Yr 6 sats - can anyone explain the 100 scaled score average

[itsabongthing]

Sorry if this is obvious but can anyone explain the sats scaled scores?

I get that 100 is the ‘expected’ level. But how do they decide the scaling.

It 100 the average? But it can’t be because then 50% would be below and 50% above.

And I’ve heard it talked about 2/3 of kids roughly getting the expected but it varies slightly year on year.

So, so they ‘fix’ the scaled score so that roughly 2/3 will be above the 100?
Or is it the other way round?

How do they decide how to scale it, if it’s not the average? But it can’t be?

Sorry I just can’t get my head round it and it’s bugging me!

From the DfE

What is a scaled score?
Tests are developed to the same specification each year. However, because the questions must be different, the difficulty of tests may vary. This means we need to convert the total number of marks a pupil gets in a test (their ‘raw’ score) into a scaled score to ensure we can make accurate comparisons of performance over time.

Pupils scoring at least 100 will have met the expected standard on the test. However, given that the difficulty of the tests may vary each year, the number of raw score marks needed to achieve a scaled score of 100 may also change. For example, if the overall difficulty of a test decreases compared to previous years, the raw score required to meet the expected standard will increase. Similarly, if the test is more difficult, the raw score required to meet the expected standard will decrease.

In 2016, panels of teachers set the raw score required to meet the expected standard. We have used data from trialling to maintain that standard for the tests from 2017 onwards.

It's not an average. It is the "pass mark" and in the raw score required to achieve that level depends on the difficulty of the test.

Ok thank you.

How do they decide how difficult the paper is?

Presumably if kids score less on average?

Sounds a bit like they expect 2/3 to achieve the expected level so the 100 will be roughly pinned to that.

Some teacher sit around for hours locked in a dark room and decide how difficult each question is! They then decide how many marks are needed for the pass mark to ensure the paper is the same difficulty as the previous year.

Sounds a bit like they expect 2/3 to achieve the expected level so the 100 will be roughly pinned to that.

There are several exam markers who have said that this is exactly how they have been told it works. Also, that this is true both for the Key Stage Two assessments and for GCSEs.

There are many who believe that the "pass" percentages are set in advance and so the mark for a scaled score of 100 is only decided upon after all the papers have been marked. That way if they want a 75% "pass" rate, they set the mark for a 100 scaled score at whatever mark would cause 75% of the children to get a scaled 100 or more.

Other people have refuted this though as this would essentially be fixing the national levels year after year.

Hi itsabongthing,

I think you may be mixing up the theory of standardized and scaled score. The former follow the laws of a normal distribution while the other sets the expected standard.

www.scaledscore.com/scaled-score-vs-standard-score.html

Thanks

I'm curious about the ins and outs of this. Is the 100 mark decided before marking the tests or after? Why not make sure the difficulty is equal each year and have a set pass mark? To me it feels like a manipulation of the figures so that gov can say 'look how policy x worked! The SATS scores have improved!'

Why not make sure the difficulty is equal each year and have a set pass mark

Because such a thing is impossible, and it certainly is attempted, which is why there's so little variation in what the raw scores correlate to on the scaled scores.

Even if you somehow made every test of equal "difficulty", a single news report the day before it was administered suddenly popularising a particular word, or subject or something would completely change the difficulty of the test.

Imagine one of the spelling test words was Potato, and then Dan Quayle came along and made it headline news...

ilovesushi,

Yes, the expected standard is always of a scaled score of 100 and above (every year). The mapping between the raw score and scaled score changes between years. So if the government wanted proof of their policy working they can manipulate this table ;-)

@itsabongthing you asked something like Is 100 the average? But it can’t be because then 50% would be below and 50% above.

The average aka the mean doesn't split the pupil population in half 50%-50%. The median value does that.

The easiest interpretation of scaled scores: scales scores are raw test scores converted to a 80-120 range, where 100 is the expected government standard. Here's the conversion table: schoolsweek.co.uk/scaled-scores-for-2019-key-stage-2-sats-announced/

I find it worrying that only two thirds (65%) achieved that expected standard in 2019 and one third has not. My interpretation is that the education system is failing one third of the children, since they didn't achieve the standard set by them.

This link shows you the average (mean) scores for 2019: 104 Reading, 105 maths, 106 spag.

www.gov.uk/government/publications/national-curriculum-assessments-key-stage-2-2019-interim/national-curriculum-assessments-at-key-stage-2-in-england-2019-interim

Thanks for all the interesting comments.

I agree that it’s not something for the government to be proud of - that only 2/3 of children are meeting their own ‘expected standard’. And not nice for 1/3 of kids and parents to end primary school being told they fall short of the expected standard.

One could think that if the proportion stays at around 2/3, that maybe their self imposed ‘expected standard’ is unrealistic and too hard!!