This blog continues our series exploring how to enable children to showcase their mathematical understanding, reasoning and problem solving in the KS2 maths SATs papers. We know that for the reasoning papers, pupils need to be able to process and manage quite a range of information for some questions, and there needs to be specific teaching of strategies to enable them to do this.

In part 1 (Measures made visual: using bar modelling to support reasoning in KS2 SATs), the bar model was used to exemplify how to solve questions from reasoning paper 2 from the KS2 SATs 2025 with a particular focus on problems involving measures. And like our blog series from 2024, the focus of part and whole remains at the forefront.

In this blog, I will explore questions from the KS2 SATs Reasoning Paper 3. Like part 1, the focus will be on questions involving measures, as there were a number of these, in different styles and formats. Like part 1, there will also be a focus on part and whole. But in this blog, I will exemplify questions where bar modelling might not be an obvious strategy choice.

When you read the examples in this blog, you may find yourself thinking that bar modelling is not a particularly efficient strategy for the questions that I have chosen. I challenge you though to disregard efficient strategies for the time being and instead consider how exploring these questions in a visual way may aid the conceptual understanding which underpins the more efficient strategies your children will be aiming for. 

Example 1:

2025 KS2 SATs – Mathematics Paper 2, question 10

Bar modelling focussing on proportion and comparison

In this question, information is shared in a data table format. The table is followed by four statements, and we need to ascertain which are true and which are false.

For the first 3 statements, I have exemplified below how different multiplicative models can be useful to support children to uncover the needed calculations.

For the first statement,  a comparison, or ratio model (as all parts are equal) is a useful visual aid.

Solution (to this statement)

The first statement is false as the elephant would need to have a mass of 7,200kg.

The comparison, or ratio model, is useful here to illustrate the relationships between numbers and equal parts where multiplicative language such as ‘three times’ is used. In this case, we could either divide the greater mass by 3 or multiply the smaller mass by 3.

Now for statement 2. 

Solution (to this statement)

The second statement is true as the hippo’s height is 1.5m which is one quarter of 6m.

Exploration of this statement using the bar model may be a useful reminder of models used to find fractions of amounts where all of the parts are equal and the whole is the total of the equal parts. In this example,  I have used division to find the answer, but I could also have used multiplication and multiplied the hippo’s height by 4 to see if this matched the given height of the giraffe.

For statement 3, a multiplicative model (involving equal parts) would not be helpful as this statement is asking us to consider the difference. 

Solution (to this statement)

The third statement is true as the rhino’s height (1.7m) is 20cm greater than the hippo’s (1.5m).

I know that models involving the difference, and/or language such as ‘taller than’ can be tricky to visualise and so the bar model, where the whole can be compared with the parts, is a useful way of reinforcing this.

Although not exemplified here, the final statement could be used to consider proportional relationships in both the heights and the masses of the animals when constructing bar models and considering how accurately we can draw them. 

Solution to the question

Statements two and three are true.

Example 2:

2025 KS2 SATs – Mathematics Paper 3, question 19

Bar modelling focussing on equal parts

The information in this question is almost already presented to us as a bar model but perhaps in a less familiar, vertical representation. By putting the information into horizontal bars, and once again using a comparison, or ratio model (as all the parts, or blocks in this case, are equal), the operations needed to solve the problem become much clearer.

Solution

The height of the smaller tower is 13.2cm.

As well as the bar model being a useful tool for this question, this question is also helpful in highlighting the importance of base facts. Here, as long as we know that 20 divided by 5 equals 4, then we can use this fact to help us to calculate that 2 divided by 5 will be 0.4. We can draw on the relationship that 2 is ten times smaller than 20 so the quotient of 2 divide by 5 will be ten times smaller than the quotient of 20 divided by 5. 

Example 2:

2025 KS2 SATs – Mathematics Paper 3, question 22

Bar modelling focussing on part and whole in the contexts of missing angles

Bar modelling might not be an obvious tool for this question but once we pull on our knowledge of interior angles in shapes and on a straight line, the usefulness of considering parts and wholes becomes clearer and helps to make a complex question much simpler.  

Angle a equals 50°.

By using different bar models to present what is known about each of the shapes including angle a and then angle b, this helps us to consider each whole in turn and gives us a good starting point in solving the problem.

As Charley states at the end of blog 1 in this year’s SATs bar modelling series, bar modelling isn’t just a useful strategy for solving these questions at the end of KS2 – Year 6 – but a strategy that could be taught, rehearsed and honed as children move through the primary phase. In addition, bar modelling is a thoroughly useful tool for assessment for learning as through building and exploring models, possible misconceptions are often uncovered. Hopefully, no matter what year group you teach in, by exploring these KS2 assessment examples, you can link these to your year group curriculum whether that be exploring part and whole in addition and subtraction in KS1 to moving into more complex multiplicative models in KS2

Further professional development

Our popular training is back this year in a new format.

Across four half-day sessions, we unpick tools for supporting children in meeting age-related expectations in maths using a range of different strategies, including the bar model!

Resources available on PA Plus, all included within the subscription and some also available as a separate purchase if you are not subscribed, which support with Bar Modelling and Year 6:

Bar modelling in maths progression

• Download sample pages
• A combination of photos of concrete resources, pictorial bar models and abstract calculations brought together to exemplify how to solve problems from across the maths curriculum.

Year 6 maths gap finder: SATS preparation toolkit

• Diagnostic summative assessment papers for early identification of gaps
• Resources to support teaching and rehearsal of arithmetic and reasoning

Year 6 SATS analysis toolkit

• a suite of resources and analysis tools to identify specific areas of learning strength and development for pupils and classes when using any past SATs papers as practice (2016-2024).

AI presents many challenges and opportunities for schools, governors & trustees who should ensure and monitor that it’s being used in ways that enhance school life and opportunities for pupils, as well as helping reduce the workload of school leaders and staff. Rebecca Jones, from our school’s HR team discusses just such an opportunity and considers the associated risks.

AI is transforming recruitment across sectors, and education is no exception. From automated screening to predictive analytics, and sample questions, AI promises faster, smarter hiring decisions. But in schools - where recruitment isn’t just about qualifications, but also safeguarding and cultural fit, its use raises serious questions. When you strip away the buzzwords and automation, one question remains: Can a synthetic decision maker keep children safe?

As someone working in education recruitment, I’ve found myself increasingly curious and concerned about the growing role of AI in hiring. The promises are bold: faster shortlisting, reduced bias, streamlined processes. But beneath the surface, I’ve started digging into the pitfalls, especially around safer recruitment, which is far more than a compliance exercise - it’s a safeguarding imperative. The more I explore, the clearer it becomes: While AI might offer efficiency, it struggles with the very things that matter most in schools - human nuance, ethical judgment, and the instinct to protect.

Let’s start with the facts:

• AI use in recruitment has tripled in the UK in the past year.
• 30% of UK employers now use AI to recruit; 43% of large companies use AI to interview candidates.
• 70% of recruiters say AI improves hiring decisions, and it cuts hiring costs by up to 71%.

There’s no denying that recruitment can be time-consuming and expensive. Some roles attract large volumes of applications, and shortlisting can be a logistical challenge. AI offers solutions:

• Speed and efficiency: Algorithms can scan hundreds of applications in minutes, flagging those that meet key criteria.
• Consistency: AI applies the same rules to every candidate, reducing variability and helping ensure fairness.
• Data-driven decisions: Predictive analytics can assess which candidates are most likely to succeed based on historical data.
• Bias reduction: AI can be programmed to ignore demographic data, potentially reducing unconscious bias.

For governors, trustees and school leaders under pressure to make good hiring decisions quickly, these benefits are appealing. AI is alluring, but here’s the problem: schools aren’t hiring for productivity - they’re hiring for trust. And trust isn’t something you can measure with an algorithm.

AI is no silver bullet, and in education, its limitations aren’t just inconvenient, they’re dangerous. Schools aren’t hiring for output; they’re hiring for trust, integrity, and safety. Recruitment in this sector is a safeguarding process, not a productivity exercise. Every appointment carries risk, and safer recruitment isn’t a checklist - it’s a mindset rooted in human judgment.

Yes, AI can flag gaps in employment and track DBS checks. But it cannot detect charm masking manipulation, interpret vague or evasive references, or ask the uncomfortable follow-up question. It can’t sense when something feels off - and in safeguarding, that instinct can be the difference between protection and catastrophe.

Worse still, AI is only as unbiased as the data it’s trained on. If past hiring decisions reflect systemic inequalities, AI doesn’t correct them, it amplifies them. And when we start relying on algorithms to make decisions that demand human scrutiny, we risk sidelining professional judgment and missing the very red flags we’re supposed to catch.

In a sector where the stakes are children’s safety, outsourcing vigilance to an algorithm isn’t just flawed, it’s reckless.

In serious case reviews where children have come to harm in school settings, time and time again it is the same failures that come to light: Incomplete checks and blind trust in processes. Between 2019 and 2021, 59 schools were judged by Ofsted to have “not effective” safeguarding. The most common failures? Poor record keeping, weak leadership, and failure to follow up concerns. In one study of 41 professionals who sexually offended against children, 92.5% were aware of their interest before age 21, and 15% chose their career specifically to abuse.

The sad truth is that some people choose education roles specifically to access children. And we think AI can screen them out?

AI has a place in education recruitment, but only if we’re brutally honest about its limits. 

It can help with admin. Elements of AI have been incorporated into the Teach in Herts website, for example, to flag gaps in employment. But it cannot replace the human responsibility to protect children. It’s crucial that school leaders retain full control over their recruitment process.

Let’s not sleepwalk into risk. AI might be the future of recruitment, but in education, we cannot afford to follow blindly. Safeguarding isn’t optional. It’s not a feature. It’s the foundation.

If you're considering AI in your recruitment process, ask yourself: Is it helping you make safer decisions, or just faster ones? Are you confident your recruitment process puts safeguarding first? Or are you trusting a system that doesn’t know what danger looks like?

Let’s keep the conversation going. Let’s challenge the hype. And most importantly, let’s keep children safe.

Further reading 

Recruitment service | HFL Education
Safer Recruitment Training 
AI for School Leaders and Business Managers
AI in Recruitment statistics UK 2025 | Latest reports & data
Safeguarding: What Can we Learn from Schools where it was Judged 'Not Effective' | Judicium Education
Safer recruitment and the importance of getting it right

A powerful resource for primary English and maths leads, available for free to all schools, is the Question Level Analysis (QLA) element of the DfE’s Analyse School Performance (ASP) site. In this blog I will highlight some of the lines of enquiry subject leaders might follow.

Of course, an important principle of all assessment activity (indeed, of any activity) is that we should only spend time doing things that are worthwhile – i.e. things that tell us something useful (that we didn’t already know) and that could inform our future decision-making or lead to some new actions.  When we look at QLA data, we must bear in mind that we are looking at information that relates to last year’s Year 6 cohort, who will in most cases have now left the school. Finding out where that cohort performed strongly and where they ‘went wrong’ is not necessarily going to be completely applicable to your current cohorts, but it could be extremely powerful if it reveals something systemic about your school – maybe about your curriculum, your pedagogy or your preparation for that peculiar week in a child’s life that we call SATs.

Accessing the QLA

Before I dive into what me might discover from QLA, a quick word about how to access it.

Every headteacher should be be able to access the ASP site via DfE Sign-in. Every school should also have a user-approver for DfE Sign-in, who is able to create logins for other members of staff and allocate them the services that they need. This user-approver might be the headteacher or might be a member of the office team. Good practice would be for this person to create individual logins for all senior leaders in the school and, I would suggest, the curriculum leads for English and maths. Provided a user has been granted access to the ASP (Named) service, they should be able to access QLA. If you log into ASP and can’t see the link to QLA, you have not been given the right level of permissions, so speak to your user-approver. You might have been given access to ASP (Anon) – which does not allow access to pupil-level data.

Once you are logged in and looking at the QLA element of ASP, you will find the data for your school, compared with national, showing performance in last summer’s tests at overall level, broken down by domain (sub-topic within the subject area) or by question.

Exploring maths QLA

The first thing we see is the bar chart, showing the percentage of correct responses across the school, with a national comparison, for each domain (pictured below). When hovering over each bar, the number of marks available for each domain along with the school and national correct response rates are revealed.

For convenience, here is a summary of the data for the 9 domains:

It should be noted that the largest number of marks available is for the Calculations domain (hardly surprising really). Nationally, 73% of the possible marks for this domain were awarded. It is also worth noting, by the way, that a very significant proportion of those marks can be gained from questions that test elements taught in Years 3, 4 or 5 of the National Curriculum. The published mark scheme indicates, for each question, which element of the curriculum, and the year group in which it should be taught, is being tested.

The second largest number of marks available is for Fractions/Decimals/Percentages, although nationally only 63% of marks were awarded in this domain.  The domain with the highest proportion of correctly answered question is Number/place value at 83%.

The biggest ‘wins’, therefore (from the point of view of attainment in the SATs which, I accept, is not the be-all and end-all of primary school life) are to be made by focusing a large part of your teaching on these areas – including building children’s knowledge, fluency and strategies for reasoning.

If the QLA figures for your school differ significantly to the national figures for these domains, ask yourself why that might be. Could it be to do with how much curriculum time was devoted to these areas? Is it to do with pedagogical approaches? Are strong foundations being laid in lower Key Stage 2  (and earlier) that are being successfully built upon in upper KS2? Are your children stronger at the non-contextualised questions (arithmetic paper) than the reasoning questions and, if so, why might that be?

It is then worth drilling down to look at the question-level data. Are there particular questions where your children’s correct response rate differs significantly from national? If there is a question where your children did not on the whole perform well, what is it about that question? Were they just not familiar enough with the mathematical content of that question, or is it content that you know they understand, but that has been presented in a way that is unfamiliar to your children? Maybe the wording of the question caught out some of your learners. Perhaps it is an area that has not had so great a share of curriculum time in your school. Perhaps it relates to knowledge taught in earlier year groups, which pupils have not had opportunities to retrieve and practise for some time.

When looking at the data for individual questions, the proportion of children who attempted the question can also be illuminating.  Looking at the figures for question attempt rate, across the two reasoning papers, the question that fewest children attempted is the final question on Paper 2 – only 87% attempted to answer. (Of course, this may be in part due to the very fact that it is the final question, hence some children might not have reached it if they ran out of time.) Let’s look at the question.

The question is classified as a ratio/proportion question - although neither of those words are used in the question. Whilst it may appear daunting to some, it’s the kind of question that lends itself very well to a bar modelling approach for solving it – and is actually quite straightforward.  (To see how a bar modelling method could have been used to approach this particular question, along with a few other examples, see this recent blog from the HFL primary maths team).

It may surprise you to learn that only 36% of pupils nationally answered this question correctly.  It could be interesting for schools to explore how their own children’s proportion of correct responses compares to that figure. If the school figure is significantly higher, that could be excellent evidence that the school’s current approach to teaching children how to solve numerical reasoning problems is effective. You can also compare your pupils’ attempt rate with the national. If there has been a strong teaching focus on solving reasoning problems, it would certainly be heartening to see a greater proportion than the national figure of 87% at least having a go at answering this particular problem, even if some don’t ultimately arrive at the correct answer.

Exploring spelling

The QLA data for spelling might also be an interesting area to explore. The table below shows the 20 words from the 2025 spelling test and their national correct response rate. 

There are a few interesting things to note about the spelling test in general. It can be seen from the above list (and this has been true for every single previous KS2 spelling test – see this blog from last year) that:

1. Quite a few words on this list involve the adding of a suffix or a prefix (or sometimes both) to a root word, e.g. rewarded, explosion, affordable, angrily, disappointed etc
2. There are a few homophones in the list (knight, passed, scene)

We can confidently predict that these sorts of words will come up each year in the spelling test, based on previous tests and on the fact that both these elements are made explicitly clear in the national curriculum for spelling. It may seem surprising, therefore, that only 24% of children nationally correctly spelled ‘disappointed’, a spelling that can easily be derived using a good knowledge of prefixes and suffixes. What the data doesn’t tell us is the nature of the incorrect answers. Maybe they are caught out by the double ‘p’, or perhaps they have incorrectly written a double ‘s’. We can’t be sure. What we can say is that if the proportion of your children correctly spelling ‘disappointed’ is significantly higher than 24%, this would indicate a strong performance in the teaching and application of morphology in spelling.

Again, it may be of limited value to over-analyse the reasons why particular children didn’t score too well in the spelling test, given that those children have now all moved on from primary settings.  However, if this QLA data reveals any interesting trends or patterns, such as consistently below average success in spelling words comprised of a root plus a suffix, this could indicate something systemic (such as the amount of curriculum time devoted to teaching this) that would be worthy of further investigation and action.

Similarly, the data might illuminate for a school where they are being particularly successful – and this is also worth unpicking, to understand the underlying causes of this success and potentially apply these elsewhere.

Exploring reading

There is of course also QLA data for the reading test paper. Opinions may vary here, but I find this element slightly less useful than those explored above. The reason for this is that the vast majority of marks on the paper are awarded for these 2 domains:

• Retrieve and record information or identify key details from fiction and non-fiction (typically about 30% of the marks)
• Make inferences from the text or explain and justify inferences with evidence from the text (typically about 50% of the marks)

The chances are, the QLA data will reveal a need to improve children’s performance on inference. However this is not entirely helpful, because the skill of ‘inference’ actually depends upon lots of inter-related skills, including a strong vocabulary knowledge and the stamina and fluency to be able to access the reading texts in the first place. More on this in this blog from our HFL primary English team. 

To conclude, as I always say, data raises questions – it doesn’t necessarily give us the answers. But this QLA is powerful data and the questions it raises could lead to some pertinent lines of enquiry which could ultimately lead to curriculum development, more focused teaching and improved outcomes for children.

If your school curriculum for maths or spelling is in need of an overhaul, take a look at our Essential Maths and Essential Spelling resources. 

It’s autumn, and while the Year 6 Reading SATs paper may feel like a distant memory, many of us already have our eyes on 2026. Looking back at the paper gives us valuable insight into how children are being asked to demonstrate their knowledge, stamina and comprehension. So, let’s take a closer look at the 2025 paper and reflect on what this means for classroom practice. 

The extracts increased in complexity as the paper progressed; while extract 1 offered a relatively gentle start, extract 3 required stamina, inference, and close attention to language.

Content domain analysis

Unsurprisingly, inference once again dominated the paper.  Almost half of the marks tested children’s inference this year, with a significant proportion of the rest of the paper focused on retrieval and vocabulary. Together, these three domains accounted for around 90% of the potential marks. This follows the usual trends. The table below provides you with a comparison to previous years:

These three domains are hugely significant for becoming a skilled reader, and much of our reading teaching should focus on developing children’s retrieval, vocabulary and inference skills. The other reading domains of course support the goal of comprehension and inference, and pupils should learn how to use and apply these comprehension strategies to build a mental model of the text. However, the 2023 Reading Framework is clear that these content domains should not be treated as a checklist for teaching reading. Instead, the framework reminds us of the importance of drawing on and weaving together a variety of strategies. When we read, we might be visualising, making predictions, and questioning – all to support inference - and often all at once. These processes are interconnected, not isolated skills to be ticked off a list.

Building on my previous blog (Reading SATs knowledge: Reading SATs power), a helpful next step is to give more attention to comprehension strategies that support children to become aware of their own thinking as they read. This means encouraging them to:

• monitor their understanding,
• notice when things don’t make sense, and
• connect new ideas with what they already know.

Encourage children to briefly summarise a paragraph after they have read it, or to note any questions in the margin of the text. Christopher Such explains this clearly in The Art & Science of Teaching Primary Reading (p. 60): ‘Research suggests that encouraging children to ask themselves questions about the text’s meaning and to summarise parts of the text can lead to improvements in comprehension. In addition, children can be taught to recognise when greater attention is required.’ This will aid children to grow into confident, thoughtful readers who can flexibly use their reading for different purposes.

A gentle start to a tough challenge

The first extract, A Life-Changing Game, gave children a fair start: a clear structure and a familiar non-fiction genre. However, cultural references to Uganda and the art of chess may have been more demanding for some children without the relevant background knowledge. We can’t prepare children for every possible eventuality, however, exposure to a rich, broad reading diet will help to power up children’s background knowledge and expose them to previously unfamiliar worlds and vocabulary.

The second extract, In the Cave, was more challenging. Children were faced with potentially unfamiliar vocabulary such as ‘beckoning’ and ‘inaudible’. The way that dialogue was used, while seeming simple at first glance, required children to keep up with the conversation as it revealed discoveries, built tension and moved the plot forward. The text was also littered with many long, multi-clause sentences. Children require plenty of practice in grappling with these; strategies such as re-reading, breaking them down into meaningful chunks, and visualising, can all help with making meaning from long sentences.

The final extract, Longbow Girl, was the most demanding. A historical fiction extract blending technical archery vocabulary along with figurative language and more of those long, multi-clause sentences - a real challenge for less confident readers. Children had to track the physical events, get to grips with the atmosphere and infer the characters’ emotions, which were sprinkled throughout. That’s a lot to balance! Inference requires the use of multiple reading processes at once. Rich dialogue about texts - in which teachers model making meaning by making connections, empathising, predicting, summarising, questioning, and children are invited to do the same – will support children to feel prepared when approaching texts of this pitch.

Reading fluency comes into play here, too. Without accurate, automatic reading and expressive phrasing, children can struggle to access and make meaning from lengthy sentences or layered passages. Fluent reading reduces the cognitive burden of decoding, allowing the reader to focus on understanding. To develop fluency, pupils require regular opportunities to read widely (a range of different texts) as well as deeply (reading the same text more than once). By doing so, they build their automaticity. They also benefit from hearing an expert model of reading and reading alongside that model (assisted reading); this supports application of prosody and phrasing which can help to make the meaning clear, paving the way for understanding. This is why reading fluency is often referred to as the bridge to comprehension. We have a number of blogs on the topic. Have a read of this one, which gives more insight into the research behind fluency and this one which explores how to teach fluency.

Precision over plausibility

This year’s mark scheme highlighted the need for accuracy, not just general understanding. Let’s dive into Question 3, which is based on A Life-Changing Game:

‘How can you tell that people in Uganda were not familiar with chess?’

At first glance, some answers may have seemed plausible but only the very specific ones were accepted:

‘There was no word for it in Phiona’s language.’

‘Chess was unusual in Uganda.’

‘She had to walk six kilometres every day to learn how to play.’

Responses such as ‘they didn’t know what it was’ or ‘they hadn’t heard of it’ were rejected. Why? Because they went beyond the text and were a little vague. The extract does not say nobody knew chess, only that it was ‘unusual’ and ‘there was no word for it.’  Children needed to tie their answers to the evidence given in the text.

Encouraging children to ‘prove it’ - ‘put your finger on it’, or ‘highlight the part of the text that tells you’ - is a simple but powerful way to build this habit. Train children to go back to the text.

Inference

As usual, inference sat firmly at the heart of the paper, with 24 marks testing this domain – even higher than in recent years. Children were asked to interpret motives, explain actions and describe character traits. Take question 37, based on Longbow Girl. Children had to explain why Merry smiled at the end of the extract. The reasons were scattered across the extract, and the mark scheme allowed several valid answers. This included her accuracy, her victory, the crowd’s respect, or proving others wrong. But to gain full marks, children had to give two distinct reasons with precision. This tempted inaccurate answers, such as the ‘she had just won ten gold coins’, which was explicitly rejected. This was a real test for some children, who may have struggled to navigate the text, comprehend it, and ignore what wasn’t relevant to answer the question. When planning reading lessons, create authentic, open-ended questions which elicit many different responses so that children get used to exploring texts in this way, drawing together ideas from across paragraphs to answer bigger questions. Ensure that follow-up questions dig a little deeper and probe children’s understanding, to help them identify nuances in what does or does not answer the question at hand.

We know that skilled readers monitor their own understanding. In a dense passage like this, that might mean noticing shifts between Merry’s actions, her emotions and the crowd’s reactions. The reader will likely need to ask themselves questions: ‘What’s happening here? What does this tell me? How do things look for Merry at this point?’ They’ll need to be identifying and summarising the key events within the extract - Merry’s fear of the bow breaking, her relief as the arrows found their mark, and her victory. When summarising, you might support children to summarise the main ideas in a paragraph to a single sentence, or the plot in just a few lines. You might use metacognitive talk to model how you ask yourself questions about the text’s meaning, e.g. ‘I’m wondering why…’ or ‘Is the author trying to tell me that…’ These different comprehension strategies can help children connect ideas and track meaning across the text. In reading lessons, incorporate teaching these strategies within rich text discussion. Demonstrate how these can be combined to make a mental model of the text and get to that ultimate goal of reading: comprehension and inference.

Words in context

Vocabulary is of course integral to understanding what we read, and there were certainly words which posed a level of challenge across the paper.

• A Life-Changing Game: children had to interpret words such as ‘anticipation’ and ‘intriguing’. This may have been a stumbling block for some.
• In the Cave: technical words such as ‘horizontal’ and ‘panel’ drew on subject and world knowledge.
• Longbow Girl: specialist archery terms sat alongside figurative expressions such as the ‘first flush of euphoria.’

Without wide reading experience and vocabulary instruction, this can quickly become a barrier. That’s why it’s so important to introduce children to a range of high-quality texts and explicitly teach new vocabulary. Pre-teaching vocabulary can be powerful, but children should also practise reading and unpicking the meaning of words in context. We can model strategies such as drawing on the clues around the unfamiliar word as, often, it’s the surrounding words that provide the extra hints needed to unlock meaning. Morphology is also key to unlocking word meanings, so model identifying root words and applying meaning of affixes to help with understanding and engage pupils in doing the same. If vocabulary teaching is on your radar, you may find this blog useful: Improving vocabulary for primary pupils.

Final thoughts

Overall, the 2025 Reading SATs paper was well-structured and fair. It rewarded reflective readers who could combine retrieval, inference, vocabulary knowledge, and give precise answers. To prepare children for reading with this pitch of challenge, both in SATs and – more importantly - wider reading experiences, we need to embed a broad, rich reading curriculum. Children should experience a wide range of genres, have the opportunity to read these with fluency and have meaningful discussions about them where they encouraged to dig a little deeper. Also, we must equip children with strategies to monitor their understanding, select evidence carefully, and tackle new vocabulary.

Come along to our Growing greater depth reading in year 5 and 6 or  HFL Education Reading Fluency Showcase 2025 to further explore effective teaching of reading at UKS2.