Standarised Age Scores (SAS) are a useful metric for determining a child's exam results, accounting for the different ages children might be when sitting the same exam.
Read on to find out what they mean, how they’re calculated, and how to use them.
What are standardised age scores?
Standardised age scoring (SAS) is used in most 11+ exams. It's a way to remove variables from test scores to compare results fairly. SAS ensures children aren't disadvantaged based on when they were born in the year.
Why are scores age standardised?
Two children in the same school year can be almost a year apart in age, giving the older child a big advantage. To balance this out and even their chances, younger pupils are given extra marks.
For example, children are exposed to more than 1,000 new words a year. The children born in September will have had a much larger opportunity to develop their vocabulary than those born in August.
Younger children can gain marks, but older children will never lose marks.
How are standardised age scores calculated?
Each child’s SAS is calculated from:
-
The number of questions they answered correctly
-
The difficulty of the questions they answered
-
Their age in years and months (e.g. 10 years and 9 months)
1. Attainment estimate
The first step in calculating an SAS is to work out an attainment estimate. This takes into account both the number of questions answered correctly and the difficulty of each question they answered correctly.
2. Age standardisation
Next, these scores are adjusted based on the child's age at the time they sit the exam. For example, let's say a child with a September birthday and a child with an August birthday both achieve a raw score of 100. The older child may end up with an SAS of 118, and the younger may be given an SAS of 121.
Find more detail below about the accuracy of SAS on Atom.
What's a good, low or average SAS?
- 80 or below is the lowest 10% of the year group
- 100 is the average SAS for the year group
- 120 or above is the top 10% of the year group
The highest SAS a child can achieve is 142 on an Atom test. This score would place them in the top 1% of children who took the test.
The distribution curve
SAS can be mapped onto a distribution curve. This shows how scores are distributed for a year group:
- Around 23% of children got a lower-than-average SAS (88 or below)
- Around 54% of children got an average SAS (89 to 111)
- Around 23% of children got a higher-than-average SAS (112 or above)
Using SAS on Atom School
We give your students an SAS for each mock test they complete on Atom School. This is the same scoring system used in the real exam, so it’s the best way to know what to expect.
The student's SAS tells you how they're likely to perform on the real exam, at their current level of knowledge. Keeping an eye on how their SAS develops over time will give you insights into their progress.
What SAS should our students aim for?
If your student is applying for selective schools, aim for an SAS of at least 115 by the time they’re ready to take the 11+.
For more competitive grammar or independent schools, aim for an SAS of 120 and above.
If they're not applying to a selective school, use the SAS as a general indicator of their progress. An SAS of 112 or above tells you they're working above average for their year group and are in a strong position to take their next steps such as starting secondary school.
SAS accuracy on Atom
How do you make sure your scoring is accurate?
Each year, children complete hundreds of thousands of tests across Atom's platforms. This data allows us to ensure the scoring on our platform is as close to the real position as possible.
We're constantly improving our platforms to give parents, tutors and teachers access to the most useful, robust data. After each exam season, we expand our data set, adding data from ~700,000 completed tests. We seek regular feedback to make sure our scoring is as accurate as possible to the results achieved in real exams.
My student only got one question wrong, but their SAS is only 130. Shouldn't it be higher?
Standardised age scores are mapped onto a distribution curve. An SAS of 100 places a child as exactly average for their year group. A score of 120 or above places a child in the top 10% of their year.
Only the top 4% of children achieve an SAS of 130+, and the top 1% achieve a score of 140–142. Standardised scores don't correspond directly to raw scores – they're calculated using factors such as the child's age and performance compared to peers.
If your child achieved an SAS of 130, this puts them in the top 4%. It shows that around 4% of children only answered one question incorrectly in this test.
My student got a different SAS on two tests even though they got the same number of correct answers. Why?
There are a few reasons why a child might get a different SAS despite the same raw score. SAS is based not just on how many questions were answered correctly, but also:
- The difficulty of the questions answered correctly
- The child's age in years and months.
On the test where they got a higher SAS, your students likely answered more difficult questions correctly.
If one or both of the tests were adaptive (like ISEB), the SAS is further impacted by any mistakes made early in the test. When taking an adaptive test, accuracy is key. Each answer will impact the difficulty of the questions your student sees later in the test.
- If a child gets easy questions correct, the test will show them more difficult questions. This gives them a greater opportunity to achieve a high SAS.
- Conversely, if a child gets easier questions wrong, the test will keep showing them easier questions until their accuracy improves.
Lastly, if your student is practising for multiple exams, you can expect them to achieve a different SAS in different types of test. This is because each SAS is calculated against other children who took the same test. The same raw score could place your students in an average position in one cohort, but above average in another cohort.
What is the highest/lowest SAS my students can achieve?
The highest SAS your students can achieve in an Atom mock test is 142, the lowest is 58. A score of 100 is average.
My student's SAS on Atom is very different to their CAT4 SAS. Why is this?
You may notice a slight difference between the scores that your students achieve in Atom mock tests vs scores they achieve in national assessments such as the CAT4.
Like the CAT4 SAS, the Atom Learning SAS has been calculated in exactly the same way by being verified every year based on analysis from a population of more than 150,000 students. However, you can expect Atom’s SAS to be 8-12 points below a CAT4 or other national test scores.
The Atom Learning SAS is lower than the CAT4 SAS because it has been standardised by, on average, higher-ability students at independent schools. The ‘average’ performance for students using Atom Learning is higher than the national average.
As an example, if a student completes an Atom mock test and a CAT4 test and achieves the same raw score (performing exactly the same in each test), their SAS on Atom will likely be ~8-10 points lower than their CAT4. If they achieved a SAS of 108 in the CAT4, their score on Atom would likely be ~100. This is because the majority of students using Atom are performing above the national average, therefore a student who is performing just above the national average (indicated by a CAT4 score of 108) is likely to be performing at an average level against the cohort of students using Atom.