This blog was written by Iain Mansfield, the author of the HEPI Occasional Paper The Impact of Selective Secondary Education on Progression to Higher Education (2019), a report written in a personal capacity prior to him taking up his current role.
The paper published today by HEPI is, in parts, a welcome contribution to the ongoing debate on academic selection. In particular, the attempt by Dickson and Macmillan to account for socioeconomic and demographic differences between selective and non-selective areas addresses a key gap in the evidence base I identified in my original paper; furthermore, the cohort studies by Sullivan provide an interesting insight into the social impact of grammar schools in the era when they were more prevalent.
Nevertheless, the paper contains a number of serious issues that raise significant concerns, and call into question its value as a definitive statement upon the subject of academic selection.
Most notably, nowhere in the paper are the most striking findings of my paper addressed: namely, the dramatic increase in propensity for children – including those from disadvantaged backgrounds – in selective areas to progress to Oxbridge, not to mention the even larger propensity for those from BME backgrounds to do so (a state school pupil with a BME background being more than five times as likely to progress to Oxbridge if they live in a selective area rather than a non-selective area). These findings were covered prominently in both the original report and accompanying press release; if the authors of today’s report do not dispute these findings, one presumes it is because they are unable to.
Other issues abound. The Sutton Trust report that found limited impact of selective schools on the attainment of children in surrounding schools is once again dismissed without explanation – a phenomenon I identified in the original report, whereby evidence that does not condemn selection is simply ignored rather than engaged with. Elsewhere in the report, Boliver and Capsada-Munsech advance the theory, without justification, that any positive effect from selective schools in terms of access to higher education is due to ‘badging’. This hypothesis appears unlikely, given the strong pressure placed upon universities by the Office for Students to diversify their impact and is, in any case, entirely unevidenced.
The cohort studies by Sullivan are one of the most interesting parts of the paper; however, they are not particularly relevant to the current debate. Sullivan finds ‘that the benefits of attending grammar school for the 1958 cohort (National Child Development Study) were balanced by the disadvantages of attending a secondary modern school, leading to no overall social mobility advantage.’ Yet Sullivan fails to mention that the tripartite system of the 1950s was structurally inequitable, with grammar schools generally funded at a higher per pupil rate than secondary moderns, with consequent impact on resources, staffing and teacher quality. Almost no modern grammar school advocate supports such a funding system and many, including myself, would argue that secondary modern schools should receive proportionately more funding, via a system such as the pupil premium or equivalent – meaning that any conclusions drawn from 1958 or 1970 are of limited value.
All of the issues above are characteristic of the unconscious bias I set out in the original paper that colours much research in this area. Boliver and Capsada-Munsech strongly dispute the suggestion that academics could be influenced by unconscious bias, despite widespread literature suggesting that such biases are pervasive in all areas of life from hiring and employment, through assessment, to project planning and consumer spending decisions. Boliver, in particular, has a track record of using statistics in questionable ways to support an ideologically-held position. In a widely cited paper last year she extrapolated the performance of a highly atypical subgroup (students with lower grades accepted to higher tariff institutions) to draw general conclusions about all students with lower grades, a flawed methodology which entirely invalidates any conclusions drawn. In the same paper, data was cherry-picked to argue in the abstract that, “academic entry requirements for disadvantaged learners can be reduced substantially without setting these students up to fail at university” despite the fact that data in the paper itself showed, ‘while students entering higher-tariff providers with AAB at A-level have a 76 percent chance of graduating with a first or upper second class degree, the figure for those entering with BCC is rather lower at 46 percent.’ This example demonstrates clearly that academic findings in education studies both can be and at times are distorted to support the preferred views of the authors.
The final section of the paper is more philosophical in nature, arguing for the introduction of a wholly comprehensive university system in the UK, a subject that the author, Blackman, has addressed before. I have the highest respect for Blackman’s principles and would be the first to acknowledge that the Open University – in the paper’s words, ‘the only truly comprehensive university in the UK’ – is of tremendous value, providing an invaluable route for people of all backgrounds to receive a second chance at higher education and transform their lives. Yet in this area he errs: the fact that the UK benefits greatly from one university such as the Open University does not mean that it would benefit more if every university were like the Open University.
Education is an engine of social mobility. For high-ability students from disadvantaged backgrounds, a good degree from a top university can have a transformative impact on their lives. Due to the highly selective nature of our most prestigious universities, they are able to tailor their courses and assessment methods accordingly, reaching a high level of academic attainment that simply would not be possible under a comprehensive university system. Accordingly, while the impact of social background is not entirely overcome, graduating from these institutions is unrivalled in opening doors to top professional careers in business, the law, academia or politics – including for those from less advantaged backgrounds.
In the one-size-fits-all system approach proposed by Blackman, this ladder would be kicked away. In a system that catered to the lowest common denominator, there would be no opportunity for those outside of the elite to demonstrate their potential through excelling in education. Accordingly, access to the top jobs would become increasingly dominated by those already inside the charmed circle: those whose parents had connections, contacts and existing networks. No doubt the children of the authors of this paper would continue do well under such a system, but many others would not.
The paper does acknowledge the widespread and persistent support (across all social classes) for academic selection. Furlong and Lunt attribute this to ignorance, saying ‘families are not aware of the evidence’, a typically elitist response characteristic of a number of other recent topical debates. If, as some papers suggest, selection increases the variance of academic performance while maintaining the mean, it is not – despite what the authors of today’s report claim – self-evident whether or not it is desirable. For many ordinary working families, a higher chance of a future child attending Oxbridge or becoming a doctor may be worth simultaneously accepting a slightly increased chance of reduced academic performance – perhaps particularly so if one lives in Southampton, Thurrock, or Rochdale, where a recent report for the Sutton Trust found that not a single student had secured an acceptance to Oxbridge in the three years studied.
Finally, perhaps the most disappointing element of the paper published today is the missed opportunity. In my original paper, I explicitly called for further research in this area, suggesting that to counter the impact of unconscious bias (on either side, ‘the research team, or its advisory panel, should be explicitly required to include at least one supporter and at least one opponent of grammar schools.’ And yet all of the authors of today’s report appear to be those who are already staunchly in favour of a comprehensive system, who consider the ‘social segregation’ caused by selection to be intrinsically morally wrong and who consider that we ‘know unequivocally’ that selection is bad.
Academic selection is a fundamentally complex subject, involving complex trade-offs that impact on different individuals in society in diverse and varied ways. Selection may be applied at different ages, on a general or a specialist basis, and both between and within schools, for example with streaming and setting. It may be applied with differing degrees of flexibility or movement between schools and under a wide variety of different funding frameworks, from the highly inequitable one prevailing in the 1950s to one significantly more progressive, in which greater resources are provided to those most in need. There are no easy answers to any of this – but it is a matter which deserves to be discussed and researched by a diverse group of individuals with different perspectives, not a closed circle of those who have already made up their minds.