Over the long summer holiday, high-achieving sixth formers are encouraged to prepare for university admissions tests that take place in the autumn.
Many find this process hard or off-putting, and give up prior to applying. This raises questions about who rises to the challenge and how this impacts participation in the top courses.
Admissions testing in context
In England, admissions testing for university entrance used to be confined to Oxbridge, where selectors used “entrance papers” to select between high numbers of similarly qualified candidates. Oxbridge admissions testing has changed form over many decades, in response to internal and external pressure to widen participation, long before that term was coined.
The 1980s version of testing was satirised in The History Boys as requiring pupils to set aside what they had been taught in school and “perform” in enigmatic ways, incomprehensible to regular teachers. The play captured the English approach to fair access at that time. The game required insider knowledge but was tolerated because it affected so few.
The business of widening participation has since become more serious: regulation, data, monitoring, action plans, access gaps and so on. We are all familiar with unfairness throughout the education system and are increasingly committed to doing something about it. School-to-university transition is a key point to set things straight because there is regulatory and media pressure to ensure that privilege does not determine access and institutions are committed to the mission.
Widening participation and admissions testing are a complicated partnership. It is often argued that tests are objective and can level the playing field, allowing talent to shine through regardless of education and social background. However, testing knowledge and skills that have not been taught in school will tend to favour the advantaged and well-resourced, as well as those confident enough to take a punt. The myriad companies offering professional test tutoring attest to this.
The testing industry
Despite the obvious tension, the number of institutions using tests, and the number of courses for which they are mandatory, has increased in recent years. Institutions facing very high numbers of applicants, including from overseas, are relying on testing as a fair and efficient way to sift, benchmark and select candidates.
As more institutions have moved towards testing, so access and efficiency concerns have created a push towards consolidation and outsourcing. There are now two broad models in play: testing for all institutions (as in medicine), and testing for the most selective institutions (as in mathematics, engineering, computer science, economics, science, law, PPE and history).
The tests for medicine (UCAT) and law (LNAT) are run by consortia of institutions. Three tests that apply across multiple subjects (TMUA, ESAT and TARA) are run by UAT-UK, a non-profit joint venture between the University of Cambridge and Imperial College London, and used by seven top English universities.
These changes have shaped the style of testing itself. All of these tests are now digitised and administered by Pearson VUE at testing centres and generally focus on multiple-choice formats rather than long-form questions. The Cambridge STEP for mathematics is the only test I know of that remains in-house and marked solely by hand.
Who regulates admissions testing?
But who is now responsible for monitoring these tests and ensuring that their effects are fair?
Admissions testing sits outside of the remit of both OfS and Ofqual. The outsourced model means that no single institution or body is accountable for ensuring that the tests themselves are fair and incorporated into admissions systems in equitable ways. This matters because fairness is not self-evident; it requires ongoing scrutiny, reflection and debate. Worryingly, there are red flags to suggest that testing constructs new and unfair barriers to entry in subjects and institutions where patterns of participation are already inequitable.
A glance at the US history of college admissions testing should be warning enough. Mandatory testing through the SAT and ACT has been heavily criticised as discriminatory in terms of class and race, resulting in test-optional policies at many colleges and ongoing controversy. In England, reports published by institutions and testing bodies show that disadvantaged students and women consistently underperform in many admissions tests. One explanation for this variation is that the tests reveal genuine differences in aptitude. Another is that they reward a particular kind of academic performance involving speed, confidence, composure under pressure and a willingness to take risks.
The position of women is particularly concerning because reliance on testing is common in subjects that are already male dominated: mathematics, computer science, engineering and economics. While it is now widely accepted that social and educational disadvantage should be considered alongside other admissions information, institutions are legally prohibited from considering sex. This means that women’s underperformance in tests is effectively baked in to admissions decisions as soon as tests become part of the system.
Another important but rarely discussed concern is the effect that testing has before an application is made. My study of mathematics admissions suggests that students measure themselves against the tests and make choices about where to apply based on their assessment of stress, risk and uncertainty. Some students interpret the tests as signalling a pressured and competitive course culture and alter their choices accordingly. These pre-application decisions do not appear in the data that monitors fairness. Tests may be shaping aspiration in ways that evade statistics and which are characterised as “choices”. Understanding these effects prior to the further expansion of testing would be wise.
What needs to be done?
In the past, admissions testing was a niche activity restricted to our two most elite institutions. As well as being responsible for the tests and monitoring their impact, Oxbridge also interviewed a high proportion of candidates to ensure a rounded picture is formed (this is still the case). This approach and accountability, combined with a high degree of media scrutiny, arguably resulted in sufficient checks and balances for the system to be unfair in a fair way. However, the new version of testing is a different beast, shaped by datafication, standardisation, consolidation, outsourcing and cost saving.
If testing is the way forward, we need to know that tests at least meet the standards laid out in the Fair Admissions Code of Practice, particularly that they are relevant to the courses for which they are required and predictive of degree outcomes. While evidence has been carefully collated for some long-standing tests, for many, the evidence is surprisingly scarce.
It is also worth ensuring that the ways tests are used in admissions systems are transparent and considered. Mechanistic approaches that rely heavily on test scores in isolation (such as screening) deserve challenge on the grounds that a single test performance cannot capture a full picture of an applicant. Regarding the impact on gender balance, institutions relying on tests in which women consistently underperform could ensure that other ways of demonstrating potential are also available. Finally, the impact of tests on pre-application choices could be an important signal of how tests are perceived by students and merits further investigation.
These concerns do not mean that testing cannot be part of a fair system or that other forms of assessment are invariably fairer. Rather, it suggests that we need to keep a close eye on the forms and effects of testing to ensure that it works to support inclusion as well as convenience.
