.
This post scrutinises the arguments advanced by Teach First in defence of reallocating Pupil Premium away from disadvantaged learners with middle or high prior attainment.
.
Background
On 29 April, Teach First responded formally to my campaign against their proposal that the Pupil Premium budget should be redistributed so that learners with low prior attainment attract double the amount available for those with middle and high prior attainment.
The original proposal was included in the Fair Education Alliance Report Card (December 2014) and repeated in a primary sector context in The Power of Reading (April 2015) published on behalf of the Read On Get On campaign.
I set out formidable arguments against this proposal in an earlier post: ‘Protecting Pupil Premium for High Attainers’ (April 2015).
It invited all the organisations listed as members of the Fair Education Alliance or supporters of Read On Get On to justify their backing for the proposal or else distance themselves from it.
To date I have pursued twelve of these organisations for a reply. Eleven have failed to respond.
The twelfth, The Future Leaders Trust provided a statement:
‘…we agree that mid- and high-attainers from poor backgrounds should not be deprived of the support that they need to succeed. FSM children who achieve Level 5 in Reading, Writing and Maths at age 11 are still significantly less likely to go on to A-levels and university than their more affluent peers….rather than trying to redistribute the existing pie, we should be campaigning for a bigger one’.
I take that to mean that they do not fully support the proposal.
Brian Lightman of ASCL sent me a response
.
.
.
He wrote:
‘ASCL is not a member of the Fair Education Alliance at this stage although we do agree with many aspects for what they are doing and are in discussion with them about what we might support and how.
However with regards to this specific point our position is similar to the one that NGA expressed. We would not be in agreement with allocating PP on the basis of prior attainment. FSM is a proxy measure which is used to identify the overall level of disadvantage in a school and therefore pupil premium allocations
We strongly believe that decisions about how to use the PP in schools should be decisions made by school leaders who are fully accountable for the impact of their decisions.’
Russell Hobby of NAHT – which is a member of the Alliance – committed to a response but has so far failed to produce it. (Quelle surprise, since NAHT has form in that department.)
.
.
The National Governors’ Association (NGA) has already confirmed its opposition to Teach First’s position.
.
.
Teach First’s argument is also opposed by John Dunford, the Pupil Premium Champion, and by the Sutton Trust.
.
The Teach First response is headed ‘Our policy position on the pupil premium’. It begins:
‘Recently, we’ve had a few questions on our policy position on the Pupil Premium, which we endorsed in the Fair Education Alliance Report Card 2015.’
This helpfully confirms that the proposal set out in the Report Card is official Teach First policy.
It is rather less helpful in failing to acknowledge the source of those questions and failing to link to the counter-arguments set out in my post.
This means that those who want to make up their own minds whether they support Teach First’s position have only one side of the argument available to them. I would have expected more generosity of spirit from an organisation as powerful as Teach First, especially when taking on a singleton blogger like me.
The remainder of this post fisks the Teach First policy position statement.
It strives wherever possible to supplement rather than repeat the substantive arguments advanced in my earlier post, so those who want to consider the case in the round do need to revisit that in addition to the material below.
.
Recommendation
The statement begins by reiterating the original recommendation, to:
‘Target pupil premium by attainment as well as disadvantage measures. This could be achieved through halving current funding per pupil for FSM Ever 6 [a deprivation measure which includes pupils who have ever been a Looked After Child or eligible for Free School Meals in the previous six years]. Half of this funding could then be re-allocated to pupils eligible for FSM Ever 6 who have low prior attainment. This would give double-weighting to those low income pupils most in need of intervention without raising overall pupil premium spend. The change of funding model would increase school accountability for ‘catching up’ pupils.’
The full implications of what is now declared as official Teach First policy are extremely unclear, because there is no modelling, in the Report Card or elsewhere, of the redistribution or its likely effects.
Indeed, when I challenged Teach First over one aspect of modelling, it admitted that none had been undertaken.
In the absence of any clarification of how the redistribution would work, this is my best guess at what the recommendation means.
One begins with an assumption that one-third of pupil premium beneficiaries are low attainers, while two-thirds are middle and high attainers. (In 2014, 67% of disadvantaged learners achieved KS L4 and above in reading, writing and maths, meaning 33% did not.)
Given a total pupil premium budget of £2.5bn per year, assuming equal shares, the low attainers get £833.33m and the middle and high attainers together get £1.67bn.
One removes half of the funding from the high and middle attainers together – so £833.33m in total, leaving an equivalent sum behind.
The sum removed is added to the low attainers’ budget giving them a total of £1.67bn, meaning they have double the amount available for the other two groups combined.
But this outcome would mean one group, half the size of the other, would also have double the funding, hence each low attainer within that group would have four times the funding allocated to each middle and high attainer.
To make the equation work, one has to divide the sum initially removed from the high and middle attainers into two, allocating £416.67m into each pot.
Then there is £1.25bn for the low attainers and an identical £1.25bn for the middle and high attainers, but there are twice as many of the latter, so each of them gets half the sum available to each low attainer.
Confused yet?
In any case, all of this is guesswork because Teach First has not yet:
- Confirmed whether this proposal applies to both primary and secondary schools though, since it is referenced in a primary context by the ‘Read On Get On.’ Report and this statement mentions the primary sector in passing , one assumes that it applies equally in both.
- Defined what constitutes low prior attainment. At secondary level for example, is it below the scaled score equivalent of Level 4b in reading writing and maths combined? Or does it count each assessment separately? Or is it achievement below Level 4, either individually or combined? What is the equivalent measure at primary level? Your guess is as good as mine.
It really behoves Teach First to be clearer on these issues than it has been to date.
However the recommendation above states clearly that learners attracting the pupil premium with low prior attainment would have ‘double weighting’, implying that those with middle and high prior attainment would find their allocations single weighted, so pitched at half this value.
So, in the absence of any further elucidation, I assume that each low attaining pupil premium beneficiary would in future receive twice as much as each middle and high attaining beneficiary.
It would be good to know the size of the premium Teach First expects to be available to each category.
One possible outcome, using the very approximate ratio above, might be:
- Low attainer primary pupil premium of £1,950 and high/middle attainer primary pupil premium of £975, compared with the present rate of £1,300.
- Low attainer secondary pupil premium of £1,402 and high/middle attainer pupil premium of £701.
Low attainers would get an additional 50% top-up while all middle/high attainers would get 75% of what they do now.
Until we know the size of the uprating and the numbers used in the calculation, we cannot quantify the redistributive impact, so Teach First has asked its supporters to sign a blank cheque (and they have done so, apparently without too much scrutiny).
.
Pupil premium as it operates now
.
The positive
The policy statement says:
‘Teach First is fully supportive of the Pupil Premium. It has been an incredibly important tool that helps to achieve our vision that no child’s success is limited by their socio-economic background. We will continue to advocate for it, and for it to be protected and enhanced. The introduction of the Pupil Premium has increased accountability for the progress of the country’s poorest children and since this was introduced, an increase in attainment has been seen in those areas where they are the minority, though they still significantly underperform their wealthier peers. We hope and expect the full impact of the Pupil Premium will become apparent as the funding beds in and those pupils who have benefitted from it complete their full school journey.’
The commitment to continued advocacy for the pupil premium to be protected and enhanced rings rather hollow, given that perhaps two-thirds of beneficiaries would have their allocations reduced to half the value of the premium provided for their low attaining peers.
One assumes that ‘protected’ means Pupil Premium should continue to be ring-fenced outside the school funding formula.
‘Enhanced’ is potentially meaningless. It stands proxy for ‘increased’ but, given the wider pressures on the national schools budget, there is little prospect of increasing the total pupil premium budget by the sum necessary to uprate low attainers’ allocations while leaving others unchanged.
This is apparently what the Future Leaders Trust would like to see, but it simply isn’t realistic.
The data supporting the claim of an increase in attainment since the premium was introduced is unsupported by evidence. What level of attainment? What measure of attainment? What size of area? How do we know the improvement is attributable to the Pupil Premium, as opposed to other factors?
In particular, does this apply to middle and high attainers? If so, what evidence is there to suggest that significantly reducing the sum available to support them will not detract from this progress?
.
The negative
The statement continues:
‘Schools are held accountable to Ofsted for their spending of the Pupil Premium – demonstrating how it has contributed to improved attainment of eligible pupils. There has not yet been a systematic review of how schools are spending the Pupil Premium, however there is some evidence from Ofsted that Pupil Premium is not always being used as effectively as it could be – in some instances plugging gaps in school budgets which have faced cuts – and that it is not always meeting the needs of those who are falling furthest behind (e.g. Chapter 6 in The Tail).’
This betrays selective use of the evidence base.
Where the funding is being used to plug gaps in the school budget (something that Teach First is also advocating at the macro level – see below) surely middle and high attainers will be suffering equally as much as low attainers, quite possibly more.
In ‘The pupil premium: How schools are spending the funding’ (February 2013), Ofsted reported:
‘Where schools spent the Pupil Premium funding successfully to improve achievement, they shared many of the following characteristics. They:
- carefully ring-fenced the funding so that they always spent it on the target group of pupils
- never confused eligibility for the Pupil Premium with low ability, and focused on supporting their disadvantaged pupils to achieve the highest levels.’
Conversely:
‘Where schools were less successful in spending the funding, they tended to have at least some of the following characteristics. They…
- …focused on pupils attaining the nationally expected level at the end of the key stage (Level 4, five A* to C grades at GCSE) but did not go beyond these expectations, so some more able eligible pupils underachieved…’
In ‘The most able students: An update on progress’ (March 2015), Ofsted said:
‘Our report in 2013 found few instances of the pupil premium being used effectively to support the disadvantaged most able pupils. In the schools visited for this survey, about a third were using the pupil premium funding effectively to target the needs of these pupils.’
Ofsted concludes:
‘… more needs to be done to develop a clearer picture of how well schools use pupil premium funding for their most able students who are disadvantaged and the quality of information, advice and guidance provided for them. Ofsted needs to sharpen its practice in this area.’
Most of the evidence I have seen on this issue suggests that the lowest attainers are more likely than higher attainers to have their needs addressed appropriately through the pupil premium.
.
The case for reallocation via both pupil premium and the NFF
My previous post argues that, to the extent that reallocation is needed, it should be undertaken solely through the national funding formula (NFF) since using pupil premium creates too much ‘collateral damage’ – in the shape of lower allocations for middle and high attainers.
Teach First asserts:
‘We believe that low prior attainment is a compounding disadvantage and should be recognised in the National Funding Formula but that there would also be value in making extra funding to low attainers explicit through shifting the emphasis onto this group in the Pupil Premium.
The re-allocation within Pupil Premium funding would incentivise schools to make more progress with their most needy low income pupils: it would focus the accountability – as well as the financial support – directly on that group of pupils most in need of intervention.’
The case for recognition in the NFF is surely built on the costs involved in raising the attainment of low attainers, whether advantaged or disadvantaged.
If Teach First want to make extra funding for low attainers more explicit, that might be achieved by introducing an additional and entirely separate low attainers’ premium which recognises the needs of advantaged and disadvantaged low attainers alike.
But it would be administratively complex for schools to administer two overlapping ring-fenced budgets. It would be more straightforward to undertake the redistribution entirely through the NFF.
Accountability is achieved fundamentally through Ofsted inspection and School Performance Tables. If Teach First believe that schools need to be made more accountable for improving the performance of disadvantaged low attainers – and they cite no evidence to show that this is necessary – those are the obvious routes.
.
Grounds for justifying the policy
I had asked Teach First to explain whether it justified the proposal on the grounds that it would divert extra funding to ‘catch-up’ or that it would redistribute wider deprivation funding between schools.
The policy statement makes clear that both are in play, but one takes precedence over the other:
- First and foremost, Teach First apparently believes that: those with low prior attainment have greater needs; that the potential benefits of investment in low attainers are more significant; and that effective interventions for them are comparatively more expensive than those for disadvantaged middle and high attainers.
- Secondly, this is assumed to be an effective method of redistributing funding away from a few (number unquantified) schools that have built up substantial funding surpluses through the combined effects of the current NFF and pupil premium, towards some (number again unquantified) which receive rather less support.
Each segment of this argument is tackled below.
.
The impact of low attainment
The statement says:
‘We believe this is important because intervention at the lower ends of the prior attainment distribution could have significant impact on later attainment. The FEA report card showed that those who fall behind early are not likely to catch up – last year only 7% of pupils achieving below expected levels aged 11 went on to get 5 ‘good’ GCSEs aged 16. And we charted how this ‘class ceiling’ can systemically hold some pupils back – having a knock-on effect on their wellbeing, employment and access to higher education.’
There is similar evidence in respect of disadvantaged high attainers, where the comparator group are those with equivalent prior attainment from more advantaged backgrounds.
In ‘Closing England’s Excellence Gaps: Part 2’ (September 2014) I set out all the research evidence I could find on the subsequent progress made by high attainers, including:
- The chances of FSM-eligible KS2 high attainers still being so at the end of KS4 are 45% lower than for other high attainers with similar prior attainment and characteristics (DfES 2007)
- 58% of FSM students within the ‘most able’ population in KS2 and attending non-selective secondary schools go on to achieve A*-B GCSE grades in English and maths, compared with 75% of non-FSM pupils, giving a gap of 17 percentage points. (Ofsted 2013)
- ‘Children from poorer backgrounds who are high attaining at age 7 are more likely to fall off a high attainment trajectory than children from richer backgrounds. We find that high-achieving children from the most deprived families perform worse than lower-achieving students from the least deprived families by Key Stage 4. Conversely, lower-achieving affluent children catch up with higher-achieving deprived children between Key Stage 2 and Key Stage 4.’ (Vignoles 2014)
Teach First continues:
‘Recent analysis of pupils’ progress has shown that – although the majority of pupils do not have linear trajectories – pupils with high prior attainment are much more likely to stay on a linear trajectory than those with low prior attainment… However, low prior-attainers at Primary and Secondary have much more varied trajectories – indicating that rapid progress is possible, despite the fact that it is often not the case – and that focus on this group could be fruitful.’
I am not quite sure what this contributes to the argument. The analysis relates to progress subsequent to KS1 attainment. As the paper notes:
‘For children achieving a Level 1C, B or A at this stage, their development is so unpredictable that most will either outperform or underperform any Key Stage Two target that might be set.’
Moreover, the percentages are low at all levels – for example, only 12% of pupils with L3C at KS1 make linear progress at all key stages.
And of course they apply to all learners and not to disadvantaged learners, so we cannot see how much variation there is as a consequence of disadvantage.
The same is true of the primary and secondary transition matrices which, amongst other things show that, in 2014:
- Of those with KS2 L5A in English or maths only half (48% in maths; 51% in English) achieved a GCSE A* grade.
- Of those with KS2 L5C in English or maths, just one in five makes only a single level of progress by the end of KS4 in English, while the same is true of almost a third of students in maths.
Perhaps more to the point, excellence gaps are wide and growing. The graph below compares the percentage point gaps between disadvantaged and all other learners at KS2 L4 and above and L5 and above in 2013 and 2014 respectively.
.
.
In 2014 the L5 gaps are larger across the board, with particularly large differences in maths and reading. In the latter, the gap at L5+ is more than twice as large as it is at L4+.
I note in passing that the Teach First model would presumably involve any disadvantaged low attainer who subsequently achieved or exceeded the expected level of performance moving from the higher level of pupil premium to the lower, otherwise the system would be inequitable. This would be complex and costly to administer.
Finally in this section, Teach First argues:
‘As well as huge personal cost, there is huge national cost to this underachievement – consultancy BCG estimated that boosting the attainment of this group could raise GDP by up to £56bn a year by 2050 (BCG, 2013)’
This is a secondary reference to a finding quoted in ‘The Tail’, which appears to be a sacred script for Teach First and the probable source of their false ideological position.
The actual wording in Marshall’s book is:
‘In a comparable study for the UK, the consulting firm BCG found that matching Finnish levels of social mobility (in terms of raising the educational outcomes of poor children) would add £6bn a year to GDP by 2030 and £56bn a year by 2050. Bringing below-average students in the UK to the national average would add £14bn a year to GDP by 2030 and £140bn by 2050.’
It doesn’t inspire confidence that Teach First has misquoted this statement in the Report Card as well as in its policy statement.
The original source is the Sutton Trust’s ‘Mobility Manifesto’ (2010). The calculations are based on PISA 2006 average scores in maths and science and based on a methodology derived by Hanushek. I shall leave it to others to comment on the reliability of the findings.
The first calculation involved estimating the benefits of matching the distribution of scores across the UK (so not just England) with those of Finland; the second with raising attainment across all socio-economic groups (based on parents’ education) to the UK average (excepting the higher than average value already recorded by the highest socio-economic group).
This is of course an entirely hypothetical model which attempts to quantify the impact of education on economic growth.
I will only note that, in ‘The High Cost of Low Educational Performance’ (2010) Hanushek also calculates the not inconsiderable benefits of improving average PISA maths and science performance by 25 points, so impacting across the attainment spectrum.
I reviewed the parallel literature on the economic benefits of investment at the top end in ‘The Economics of Gifted Education Revisited’ (March 2013).
In light of that, it seems to me there is a reasonable case for arguing that investment at the top end would yield commensurate benefits.
Hanushek himself recognises that:
‘Importantly, the relative size of the effects of performance at the bottom and at the top of the distribution depends on the specification, and further research is needed to yield more detailed predictions. Even so, the evidence strongly suggests that both dimensions of educational performance count for the growth potential of an economy.’
.
The impact on Goal 5
My original post pointed out that the Fair Education Alliance was also pursuing another goal to:
‘Narrow the gap in university graduation, including from the 25% most selective universities
The Fair Education Alliance is committed to closing the graduation gap between young people from low income backgrounds and those from high income backgrounds. Our goal is for at least 5,000 more pupils from low income backgrounds to graduate each year, with 1,600 of these young people graduating from the most selective universities.’
I argued that reducing pupil premium for middle and high attainers would make this much harder to achieve, especially the highlighted final phrase, because it would reduce the chances of such learners achieving the grades necessary for admission to such universities.
Teach First’s policy statement says:
‘We see this recommendation as focusing on a different part of ‘the gap’ from Impact Goal 5 (the gap in university access) recommendations – this policy is about raising the attainment at KS2 and KS4 (our Impact Goals One and Two) for some of the nation’s most vulnerable children.’
This is risible I’m afraid, since a corollary of rationing pupil premium in this fashion is that exactly those disadvantaged learners most likely to proceed to selective universities will lose funding, while those least likely to do so will gain.
The reference to ‘vulnerable children’ introduces a whole new dimension, only for it to disappear as rapidly. Because if we are talking about funding for additional needs, perhaps SEN or behavioural, a range of additional considerations (and funding streams) apply.
Teach First continues:
‘We know that the kind of intensive interventions needed to raise attainment can be expensive and that working to change a pupil’s trajectory is likely to be harder than to ‘keep pupils on track’. We also know that there are an array of inexpensive projects working with schools who can boost the non-cognitive and academic skills of those pupils already on positive trajectories – such as debatemate, The Brilliant Club and our own Futures programme. Hence our recommendation that Pupil Premium funding is redistributed to give greater weighting to low prior attainment and the more expensive interventions required there to change a child’s life.’
Hang on, weren’t we told earlier that the majority of students don’t have linear trajectories?
I would like to see evidence that it is necessarily harder to move, for example, a secure L3 to a L4 than it is to move a secure L5 to a L6. My experience suggests that interventions to raise disadvantaged attainment at the top end may need to be equally intensive as those lower down, especially when the focus is admission to selective universities.
On top of pupil premium, there is additional investment in catch-up, including over £50m a year (£500 per pupil) for the Catch-up Premium and the £50m annual topslice from the pupil premium budget for end of KS2 summer schools, also heavily focused on catch-up.
I have called for a similar £50m topslice to support intensive provision for disadvantaged high attainers seeking admission to selective universities.
In their parallel response to that post, Teach First says:
‘The single biggest factor linked to HE access is prior attainment. The Russell Group highlight that, of 15-year-olds on Free School Meals in 2007, only 0.3% achieved 3As or equivalent in their A-levels two years later – a huge barrier for progression to the most selective universities.’
In this response, however, it all seems much more straightforward. There ‘are [sic] an array of inexpensive projects’ that can sort this out. (Do English teachers now consider an array to be plural?)
Unfortunately it’s not that simple. I believe debatemate and The Brilliant Club are both Teach First spin-offs (run by alumni). While debatemate is a member of the Fair Education Alliance, The Brilliant Club is not. While debatemate is focused on developing speaking and listening and critical thinking skills, The Brilliant Club is dedicated principally to placing PhD students in schools.
No doubt both are valuable niche programmes and there are dozens more like them, offered by commercial, third sector or university providers. Some are free, some relatively cheap, others more expensive.
The problem is that disadvantaged students aiming for selective universities need a coherent, long-term support programme that addresses their particular strengths and weaknesses. This is increasingly recognised in the national strategy for access.
They also need support from their schools to secure that provision, drawing on a range of different providers to supply the elements they must combine to generate a holistic programme. That’s precisely what my proposed £50m pupil premium topslice would achieve.
It would support a personal budget of £2,000 a year (almost exactly the same as the illustrative higher rate pupil premium for low attainers above) for some 5,000 high attaining pupil premium eligible learners.
It would be designed to increase significantly the number of such students progressing to high tariff universities, including Russell Group institutions and especially Oxbridge.
No sign of Teach First support for this of course.
.
Redistribution of funding
Reverting to its secondary reason for reallocating pupil premium, Teach First argues:
‘A secondary effect of this Pupil Premium change is that it might better recognise the compound disadvantage of growing up in a low income home in an area with a history of educational under-performance.
The Free School Meals (FSM) measure of disadvantage in the UK is not fully progressive or entirely comprehensive. For example, the binary FSM/non-FSM to dictate funding does not allow for recognition of low-income families who just miss the eligibility criteria for Free School Meals; the national funding formula does not currently compensate for geographical isolation and high transport costs which can compound low incomes in parts of the country. Consequently – due to the combination of a high intake of pupils attracting the Premium and a currently unequal national school funding formula – there are a small number of very successful schools building up surpluses. Meanwhile some schools with arguably greater need, where pupils suffer different socioeconomic disadvantages that affect their attainment, are receiving comparatively little extra funding. This hampers their ability to deal with the challenges that their students face and to prevent those vulnerable pupils from falling behind their peers. Those areas struggling to raise the attainment of their deprived pupils would most benefit from this double-weighting for their pupils who have fallen behind.’
My previous post argued strongly that any redistribution of this nature should be undertaken through the NFF and not the pupil premium.
Teach First is perfectly at liberty to lobby for changes to the Formula that would achieve its desired outcomes, though it seems that only ‘a small number’ of schools have built up surpluses.
There is no reason in principle why the NFF should not take account of aspects of disadvantage not caught by ‘ever 6 FSM’ (or indeed the other routes to pupil premium), or reflect sparsity factors.
Pupil premium reallocation might be a ‘quick fix’ for this problem but, as noted above, the collateral damage is too great. It drives a coach and horses through the principle that every ‘ever 6 FSM’ learner attracts the same rate of support. As such, it is not to be tolerated.
.
Conclusion
This policy position is fundamentally inequitable, predicated as it is on the mistaken ideological assumption that a low attainer’s needs must necessarily outweigh and be prioritised over those of a high attainer with the same level of disadvantage.
Teach First will surely nail their colours to this mast and sail away into the sunset. In doing so, they confirm the existence of the bias I already suspected.
But, in the words of the Report Card itself, we need ‘a fair education for all’ supported by the ‘sound moral argument for giving every child an equal chance to succeed‘. Success should not mean all learners achieving the same outcomes. The success of one group should not be at the expense of another.
Nothing in Teach First’s new line of argument has persuaded me that high attainers’ chances of success will be protected if their pupil premium is reduced in this way. The same goes for the ‘squeezed middle’.
At bottom, this is nothing more than robbing Peter to pay Paul.
So I call again on the members of the Fair Education Alliance and supporters of Read On Get On to justify their commitment to this ill-conceived and ill-formed idea.
Or else make clear that they no longer support it.
.
.
GP
April 2015
The measure was addressed briefly in 
Chart 1: PISA resilience in science for selected jurisdictions by gender (PISA 2006 data)
Chart 5: PISA resilience in maths for selected jurisdictions by gender (PISA 2012 data)
Gifted Phoenix 