KS2 SATs 2024: Maths paper analysis

“Question 17, though!”

So, we're (finally) free to discuss the rollercoaster ride that was the 2024 KS2 Maths SATs papers. We know lots of you look forward to Sophie B’s annual breakdown (not that kind) of the content, conversations, and controversies of the Arithmetic and Reasoning papers - SATs Springboard has the exclusive.

Buckle up for another candid and insightful journey through the numbers!

This year was the 3rd set of KS2 SATs papers since the hiatus due to Covid, where we didn’t see any 2020 or 2021 papers. We must remember (for most of us, it’s hard to forget!) that this year’s cohort of Year 6s missed a healthy chunk of Year 2 and Year 3 and therefore probably have some fundamental gaps in their learning. Definitely worth taking that into account when you look through your school’s results in July.

Maths reasoning SATs predictions vs reality

In February, we made some basic predictions about this year’s maths papers:

“Like previous years, it is still the case that in every set of maths papers so far, over half the content has come from the curricula of Years 3-5… Theoretically, based on the pass marks of almost all the previous papers, children should be able to meet the expected standards by almost exclusively knowing content from Years 3-5.”

We are still seeing over half the questions on the maths paper from the Years 3, 4 and 5 content domains. As long as the ‘pass’ mark is below 56% (and it hasn’t been higher than 55% previously), it would still be the case that, in theory, a child could achieve EXS without knowing any Year 6 content.

Please note that data from the 2016 tests has now been removed from all data tables. Not only were the tables beginning to get a bit busy from all the numbers (!) but, as the first paper under the new curriculum, data from the 2016 paper provided too many anomalies across the board, proving to be unhelpful when analysing data trends.

This is the highest amount of Year 4 content we’ve seen since 2017, and the lowest amount of Year 5 content we’ve seen ever!

Another prediction: “In theory, children should be able to meet the expected standards by exclusively knowing content from the calculations and FDP content domains. It’s probably safe to assume that the same topics will appear in a similar proportion in 2024.”

The proportion of this maths paper using the Calculations domain is the highest it’s ever been, with 41% of the questions requiring children to use addition, subtraction, multiplication and/or division. The representation of the other content domains in this paper have stayed in a similar proportion to recent years.

If the pass mark is below 67% (and I would be incredibly surprised if it were anywhere close to that!), the children still could, theoretically, be able to ‘pass’ the paper by exclusively knowing content from the calculations and FDP content domains.

Please note: not all content domains are created equal

It’s worth noting that although there are nine content domains in the KS2 mathematics test framework, they are not all weighted equally. Each content domain has a varying amount of substrands, ranging from 3 in both Geometry (position and direction) and Statistics, to 12 substrands in Fractions, decimals and percentages (where F1 is ‘recognise, find, write, name and count fractions’, up to F12 which is ‘solve problems with percentages’).

Each substrand is then split across various year groups to create different objectives; for example, the aforementioned substrand F1 is divided into four objectives: three in Year 3 (e.g. 3F1a: count up and down in tenths) and one in Year 4 (4F1: count up and down in hundredths). Calculations and FDP, the two weightiest content domains in terms of mark allocation, each have 39 objectives and 40 objectives respectively.

This goes some way to explaining the distribution of marks across the content domains (i.e. the most marks are allocated to these domains because they contain the most content. However, this doesn’t always follow as Measurement, a domain containing 44 objectives, has been allocated under 10% of the total marks for the last four years!).

I wanted to be really thorough when it comes to content domains this year. I took every single content domain into account. When a question had three content domains allocated, I allocated it ⅓ of a point. Where a question had 2 content domains, I allocated a half each.

What does this mean for next year?

Keep drilling the four operations, and fractions, decimals and percentages – you can pretty much guarantee that these topics will make up the majority of any future set of SATs papers. And, as always, nail all the maths curriculum content from Years 3-5 (this will, obviously, also give them a brilliant foundation to any new Year 6 content).

Arithmetic predictions vs reality

Our prediction: "It may also be worth noting that for the last two years, we haven’t seen a question on a mixed number multiplied by an integer – maybe this year?"

This year, indeed! A question like Q35 hadn’t appeared since 2019 - until now!

As always, we had two long multiplication questions worth 2 marks each, and two long division questions worth 2 marks each. There were no division questions with remainders this year, unlike last year’s (the first division with remainders we’ve ever had in an arithmetic paper). There were also no ‘fraction of an amount’ questions, whether presented with a multiplication symbol or as words (e.g. ⅓ x 15 vs ⅓ of 15).

What does this mean for next year?

Thinking about the 2025 SATs arithmetic paper - being able to use known facts to solve multiplication and division problems has always been quite popular in the arithmetic papers, so keep that up in your arithmetic sessions.

Reasoning paper analysis

Reasoning paper analysis I will openly say that I was very much not a fan of paper 2. There were 27 questions, which is the most we’ve ever seen in a reasoning paper (second to 2023’s paper 2, which had 26). Some people might argue that, overall, the allocated marks are the same, so whilst there may have been more questions, each would require less work (as there would have to be more 1-mark questions). To anyone who might argue this, I couldn’t disagree more.

Solving three 1-mark questions, especially questions such as those in paper 2, can be much more cognitively demanding than answering one 3-mark question, and therefore take more time. Whilst a 3-mark question would typically involve two or three steps to complete, the different calculations required are all within the same maths topic and part of the same mathematical process. In this paper, many of the 1-mark questions involved more than one step, meaning children could get part of the question correct and still not be rewarded for any of their effort. For example:

This question required 3 correct answers for 1 mark.

This question requires the children to complete 2 steps, or processes (firstly that 7 = ¼, then that 7 x 4 = 28), for 1 mark.

Historically, you might expect questions like these to be worth 2 marks, with 1 mark still available for a partially correct answer.

There were two particular questions in this paper that I felt were actually rather unkind.

The return of Chen: back with a vengeance in Q17

Chen may be back (as a girl this time), but I don’t welcome their reappearance for this question! For 2 marks, the children are expected to solve a multi-step addition and subtraction problem. However, there are seven numbers included in the question, four of which are completely irrelevant to the calculation. I had to re-read it a few times to double check I hadn’t misread it, mainly because I was convinced they wouldn’t include any unnecessary information. The only reason they could possibly have done this is to catch the children out.

Lookalike alert!

Using the inverse to solve addition and subtraction

A very similar LbQ question asks pupils to work out how much the grapes cost, when a receipt shows 3 items (tangerines 52p, apples 79p, grapes?) and a total cost of £2.20. According to LbQ’s SATs Springboard data, 51% of pupils can solve this problem correctly on the first attempt. The top misconception for this question was 131p, showing that the pupils had correctly completed the first step but then had gone no further.

I always turn to my other half when I come across maths-related issues (would you believe there has been more than one maths-related issue in my life?) - he is currently studying an MSc in Maths and said that he hadn’t, in any under- or postgraduate maths coursework or exams, seen any deliberate deception of confusion in a question. If this isn’t common in further education, why are we (they) doing this to children?

Mark McCourt (Complete Mathematics, Emaths) shared his thoughts on the paper during SATs week (it felt very validating knowing my opinions were shared by someone of his expertise!):

“Here are 3 translations on a coordinate grid.” WHERE? I flipped the pages myself a few times to check I wasn’t missing anything. Has anyone ever taught position and direction without the visual of a coordinate grid? We encourage children to use resources and aids to support them in maths, so why, when it comes to a maths topic that is supposed to be visual, is it made into such an abstract concept?

My (and Mark’s!) frustration was shared by @lifeatthenest and @MoreMorrow too.

Lookalike alert!

Translation using coordinates

In our Spring insights blog, translation using coordinates was highlighted as a weak area nationally (only 39% of all responses on Learning by Questions were correct the first time, and that was with the help of an image!).

I generally felt that paper 3 was a lot fairer, with fewer questions and nothing to seemingly catch the children out.

You may disagree with my thoughts on these papers, so let’s have a look at something nobody can argue - the data!

Content domain weighting data

Previously we said, "There have been some content domains that haven’t been represented for two years in a row – we could therefore assume that they may be a priority to appear this year. The content domains that haven’t appeared in a reasoning paper since before Covid are:

• Number and place value problems

• Roman numerals

• Scale factors

• Reading scales

• Money

• Volume

• Common shapes

• Patterns in position and direction problems

Worth noting that there are three content domains above that have never appeared in a SATs paper, as they are part of the KS1 content domains:

• Money (M3)

• Recognise and name common shapes (G1)

• Position and direction patterns (P1)

Number and place value problems (N6) has made its first appearance since 2018.

And this year, we had a question about Roman numerals, which we haven’t had since 2017!

Lookalike alert!

Roman numerals

Roman Numerals appear in the SATs Springboard diagnostic assessments. During Autumn 23 and Spring 24, only 39% of Y6 pupils were able to correctly answer this question: CLXI = ? The top wrong answer was 1511, where the pupils misinterpreted the C as representing 1,000 and L as representing 500 (likely due to the place value of familiar 4-digit numbers). Later in the year, there was a good improvement to 56% of Y6 pupils getting a similar question correct at the first attempt.

What does this mean for next year?

These content domains haven’t been represented for three years in a row now – we could therefore assume that they may be a priority to appear next year. The content domains that haven’t appeared in a reasoning paper since before Covid are:

• Scale factors (R3)

• Estimate, measure and read scales (M2)

• Volume (M8)

M1 (compare, describe and order measures) hasn’t been represented for two years in a row – this may also be in 2025’s paper.

Maths pass score

The “pass mark” (a scaled score of 100) for maths has always been between 51-55%, and the “greater depth” boundary (a scaled score of 110 is commonly accepted to be the standard required for this) between 85-87% - I’d expect both grade boundaries to be something similar this year.

We’re expecting to see something in the 70s for the national percentage of pupils meeting EXS.

Join us again on Tuesday 9th July, when this year’s results and raw score to scaled score conversion tables will be released. I’ll be holding my (bated) breath for the next six weeks!

You can check out my analysis of the 2024 English papers here.

After making real waves in Year 6 classrooms this year, the SATs Springboard is back, bigger and better for the 2024 - 25 academic year!

Grab yourself a FREE, no-obligation SATs Springboard trial and propel your Year 6 class from where it is to where it needs to be.