I'll try summarising these terms since the NESA post is quite long:
Atar stands for 'Australian Tertiary Admissions Rank. Atars are given in increments of 0.05 so for example 75.45, 92.00, 96.55 etc.
The process for calculating ATAR is as follows:
- NESA gives UAC raw HSC marks (Ill get into what RAW and scaled marks are in the next section.)
- HSC marks are 'scaled'
- ONLY the best 10 units a student studies (with 2units taken from English and 8 taken from other subjects)
- Using NESA's magic formula they spit out a mark out of 500 for students (Usually students have no clue what their mark is)
- NESA then ranks students based on these marks and ATAR's are produced
A very important concept to note is that Atar's are ranks and NOT marks. So, for example, let's say you get a 92.00 atar. This literally means you beat 92% of people in your cohort. Cohorts are state-wise, so NSW, ACT, SA, WA, QLD etc.
For some states, it may be easier to achieve a higher atar compared to other states due to the relative number of students. For example, only 11856 students completed their SACE exam (SA equivalent of HSC) whereas a whopping 54,894 students completed their HSC in 2020. Therefore to say get a 92.00 atar one would need to beat 10,907 students in SACE, whereas the same student would need to beat 50,502 students in NSW, almost 5 times the amount.
What Marks Count and how important are they?
Firstly a common myth is that year 11 marks count towards your HSC this is simply NOT true. The only marks that count towards your HSC are your year 12 marks. Your year 12 marks are divided as follows:
50% of your marks are from internals (that is, your in-school assignments/tests/quizzes etc. )
50% of your marks are from externals (That is, the one final HSC exam you do sometime October-November)
As you can see whilst your internal marks are important really the bulk of your marks come from your external exam. For example let's say you get full marks for all your inschool exams but bomb your final HSC and get a say 50%, you are left with a 75% in total (despite getting 100% for a years worth of inschool assignments!). It's not as simple as this but hopefully that highlights the point.
Now we are getting to the juicy stuff. See NESA faces a problem. Let's say we a student from say James Ruse and a student from a school ranked 500, both got 90% for their internal math test. Achieving a 90% in James Ruse is vastly more difficult than achieving a 90% in the rank 500 schools, yet both these students on paper have the same mark right?
This is where moderation comes in.
The process for moderation is:
- Raw marks are calculated and a mean and standard deviation is produced for INTERNALS (Raw marks generally follow a normal distribution)
- Mean and standard deviation of INTERNAL is moderated to follow that of the EXTERNAL mean and standard deviation
- The top Internal score is replaced by the top external score and usually, the bottom internal score is replaced by the bottom external score
Now, this may seem a bit confusing so let's look at an example.
Say Student A who was rank 1 for internal got a 100% internal and a 80% external and student B who was rank 5 got a 60% internal and 90% external (which is the highest external). Then Student A's internal becomes 90% (the highest external) and their 80% remains the same, so in total, they get an 85% average. Notice how the raw marks for internals are near meaningless since it will get replaced anyways by the external mark. So the important thing to note is RANKS are what are important for internals. Also, note that your external mark does NOT change depending on the strength of your cohort. Also, refer to this table by NESA:
Finally Scaling addresses the second problem NESA has. Some subjects are inherently more difficult than others, for example, Extension 2 maths is clearly more difficult than advanced maths, so a raw 90% for ext 2 is much more difficult to achieve than a 90% in Advanced. Therefore some adjustments must be made to account for this. For example, a 91 raw mark in advanced scaled to a 79 and a 93 raw mark in extension 2 remained the same at 93. This shows the relative difficulty of subjects. Note that scaling is entirely dependent on how the cohort performs. Also, note that moderation occurs after scaling.
This is a quick summary of the process, let me know of any errors and hope this was useful.