schools. For the success rate of L5 FSM to match that of other L5 would require 200 more FSM children a year to enter grammar schools. Just one more in each school would virtually close the gap….

….The recommendations of the report are not, as claimed, either new or radical. All are areas that had already been identified by GSHA as options to aid access and represent practices that are already adopted by schools. This work, however, is usually carefully presented to avoid promotion of a coaching culture.

It is unfortunate that the press briefing both contributed to reinforcing the false stereotyping of grammar schools and failed to signal initiatives taken by grammar schools.’

There is evidence here of retaliatory ‘statistical machination’, together with a rather defensive attitude that may not bode well for the future.

On the other hand HMCI Wilshaw was characteristically forthright in the expression of an almost diametrically opposite opinion.

In December 2013 he is reported to have said:

‘Grammar schools are stuffed full of middle-class kids. A tiny percentage are on free school meals: 3%. That is a nonsense.

Anyone who thinks grammar schools are going to increase social mobility needs to look at those figures. I don’t think they work. The fact of the matter is that there will be calls for a return to the grammar school system. Well, look what is happening at the moment. Northern Ireland has a selective system and they did worse than us in the [international comparison] table. The grammar schools might do well with 10% of the school population, but everyone else does really badly. What we have to do is make sure all schools do well in the areas in which they are located.’


The Laws Speech

Liberal Democrat Education Minister David Laws made clear the Government’s interest in reform with his June 2014 speech, already referenced above.

Early on in the speech he remarks that:

‘The debate about grammar schools seems to have been put in the political deep freeze – with no plans either to increase or reduce the number of what are extremely popular schools in their localities.’

With the benefit of hindsight, this seems rather ignorant of (or else disrespectful to) UKIP, which had nailed their colours to the mast just three weeks previously.

Laws acknowledges the challenge thrown down by Wilshaw, though without attribution:

‘Are you, as some would have it, “stuffed full of middle-class kids”?

Or are you opening up opportunities to all bright children regardless of their background, or can you do more?

Why is entry to grammar schools so often maligned?’

He says he wants to work with them ‘openly and constructively on social mobility’, to ‘consider what greater role they can play in breaking the cycles of disadvantage and closing the opportunity gap’, while accepting that the Government and the primary sector must also play their parts.

He suggests that the Government will do more to increase the supply of high attaining disadvantaged learners:

‘…a key barrier is the low level of free school meal pupils achieving level 5, typically a proxy for pupils you admit. So this is not just a challenge for grammar schools, but for the whole education system…

….My promise to you, alongside my challenge to you, is that this government will do everything in its power to make sure that more children from poorer backgrounds achieve their full potential.’

He lists the policies that:

‘Taken together, and over time…will start to shift the dial for poorer children – so that more and more reach level 5’

leading of course with the pupil premium.

He also proposes aspirational targets, though without any timescale attached:

My ambition is that all selective schools should aim for the same proportion of children on free school meals in their schools as in their local area.

This would mean an additional 3,500 free school meal pupils in selective schools every year, or an additional 35,000 pupils over 10 years.’

In relation to the flexibilities in the Admissions Code he adds:

I am pleased to be able to say that 32 grammar schools have implemented an admissions priority for pupils eligible for free school meals this year….

We in the Department for Education will fully support any school that chooses to change its admissions criteria in this way – in fact, I want to see all grammar schools give preference to pupil premium pupils over the next few years.’

Similarly, on coaching and testing:

‘…I really welcome the association’s work to encourage a move to entry tests that are less susceptible to coaching, and I am heartened to hear that at least 40% of grammar schools are now moving to the introduction of coaching resistant tests.

Again, I hope that all grammar schools will soon do so, and it will be interesting to see the impact of this.’

And he adds:

I want all schools to build on the progress that is being made and seek to close the gap by increasing parental engagement, and stronger working with local primaries – with a focus on identifying potential.’

So he overtly endorses several of the recommendations proposed by the Sutton Trust seven months earlier.

A Sutton Trust press release:

‘…welcomed the commitment by Schools Minister David Laws, to widening access to grammar schools and making the issue a priority in government’.

This may be a little over-optimistic.

A Collaborative Project Takes Shape

Laws also mentions in his speech that:

‘The GSHA will be working with us, the Sutton Trust and the University of Durham to explore ways in which access to grammar schools by highly able deprived children might be improved by looking more closely at the testing process and what may be limiting the engagement of pupils with it.’

The associated release from the Sutton Trust uses the present tense:

‘The Trust is currently working with the King Edward VI Foundation, which runs five grammar schools in Birmingham, Durham University, the Grammar School Heads Association and the Department for Education to target and evaluate the most effective strategies to broaden access to grammar schools.

A range of initiatives being run by the Foundation, including test familiarisation sessions at community locations, visits from primary schools and support for numeracy and literacy teaching for gifted and talented children at local primary schools, will be evaluated by Durham University to understand and compare their impact. The resulting analysis will provide a template for other grammar schools to work with.’

We know that Laws had been discussing these issues with the grammar schools for some time.

When he appeared before the Education Select Committee in February 2014 he said:

‘We are trying, for example, to talk to grammar schools about giving young people fairer access opportunities into those schools.  We are trying to allow them to use the pupil premium as a factor in their admissions policy.  We are trying to encourage them to ensure that testing is fairer to young people and is not just coachable. ‘

The repetition of ‘trying’ might suggest some reluctance on the part of grammar school representatives to engage on these issues.

Yet press coverage suggested the discussions were ongoing. In May the GSHA Newsletter states that it had first met Laws to discuss admissions some eighteen months previously, so perhaps as early as November 2012.

It adds:

‘We are currently working on a research project with the DfE and the Sutton Trust to try to find out what practices help to reduce barriers to access for those parents and students from deprived backgrounds.’

A parallel report in another paper comments:

‘The grammar school heads have also gone into partnership with the education charity the Sutton Trust to support more able children from middle and lower income backgrounds applying to selective schools.

Other ideas being considered include putting on test familiarisation sessions for disadvantaged children – something they have missed out on in the past.’

While an entry on CEM’s website says:

‘Access Grammar:

This project seeks to look at ways access to grammar schools for highly able children from non-privileged backgrounds can be improved. The project will identify potential target cohorts in the study areas for a range of outreach interventions and will look to evaluate these activities. For this project, the CEM Research and Evaluation team are working in collaboration with the Sutton Trust, Grammar School Heads Association, King Edwards Foundation and the Department for Education.

Start date: January 2014
End date: January 2017.’

So we know that there is a five-way partnership engaged on a three year project, The various statements describing the project’s objectives are all slightly different, although there is a clear resemblance between them, the aims articulated by Laws and the recommendations set out by the Sutton Trust.

But I searched in vain for any more detailed specification, including key milestones, funding and intended outcomes. It is not clear whether the taxpayer is contributing through DfE funding, or whether the Sutton Trust  and/or other partners are meeting the cost.

Given that we are almost a year into the programme, there is a strong case for this material to be made public.


Progress on Admissions Criteria

Of the issues mentioned in the Sutton Trust’s recommendations – tests and test preparation, admissions flexibility, outreach and partnership with primary and non-selective secondary schools – those at the front of the list have been most prominent (though there is also evidence that the King Edward’s Foundation is pursuing reform across a wider front).

The GSHA’s May 2014 newsletter is less grumpy than its predecessor, but still strikes a rather defensive note.

It uses a now familiar statistic, but in a slightly different fashion:

‘The actual number of students with Level 5s in their SATs who either choose not to apply to a grammar school or who apply but do not receive a place is reckoned by GSHA and the DfE to be two hundred students a year; not the very large number that the percentages originally suggested.’

This is the third time we have encountered this particular assertion, but each time it has been articulated differently. Which of the three statements is correct?

The GSHA is also keen to emphasise that progress is being made independently through its own good offices. On admissions reform, the article says:

‘A significant number of schools 38 have either adopted an FSM priority or consulted about doing so in the last admissions round. A further 59 are considering doing so in the next admissions round.’

The GHSA was also quoted in the TES, to the effect that 30 grammar schools had already been given permission by DfE to change their admissions policies and would so with effect from September 2015, while a further five or six had already introduced the reform.

A November 2014 PQ reply updates the figures above, saying that 32 grammar schools have already prioritised disadvantaged learners in their admissions arrangements and a further 65 ‘intend to consult on doing so’.

That leaves 66 (40%) which are not giving this active consideration.

The Chief Executive of the GSHA commented:

‘“You won’t notice a dramatic change in schools themselves because the numbers are quite small…This is reaching out at the margins in a way that won’t deprive other people of a place. The real need is to raise the standard among free school meals pupils at Key Stage 1 and Key Stage 2, that’s the key issue.

“What we are looking at in the meantime is what we can do to help these free school meals pupils who want to come to grammar school.”

Mr Sindall said that many of the country’s 164 grammar schools would not change their policies because competition for places was less fierce and it would be unnecessary. Many schools were also increasing outreach programmes and some were running eleven-plus familiarisation sessions to help prepare poorer children for the test, he added.’

There is evidence here of a desire to play down the impact of such changes, to suggest that the supply of disadvantaged high achievers is too small to do otherwise.

The data analysis above suggests that almost all selective schools need to address the issue.

Between them, the various press reports mention admissions changes at several schools, including Rugby High, South Wilts, ‘a series of Buckinghamshire grammars including Sir William Borlase’s, Dr Challoner’s  and Aylesbury Grammar’, as well as the King Edward’s Foundation Schools in Birmingham.

I checked how these changes have been embodied in some of these schools’ admissions policies.

The reports indicated that Rugby was:

‘…going even further by reserving a fixed number of places for FSM-eligible children, so potentially accepting pupils with lower entrance exam scores than other applicants.’

Rugby’s admissions arrangements for 2015 do indeed include as a second overall admissions priority, immediately following children in care:

‘Up to 10 places for children living within the priority circle for children in receipt of Free School Meals whose scores are between one and ten marks below the qualifying score for entry to the school.’

South Wilts included FSM as an oversubscription criterion in its 2014 admission arrangements, replacing it with pupil premium eligibility in 2015. However, in both cases it is placed third after children in care and those living in the school’s designated [catchment] area.

Sir William Borlase’s goes one better, in that its 2015 admissions policy places children eligible for free school meals immediately after ‘children in care’ and before ‘children living in the catchment area of the school’, though again only in the oversubscription criteria.

The King Edward’s Foundation is pursuing a similar route to Rugby’s. It announced its intention to reform admissions to its five Birmingham grammar schools in April 2014:

‘The Government wishes to improve the social mobility of children in the UK and has urged selective schools to consider how their admission policies could be changed to achieve this. The King Edward VI Grammar Schools have applied to the Department for Education which can allow them to give preference in their policies, to children who are on free school meals, or have been at any point in the last six years…

… In addition the grammar schools will be offering familiarisation sessions which will introduce children from less privileged backgrounds to the idea of attending a grammar school and will encourage them to take the 11+.

All of the Grammar Schools have set themselves a target of a 20% intake of children on free school meals (Aston has already achieved this and has a target of 25%). The expansion of the grammar schools which was announced earlier this year means that these additional children will simply fill the additional space.’

According to the 2013 Performance Tables, the FSM rates at each of these schools in January 2013 were:

  • Aston – 12.9%
  • Camp Hill Boys – 3.6%
  • Camp Hill Girls – 5.3%
  • Five Ways – 2.6%
  • Handsworth Girls – 6.3%

There must have been a major improvement at Aston for the September 2013 admissions round. As for the other four schools, they must increase their FSM admissions by a factor of between 4 and 8 to reach this target.

I wonder whether the targets are actually for ‘ever 6’ admissions?

In the event, the Foundation’s applications encountered some difficulties. In July the Admissions Adjudicator was obliged to reject them.

A parent had objected on the grounds that:

‘…it is necessary to request financial information from parents to achieve this priority which is contrary to paragraph 1.9(f) of the School Admissions Code.

… The objector further feels that it is unclear, unfair and unreasonable to use the pupil premium to differentiate between applications when the school is oversubscribed.’

The Adjudicator found in favour of the parent on the technical grounds that, although the schools had applied for variations of their funding agreements to permit this change, they had only done so retrospectively.

However, in each case:

‘The school is now entitled to give priority to girls eligible for the pupil premium as the funding agreement has been amended.’

By August the Foundation was able to state that the issue had been resolved:

‘Children applying for a place at any of the King Edward VI Grammar Schools must now achieve a minimum “qualifying score” in the test to be eligible for entry.

Any Looked After Child or previously Looked After Child (a child who is or has been in the care of the Local Authority) who achieves the “qualifying score” will be given priority for admission for up to 20% of the available places (25% at Aston).

Children eligible for Pupil Premium (those who have been registered for Free School meals at any point in the 6 years prior to the closing date for registration, 11 July 2014) who achieve the “qualifying score” will also be given priority for admission.

After this allocation, children not eligible for the Pupil Premium but who achieve the “qualifying score” will be admitted by rank order of scores until all places are filled.’

The Foundation has published an interesting FAQ on the new arrangements:

‘Q5. Will this mean that if you are poor you won’t have to score as high in the 11+ admission tests?
A. That is essentially correct – up to 20% of places (25% at Aston) are set aside for pupil premium children who achieve “a qualifying score”. This qualifying score will be set before the test in September after we have reviewed data in order to ensure that children who achieve the score can flourish in our schools.

Q6. Why don’t you want the cleverest children at your school anymore?
 We want our schools to represent the City of Birmingham and the diverse backgrounds that our children might come from. We believe that there are clever children out there who just don’t have the same opportunity to succeed as those from more privileged backgrounds and we want to try to do something about that.’

It acknowledges the magnitude of the challenge ahead:

‘John Collins, Secretary to the Governors of the charity The Schools of King Edward VI in Birmingham said “This is a hugely challenging target which we do not expect to achieve in the first few years of the initiative, as currently there are relatively few free school meal pupils who apply to take the test. These low numbers are something we are trying to address with our “familiarisation” programme which seeks to encourage bright children from less privileged backgrounds to take the test.”’

Also in July the Government opened up the same possibility for grammar schools that are not academies by consulting on amendments to the Admissions Code to permit this.

In October this was confirmed in the Government’s response to the consultation which stressed it was being introduced as an option rather than a universal requirement.


Progress on 11+ Test Reform

The new-style 11-plus tests developed by CEM have not had a universally positive reception. Much of the attention has been focused on their adoption by Buckinghamshire grammar schools.

The GSHA’s May 2014 newsletter notes that ‘some schools in the Midlands’ have been using CEM tests for five years. From 2015, 40% of grammar schools will be using these tests, which are:

‘…designed to be immune to the influence of coaching’


‘The analysis of data from Buckinghamshire (a wholly selective area which has recently switched to the CEM Centre tests) will provide us in time with valuable hard data on the large scale impact of the change over time.’

Back in February 2014 an Observer article had already cited positive feedback from Buckinghamshire:

‘Last autumn, a handful of education authorities in England introduced an exam designed to test a wider range of abilities – ones that are already being taught in primary schools, rather than skills that can be mastered through home tutoring – to make the selection system fairer.

Provisional results indicate that a more diverse selection of pupils passed this test, and headteachers say they feel the change has made a difference.

Ros Rochefort, headteacher at Bledlow Ridge primary school in Buckinghamshire…said that this year, for the first time in her career, the test has delivered a fair result. “All the kids who got through were expected to pass and, as usual, there are a couple of appeals coming through. All our very able children were selected….

…. Philip Wayne, headteacher at Chesham grammar school and chairman of the Bucks Grammar School Heads Association, has welcomed the changes and says he is “very confident” that the new test will avoid the current situation, in which many pupils who won places at his school with the help of intensive tutoring struggle to keep up with lessons once they arrive.’

However, there were contemporary reports that the 2013 tests led to a 6% fall (110 fewer pupils) in the proportion of places awarded to children from in-county state primary schools, even though 300 more pupils applied.

In September this was further developed in a Guardian story:

‘According to the data, a child from a Buckinghamshire private school is now more than three and a half times more likely to pass the 11-plus than a child from one of its state primaries….

…FOI requests to the eight secondary schools in Wycombe, which includes some of the most deprived and diverse wards in the county, suggest that children on free school meals and of Pakistani heritage have been less successful this year. ‘

A local pressure group Local Equal and Excellent has been trying to gather and analyse the data from the initial rounds of testing in 2013 and 2014 (ie for admission in 2014 and 2015).

Their most recent analysis complains at refusals to publish the full test data and contains an analysis based on the limited material that has been released.

In November 2014, the matter was discussed at Buckinghamshire’s Education, Skills and Children’s Services Select Committee.

The ‘results and analysis’ paper prepared by Buckinghamshire’s grammar school headteachers contains many words and far too few numbers.

The section on ‘Closing the gap’ says:

‘One local group has claimed that children from poorer backgrounds and BME have ‘done worse’ in the new Secondary Transfer Test. It is not specified what ‘worse’ means; however it is not reliable to make statements about trends and patterns for specific groups from a single year’s data and as stated above the data that has been used to make such claims is a small subset of the total and unrepresentative. To substantiate such claims a detailed analysis of additional information such as the current attainment of the children concerned would be needed. We are currently considering how a longitudinal study might be achieved.’

This is overly defensive and insufficiently transparent.

There is some disagreement about whether or not the new test is less amenable to coaching.

The ‘results and analysis’ paper says:

‘There is no such thing as a ‘tutor proof’ test. However, the new tests are less susceptible to the impact of specific test tutoring because they are aligned to the National Curriculum which all children study. Additionally, the questions in the new test are less predictable than in the previous test because they cover a wider range of topics and there is a broader range of question types – points acknowledged and welcomed by primary headteachers’.

Conversely, the pressure group says:

‘The new 11-plus, devised by the Centre for Evaluation and Monitoring (CEM) at Durham University, is supposed to rely less heavily on verbal reasoning and be more closely allied to the primary curriculum. Practice papers for the CEM test are supposed to be less readily available…

But… the fact that it is modelled on what can be taught in schools means the CEM test is more amenable to coaching… if children can’t be taught to get better in maths, why are we teaching it in schools? Practice will make anyone better and I see no sign that tuition has tailed off at all.’

Elsewhere there is evidence that 11+ testing is not immune to financial pressures. North Yorkshire is presently consulting on a plan to scale back from a familiarisation test and two sets of two full tests, with the best results taken forward.

Instead there would be a single set of tests taken by all candidates on the same day at a single venue, plus sample booklets in place of the familiarisation test. A system of reviews, enabling parents to provide supporting evidence to explain under-performance, would also be discontinued.

The reason is explicit:

‘The cost of administering an overly bureaucratic system of testing is no longer sustainable in the light of very significant cuts in public expenditure.’

Even though the draft impact assessment says that the Council will consider applications for support with transport from rural areas and for those with low incomes, there is some unacknowledged risk that the new arrangements will be detrimental to efforts to increase the proportion of disadvantaged learners admitted to these schools.


How Best to Close Excellence Gaps


What to do with the status quo

The next Government will inherit:

  • The Access Grammar reform project, outlined above, which is making some progress in the right direction, but needs closer scrutiny and probably more central direction. There is an obvious tension between Laws’ aspiration that all grammar schools should ‘give preference to pupil premium pupils over the next few years’ and the GSHA position, which is that many schools do not need to change their policies. It will be important that the changes to admissions arrangements for the 163 schools are catalogued and their impact on admissions monitored and made public, so that we can see at a glance which schools are leading the pack and which are laggards. A published progress report against the Sutton Trust’s six recommendations would help to establish future priorities. Greater transparency about the project itself is also highly desirable.
  • A small cadre of selective 16-19 free schools. It will need to articulate its position on academic selection at 16+ and might need to take action to ensure a level playing field with existing sixth form colleges. It might consider raising expectations of both new and existing institutions in respect of the admission of disadvantaged learners, so securing consistency between 11+ selection and 16+ selection.
  • Flexibility within the Admissions Code for all grammar schools – academies and LA-maintained alike – to prioritise the admission of disadvantaged learners. It may need to consider whether it should move further towards compulsion in respect of grammar schools, particularly if the GSHA maintains its position that many do not need to broaden their intake in this fashion.
  • Flexibility for all grammar schools to increase Planned Admission Numbers and, potentially, to submit proposals for the establishment of Satellite institutions. The approval of such proposals rests with the local authority in the case of a maintained school but with the Secretary of State for Education in respect of academies. An incoming government may need to consider what limits and conditions should be imposed on such expansion, including requirements relating to the admission of disadvantaged learners.

It may be helpful to clarify the position on satellites. The Coalition Government has confirmed that they can be established:

‘It is possible for an existing maintained grammar school or academy with selective arrangements to expand the number of places they offer, including by extending on to another site…There are, however, limitations on that sort of expansion, meaning it could only be a continuation of the existing school. The school admissions code is written from a presumption that those schools with a split site are a single school’ (Hansard, 16 February 2012, Col. 184W).

In December 2013, a proposal to establish a grammar school annexe in Sevenoaks, Kent was rejected by the Secretary of State on the grounds that it would constitute a new school:

‘Mr Gove’s legal ruling hinged on the issue of a girls’ grammar school being the sponsor of a Sevenoaks annexe for both girls and boys. The planned entry of Sevenoaks boys to the annexe lead Mr Gove to rule that the annexe’s proposed admissions policy was sufficiently different to the sponsor school’s girls-only admissions policy to constitute a wholly new grammar school.’

But a revised proposal was submitted in November 2014 for a girls’ only annexe. Moreover, the local authority has committed to exploring whether another satellite could be established in Maidenhead, acknowledging that this would require the co-operation of an existing grammar school.

The timing of the decision on the revised Sevenoaks proposal ensures that selection will remain a live issue as we approach the General Election

Further options to promote between-school selection

There are several options for strengthening a pro-selection policy further that would not require the removal of statutory constraints on opening new 11-18 grammar schools, or permitting existing schools to change their character to permit selection.

For example:

  • Pursuing the Wilshavian notion of organising schools into geographical clusters, some with academic and others with vocational specialisms, and enabling learners to switch between them at 14+. In many areas these clusters will incorporate at least one grammar school; in others the ‘academic’ role would be undertaken by high-performing comprehensive schools with strong sixth forms. The practical difficulties associated with implementing this strategy ought not to be underplayed, however. For example, how much spare capacity would the system need to carry in order to respond to annual fluctuations in demand? How likely is it that students would wish to leave their grammar schools at 14 and what tests would incomers be expected to pass? Would the system also be able to accommodate those who still wished to change institution at age 16?
  • Vigorously expanding the cadre of post-16 selective free schools. There is presumably a largely unspent budget for up to twelve 16-19 maths free schools, though it will be vulnerable to cuts. It would be relatively straightforward to develop more, extending into other curricular specialisms and removing the obligatory university sponsorship requirement. Expansion could be focused on clones of the London Academy of Excellence and the Harris Westminster Sixth Form. But there should be standard minimum requirements for the admission of disadvantaged learners. A national network might be created which could help to drive improvements in neighbouring primary and secondary schools.
  • Permit successful selective post-16 institutions to admit high-attaining disadvantaged students at age 14, to an academic pathway, as a parallel initiative to that which enables successful colleges to take in 14 year-olds wishing to study vocational qualifications. It may be that the existing scheme already permits this, since the curriculum requirements do not seem to specify a vocational pathway.

UKIP’s policy, as presently articulated, is merely enabling: few existing schools are likely to want to change their character in this fashion.

One assumes that Tory advocates would be satisfied with legislation permitting the establishment of new free schools that select at age 11 or age 14. It seems unlikely that anyone will push for the nuclear option of ‘a grammar school in every town’… but Conservative Voice will imminently reveal their hand.


Further options to promote within-school selection

If the political preference is to pursue within-school provision as an alternative to between-school selection there are also several possibilities including:

  • Encouraging the development of more bilateral schools with parallel grammar and selective streams and/or fast-track grammar streams within standard comprehensive schools.
  • Requiring, incentivising or promoting more setting in secondary schools, potentially prioritising the core subjects.
  • Developing a wider understanding of more radical and innovative grouping practices, such as vertical and cluster grouping, and trialling the impact of these through the EEF.

It would of course be important to design such interventions to benefit all students, but especially disadvantaged high attainers.

The Government might achieve the necessary leverage through a ‘presumption’ built into Ofsted’s inspection guidance (schools are presumed to favour the specified approach unless they can demonstrate that an alternative leads consistently to higher pupil outcomes) or through a ‘flexible framework’ quality standard.


A national student support scheme

The most efficient method of supporting attainment and social mobility amongst disadvantaged high attainers is through a national scheme that helps them directly, rather than targeting the schools and colleges that they attend.

This need not be a structured national programme, centrally delivered by a single provider. It could operate within a framework that brings greater coherence to the existing market and actively promotes the introduction of new suppliers to fill gaps in coverage and/or compete on quality. A ‘managed market’ if you will.

The essential elements would include:

  • This supply-side framework, covering the full range of disadvantaged students’ learning and development needs, within which all suppliers – universities, third sector, commercial, schools-based – would position their services (or they would be excluded from the scheme).
  • A commitment on the part of all state-funded schools and colleges to implement the scheme with their disadvantaged high attainers (the qualifying criterion might be FSM or ‘ever 6’) – and to ensure continuity and progression when and if these students change institution, especially at 16+.
  • A coherent learning and development programme for each eligible student throughout Years 7-13. Provision in KS3 might be open access and light touch, designed principally to identify those willing and able to pursue the programme into KS4 and KS5. Provision in these latter stages would be tailored to individuals’ needs and continuation would be dependent on progress against challenging but realistic personal targets, including specified GCSE grades.
  • Schools and colleges would act as facilitators and guides, conducting periodic reviews of students’ needs; helping them to identify suitable services from the framework; ensuring that their overall learning programmes – the in-school/college provision together with the services secured from the framework – constitute a coherent learning experience; helping them to maintain learning profiles detailing their progress and achievement.
  • Each learner would have a personal budget to meet costs attached to delivering his learning programme, especially costs attached to services provided through the framework. This would be paid through an endowment fund, refreshed by an annual £50m topslice from the pupil premium budget (analogous to that for literacy and numeracy catch-up) and a matching topslice from universities’ outreach budgets for fair access.
  • Universities would be strongly encouraged to make unconditional offers on the basis of high quality learning profiles, submitted by students as part of their admissions process.
  • There would be annual national targets for improving the GCSE and A level attainment of students participating in the scheme and for admission to – and graduation from – selective universities. This would include challenging but realistic targets for improving FSM admission to Oxbridge.



The current political debate is overly fixated on aspects of the wider problem, rather than considering the issue in the round.

I have set out above the far wider range of options that should be under consideration. These are not necessarily mutually exclusive.

If I were advising any political party inclined to take this seriously, I would recommend four essential components:

  • An enhanced strategy to ensure that all existing selective schools (including 16+ institutions) take in a larger proportion of high-attaining disadvantaged learners. Approval for expansion and any new schools would be conditional on meeting specified fair access targets.
  • Development of the cadre of 163 grammar schools into a national network, with direct responsibility for leading national efforts to increase the supply of high-attaining disadvantaged learners emerging from primary schools. Selective independent schools might also join the network, to fill gaps in the coverage and fulfil partnership expectations.
  • A policy to promote in all schools effective and innovative approaches to pupil grouping, enabling them to identify the circumstances in which different methods might work optimally and how best to implement those methods to achieve success. Schools would be encouraged to develop, trial and evaluate novel and hybrid approaches, so as to broaden the range of potential methods available.
  • A national support scheme for disadvantaged high attainers aged 11-19 meeting the broad specification set out above.

Regrettably, I fear that party political points-scoring will stand in the way of a rational solution.

Grammar schools have acquired a curious symbolic value, almost entirely independent of their true purpose and largely unaffected by the evidence base.

They are much like a flag of convenience that any politician anxious to show off his right-wing credentials can wave provocatively in the face of his opponents. There is an equivalent flag for abolitionists.  Anyone who proposes an alternative position is typically ignored.



November 2014

The Politics of Setting

I had been intending never to revisit the difficult topic of setting, secure in the knowledge that I could not improve on my earlier treatment of the pros and cons.


Irrelevant picture of Norway by Gifted Phoenix

But recent developments have caused me to reconsider, led me to address the issue from a different perspective.

My previous post attempted an objective and balanced statement of the educational arguments for and against, drawing on the research evidence and taking account of all learners, regardless of their attainment.

This one explores how setting – just one option within the far wider range of so-called ‘ability grouping’ strategies – has been reflected in government policy and party political policy documents since 1997, culminating in the position we have  reached as the parties begin to firm up their 2015 General Election manifestos.

The post begins with brief notes on terminology and incidence.

The substantive text is organised into four sections:

  • How Labour government positions on setting developed and fluctuated between 1997 and 2010.
  • How the Conservative party turned to setting while in opposition.
  • How Coalition Government policy on setting has shifted since May 2010.

It concludes with a summary of the position we have reached as we approach the next Election, together with some ideas for how we might move forwards more constructively.

In case you prefer to read selectively I have included links to the relevant section from each of the bullet points above.



I take setting to mean grouping learners in a discrete class, ostensibly selected with reference to prior attainment in a specific subject.

It is distinct from streaming, where the selection – which may well be generic and ability-based – applies to teaching across a range of different subjects. The learners in a higher stream may not be higher attainers in each of these subjects.

One sometimes also encounters references to banding which is broadly synonymous with streaming, except that streaming tends to refer to a single class, while bands may include more than one class. It may therefore be a less differentiated form of streaming.

Both setting and streaming are within-school selection strategies, which may be adopted by selective or comprehensive schools. They may be perceived as viable alternatives to between-school selection which is no longer regarded as politically feasible by Labour, Conservatives or Liberal Democrats.

There is, however, continuing pressure from the right of the Conservative party and recently from UKIP for the restoration of grammar schools. The Coalition government has opened up the prospect of satellite establishments and overseen the introduction of several selective post-16 institutions. This might be viewed as the thin end of the wedge.

It has not always been possible to isolate the approach to setting since there is often a tendency to brigade it with streaming and/or a wider range of grouping strategies, occasionally including various approaches to within class grouping.

Sometimes these distinctions are clear and sometimes they are less so. To take a prominent example, the relevant entry in the Education Endowment Foundation’s Teaching and Learning Toolkit is not a model of clarity.

Called ‘Setting or Streaming’, it discusses effectiveness initially in terms of ‘ability grouping’ (first paragraph).

Clarity is not improved by the inclusion of the American terms for streaming (tracking) and setting (regrouping).

Nor is it clear whether ‘ability grouping’ is intended as a synonym for ‘setting or streaming’ or whether it has a broader scope.

The second paragraph reverts to ‘setting or streaming’ before discussing a wider range of interventions targeted at gifted and talented learners including several accelerative measures. One of these – promotion – is not necessarily a grouping strategy, at least as I understand the term.

The next three paragraphs relate to low attainers. The third focuses on ‘ability grouping’, although there is one reference is to ‘setting or streaming’, the fourth discusses both ‘setting’ and ‘ability grouping’, while the fifth mentions only ‘ability grouping’.

This terminological imprecision is confusing and unhelpful, especially when it appears in a text that purports to present the available research evidence clearly and unambiguously.


How prevalent is setting?

There are few recent and reliable statistics available on the incidence of setting.

Statistics deposited in the Commons Library in March 2012 (Dep 2012-0434) provide Ofsted data on the percentage of lessons observed in secondary schools that were either setted or streamed/banded for every year from 1996/97 to 2002/03, excluding PE.

In 2002/03, 40% of all secondary lessons observed were setted and 4% were streamed or banded.

From 2003/04 to 2010/11, the table provides percentages of lessons observed that were setted, streamed or banded, for ‘lower’, ‘average’ and ‘upper ability’ learners respectively.

In 2010/11, the average percentages across all year groups were 12% for average ability, 16% for lower ability and 17% for higher ability.

The reply to a PQ from July 2011 provides 2009/10 data, for maths, English and science in primary and secondary schools respectively. The percentages relate to ‘classes setted, streamed or banded by ability where pupils are placed within an ability range within the school’.

The average figures across all year groups are set out below. For primary schools I have included Year 6 percentages in brackets:

  • Maths primary 19% (34%)
  • English primary 11% (19%)
  • Science primary 2% (3%)
  • Maths secondary 71%
  • English secondary 58%
  • Science secondary 62%

A 2014 study of primary practice found that:

Approximately 17% of the pupils studied, who were born in 2000-2001, were in ability streams. Some 8% of the total group were in the top stream, 5% in the middle and 4% in the bottom stream.

Last year Ofsted estimated that, excluding PE, some 45% of secondary lessons were set or streamed. The TES story containing these figures notes:

‘The Department for Education was unable to produce statistics on how many students are set or streamed. Ofsted produced limited data based on lessons it had inspected… but stressed that “there is no way of using this data to draw out national conclusions in any way”….

…In comments accompanying Ofsted’s figures, Sir Michael noted that, since 2005, its inspections have not involved observing all teachers in a school. Lessons that were seen were not “necessarily representative” of the school or system as a whole, he said.

….”It is not possible to deduce from inspection data the proportions of pupils nationally who are taught in setted/streamed classes or in mixed-ability groups,” the chief inspector said.’

We can only conclude that a significant proportion of secondary students and older primary learners is setted and that that this practice is most prevalent in the core subjects. It is unclear whether these percentages are now increasing, stable or declining. 

It would be highly desirable to obtain more accurate figures through the School Census, if only to track the influence of the presentation of the evidence base in the Toolkit.


Part One: The evolution of Labour government policy from 1997 to 2010


First Labour Government

In 1997 the incoming Labour Government published its White Paper ‘Excellence in Schools’. The chapter on ‘Modernising the comprehensive principle’ said:

Mixed ability grouping… requires excellent teaching and in some schools has worked well. But in too many cases it has failed both to stretch the brightest and to respond to the needs of those who have fallen behind. Setting, particularly in science, maths and languages, is proving effective in many schools. We do not believe that any single model of grouping pupils should be imposed on secondary schools, but unless a school can demonstrate that it is getting better than expected results through a different approach, we do make the presumption that setting should be the norm in secondary schools. In some cases, it is worth considering in primary schools. Schools should make clear in reports to parents the use they are making of different grouping approaches. OFSTED inspections will also report on this.

The clear implication is that, where the quality of teaching is not excellent, setting is likely to prove relatively more effective than ‘mixed ability grouping’, particularly in science, maths and languages.

Setting will not be made compulsory in secondary schools, but there is a presumption that it should be ‘the norm’, presumably in all subjects but certainly in science, maths and languages, unless schools can show ‘better than expected results’ through a different approach. In primary schools, setting should be considered in some unspecified cases.

Ofsted will check what schools are doing (and presumably validate or otherwise any claim of ‘better than expected results’, although the precise meaning of this term is not explained).

The text also says that the Department will publish guidance and exemplification of best practice, taken from this country and abroad ‘in organising classes to meet the different abilities of pupils’. There is a list of strategies in which it has particular interest, including:

  • ‘target-grouping, where pupils are grouped by ability for part of the week and groups are altered in line with regular assessment;
  • fast-tracking, where pupils are encouraged to learn and take qualifications ahead of their age cohort.’

Early in 1999, Ofsted published a survey on setting in primary schools. I cannot source the text online, but contemporary reviews, such as this from the LGA, show that it was strongly supportive of the practice:

Setting, rather than streaming, in primary schools provides a powerful lever for raising standards, so long as it is carefully implemented and properly managed, say Her Majesty’s Inspectors from OFSTED.

A new survey of the practice of setting – grouping children by ability for specific subjects – uses evidence from OFSTED inspection data, from a questionnaire and from focused inspections by HMI. It endorses the government’s view that setting is well worth considering.

‘Where teachers understand its potential and modify their teaching techniques accordingly, setting can be a very successful way of organising teaching groups,’ HMI say in the report Setting in Primary Schools, published today by OFSTED.

They point out that setting does not, by itself, guarantee success in raising standards nor can it compensate for poor teaching. However, evidence from school inspections suggests that the quality of teaching in setted lessons in the three core subjects is slightly better than in lessons with the full ability range.’

This introduces two important themes – that the efficacy of setting is dependent on:

  • it being implemented and managed effectively and
  • the appropriate adaptation of teaching techniques.

In September 2000 a DfEE research report on ‘Innovative Grouping Practices in Secondary Schools’ summarises the advantages and disadvantages of ability grouping more generally, but consists mainly of extended case studies of contemporary innovative practice.

The introduction sets the context thus:

The challenge now is to find ways of grouping pupils and developing pedagogy that capitalises on the advantages and minimises the disadvantages outlined above. In other words, how can schools develop grouping plans to achieve the best attainment outcomes for pupils while minimising any negative impact?

This rather more pragmatic approach reappears in subsequent guidance documents, but was set aside when government policy was articulated.


Second Labour Government

The 2001 Green Paper ‘Schools Building on Success’ reverts to a bullish reference to setting in the section on KS3:

We want to see further increases in the extent of setting within subjects including express sets to enable those who are capable of doing so to advance beyond the levels set for their age and to take Key Stage 3 tests early.’

But this does not survive into ‘Schools Achieving Success’, the White Paper published the same year, which makes no reference to setting specifically or ‘ability grouping’ more generally..

A roughly contemporary PQ reply also hedges its bets:

The Government supports a flexible approach to pupil grouping, including setting by ability where appropriate’.

The sentence is vacuous because deliberately imprecise. Essentially it expresses the government’s preference for schools to decide their own approaches.

It seems that there is growing indecision over which line to take. Should the government opt for consistent and wholehearted endorsement, full devolution of responsibility to schools, or a middle path that focuses on developing and disseminating effective practice to meet the needs of different settings?

This is of course redolent of wider contemporary debate about the role of the government in determining education policy and practice.

Setting is not mentioned in the ‘Five Year Strategy for Children and Learners’ which appeared in 2004.


Third Labour Government

Setting makes a significant reappearance in the October 2005 White Paper ‘Higher Standards, Better Schools For All’:

‘Grouping students can help to build motivation, social skills and independence; and most importantly can raise standards because pupils are better engaged in their own learning. We have encouraged schools to use setting since 1997. Putting children in different ability groups within a class is commonplace in primary schools. Ofsted reports show that the proportion of Key Stage 3 lessons which are set has risen since 1997 to over a third now, with greater rises in English and maths. The significant majority of English, science and modern foreign language lessons in secondary schools, and about nine in ten maths lessons are already organised by setting.

It will continue to be for schools to decide how and when to group and set by ability. But we will encourage more schools to adopt such grouping and help them to learn from the innovative practices that some schools are already employing without lowering expectations for pupils in lower ability groups or limiting choices in the curriculum. We will publish, in the New Year, independent research into current best practice.

The first emboldened point implies a consistency that is not fully reflected in the narrative above, in that the encouragement for setting seems to have waned somewhat between 2001 and 2004.

The second emboldened section makes it clear that schools remain free to determine their own approaches. The presumption in favour of setting has gone by the wayside and the government will focus instead on encouragement through the continuing promotion of innovation and best practice.

Shortly afterwards, the research report ‘The Effects of Pupil Grouping: Literature Review’ appeared.

Back in 2010 I summarised its key findings thus:

  • 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.’

A contemporaneous TES story argues that the report undermines the government’s position by offering too little support for setting:

‘Setting pupils by ability, one of the most widely-trailed parts of last week’s white paper, has few benefits, a study funded by the Department for Education and Skills has concluded.

There is no evidence that streamed or set classes produce, on average, higher performance than mixed-ability classes, said the report. It also found that setting pupils is already widespread, particularly in maths….

It says the debate between setting and mixed-ability teaching has become polarised and does not reflect what happens in schools where a wide range of ways of grouping pupils is used….

…The review concluded: “There are no significant differences between setting and mixed-ability teaching in overall attainment … but … low-achieving pupils show more progress in mixed-ability classes and high-achieving pupils show more progress in set classes.’

This provides the spur for a renewed effort to push beyond the polarised debate, to refocus on helping to develop solutions to fit particular needs and circumstances

In 2006, DfES published ‘Pupil Grouping Strategies and Practices at Key Stage 2 and 3: Case Studies of 24 Schools in England’, a companion piece to the 2005 study.

The impact of grouping on pupil attainment were summarised thus:

  • Schools identified that the use of setting enabled them to tailor teaching for different ability pupils in order to impact on their understanding and achievement. However, the research did not find evidence to corroborate these expected achievement gains.
  • In secondary schools that adopted mixed ability or part mixed ability grouping approaches, the rationale given by teachers and senior managers tended not to make reference to attainment but rather to focus on the benefits in terms of social awareness and inclusivity. 
  • In primary schools, which used mixed ability as the predominant organisational grouping, pupils were often seated around tables on the basis of ability and it was not possible to differentiate attainment outcomes that related directly to setting or mixed ability from these observations.’

So advocates of secondary setting could not demonstrate stronger attainment overall, while advocates of secondary mixed ability teaching were not primarily concerned with the impact on attainment.

In the primary sector it was not possible to distinguish a differential impact on outcomes from either option.

In September of the same year, the National Strategies produced ‘Grouping Pupils for Success’, useful guidance for schools deciding on the most appropriate grouping strategies.

The introduction says that it:

‘…moves on from the old ‘for and against’ debates about grouping to a more sophisticated understanding of what it means to group pupils for success.’

Suggestions relating specifically to setting include:

  • ‘Make a careful match of individual teacher strengths with the nature of sets, for example placing a teacher experienced in challenging low attainers with the lowest set or band, to lift attainment.
  • Avoid ‘teaching to the middle’ in mixed-ability classes.
  • Monitor pupils’ learning to ensure that pupils have opportunities to demonstrate higher attainment, for example in tiered papers in the National Curriculum tests, and that access to the curriculum and resources are not limited by assumptions about ability level.
  • Ensure that teaching in top sets creates a learning atmosphere in which it is acceptable to make mistakes, to ask for clarification or repetition.
  • Develop inclusive teaching approaches, for example through differentiated questioning or the use of within-class groupings.’

It summarises the research on setting and mixed ability grouping respectively in the two tables reproduced below.


2014 Setting Capture 1

2014 setting Capture 2


Effective Teaching and Learning for Pupils in Low Attaining Groups’ (2007) takes the same line as the previous studies in arguing that:

‘…the polarisation of the grouping debate does not reflect the available evidence….Rather than pointing towards the overwhelming superiority of one form of grouping over another, it suggests that different forms of grouping are effective for different ‘types’ of pupils, in relation to different kinds of outcomes.’

But it continues:

‘The decision, therefore, about whether to group by attainment, either has to be seen as a matter of principle, where empirical evidence is of limited relevance, or else has to be regarded as one that is complex and may even be too close to call.

Nevertheless, the authors contribute some further empirical evidence to the debate, notably concerning the characteristics of pupils in low attaining sets:

  • ‘The analysis of data on pupils’ allocation to groups confirms prior attainment as the main, albeit a relatively poor predictor of set placement, for example, with over half the pupils with low prior attainment in English ending up in middle or high sets. Although prior attainment remains statistically significant, setting decisions are clearly not made on this basis alone.’
  • ‘Social class is a significant predictor of set placement. Pupils from higher socio-economic status (SES) backgrounds are more likely to be assigned to higher sets and less likely to be assigned to lower sets.
  • Special Educational Need (SEN) is a significant predictor of set placement (after controlling for social class and prior attainment), with these pupils concentrated in the low attainment sets. Less than 10% of pupils in the highest sets have SEN. This suggests that SEN and low attainment are seen as closely related or overlapping and that set placement may also be confounded by the effect of behaviour.
  • Ethnicity was a weaker significant predictor of set placement, (after controlling for social class and prior attainment), with pupils of Bangladeshi origin being slightly less likely to be selected for the higher sets.
  • Gender was not a significant predictor of set placement (after controlling for social class and prior attainment), except in Key Stage 2 literacy where, against recent trends, females were more likely to be placed in a low set. Overall, males are slightly overrepresented in the low sets and under-represented in the middle sets but this difference was not statistically significant.
  • Other factors including teacher assessments, teacher judgements and pupil characteristics such as behaviour are likely to influence set placement. Some schools allocated pupils with behavioural difficulties to high sets irrespective of prior attainment because they believed that the classroom context provided in these groups would promote positive behaviour. Other schools allocated these pupils to lower sets because they were smaller and provided higher staff ratios.’

Also in 2007, ‘The Children’s Plan’ included a section entitled ‘Good classroom practices – better use of grouping and setting’.

Essentially this replicates the approach taken in the 2005 White Paper, though the drafting is far more convoluted and so far less clear:

‘Improved understanding of each child’s progress should also lead to more effective use of group teaching. Since 1997 we have been encouraging schools to use ‘setting’ (teaching groups of pupils by ability in a particular subject rather than across a range of subjects) and other forms of pupil grouping, and we continue to encourage these practices.

Using setting and groups to teach children of similar abilities and interests can bring real educational benefits. But where it is poorly implemented, for example through ‘streaming’ (where pupils are grouped across a range of subjects based on general rather than subject-specific assessment) it can be socially divisive and detrimental to all but the highest achieving pupils. Grouping can also be used more effectively in the classroom – in particular, through proven approaches to in-class grouping by need, and guided group work when the teacher coaches a small group to apply immediately what they have been learning in the main part of the lesson. We will promote this best practice as standard practice.

Under this new formulation, there is recognition that there can be effective practice in setting and mixed ability grouping alike.

The final sentence potentially embodies a slightly different approach, by introducing the notion of promoting ‘standard practice’, but no further details are provided about what exactly this will entail.

Then Labour appears to lose interest in setting. A 2008 publication from DCSF ‘Personalised Learning: A Practical Guide’ includes a chapter on ‘Pupil Grouping’ but it says almost nothing about setting. It is as if the authors are keen to move on from what has become a rather sterile debate.

A PQ from March 2009 uses the Children’s Plan formulation:

‘Analysis of research suggests that no single model of pupil grouping will be of benefit to all pupils all of the time. For example, there is some evidence that being taught in a mixed ability class can be beneficial for low attainers, but that ability-based classes can be beneficial for high attainers.

We promote setting — the grouping of pupils according to their ability in a particular subject — as an effective way of ensuring that individual pupils are receiving personalised help appropriate to where they are in their learning. Similarly, we promote effective pupil grouping practices, and guided work, as tools for delivering the most appropriate curriculum to each individual in mixed ability classes.

We do not promote streaming—where pupils are assigned to classes on the basis of an overall assessment of their general ability and pupils remain in their streamed classes across the majority of subjects—as it assumes that children will have the same level of ability in all subjects.’

But, three months later, the 2009 White Paper ‘Your child, your schools, our future’ has nothing to say on the subject, and references to setting are conspicuously absent from the Pupil and Parent Guarantees. Ministers have apparently decided that schools are best left to their own devices.

Setting did not appear in Labour’s 2010 Election Manifesto either.


Part 2: The Conservatives in opposition

In January 2006, just at the time when Government research reports were discussing the polarised nature of debate on setting and advocating a more nuanced approach, David Cameron and David Willetts (then Tory education spokesman) both made statements that explicitly supported setting.

One report has Cameron saying:

‘I want no child held back, so my priority is not selection by ability between schools but setting by ability within schools, because every parent knows that a high quality education means engaging children at the right level.’

Another attributes to him the statement:

‘I want the Conservative Party to help me campaign in setting by each subject in every school so that we actually do what I think is common sense which is to help stretch the brightest pupils and help those who are in danger of falling behind…There’s a real case for more selection within schools rather than selection between schools.’

“The government is getting into a mess over the issues of selection and admissions.”

It seems that Cameron has identified support for setting as a means of distancing himself from calls within his party for the introduction of more selective schools.

Willetts said:

What I shall be looking for in the months ahead is how best to spread setting, and I would not rule out using central government more in this area…The evidence that setting works is powerful indeed, and yet you still have more than half of lessons not taught in sets, where you can target your teaching methods to children with a particular level of skill.’

Another report has a slightly different version:

We are not saying that an edict will go out from the Department for Education that schools are instructed to set in all circumstances but the empirical evidence is that it works.

I would not rule out ministers getting involved in the way schools organise setting, but our instincts are to cut back rather than add to central bureaucracy and direction.

One can see writ large the tension between the mutually exclusive desires for prescription and autonomy. Willetts is leaving open the possibility of central direction of some sort.

In the event, it was decided that Ofsted would be the enforcer. The November 2007 Conservative Green Paper ‘Raising the Bar, Closing the Gap’ made this clear:

‘While every pupil must be given the opportunity of a good education, we also recognise that each pupil should be given the opportunity to learn in accordance with their particular aptitude and ability, so that the brightest pupils continue to be stretched at the same time as pupils who might be struggling are given extra support.

We believe that setting by ability is the only solution to achieving this ambition. Labour’s 1997 manifesto acknowledged the importance of setting and implied that the amount of setting in schools would be increased significantly. This has not taken place.

… We believe that school children learn more effectively when taught with children of a similar ability. We also believe setting contributes to better behaviour. We will therefore alter guidance to Ofsted to ensure that schools – particularly those not performing at high levels – set all academic subjects by ability.’

Contemporary press reports remind us that Cameron had originally spoken of ‘a grammar school stream’ in every school, but streaming was set aside in favour of setting:

‘The Tories will now make clear that streaming need only apply in smaller schools with limited timetabling.’

Cameron has decided that setting is ‘the only solution’ and the inspection regime will impose this on all schools (the document is not clear whether primary schools are included). There is no explicit exemption for those ‘performing at high levels’ although they will be a lower priority.

This new position is a restatement of the Labour position of 1997.

Hansard shows that new opposition spokesman Michael Gove maintained an interest in the issue until at least summer 2009.

In May 2008 he requests the latest data on the extent of setting and DCSF’s guidance on the issue. Ofsted answers the first point while the reply directs him to the materials referenced above.

In July 2009 he again asks for updated data on the incidence of setting.

But the enforcement of setting through Ofsted was clearly set aside when the time came to consider the content of the 2010 Tory Election Manifesto.

For the time being at least, it seemed that the alternative attractions of full autonomy for schools had triumphed.


Part 3: The evolution of Coalition policy on setting


2010 to 2014

Indeed, the 2010 Schools White Paper made a virtue of schools’ autonomy in such matters.

‘We will expect schools to set their own improvement priorities. As long as schools provide a good education, we will not mandate specific approaches.

We…believe that it is often effective to incentivise improvement and innovative ideas, rather than to mandate a uniform approach.’

But it takes some time for any evidence of this approach in relation to setting.

A PQ from July 2011 asking for data and current guidance to schools elicits the statement that, while there is no guidance,

Case studies showing the effective use of setting in schools will be made available on the department’s website shortly.’

The task was not a high priority. Eight months later, in March 2012, the answer to the next PQ on the topic confirms that the material has now been published.

These case studies were not transferred to, but have been archived and are still available.

The covering article, dated 26 April 2012, reads:

Setting and other forms of pupil grouping are ways of tailoring teaching and learning for mixed-ability classes which can help raise standards.  When setting is done well it can be an effective way to personalise teaching and learning to the differing needs of groups of pupils.’

There are five case studies in all, two of secondary and three of primary schools. Each is no more than a page in length. Compared with some of the guidance produced by Labour administrations they are of relatively limited value.

But the issue was soon stirred up by the intervention of HMCI Wilshaw.

In September 2012, he referred to the issue obliquely, but in such a way that his comments could be interpreted to fit the very different political perspectives of the newspapers that carried his comments.

I can find no official Ofsted record of what he said.

One report offers this version:

‘Heads have got to make up their mind. If they want mixed-ability, then they have got to make sure there’s differentiated teaching. And we will be very critical when we inspect schools, particularly in the secondary sector, if we see mixed-ability without mixed-ability teaching.

He added: ‘If you have got a youngster with low basic skills sitting alongside a youngster with Oxbridge potential, then it is really important that that’s taken into account.’

Another provides a slightly different interpretation:

‘”Where there are mixed-ability classes, unless there is differentiated teaching… it doesn’t work,” he said, adding that effective differentiated teaching was “hugely difficult” to achieve.

He said mixed-ability classes could be an “article of faith” for schools who were not concerned enough about good practice and were doing something “more akin to social engineering”. In those cases Ofsted inspections would be “very critical”.’

A third suggests that Wilshaw expressly put some distance between his remarks and the setting controversy:

This is not a judgment on mixed ability as opposed to setting or streaming, it is saying where there are mixed ability classes unless there is differentiated teaching to groups of school children in the class, unless there are individual programmes of work, it doesn’t work,” he said,

“It is absolutely critical that if you have a youngster with low grades at school who struggles with literacy and numeracy sitting alongside a youngster with Oxbridge potential then it is really important that is taken into account and they are taught by people who are experienced in good teaching of mixed ability classes.”’

Conversely, a fourth was much more bullish:

‘Inspectors will now be critical of schools that do not differentiate between high and low achievers.

This could lead to schools falling into the new category of ‘requires improvement’ (which replaces the old ‘satisfactory’ description), or even being labelled ‘inadequate’…

Ofsted cannot force schools to adopt setting – grouping pupils according to their academic ability in single subjects – or streaming, where ability groups cover most or all subjects.

However, Sir Michael’s intervention is likely to make headteachers rethink their practice of mixed ability classes for fear of being marked down in future inspections

‘It’s a combination of low expectations of what these youngsters can achieve, that their progress is not sufficiently tracked, and what I would call and have done ever since I have been a teacher the curse of mixed-ability classes without mixed-ability teaching,’ he said.

The former head said mixed-ability classes did not work ‘unless there is differentiated teaching to groups of schoolchildren in the class’ and ‘individual programmes of work’….

…Many schools had recognised this and ‘moved towards setting arrangements’, he said.

It seems as though everyone heard what they wanted to hear.

Wilshaw’s fundamental point seems to echo the 1997 White Paper and some of Labour’s guidance material reviewed above.

His principal argument is that mixed ability settings require mixed ability teaching, and that effective mixed ability teaching is a difficult skill to master. He implies, but does not state explicitly, that teaching a narrower range of ability is comparatively easier.

He suggests that Ofsted will look askance at schools that adopt mixed ability teaching on ideological grounds,  that cannot justify it in terms of their learners’ achievement, or where the quality of teaching is insufficient to support it.

Seven months later, in April 2013, a peculiar story appeared in the TES called ‘Conservatives abandon pledge to enforce ability grouping’:

‘The practice of grouping classes by ability has long had strong backing from the top. Ofsted, the education secretary, the prime minister and their Labour predecessors have all encouraged schools to use setting in more lessons.

But, despite their rhetoric, Conservative ministers have quietly dropped a pledge to enforce setting by ability…

… Last September, Ofsted chief inspector Sir Michael Wilshaw appeared to support the call, warning that some students were being held back by “the curse of mixed-ability classes without mixed-ability teaching”, adding that such teaching was “hugely difficult” to achieve.

But the government has now said that it does not advocate setting. “It is for schools to decide how best to organise teaching – including whether to group and set pupils by ability – as they know exactly what their students need,” a spokesman said.

And Ofsted says it “doesn’t have a view on whether setting or streaming is a good idea or not”. A spokeswoman for the inspectorate also revealed that Conservative ministers had not asked Ofsted to enforce setting.’

This is odd, since any Conservative adherence to the enforcement of setting would date back to 2007. I can find no more recent commitment than that. So why overtly drop a policy that no-one could reasonably have assumed the Conservatives still to advocate?

It is as if ministers are determined to re-impose the position on autonomy reached back in 2010, which has been compromised by Wilshaw’s insistence on linking schools’ decisions to Ofsted’s assessment of their performance.

Given more recent history, it is also conceivable that ministers were on the receiving end of pressure from the Prime Minister’s Office to adopt a more interventionist approach. Perhaps this was their way of distancing education ministers from such pressure.

But Ofsted’s alleged neutrality on the question of setting was soon called into question again when in June 2013 it published  ‘The Most Able Students’.

This developed HMCI Wilshaw’s theme:

‘In around a third of the schools visited, students were taught mainly in mixed ability groups throughout Key Stage 3. Where setting by ability occurred at an early stage, this was usually only for mathematics. Sets were introduced at various times for English and science, but often only in the later stages of Key Stage 3.

For most other subjects, mixed ability classes were retained throughout Key Stage 3. In the very best schools, this did not appear to have a detrimental impact on students’ progress because the teaching was carefully planned and well matched to the most able students’ needs. In the less effective schools, the work was pitched at the level of the average-attaining students. It was not challenging enough for the most able and their progress was insufficient…

…It was evident in some of the schools visited that school leaders had responded to recent research findings about mixed ability teaching, particularly in Key Stage 3. Eight of the schools had moved recently towards grouping by ability, particularly in English, mathematics and science. Some other school leaders recognised that their earlier grouping arrangements had not always promoted the best outcomes for the most able students. They indicated that they were moving away from mixed ability teaching to setting, streaming or banding in most subjects. Schools’ data shown to inspectors during the visits indicated that these moves were beginning to have a positive impact on outcomes for the most able students.’ 

Although Ofsted may not have an official view on the desirability of setting, it is abundantly clear that schools are encouraged to consider it where the quality of mixed ability teaching provided is not sufficiently strong to secure commensurate outcomes for able learners.

The current iteration of the inspection handbook says:

Inspectors should consider how effectively pupils are grouped within lessons and across year groups. For example:

  • where pupils are taught in mixed ability groups/classes, inspectors will consider whether the most able are stretched and the least able are supported sufficiently to reach their full potential. 
  • where pupils are taught in sets, inspectors will consider how leaders ensure that pupils in lower sets are not disadvantaged or that teachers take into account that pupils within a set may still have very different needs. ‘

The associated grade descriptors for an inadequate school mention:

‘The organisation of the curriculum and classes is resulting in some pupils achieving less well than they should.’

This carefully balanced approach makes it clear that inspectors will consider equally seriously the efficacy of sets and mixed ability groups.

Schools are likely to be pulled up if their mixed ability settings are insufficiently challenging for high attainers, but will also be challenged if their sets are holding back low attainers or if teaching within sets is insufficiently differentiated.


Developments in Autumn 2014

This careful balance was once more disturbed.

On 3 September, the Guardian reported that:

‘Compulsory setting according to ability in England’s secondary schools is to be proposed by the education secretary, Nicky Morgan, in her first big initiative since she took the role in July. She is due to make the announcement as early as today.’

This shock introduction was immediately undermined by a subsequent paragraph indicating that setting would not be compulsory after all, though it would be incentivised through the inspection regime:

‘It is expected that Morgan will ask the education watchdog, Ofsted, to implement and enforce the measure, probably by making it a condition of receiving an outstanding rating‘.

So the strategy would be to prevent schools from receiving the top inspection rating if they had not adopted setting.

The piece also stated unequivocally that the policy had been ‘cleared with Downing Street’, although ‘The Department for Education gave no comment after being contacted.’

This implied that the source of the story was the Prime Minister’s Office.

Just six hours later, the same paper carried a second report featuring comments made by Morgan in a Commons debate that same afternoon:

‘Richard Fuller: The Secretary of State has faced a number of confusing interventions from Opposition Members, one of which repeated something that was said in The Guardian today, which was that she was about to announce a policy a compulsory setting. Will she take this opportunity to say whether she is going to do that?

Nicky Morgan: Let me confirm for the benefit of the House that there is absolutely no truth in those rumours. There are some people outside this House who have a rather unhealthy interest in speculating about what I am or am not about to announce. They would be better served if they spent less time on Twitter and talking to journalists, and more time reflecting on the importance of the policies and reforms that have already been implemented by this Government.’ (Hansard, 3 Sep 2014, Col 357)

So as not to appear entirely wrong-footed, the Guardian cited Dominic Cummings in support of its original story:

Gove’s former special adviser Dominic Cummings said he had been told Cameron wanted to back compulsory setting.

He added on Twitter: “I was told by No 10 and two others in Whitehall a version v close to the Guardian story. Some had warned internally it was mad.” He also suggested there was a launch plan prepared inside No 10.’

Cummings’ Twitter feed on the day in question is instructive:



A BBC report included comment from the Lib Dems, confirming that they would not support a Coalition policy along these lines:

‘A senior Liberal Democrat source also distanced the party from any such proposal.

“This has not been agreed by the Liberal Democrats and is not government policy. We do not think it would be appropriate to tie schools’ hands in this way.”’

And Labour in opposition took a similar line:

‘Labour’s shadow education secretary Tristram Hunt had called on the education secretary to reject political involvement in such school decisions.

“I believe that excellent heads and great teachers know better than Westminster politicians how to deliver the best schooling for all pupils.

“We thought there was political consensus on the importance of school autonomy.’

The Cummings version lends support to the idea that some sort of enforcement of setting remains under consideration for inclusion in the Conservative Election Manifesto for 2015.

It might once again help to pacify those in the Party who seek a renewed commitment to selective education. Conservative MPs will be acutely aware of UKIP’s declared policy:

‘Existing schools will be allowed to apply to become grammar schools and select according to ability and aptitude. Selection ages will be flexible and determined by the school in consultation with the local authority.’

I could find no explicit statement to the effect that a commitment to introduce setting would definitely not be in the 2015 Manifesto. The final paragraph of a related TES story claimed this was the case, but this is not supported elsewhere.

While there was no reference to setting in Morgan’s speech to the Conservative Party Conference, the idea has subsequently reappeared in a different guise.

On 12 October the Conservative party let it be known that their Manifesto would include plans to enable Regional Schools Commissioners to intervene directly in the operation of any school rated inadequate by Ofsted, whether or not an academy.

The briefing made an explicit link with setting:

‘A Conservative spokesperson said the new powers would be developed in “consultation with Ofsted and the Education Endowment Foundation”, but a “menu of options” might include forcing schools to put children into classes based on ability, or ‘sets’ as they are also known.’  (Academies Week).

So, rather than making setting a condition of an ‘outstanding’ Ofsted rating, this possible new approach is to empower RSCs to impose setting on inadequate schools.

Whether the inclusion of setting in the menu of options would survive the consultation process is open to question – and presumably RSCs would also be reluctant to impose it without hard evidence that it would radically improve the performance of an inadequate school. Such evidence would be hard to find.

Perhaps this is a method of parking the issue:  giving No 10 the impression that enforcement of setting is part of the agenda when in fact it is not.

Meanwhile, the DfE has restated its existing commitment to giving schools autonomy in this matter. On 30 October, a Conservative MP tabled a PQ:

‘Andrew Rosindell (Romford):

To ask the Secretary of State for Education, what steps her Department is taking to ensure that children at secondary school are being efficiently grouped according to their academic ability.

Answered by: Mr David Laws

The Department for Education believes that individual schools are best placed to determine whether and how to group children by academic ability. There are many different models of pupil grouping, and schools themselves are best able to respond to their individual circumstances to meet the needs and capabilities of their pupils.

Note that the reply refers to the DfE’s belief rather than the Government’s position.

This suggests that we may not have heard the last of the matter, especially if setting remains part of the Prime Minister’s strategy for buying off the siren voices calling for renewed commitment to grammar schools.


Part 4: The Education Endowment Foundation’s Evidence Base

The Education Endowment Foundation (EEF) exists to improve the achievement of disadvantaged learners. The website says:

‘We aim to raise the attainment of children facing disadvantage by:

  • Identifying and funding promising educational innovations that address the needs of disadvantaged children in primary and secondary schools in England;
  • Evaluating these innovations to extend and secure the evidence on what works and can be made to work at scale;
  • Encouraging schools, government, charities, and others to apply evidence and adopt innovations found to be effective.’

But, confusingly, it has also been designated jointly with the Sutton Trust as a What Works Centre for improving educational outcomes for all school age children:

‘The What Works centres will summarise and share research with local decision-makers, helping them to invest in services that deliver the best outcomes for citizens and value-for-money for taxpayers.

In the EEF’s case, decision-makers include teachers and school-leaders, parents and governors, researchers and policy-makers. They are the primary audience for our Teaching and Learning Toolkit, an accessible summary of educational research which provides guidance for teachers and schools on how to use their resources to improve the attainment of disadvantaged pupils. ‘

See the logical disconnect? The principal tool used by the EEF/Sutton Trust to inform decision makers about what works well with all learners has been designed to inform decisions about what works well with disadvantaged learners.

This is particularly problematic when it comes to setting.


 The Teaching and Learning Toolkit

The EEF’s website describes the Toolkit as follows:

‘The Sutton Trust-EEF Teaching and Learning Toolkit is an accessible summary of educational research which provides guidance for teachers and schools on how to use their resources to improve the attainment of disadvantaged pupils.

The Toolkit currently covers 34 topics, each summarised in terms of their average impact on attainment, the strength of the evidence supporting them and their cost.’

One of the 34 topics is ‘Setting or streaming’. This pairing is potentially problematic since the subsequent commentary does not consistently distinguish the impact of one from the other.

I have already described above how the guidance switches between setting, streaming, ability grouping and wider gifted and talented provision.

When it comes to quantification, the Toolkit arrives at an average impact measure of -1 month – ie in terms of average pupil progress over a year, the impact of ‘setting or streaming’ on disadvantaged learners is negative.

The description of the Toolkit notes:

‘Most approaches included in the Toolkit tend to have very similar average impacts on pupils with different characteristics. However, where the research summarised suggests that an approach has a different average impact on the learning of pupils from disadvantaged backgrounds compared to the learning of their peers, the Toolkit’s ‘headline’ average impact figure refers to the former.’

The section describing the impact of ‘setting or streaming’ begins:

Overall, ability grouping appears to benefit higher attaining pupils and be detrimental to the learning of mid-range and lower attaining learners. On average, ability grouping 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.’

It continues:

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.’

No reference is made to the plight of disadvantaged high attainers, who might be expected to benefit commensurately.

The impact of setting and streaming remains undifferentiated.

The next section of the commentary considers a wider range of grouping interventions targeted on gifted and talented learners. This does not seem directly relevant to the narrower case of ‘setting or streaming’.

The final section of the commentary is concerned with low attainers (and so, by implication, includes the majority of disadvantaged learners).

It says:

Low attaining learners fall behind by one or two months a year, on average, when compared with the progress of similar students in classes without ability grouping. It appears likely that routine setting or streaming arrangements undermine low attainers’ confidence and discourage the belief that attainment can be improved through effort. Research also suggests that ability grouping can have a longer term negative effect on the attitudes and engagement of low attaining pupils. It should be noted that there are some exceptions to this average, where ability grouping has benefitted all learners. Further study could be undertaken to understand what happened differently in these examples.

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 ability grouping 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.

Though the average impact of ability grouping on low attaining pupils is negative, evidence suggests that certain types of ability grouping are more effective than others. Some studies have shown that reducing the size of the lowest attaining groups and assigning high-performing teachers to these groups can be effective, as can providing additional targeted catch up support.’

So the text suggests that high attaining learners make between one and two months more progress in sets/streams, while low attaining learners fall behind by the same amount. There is, therefore, between two and four months’ difference in the impact on high and low attainers respectively.

But this commentary:

  • Is not providing sufficiently accurate information to enable us to distinguish the impact of setting alone, as opposed to ‘setting or streaming’ together. 
  • Neglects the interests of high-attaining disadvantaged learners who are assumed to be an insignificant minority. 
  • Is fundamentally unclear, a particularly heinous crime considering the purpose of the Toolkit.


The KCL Study

One of the projects funded by the EEF is examining Best Practice in Grouping Students. The four-year project began in 2014 and continues through to spring 2018. It is co-ordinated by a team based at King’s College London and is receiving a total of £1.184m. The evaluation has been assigned to the NFER.

The project summary on EEF’s website distinguishes two parallel strands:

  • A randomised control trial of an intervention ‘which trains schools in a best practice approach to setting’ and is focused on English and maths in Years 7 and 8. The trial begins in September 2015, but results are not expected until spring 2018. It will be conducted in a sample of 120 schools, randomly allocated to conduct the trial or form part of the control group.
  • A pilot of an intervention ‘to introduce mixed ability teaching to secondary schools’ and ‘examine whether it is possible to overcome the common barriers to mixed ability teaching’. The intervention will be developed initially with three schools but the pilot will subsequently be extended to ten. Results are due in spring 2017.

One of the descriptions on the King’s College site suggests that the focus is explicitly placed on lower sets and low attainers:

The project addresses the needs of pupils in low ‘ability’ sets and streams, wherein research has identified socially disadvantaged pupils are strongly over-represented.

The project draws on substantial existing research evidence (concerning the educational outcomes for young people in low sets and streams, and the related poor practice often associated with such groupings), as illustrated in the Education Endowment Foundation/Sutton Trust Toolkit and elsewhere. The evidence from the literature concerning existing bad practice and detrimental outcomes associated with low ‘ability’ groups establishes areas for potential improvement, which will be applied via the interventions.’

It adds that the trial will measure the impact on pupil attainment, noting that the developmental phase:

‘Will also allow us to research why schools and policy-makers appear so wedded to ‘ability’ grouping, and what might persuade the adoption of an evidence-based approach.’

A second description confirms the focus of the setting strand on lower sets:

One, on Best Practice in Setting, seeks to remedy the detrimental elements identified by research to be associated with low sets.’

This also mentions that a pilot study – it is not clear whether this is of one strand or both – is being undertaken in six schools in the current academic year. It states that the full study will involve around 100 schools (rather than 120) and will be completed in 2017 (rather than spring 2018).

The exclusive emphasis on low sets is directly contradicted in a TES story about the project:

The Education Endowment Foundation, which commissioned the study, said the research aimed to address poor practices in both low and high sets.’

Is there a difference of opinion between KCL and the EEF? It would be helpful to know the truth, since there is otherwise strong reason to believe that the needs of high-attaining disadvantaged learners will be neglected.

NFER’s description of its evaluation currently says that the protocols for both strands are not yet agreed. Hopefully they willl be clear on whether the operation of higher sets – and the impact on disadvantaged learners within them – is also part of the agenda.


What Makes Great Teaching?

On 31 October 2014, the Sutton Trust published ‘What Makes Great Teaching?

One assumes that it has done so in its role as partner with the EEF in the What Works Centre, rather than as a charity supporting social mobility.

The press release dictated much of the media coverage. Provocatively headed:

‘Many popular teaching practices are ineffective, warns new Sutton Trust report’

It begins:

‘Lavish praise for students is among seven popular teaching practices not supported by evidence, according to a new Sutton Trust report which reviews over 200 pieces of research on how to develop great teachers.

What Makes Great Teaching, by Professor Rob Coe and colleagues at Durham University, warns that many common practices can be harmful to learning and have no grounding in research. Examples include using praise lavishly, allowing learners to discover key ideas by themselves, grouping students by ability and presenting information to students based on their “preferred learning style”.’

Later on the press release lists seven ‘examples of strategies unsupported by evidence’. Third in the list is:

Grouping students by ability. Evidence on the effects of grouping by ability, either by allocating students to different classes, or to within-class groups, suggests that it makes very little difference to learning outcomes. It can result in teachers failing to accommodate different needs within an ability group and over-playing differences between groups, going too fast with the high-ability groups and too slow with the low.’

The Report itself does not include this list in its executive summary. It appears in a section called ‘Examples of Ineffective Practices’. But the text repeats more or less verbatim the claim in the press release

The following are examples of practices whose use is not supported by the research evidence…

Group learners by ability

Evidence on the effects of grouping by ability, either by allocating students to different classes, or to within-class groups, suggests that it makes very little difference to learning outcomes (Higgins et al, 2014). Although ability grouping can in theory allow teachers to target a narrower range of pace and content of lessons, it can also create an exaggerated sense of within-group homogeneity and between-group heterogeneity in the teacher’s mind (Stipek, 2010). This can result in teachers failing to make necessary accommodations for the range of different needs within a supposedly homogeneous ‘ability’ group, and over-doing their accommodations for different groups, going too fast with the high-ability groups and too slow with the low.’

The first reference to grouping by ability making ‘very little reference to learning outcomes’ is to the Toolkit, though the report’s bibliography attributes it to 2013 not 2014. The second reference – ‘Stipek 2010’ – inexplicably appears in the bibliography under D rather than S.

As far as I can see, this is a reference to an article – an excerpt from a 2002 book called Motivation to Learn: Integrating Theory and Practice – that cites a series of other studies dating between 1976 and 1998.

Is the opening sentence an accurate description of what the Toolkit says?

As we have seen, the Toolkit considers ‘setting or streaming’ – though it also mentions a range of other strategies targeted at gifted and talented students – but it doesn’t discuss substantive evidence relating to within-class groups.

The only reference to them comes at the end of the Toolkit entry, in the section ‘What should I consider’. It says:

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

But that doesn’t support the statement above. (Nor, for that matter, is it supported by the evidence in the earlier parts of the Toolkit text.)’

The differential impact of setting on high and low attainers is not mentioned.

How might this statement be improved to reflect the evidence? It might say:

  • When discussing evidence on the effectiveness of ability grouping, it is important to distinguish the impact of setting, streaming and within class ability grouping respectively.
  • It is also important to distinguish the differential impacts on high attainers and low attainers respectively. Great care should be taken to clarify whether the discussion relates to all learners or only to disadvantaged learners. The subset of low attainers ought not to be regarded as analogous with the subset of disadvantaged learners.
  • The evidence suggests the overall impact of setting or streaming – ie one or the other – on low attainers is negative (one to two months) whereas the impact on high attainers is positive (one to two months). There is therefore a difference of up to four months’ progress between high and low attainers respectively.
  • There is less evidence on the differential impact of setting and streaming respectively. What we do know is x.
  • The impact of setting varies according to prior attainment of the learners, the subject of study and how well it is implemented. The available evidence suggests that setting is most likely to be successful under the following conditions….and, conversely, is least likely to be successful when….


Where we are now – and future prospects


The Evidence Base

The arguments about the advantages and disadvantages of setting have long been polarised – and there is little evidence to suggest that this will change as we head into 2015.

The EEF/Sutton Trust nexus purports to stand for evidence-based pedagogy, but both arms of the partnership are too vague and unspecific in how they present this evidence.

Because they take short cuts, it is too easy to interpret their coverage as overwhelmingly negative towards setting. A more careful and nuanced presentation would highlight the different contexts where setting might be more and less likely to operate effectively.

As things stand, the standard bearers for evidence-based practice seem more inclined to perpetuate the polarisation of views, rather than promoting a more sophisticated understanding of the issue.

This may be a deliberate reaction to the unevidenced espousal of setting by politicians, or it may just be insufficient attention to detail.

Substantive amendment of the toolkit entry – along the lines set out above – is devoutly to be wished for.

And it should be accompanied by a commitment to produce and update accurate data about the incidence of setting by sector, type of school, subject and pupils’ prior attainment. The Schools Census is the perfect vehicle.

One hopes that the results from the KCL study will be more carefully presented, but the absence of evaluation protocols and the disagreements over the focus of the study are a cause for concern. The protocols should be finalised and published forthwith.

The KCL study is unlikely to reveal that best practice in setting has a substantial impact on improvements in the performance of disadvantaged learners, even the high attainers.

But everything is relative: hardly any of the studies of other interventions so far completed by the EEF have identified a significant positive effect.

I would be satisfied with confirmation of a limited but positive impact on the performance of disadvantaged high attainers, combined with recognition that any negative impact on disadvantaged low attainers can potentially be eliminated through effective practice.

Some recommendations for the implementation of hybrid approaches – perhaps combining a single top set with several parallel mixed ability groups – wouldn’t go amiss.

Any evidence that does emerge from the KCL study – positive or negative – will not appear until well after the 2015 Election.

For the future, we anticipate keenly the pronouncements on setting that will emerge from a College of Teaching and/or from the Liberal Democrat’s independent Education Standards Authority. There is no reason to believe that they will be any more inclined to withstand the ideological pressures than their predecessors.


The Policies

Labour and Liberal Democrat politicians seem wedded to the notion of full autonomy for schools, though their parallel enthusiasm for the new entities mentioned above might tend to undermine this position.

It is not clear whether schools would be protected as much from the setting-related pronouncements of a College of Teaching as they would from the predilections of a government minister.

As for the Conservatives, they seem caught on the horns of a dilemma. Do they too opt for autonomy and the virtues of the market, or do they get on the front foot and develop a more robust alternative to UKIP’s espousal of selection?

They could commit to more selective schools, if only of the post-16 variety. They might even push back the concept to 14+, perhaps enabling the strongest sixth form colleges to accept 14-16 year-olds just as FE colleges can.

They might develop a new cross-school support system for high attainers, especially disadvantaged high attainers. They need look no further than posts elsewhere on this blog for ideas as to how it might be constructed.

They should avoid at all costs the Sutton Trust open access wheeze, which directs a substantial taxpayer subsidy towards independent schools while denuding the state sector of high attaining learners.

Or they might continue to refine the idea of a grammar stream in every school.

The research evidence against streaming seems to me more damning than the evidence against setting, though this is often obscured by the tendency to roll them up together.

That said, several comprehensive schools operating in direct competition with grammar schools seem to have introduced a grammar stream. It would be possible to promote this practice in this subset of comprehensive schools – whether through Ofsted or otherwise – and to develop practical guidance on effective practice.

Setting might well remain the path of least resistance, although compulsory setting across the board would be too much of a straitjacket, restricting the flexibility of those schools that perform outstandingly with mixed ability teaching.

So some sort of selective imposition is necessary. The Ofsted inspection regime is the only effective lever remaining in the hands of central government. The Inspection Handbook might be altered to reinstate a presumption of the kind advanced by Labour in 1997 – and this might be weighted towards schools in either the higher or the lower performance categories. In either case the presumption might be confined to the core subjects.

But even this could only be carried forward in the teeth of opposition from the education profession, so would have the potential to reduce still further the quantum of Tory teacher votes.

The more recently suggested fall-back – adding setting to a menu of possible school improvement interventions managed through the Regional Schools Commissioners – is weak by comparison. So weak that it is tantamount to kicking the whole idea into the long grass.

There are precious few alternatives. Perhaps the only other realistic proposition is to revert to presentation of the evidence base, but to develop this into substantive guidance that schools are expected to consider before determining their approach to ability grouping – much more substantive than the half-hearted case studies published in 2012.

If it were my call, I would construct a ‘flexible framework’ quality standard that defines the broad parameters of effective practice while permitting schools significant flexibility over interpretation of those in practice.

This would align with and reflect the available research evidence on the most effective approaches to setting, including advice on when setting is most likely to work and when it might be preferable to select an alternative strategy.

I would incorporate the standard into supplementary guidance for Ofsted inspectors, to ensure that inspection judgements on setting are fully consistent with it.

And I would strongly encourage schools to use it within their own improvement planning processes – and through peer-to-peer assessments undertaken by teaching schools, national leaders of education and other elements of the emerging self-improving schools system.

I would combine this with a national support programme for disadvantaged high attainers in Years 7-13, generously funded through matched topslices from the Pupil Premium and HE fair access budgets.


The Politics

With May and Johnson already on manoeuvres, not to mention continued pressure from the Brady/Davis camp, Cameron may be even more inclined to press ahead, even in the teeth of opposition from some DfE ministers.

In June 2013, polling suggested  (p8) that 43% of all voters – and 66% of Tory voters – agreed that:

‘The Government should encourage more schools to select by academic ability and build more grammar schools’.

It seems highly likely that, without any viable alternative, the Tories will haemorrhage right wing votes to UKIP over this issue. But there is a trade-off with teacher votes. Are they capable of defining an acceptable middle way, or are they doomed to fall between two stools?

They might at any rate consider some of the ideas set out above.



November 2014