In the last few weeks, the Australian Tertiary Admission Rank (ATAR) has come under renewed criticism. A paper from the Mitchell Institute started this, and Swinburne University VC Linda Kristjanson followed up with an op-ed on one of their alternative entry systems.
These critiques have a history. More than 20 years ago, then education minister Amanda Vanstone gave a speech attacking the ‘tyranny of the TER’, an ATAR predecessor called the Tertiary Entrance Rank. ATAR’s decline has long been championed. But despite what the Mitchell paper suggests, ATAR is still used in the vast majority of school leaver university admissions (Mitchell’s numbers are due more to the rise of other academic results, such as past higher or vocational education, than non-ATAR school leaver admissions, although these are also increasing).
Some form of ranked school results has been used for university admissions since the 1960s. Previously it was enough to ‘matriculate’ (complete Year 12), but demand for higher education was exceeding the places available, and so school results were used to allocate places to applicants. But this history shows that there is an alternative – admit everyone who meets some basic threshold, and then cull them. University policy reviews from the 1950s and 1960s report fail rates that are very high by today’s standards.
There is something to be said for this admit-and-attrit approach to selection – let applicants try university and see how they do. It avoids proxies based on an imperfectly-analogous environment such as school, or admissions tests that can potentially be gamed or only measure some attributes needed for success at university. To a substantial extent, as a forthcoming Grattan report will show, trial-and-error selection after enrolment is still a major way of deciding who gets a longer-term university place.
But issues of fairness, efficiency, and student protection work in favour of retaining an active role for universities in selecting students.
Even if the higher education system can meet overall demand, there will always be capacity issues in particular courses and institutions. We need some system of sorting out who gets in. We could use non-academic means of allocating access, such as pricing, zoning, lotteries or queuing – several of these are used in the school system. But none are likely to win widespread support in higher education. There is still a strong sense that academic criteria should be used for admission to academic courses, even if they are supplemented with other indicators. Some or all of the alternative mechanisms would increase inefficiencies such as weak matching between courses and students, difficult-to-teach wider ranges of abilities in the classroom, and delays in commencing studies.
Many ATAR critics are concerned with ATAR obsession in schools. It does seem excessive in some cases. But by weakening the entry requirement to just a minimum threshold of school completion, all non-academic methods of admission would lead to university students who were less well-prepared for their studies, as the incentive to work hard in Year 12 would diminish. Especially in the courses that build on school knowledge, university students would be more likely to fail as a result. Soft entry schemes do not abolish academic selection – they just delay it until the first year of university.
All non-academic admission methods would also absolve the universities of any responsibility to protect prospective students from poor choices that have a high risk of not bringing benefits worth the time and money spent.
So we are back to looking for selection tools relevant to an academic course, and from a university perspective it is hard to go past a number paid for by the school system – ATAR. And despite the stresses of ATAR for students, it generally avoids the double stress of school and another admission test such as the SAT in the United States or the very tough Chinese higher education entrance exam. If we stop producing school results in ways that are useful to universities, at least the academically elite universities will develop additional onerous admission tests.
As the Mitchell paper notes, ATAR is reasonably predictive of success at university. Indeed, for above-60 ATAR school leavers their ATAR is a rough guide to their prospects of completing a degree. Someone with a 60-something ATAR has a 60-something per cent prospect of getting a degree, someone with a 70-something ATAR has a 70-something per cent prospect of getting a degree, and so on. As the chart below shows, most school leavers with ATARs are at the higher end of the range, and so also have reasonably good prospects of successfully completing a degree.
At these higher levels, ATAR is being used principally as a way of allocating the scarce resource of student places in particular courses and universities. It can still seem arbitrary at the entry cut-off (or whatever we are supposed to call it these days), but school results are recognised as reasonable and fair in an academic context. Many universities give bonus ATAR points for selection purposes, or require lower ATARs, for students from disadvantaged backgrounds. This does not violate the broad principles being used in selection, both because it recognises that the ATARs of students from disadvantaged backgrounds tend to under-predict their academic ability (empirical evidence and potential explanations here), and because equity is an established university mission and policy goal.
At the lower levels the issues, and the metrics, are more complex. Here, I don’t think the main issue is a fair allocation of a scarce benefit, a place in a course or university for which demand significantly exceeds supply. Rather, the biggest issue is that there is an elevated chance that the benefits of enrolling won’t be worth the costs in time and money.
Below 60, ATAR is less predictive of completion prospects – lower than 60-plus, but on the evidence to date a 40-something ATAR is not more risky than a 50-something ATAR. They all have a little better than a 50:50 chance of completing.* ATAR tells us that the risk is high, but doesn’t tell us how to work out which students have a reasonable chance of success and which do not.
Universities have various alternative entry schemes for students with lower ATARs. In principle, looking for ways of trying to work out which side of the 50:50 a prospective student might be on is sensible. But universities rarely publish evidence showing that these schemes add predictive value. The four, six and nine year complete-or-retain statistics for students with 30-49 and 50-59 ATARs don’t show any clear trend, suggesting that in aggregate we are not getting any better at picking the lower-ATAR students who will complete.
If anything, the situation with lower-ATAR students is getting worse. For the 50-59 ATAR group, fewer of those who are not going to get a degree or still be studying after four years are leaving in first year, meaning than admit-and-attrit is becoming less successful at exiting students quickly when things aren’t working out.
It’s mainly 50-something ATAR students taking things into their own hands that limits losses. More than three-quarters of the school leavers with a 50-something ATAR who apply get an offer. But many of them don’t apply in the first place. In a forthcoming Grattan report, we will publish research showing that almost every school leaver with an ATAR above 80 applies, but shares decline after that with a small majority of the 50-something ATARs applying. Of the 50-something ATAR students who do receive an offer, a third don’t accept. And of those who do accept, more leave prior to the census date at which they become liable to pay student contributions.
Of course, lower-ATAR students would be aware of their general academic aptitude prior to receiving their ATAR. But the ATAR does make where they stand clear, and so is likely to be a useful input into their decision making.
Nobody believes that ATAR is the only potential selection tool for school leavers, and universities have always used other information where it is relevant, readily available, and cost effective. It would be great if we could find cheap and reliable ways of giving better advice to lower-ATAR school leavers. But I find it odd that the most useful indicator we have of who should get a university place gets such a bad press.
* The main reason for these similar chances is that although ATAR is a straight sequence of numbers ranking everyone in an age cohort, the underlying distribution of academic ability and study scores is in a bell curve. There are lots of people in the middle with quite similar underlying results. ATAR exaggerates the differences between them (and under-states at the high end – the study score gap between an ATAR of 99.95 and 89.95 is much larger than between 49.95 and 59.95). In this sense, using the underlying study scores might clarify things more than ATAR during the entry process. But study scores are harder to compare across different Year 12 results – across state borders, at different times, or alternatives like the IB. That’s why in the 1990s most states moved to ATAR-antecedent cohort ranking systems.