Augar and the dark side of Robbins

The Augar Report last week sought to make its impact look as profound as that of the Robbins Report in 1963. In doing so, it leads observers to look back and make some comparisons. The emergence of the 'Robbins Principle' that Higher Education “should be available to all who were qualified for them by ability and attainment" was founded on what was known at the time about ability and attainment. It was assumed by many that expansion of university provision would be limited by the availability of suitable candidates. But the "pool of ability" could not be accurately determined. In contrast, Robbins was right in expecting no "shortage of potential ability". However, a dark side emerged in the process with consideration of "inherited potential". In 2019, Augar does not consider how far we have advanced since 1963 in determining the 'pool of ability' to assess future student numbers.

The Post-18 review of education and funding: independent panel report, known as the Augar Report, emerged last week and is attracting considerable ongoing interest and analysis. TEFS initial response last week, ‘Augar stirs up the system: The ripples will go far beyond his remit’, concentrated on some unintended consequences in a febrile political environment. Augar also was keen to make reference to the Robbins Report of 1963 with,

“The present government’s review is the first since the Robbins report in 1963 to consider both parts of tertiary education together.” 

It indeed compares favourably with Robbins in its size and its well considered detail. However, the aims and outcomes are very different. Robbins did not include, “as customers have at least £7,500 available for fees and resources to feed and house themselves for at least three years.” That is the Augar conclusion alongside ensuring that there are no costs ‘up front’ but much more to be paid back in time.

Robbins and the seismic shift in social mobility.

It is worth revisiting in a little more depth the background to the Robbins Committee on Higher Education Report, 1963 and why the committee he chaired became determined to open up the universities so that places should be  "available to all who were qualified for them by ability and attainment". This became known as the 'Robbins Principle'.

This was no accidental choice of words and it set the scene. The report itself is 335 pages and it was backed up by a vast amount of evidence that is lodged in the National Archives and elsewhere. The enormous vision of the work was only outstripped by the enormity of its implementation. Its conclusions were accepted just after the resignation of the Prime Minister and acted upon by the incoming Labour government a year later.  The extent and depth of the work is impressive and it makes the reader in 2019 seek to compare it to later reports such as those of Browne in 2010 and now Augar in 2019 when history is repeating itself. The task of implementing Augar will fall to a new government as news of a delays in the wider spending review emerge.

The future demand for higher education and the places needed to meet it.

This was the title of Robbins CHAPTER VI and was central to the Higher Education remit of Robbins and the outcomes from the report. Robbins was asked to consider the scale of higher education in this country up to about 1980. There were two logical approaches to the task. Firstly, to estimate what output is required from higher education in order to meet the country's needs for qualified people. Although ‘manpower’ was used in the report, there was reference to male and female throughout. Secondly, to estimate the number of places required to meet the demand for entry from suitably qualified applicants. It was decided early on that the second approach provided the sounder basis of estimation of numbers and outcomes. This then set in train a number of logical consequences. The numbers taking qualifying examinations across the UK was rising and the demand would rise further as the population continued to increase fast. Since most were already born, the imperative was obvious. The key statistical evidence that was crucial to the decisions was coordinated by economists Claus Moser, Richard Layard and statistician Joseph Irwin. They were key to the report and their overview, published in 1964 [1], provides the best insight into the basis for the decisions made.

The ‘potential ability’ question that Augar ducks.

Augar steadfastly avoided the issue of qualified by “ability and attainment”. The idea of setting minimum attainment grades for university entry was dropped fast and passed on for others to address. 'Ability' as a criterion was avoided from the outset. By contrast, this question was hiding in a dark corners throughout the Robbins deliberations. Many observers at the time incorrectly and naïvely thought that an expansion of university provision would be limited by the availability of suitable candidates. In some ways that might have been correct in practical terms. A school leaving age of fifteen at the time led to many not attaining GCE qualifications and Grammar schools only took a minority of pupils based on an 11-plus exam. However, by the time of Robbins, the exam itself had been shown to be badly flawed and did not predict the ability or potential of many pupils. Far too many were ‘misallocated’ [2] and were excluded at that point. Also many failed to attain A Level results high enough. It was not widely known at the time that the A-Level marking was ‘rigged’ to allow only a small percentage to attain the higher grades of C and above needed. This was a fixed percentage and was only revealed at the Joint Matriculation Board as late as 1977 (see TEFs 17th August 2018 ‘A-Level Playing Field or not: Have things changed over time?’). This underhand strategy limited access to the ‘best’ universities despite a belief that A-Level grades did not necessarily predict degree outcome for many students. This idea persists to recent times.

Although it is currently concluded that A and A* grades at A-level are good indicators that a First degree will be obtained, some studies fail to emphasise that a significant proportion of students with lower A-level grades gain Firsts. In one study from 2015, almost 15% of students at Russell Group Universities with no A/A* A-Level got firsts [3]. These are considerable numbers making major gains at university and this might be better acknowledged.

Limiting numbers.

Hidden in the Robbins report is one reason proposed why the numbers of students suitable for university might be limited over the subsequent 20 years.

“The crucial question is whether, within the next twenty years, the growth in the output of qualified school leavers is likely to be limited by a scarcity of inherited potential ability. This raises the fundamental issue of whether the pool of ability, as it is usually called, can be measured and, if so, how.”

However, the evidence in Robbins led to the conclusion that “there is no risk that within the next 20 years the growth in the proportion of young people with qualifications and aptitudes suitable for entry to higher education will be restrained by a shortage of potential ability". The history after that reveals this was clearly the case and the potential of the country was suddenly unlocked. The reasoning and statistical evidence that led to this observation was complex but inexorable in its logic [1].

The flawed contemporary view.

Is well represented by the authors of Social Mobility and its enemies [4] (reviewed by TEFS 1st October 2018 ‘Social Mobility and its Enemies – A review: Finding the Lost Souls’). They, make reference to the ‘inheritance’ issues but wisely shy away from the most controversial implications of genetics and inheritance.

“The complex interactions between children’s environments and their genetic inheritance remains a hotly disputed area with an extensive scientific literature”

The authors cite ‘twin’ studies from 2015 that “points to some genetic heritability in education and life outcomes”. The understated ‘twin’ studies are discussed in greater detail in a review by Robert Plomin of King’s College London and Ian Deary of Edinburgh University from 2015 [5]. However, the advent of rapid sequencing of DNA and mass screening of genes means that our understanding is advancing very fast. There have been more recent studies that reach similar conclusions and strengthen the case for a genetic influence and this is explored a little more below. However, to do this justice would require a much more detailed treatise by TEFS.

Instead the authors concentrated on an extensive study of surnames that persist as advantaged people for generations. It is admitted that the study is “based on such small numbers that it is hard to generalise these patterns to the entire population”. The study in 2014 by historian Gregory Clarke in the book ‘The Son Also Rises: Surnames and the history of social mobility’ acknowledges that “by and large, social mobility has characteristics that do not rule out genetics as the dominant connection between the generations”.

This is a belief that is still widely held amongst the wealthy and advantaged and the modern aristocracy. Indeed, a few years ago at a wedding event a senior executive from a very well healed aristocratic family  confirmed with confidence “of course our children (he was including mine) will excel as they come from excellent genetic stock”. I did not feel it the right time or place to inform him that my mother had been a dinner lady and my father a labourer in a foundry in Coventry. I was the first from our family to enter a university with the intention of getting a degree. My father had been a labourer in the 1950s building what is now Coventry University. The idea of ‘genetic determinism’ as a convenient excuse to continue to seek advantage for their children is beginning to wear thin. The outcome of higher education expansion after 1963 is clear proof of a wider pool of ability amongst the various strata of society.

The genetic question.

Ability was measured with IQ tests in the time of Robbins and was considered a proxy for ‘inherited potential’. However, it had become increasingly obvious that the outcome for individuals was biased by background. Robbins was not able to come to terms with the rapidly advancing field of Genetics and Molecular Biology but must have been aware of its possibilities. Would Robbins have avoided its consequences in 2019?

Robbins had only a rudimentary appreciation of the basis of inheritance despite mentioning it several times. The chemical basis of inheritance, DNA, had only been discovered in 1928 by a Microbiologist, Fred Griffith in the UK and confirmed by Oswald Avery, Colin MacLeod and Maclyn McCarty in the USA in 1944. The chemical structure of DNA was revealed by the Work of Franklin, Watson and Crick in the UK in 1953. The relationship between DNA and function of genes was proposed by 1957. The crucial experiment by Matthew Meselson and Franklin Stahl in the USA in 1958 confirmed how DNA replicated itself in cells.

Robbins must have been aware of these advances and wary of the potential implications of the sequencing of DNA in genes associated with ‘inherited potential’ in the future. But for him that was a far off concern. However, by 2019 our knowledge about genes and their action has advanced fast. The irony is that a significant contribution to the basic research was in the UK and was conducted in the new universities by students and staff from modest backgrounds. The DNA sequencing chemistries got off to a slow start in the 1970s but advanced in the 1980s and 1990s. The first completed sequence map of the human genome was released on 14th April 2003. Now we have more information that links measures of intelligence with genes.

The ‘twin’ studies cited in 'Social Mobility and its Enemies' observed that “genetic influence on individual differences in intelligence is substantial” but also that the heritability was caused by many genes each having a very small effect. This idea has been tested subsequently by whole genome DNA studies on very large numbers of humans. This approach is called Genome-wide Complex Trait Analysis (GCTA). A fairly accessible review from 2018 is by Robert Plomin and Sophie von Stumm from 2017 [6].

Drawing lots for the right permutation.

The discovery of a large number of genes associated with abilities and measured intelligence has expanded in recent years. The genetic measurements are certain at a molecular level. However, how ‘cognitive ability’ or intelligence’ is measured is a major constraint. The studies use genome-wide association study (GWAS) methods to link specific gene sequence variations (known as single nucleotide polymorphisms or SNPs) to observable traits of individuals. This approach is used more often to determine if common diseases are linked to genetic factors. Measuring ‘intelligence’ is more difficult and is open to criticism.

In 2017, Suzanne Sniekers and twenty nine co-workers reported screening 78,308 individuals where at least 22 genes, of which 11 were new discoveries, could be associated with intelligence [7]. This was determined by using a 13 point questionnaire. Crucially the genes identified were predominantly expressed in brain tissue. In 2018, an even larger study of 269,867 individuals [8] revealed that at least 205 genomic loci (190 new) and 1,016 genes (939 new) could be inked to intelligence. Again many of these genes are strongly expressed in the brain. More were added as the results were released whilst the Augar panel made their deliberations unaware of the implications. Before the year was over, James Lee of the University of Minnesota and seventy seven co-workers reported that the number of SNPs had expanded to at least 1,271 in a screening of 1.1 million people [9]. They used a ‘multi-trait’ approach, that considered educational years, cognitive performance and mathematical ability, to make the associations.

It is likely that we are only at an early stage of discovery and that the numbers of genes are very much underestimated as the biochemical basis of the complexities off human behaviour are revealed. The biochemical functions of many of the gene products is still uncertain but will be revealed as we better determine brain function. However, more functional links are emerging this year [10]. We are entering a new biological age as genetics and social policies collide. With this will come immense dangers in how we use the information nature has revealed to us through science.

What we must consider.

Some things are certain, such as direct measurements relating to human biology. This includes the chemistry of DNA and the sequences determined in laboratories. They reveal that, if not twins or clones, all humans are different regardless of their parents. This is simply because of the random assortment of the chromosomes that the DNA sequences are arranged in. The chances of one human being exactly the same is almost zero unless they are identical twins or cloned. For every child, the 'pack' of 23 pairs of chromosomes is 'shuffled 'to give a different individual each time. The number of genes associated with cognitive ability and intelligence is large and these genes are shuffled along with the chromosomes. There is no one gene involved and it is the vast number of possible permutations of the mixed genes that is important.

It is also known that DNA can be altered by cells adding methyl groups during development. This is a so called ‘epigenetic effect’ whereby the expression of some genes gets ‘turned down’ to become less effective. The idea that such changes might be inherited is gaining credence over time and it does affect children as they develop to become adults. One such study looking at the effects on intelligence [11] has been widely reported. We must also now consider that changes in the educational system, overall exposure to stress, traumatic experiences, nutrition, and poverty all combine to affect how children develop. Being dealt a ‘good hand’ in the shuffling of genes may not be enough alone. It needs nurture to develop successfully.

What is less certain is how intelligence is measured and what it really means. While we may have some indication of the ability of individuals through tests and educational attainment, we have currently no biochemical assessment of their potential through DNA analysis. But this may be only a short hop away and is likely to yield many anomalies if used alone. There are many reasons to be cautious

A way forward.

Those from disadvantaged backgrounds fear that this will merely confirm that they are mostly of a lesser ability. Meanwhile those from advantaged backgrounds fear that it will confirm that their attainment is an illusion not backed by genetic evidence. Either way it will be unpopular.

The logical way forward is to accept that the vast number of permutations of genes means that there will always be a wide spectrum of ability. But we know that already through education and attainment measures. In a recent article in the New Scientist [12], Robert Plomin was quoted as “DNA is more important to a child’s personality, exam results and future income than the way they are brought up – but that’s good news, says geneticist”. But he also went further with,

“I actually would prefer if we didn’t have any selective schools and we just had good community schools. But if you’re going to select, you should take genetics into account, and I think eventually we will. The best predictor for exam results is the polygenic score for educational achievement, which adds up the results from hundreds of genetic tests. We can today predict 15 per cent of the variance of GCSE scores with DNA alone. DNA is an objective, unbiased predictor. You can’t tutor your child to get them a better polygenic score for intelligence”.

If DNA modifications and epigenetics are also considered, then it would be illogical to select out people for education on the basis of DNA sequence testing alone. Instead it would be better to aim to offer all children the same opportunities equally. Then adhere to the Robbins Principle and offer a chance to anyone with ability. We should make sure they all reach their optimum level of attainment. In the Augar Report ‘Annex Methodology for Assessing Costs of Panel Recommendations’ a key assumption is made that there would be “No behavioural change in student participation”. This falls well short of the analysis needed to predict student numbers as widening participation improves. But there is hope that the full potential of all children is realised as we learn more.

Mike Larkin, retired from Queen's University Belfast after 37 years  teaching Microbiology, Biochemistry and Genetics.

References.

[1] Planning the Scale of Higher Education in Britain: Some Statistical Problems. C. A. Moser, P. R. G. Layard and J. O. Irwin. Journal of the Royal Statistical Society. Series A (General) Vol. 127, No. 4 (1964), pp. 473-526

[2] A. Yates & D. Pidgeon, Admission to grammar schools (NFER, London, 1957), pp. 63-4, table V I, p63, pp. 66 and P. E. Vernon, Secondary school selection (Methuen, London, 1957), pp. 77, 169

[3] The role of the A* grade at A level as a predictor of university performance in the United Kingdom. Carmen Vidal Rodeiro & Nadir Zanini Oxford Review of Education, 2015. Vol. 41, No. 5, 647–670.
See also HESA September 2015/21 Issues paper Differences in degree outcomes: The effect of subject and student characteristics.

[4] Social Mobility and its Enemies. Lee Elliot Major and Stephen Machin 2018 Penguin Books https://www.penguin.co.uk/books/306476/social-mobility/9780241317020.html

[5] Review ‘Genetics and intelligence differences: five special findings’ by R Plomin and IJ Deary, Molecular Psychiatry volume 20, pages 98–108 (2015). 

[6] ‘The new genetics of intelligence’. Robert Plomin and Sophie von Stumm. Nature Reviews Genetics volume 19, pages 148–159 (2018)

[7] ‘Genome-wide association meta-analysis of 78,308 individuals identifies new loci and genes influencing human intelligence’ Suzanne Sniekers et al. Nature Genetics volume 49, pages 1107–1112 (2017)

[8] ‘Genome-wide association meta-analysis in 269,867 individuals identifies new genetic and functional links to intelligence’. Jeanne E. Savage. Nature Genetics 50, 912–919 (2018)

[9] ‘Gene discovery and polygenic prediction from a genome-wide association study of educational attainment in 1.1 million individuals’. James J. Lee et al.

Nature Genetics 50, 1112–1121 (2018)

[10] A combined analysis of genetically correlated traits identifies 187 loci and a role for neurogenesis and myelination in intelligence’ W. D. Hill et al. Molecular Psychiatry 24, 169–181 (2019)

[11]. Epigenetic variance in dopamine D2 receptor: a marker of IQ malleability? Kaminski, J.A. et al. Translational Psychiatry 8, Article number: 169 (2018).

See also: ‘Epigenetics and IQ: the mechanism behind environmentally-induced effects on cognitive performance’ By Siobhán Dunphy. European Scientist 24.09.2018

[12] ‘The parenting myth: How kids are raised matters less than you think’ New Scientist 22nd May 2019. https://www.newscientist.com/article/mg24232310-800-the-parenting-myth-how-kids-are-raised-matters-less-than-you-think/