The case for a British SAT

Monitoring universities and promoting meritocracy

The American SAT, introduced in 1926, marked the first major intrusion of psychometrics into the lives of ordinary people, at least outside of the Army. It contains two sections — Reading and Maths — and both are multiple-choice and marked by a computer, completely removing any subjectivity from the marking. Raw scores are then converted to ‘scaled’ scores out of 1600 (each section is scored between 200 and 800), with a mean of around 1050 (somewhat higher than the 1000 which would be required for a perfectly normally distributed set of test scores) and a standard deviation of around 217.

It has never been a popular test, and it is perhaps unsurprising that there are few advocates for copying such an explicitly inegalitarian (yet also rather philistine) test in Europe. Each year, the SAT gives nearly two million people a single number that purports to tell the world how able they are in not just some, but all intellectual fields — this from a test that tests nothing much specific, and takes just three hours to finish! The curious survival of such a test speaks to its very real benefits.

The SAT was designed to replace the diverse, subjective, and often frankly unprofessional entrance exams for the Ivy League universities. It was pioneered by James Bryant Conant, one of Harvard’s most transformative presidents. Conant wanted to find a better way for Harvard to attract America’s best and brightest, regardless of background. In this task, he enlisted the help of the patrician Henry Chauncey — the founder of ETS, which administers the SAT to this day. The SAT was, in effect, used as a battering ram by the insurgent meritocrats — Conant himself was from a modest background — who wanted to destroy the foundations of the old, quasi-aristocratic Ivy League, where being a ‘good chap’ mattered as much (if not more) than academic excellence. No longer would a small number of New England boarding schools entirely dominate America’s oldest and most prestigious colleges. In some sense, we could even say that the SAT helped create modern American higher education.

The SAT slowly but surely won more and more advocates, especially during the rapid expansion of higher education following Second World War. At first, it was mostly just used by Harvard to allocate scholarships. But today, almost all applicants to almost all American universities will take either the SAT or its little brother, the ACT. Although during the Pandemic there were attempts to move away from the test by radical egalitarians — who rightly viewed the test as an obstacle to their social engineering work — this is now widely seen as a failed experiment and has mostly been reversed. This is unsurprising: SAT scores continue to be one of the best predictors of performance in college.

In Britain, by contrast, only a very small proportion of university applicants will take (to use the psychology jargon) an ‘aptitude test’. This means that the test is designed to probe an applicant’s intellectual ability to assimilate new concepts. This can be contrasted with an ‘achievement test’, which will test an applicant’s knowledge of a certain subject area. In some sense, we could say that while an ‘aptitude test’ usually looks more for future potential, an ‘achievement test’ usually looks more for current ability. And we could say that, in general, while an ‘achievement test’ not only encourages but requires prior preparation, the ideal ‘aptitude test’ seeks to be as untrainable as possible.

In reality, there is a use for both of these types of test, both in university admissions and in society more generally. A society which had no time at all for the ‘achievement test’ would almost certainly be a philistine one (though perhaps some readers may welcome this). It would also be an extremely dull one — although an overemphasis on ‘achievement tests’, with their tendency to encourage rote learning, may well have similar results; a kind of horseshoe effect. Moreover, in the real world, we do actually care more about ‘achievement’ than we do about ‘aptitude’: it is all very well that a clever person could learn about brain surgery if they really wanted to, assuming they were diligent enough, but we wouldn’t want someone with high SAT scores but no degree to be operating on us. If we are to introduce a British version of the SAT, it should be taken alongside A-Levels, rather than wholly replacing them (except for a small minority of applicants — see below).

It is also rare that a test can so easily be put into one or another box. In reality, most ‘aptitude tests’ are at least partly trainable, and there are often strong societal pressures (due to ‘striverism’) to make them more so. Nowadays, some ‘aptitude tests’ deny that they are ‘aptitude tests’ at all, most notably the American SAT — which used to be an acronym for ‘Scholastic Aptitude Test’, but now (officially) is no longer an acronym. The reality is that today’s SAT is at least somewhat trainable; nonetheless, the test to this day probably has much more severely diminishing returns on inputs of both time and money than any other element of the American college admissions process.

Probably the only tests widely used in Britain that qualify for the label of ‘aptitude tests’ are the LNAT (used for Law applicants at ten high-ranking British universities) and the UCAT (mandatory for Medicine and Dentistry applicants at all British universities). It is telling that these tests are mostly used to help differentiate between candidates for the most oversubscribed courses in the country. The other main test used by British universities, the STEP (a mathematics test used by Cambridge, Imperial and Warwick for applicants for certain degrees), is definitely more of an ‘achievement test’ than an ‘aptitude test’. Instead, nearly all home applicants to English universities (especially outside of Oxbridge) will be judged almost solely by their A-Levels, alongside a personal statement that generally holds little weight so long as it is vaguely literate.

Adding an SAT score to the mix would give useful information for the universities in making admissions decisions. I envisage that a British version of the SAT would have four one-hour sections. These would be Verbal, Non-Verbal, Quantitative, and Writing. These could be weighted as appropriate (many readers may want to reduce the weight of the Writing section, which is inherently more subjective than the other sections).

The Verbal section would be similar to the existing LNAT, mostly containing reading comprehension questions, perhaps with some added verbal analogies questions. The Non-Verbal section would contain Raven’s Progressive Matrices. The Quantitative section would not require much knowledge of mathematics — certainly nothing beyond GCSE — but would focus upon problem solving and (most importantly) the demonstration of a very basic competency with statistics. The Writing section would be similar to the Writing section in the American GRE, but easier, as it is targeted at school leavers rather than university graduates.

The new test would be flexible: there would be multiple sittings scheduled over the year. Three of the four sections would be multiple choice, and could be marked by a machine, saving time and money. The existence of the Quantitative section should address many of the concerns of those who want to make everyone learn Maths to eighteen without needing to recruit enormous numbers of new Maths teachers. The body responsible for administering the new test would require professional psychometricians, much like ETS (which administers the SAT and GRE); this would not be a small operation, but it would not be prohibitively expensive one either.

---

Introducing a British version of the SAT, henceforth ‘BSAT’, would have a number of important benefits. These can be put into two main categories: firstly, monitoring universities, which presently have power (or at least money) but no real responsibility; and secondly, promoting meritocracy, as the creators of the American SAT always intended.

Firstly, universities — defending themselves against accusations that degrees are a waste of taxpayer (or student) money, and that they fail to teach any tangible skills — often allege that they improve various intangible skills: reading comprehension, writing ability, critical thinking, and so on. The BSAT would help us to better verify these claims, rather just giving them money and taking them at their word.

To do so, it would not be necessary to retest the entire cohort: a representative sample would do, especially if there were suitable financial incentives for improved performance. The results would then be compared: firstly, with the BSAT results of the graduates before they matriculated; and secondly, with the results of non-graduates with similar SAT scores who were also retested (with equivalent financial incentives) at 21. (With very high BSAT scores, it might prove challenging to obtain a big enough sample for the non-graduate group.) This would not be a small or inexpensive operation, but the costs would still be tiny compared to the sums of money that are at play in higher education.

My expectation is that we would see very little (albeit perhaps not zero) improvement in performance among graduates in either the Verbal or Non-Verbal sections above the non-graduate cohort, but that we might see some meaningful improvement in the Writing section. The Quantitative section would be the wildcard.

It would become harder for universities to justify admitting home applicants who are only dubiously literate — taking public funds to print credentials that are not worth the paper they are written on — as extremely low BSAT scores would make these cynical decisions more transparent than merely referring to a student who got bad grades in their specific A-Level subjects of ‘Psychology’, ‘Politics’, and ‘Physical Education’; after all, maybe they just had a bad day, or maybe they just didn’t enjoy these subjects.

Universities also use the differences between the systems of secondary education around the world, which are inherently difficult to compare, to justify accepting international students (who offer higher fees) on what at least appears to be more lax terms than home students. This is perhaps not true for all courses, but there are good grounds for suspicion, especially in light of the recent scandal around foundation years. The international students are rarely paying for the quality of the course in any case: they are in fact just using the university as a visa printer. Obviously, it would not be reasonable to expect all applicants to British universities, regardless of where they are educated, to take A-Levels. It would, however, be perfectly reasonable to expect all applicants, regardless of nationality, to take the BSAT, as is already the case with international applicants to American universities and the SAT. This would help make home and international applicants more easily comparable, and would make the widespread practice of accepting international students with lower grades than their home counterparts much more obvious.

Secondly, the BSAT would help promote meritocracy. It would give universities additional information on high-performing applicants, helping them make fairer admissions decisions. Presently, due to decades of relentless grade inflation, the number of students with AAA or higher at A-Level has reached extraordinary levels, approaching a quarter of those taking A-Levels in the 2021/2 academic year. Even the A* — which itself was introduced in order to provide more differentiation between top applicants — is becoming less and less effective at differentiating applicants, especially in subjects like Maths, where the A* is now very common. It is difficult to know how exactly universities are actually choosing between this sea of outwardly identical applicants at present — especially outside of Oxbridge, which at least has interviews, whatever one might think about the system — but it seems likely that it is a mixture of randomness, ‘vibes’, and (perhaps most insidiously) not-so subtle attempts to meet various formal or informal diversity targets. In fact, this is a situation that the universities are probably quietly rather happy with, as it allows them to make their blatant social engineering plausibly deniable.

The SAT in the United States, by contrast, still strongly differentiates between the best applicants, even those who score in the top decile: a score of 1350 puts you at the 90th percentile, while a score of 1530 puts you at the 99th percentile. The BSAT would also likely prove more resistant to the scourge of grade inflation than A-Levels have in recent decades. Dominic Cummings notes that a candidate in A-Level Maths who would have received a U in 1988 would receive a B/C by 2006, and presumably even higher a grade in 2024; in other subjects, the effect was less dramatic but still substantial, at least one grade. By contrast, in the United States, combined average SAT scores have hardly budged in five decades: while average Maths scores have increased (especially since the ’90s), this has been mostly offset by a fall in average Reading scores.

The Pimlico Journal has consistently condemned ‘striverism’: in short, the belief that merit should be determined by effort alone, not innate ability — an ideology that would certainly be undermined by the BSAT. Such ideas have wasted money and almost certainly have suppressed fertility rates; they have also spread a noxious educational culture across Britain, centred around Michaela School’s ‘tough-love progressivism’. Reducing the efficacy of such pedagogical tactics by reducing the trainability of important examinations is thus crucial, and would be partially achieved by this scheme.

But the BSAT would not only be useful in both differentiating between and rewarding the best applicants. It also would be a far more effective system at locating genuine ‘diamonds in the rough’ than our system at present. Although the importance of teaching in academic performance at present is usually grossly overstated, there is still some truth to the idea that, for A-Levels — especially with A-Levels becoming more and more trainable, as mark schemes become increasingly rigid and transparent — bad teaching does still matter to a certain extent. Culture also matters: you cannot do well in A-Levels if you do not revise enough. A genuine hothouse, whether state or private, and whether spearheaded by teachers or parents, might well be able to get even deeply mediocre students sufficient grades for a course at a Russell Group university nowadays.

Presently, the system works by identifying ‘underprivileged’ students by some means; they are then given ‘contextual offers’, which reduce the required A-Level grades. Yet A-Levels were never intended to be used to find those with high but as yet unfulfilled potential: as mentioned earlier, A-Levels are achievement, not aptitude tests. Perhaps those students who needed the contextual offers, however ‘underprivileged’, were just not very clever in the first place, just like the very many ‘privileged’ students who would also be unable to meet the standard offer.

If we want a system that places some emphasis upon the finding of ‘diamonds in the rough’, the BSAT would be much more helpful for this purpose: the combination of a very bad school and excellent BSAT scores would be the two linchpins of an improved contextual offer system once the BSAT is implemented. For the minority of students with contextual offers, A-Levels would be removed from the offer altogether (or set at an extremely low level), but the required BSAT score would be identical to (or perhaps actually higher than) the majority of students with standard offers.

---

Image credits: Tdorante10, Creative Commons Attribution-Share Alike 4.0

This article was written by Nigel Forrester, our editor-in-chief. Have a pitch? Send it to pimlicojournal@substack.com.

If you enjoyed this article, please consider subscribing. If you are already subscribed, why not upgrade to a paid subscription?