KS2 SATs 2025: GPS papers analysis

SATs expert Sophie Bartlett is back with her analysis of the GPS papers for 2025. We’ve got grammar ‘disguises’ and potential accent trip ups…

Sophie Bartlett is back with her analysis of the GPS SATs papers for 2025. We’ve got grammatical ‘disguises’ and potential accent slip ups as well as the usual data, data, data that we know and love.

If you would like a slides version of this blog as well as reading and maths papers analysis to show to colleagues and friends, sign up to our mailing list and get them sent to your inbox.

Well, at risk of feeling like that meme of Homer Simpson retreating into the bushes to avoid any sort of backlash… those GPS papers seemed alright? No? Although to be honest, I never have any sort of hugely visceral response to the GPS papers. They tend to be a bit bog-standard and - well - boring. I personally thought this year’s were fair, and Paper 1 had some great juicy questions in it towards the end! Let’s dig our teeth in, shall we?

Grammar, punctuation and spelling Paper 1 analysis

Let’s start by breaking down the content domains, and then look at some of the questions that caught my attention.

The facts and figures

In Paper 1, children are tested on seven content domains.

The English grammar, punctuation and spelling test framework details the profile of marks by content area.

  • Punctuation (G5) is assigned 10-20 marks (which makes up 20-40% of Paper 1)

  • Vocabulary (G6) is assigned 3-7 marks (6-14% of Paper 1)

  • Grammar (the other five content domains) is assigned 25-35 marks (50-70% of Paper 1)

This can be seen in the analysis of each SATs paper below.

Within these seven content domains, there are 51 strands, ranging from G1.1 (nouns), to G7.4 (subjunctive). The proportion of marks assigned to each of these individual content domains makes sense as the most represented (G1 and G5) are also the largest in the curriculum (G1 and G5 have 12 and 15 strands within them, respectively).

G1 and G5 collectively make up over half of the SPAG paper. This proportion has also consistently been lower than the amount of marks needed to pass the paper meaning that, theoretically, you could achieve EXS by correctly answering questions from only these two content domains. As long as this year’s pass mark isn’t above 56% (and it’s never been higher than 54%), this remains true.

These are the individual substrands that are most highly represented:

  • Punctuation for parenthesis (G5.9): 17 questions since 2017

  • Synonyms and antonyms (G6.1): 15 questions since 2017

  • Apostrophes (G5.8): 14 questions since 2017

  • Verbs (G1.2): 14 questions since 2017

In fact, verbs appeared the most in 2025’s paper - more than any other substrand.

The GPS Paper 1 questions

There were a few examples of testing children’s true understanding of the concept at hand, rather than just using ‘tricks’ - for example, questions 12 and 34. To spot fronted adverbials, children often just look for the first part of the sentence that’s separated by a comma; however, they could have been easily misled by the use of the comma in the compound sentence here. When identifying relative clauses, a common mistake is to choose whatever is in the middle of a sentence - often between parentheses. However, this relative clause has been ‘disguised’ (although not really, if you have a genuine understanding of a relative clause!) at the end of the sentence, with the main clause being in the middle.

This leads us nicely onto question 22. No more relying on the comma to help spot the fronted adverbial - they all have a comma! There are two used to list items, and two used to separate clauses. This is where even I had to use a process of elimination - I knew one was clearly a main clause, and the other a subordinate clause, so it had to be the subordinate clause… but I have to admit, teaching a subordinate clause as a type of fronted adverbial is something I’ve rarely done. I tend to lean towards using phrases as fronted adverbials, so this is definitely a learning curve for me! Anyone else?! Someone come and join me on this hill of grammatical incompetency!


Related content:

Get this blog and the GPS & maths analysis in slide format - sign up to our mailing list, today

KS2 SATs 2025: reading paper analysis

KS2 SATs 2025: maths papers analysis


Question 20 perhaps has me a bit biased as a Southerner. In my (and many other) accent/s, the word ‘class’ read as a root word (‘cl-ah-ss’) changes pronunciation when the suffix ‘ify’ is added (‘cl-a-ssify’). I just wonder whether some other children may have been thrown by this as none of them would have sounded ‘right’ - but then again, if they knew the word, they’d know the answer!

Question 21 was your classic “the first answer sounds right/is nearly right, so I’ll tick it and move on”... I reckon that will have tripped quite a few children up!

Don’t get me wrong, I love a good question on sentence form - however, I did feel question 26 was a little mean. Putting the word ‘Don’t’ in front of what would have been an obvious command (starting immediately with an imperative verb) may have caught a few children out. However, I can hear you all saying “but ‘Do’ is an imperative verb Sophie, aren’t you supposed to be good at grammar?” And yes, whilst I do agree, it’s hidden in a contraction, so I’m just sympathising with those 11-year-olds who aren’t as grammar-savvy as the rest of us!

The last few difficult questions really treated us. I have to say, I thought they were great questions - question 44 in particular (did all the possessive nouns trip your children up?!) I love a question where even I have to spend more than a couple of seconds searching for the answer - and it’s not a deliberately mean one!

Almost all of the 51 substrands have been represented in every Paper 1 so far. Verbs in the perfect form (G4.1b) hadn’t appeared in the previous two papers, and question 46 was the first time we’ve seen subjunctive verb forms (G4.3) since 2018!

Grammar, punctuation and spelling Paper 2: spelling analysis

Do you think if the children sat this test in December, they’d all be more likely to get one of the spellings correct? ‘Reindeer’ seemed a bit out of left field, although it does meet the rules of the spelling strand S52 (words with the /eɪ/ sound spelt ei, eigh, or ey).

The spelling tests contain 20 words, testing the children on 24 KS2 objectives – these are split between the Year 3/4 spelling rules and the Year 5/6 spelling rules, although there has always been a few more from Lower Key Stage 2.

See the analysis of all the SATs spelling tests below.

  • An italic word means it appears under more than one content domain

  • A blue word is one of the first three of the test – considered easier to spell

  • A red word is one of the final three of the test – considered harder to spell

A few weeks ago, we predicted what might appear:

“The most common spelling domains (those that have appeared the most since 2017) are as follows: S38 (adding suffixes beginning with vowel letters to words of more than one syllable); S60 (words with ‘silent’ letters); S43 (the suffix –ly); S47 (endings that sound like /ʃən/, spelt –tion, –sion, –ssion, –cian).”

Again this year, we had at least one spelling from each of these domains: rewarded, developed, affordable, knight, angrily and suggestion.

“There are three content domains that didn’t appear in 2024, but were consistently represented from 2018 – 2023, so we could assume that they would be more likely to appear this year: S51 (words with the /s/ sound spelt sc); S58 (words with the /i:/ sound spelt ei after c); S59 (words containing the letter string ough).”

And so they did! With scene, receiving and through. This is the third time that a word derived from ‘scene’ has appeared in the SATs - we had ‘scenic’ in 2022 and ‘scenario’ in 2023.

“And lastly, just for fun, here are some specific word predictions for this year (seeing as last year I managed to predict one correctly!). Here are some content domains where there aren’t many words to fit each rule… S40: double or country; S48: scheme, chemist or echo; S50: antique.”

No such luck for me this year! My word guesses were trumped by flourish, anchor and plaque (the first and last of which I think are particularly tricky for an 11-year-old!)

There were three homophones on this year’s paper (through, passed and scene) - the most we’ve ever ‘scene’ compared to ‘passed’ papers! (Ha - did she really just ‘through’ those in there?!)

The GPS papers results

To achieve the expected standard in GPS, the pass mark has previously been just over 50% (between 50-54%); this is roughly equivalent to getting half of each paper correct (25-28/50 in Paper 1 and 10/20 in Paper 2).

To achieve what is generally accepted as GDS, children have had to score between 76-80%; this is equivalent to only losing 14-17 marks overall (out of 70).

On Tuesday 8th July, we shall find out what this year’s mark boundaries are!

While the GPS papers may not have delivered any major curveballs, there were certainly a few sneaky moments that tested children’s genuine understanding rather than just their ability to memorise grammatical 'tricks'. There’s never much drama with GPS, but the detail in the substrands and some cleverly crafted questions gave us plenty to sink our teeth into. Only a little over six weeks to go until we see what the pass marks have in store for us!

Looking for a helping hand for your year 6s? You can trial LbQ’s SATs Springboard for free in your school and get access to GPS (and maths!) diagnostics and resources designed to plug gaps for the individual.