How predictive is ATAR of university results?

In response to my Grattan Institute colleague Ittima Cherastidtham’s op-ed supporting ATAR, Victoria University VC Peter Dawkins and Professor Yong Zhao argue in The Australian that

“The focus on maximising the ATAR through Year 12 exams, however, tends to lead to coaching of exam technique, so students memorise answers to questions that are designed to promote critical thinking.”

Coaching can boost student results. I suspect it is one reason that students from private and selective government schools tend to slightly under-perform at university relative to students with the same ATARs from non-selective government schools.

If it was just coaching that explained ATARs, they would not have any predictive value for future academic performance at university, which does not offer school-level hand-holding, and at which students take sometimes quite different subjects. But ATAR does have predictive value.

As the chart below shows, as ATARs go down students become more likely to fail half or more of their subjects in first semester – a fail rate that will send them to the unsatisfactory progress committee unless improved.

low ATAR fail rate

The chart above analyses all students with a recorded ATAR between 2006 and 2016, which may not take account of improvements in later years. The data is more volatile when examined by individual years, but overall there is no trend towards better outcomes, as the chart below shows.

atar_blogpost (005)

Although ATAR is identifying different levels of general academic ability, it is possible that additional or other information would improve on it.

The Dawkins and Zhao article notes that Victoria University is using alternatives such as  study scores for individual subjects. It is quite plausible that having done particular school subjects that are directly relevant to a university course, and having done well in them,  would have more predictive power than an ATAR that is a composite of different subject results. There is some published evidence for the subject link (eg here), but so far as I am aware not anything that specifically looks at whether it helps predict outcomes for low-ATAR students.

We are still a long way from an evidence base that supports the conclusion that ATAR is less reliable than alternative admission systems.


One thought on “How predictive is ATAR of university results?

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s