Ask a child to work out 23 × 4 and they reel it off. Wrap the same calculation inside a story about cupcakes at a party and they stall, sigh, and announce they "can't do it." The maths didn't change — the reading did. Word problems are where a quiet, second skill decides everything, and the research on what that skill is can transform how we help.
A ten-year-old can tell you instantly that six eights are forty-eight. Then she reads: "A baker puts cupcakes into boxes of 8. She fills 6 boxes and has 5 cupcakes left over. How many cupcakes did she bake?" — and she's gone. She stares. She guesses "8 minus 6." She writes "5." She insists she's "rubbish at word problems," and in a sense she's right — but not for the reason she thinks. The arithmetic she needs is arithmetic she owns cold. What defeated her happened before any calculation: she couldn't turn the sentence into a picture.
If you have a child who can calculate but crumbles the moment a question has words in it, you are not looking at a maths problem in the usual sense. You are looking at the meeting point of two skills — reading and arithmetic — and at the quiet, under-taught skill that joins them. This article unpacks what researchers have learned about why word problems are so hard, and how to teach a child to read one the way a mathematician does. It builds on our piece about why bright children "blank", because much of what happens here is a story about a small mental workspace getting overloaded.
The pattern is unmistakable once you know it. A child does a page of bare calculations cleanly, then meets a worded version of the very same maths and seizes up. They often do one of three revealing things. They grab the numbers and combine them in some plausible way ("there's an 8 and a 6, so… 8 minus 6?"). They hunt for a keyword they've been taught ("'left over' — does that mean take away?"). Or they simply freeze, because the sentence hasn't become anything they can hold in their mind.
What none of these children are doing is the thing successful solvers do automatically: building a little mental model of the scene — boxes filling with cupcakes, a small pile left on the side — and then reading the maths off that picture. That gap, between manipulating symbols and understanding a situation, is the whole game.
Here is the idea that reframes everything, and it's worth pausing on.
A bare sum is one task: calculate. A word problem is two tasks stacked on top of each other. First you must comprehend — read the words, work out what is happening, and decide what the question is actually asking. Only then do you calculate. Most children, and most worried parents, focus entirely on the second task. But the research is strikingly clear that the first task is where the wheels come off.
Think of a word problem as a recipe written in a foreign language. The cooking itself — chopping, stirring, timing — your child can do beautifully. But the instructions arrive in a language they read slowly and uncertainly, so by the time they've struggled through the sentence, half the meaning has slipped away and they're left clutching a few stray ingredients (the numbers) with no idea what dish they're making. The problem was never the cooking. It was the translation.
This is why a child can be genuinely good at arithmetic and still drown in word problems, and why "do more sums" almost never fixes it. You can't drill your way out of a comprehension bottleneck by practising the part that wasn't broken. The repair has to happen at the reading-and-understanding step — the hidden first task.
This is well-trodden research ground, and a few findings recur across labs and countries with unusual consistency.
In a now-classic study, Hegarty, Mayer and Monk (1995) tracked how successful and unsuccessful problem-solvers actually read word problems — including where their eyes lingered. The unsuccessful solvers used what the researchers called a "direct translation" strategy: they plucked out the numbers and keywords and stitched them into a calculation, largely ignoring the meaning. The successful solvers built a "problem model" — a mental representation of the situation the words described — and worked from that. Same words, two completely different reading strategies, and one of them reliably wins.
Working with around 225 nine- and ten-year-olds, Vilenius-Tuohimaa, Aunola and Nurmi (2008) found that performance on maths word problems was strongly related to reading comprehension — and crucially, the link held even after accounting for how fast and fluently the children could read the words on the page. In other words, it isn't just about decoding the text quickly; it's about understanding it. The two skills, they suggested, draw on shared reasoning abilities.
This isn't a momentary effect. In a longitudinal study, Björn, Aunola and Nurmi (2016) found that children's text comprehension in primary school predicted their maths word-problem skill in secondary school, even after controlling for early reading fluency and basic calculation. The comprehension a child builds at nine is still shaping their maths at fourteen. That's a powerful argument for treating reading comprehension as a maths investment, not just an English one.
Researchers who study this, such as Anton Boonen and colleagues (2016), have gone as far as to argue that maths teaching should explicitly include reading-comprehension strategy training — teaching children to visualise the situation rather than hunt for surface cues. Their work points to a specific skill that separates strong solvers: the ability to build a visual-schematic representation (a meaningful picture of the relationships) rather than a vivid-but-useless mental image of, say, the cupcakes themselves. It's not picturing prettily; it's picturing the structure.
Several forces combine, and none of them is a verdict on your child's intelligence.
Comprehension competes for the same workspace as the maths. Holding a sentence's meaning in mind and carrying out a calculation both draw on working memory — that small mental desk where we juggle information for a few seconds. Reviews of the field (Raghubar, Barnes and Hecht, 2010) show how heavily maths leans on this limited resource. A child wrestling to understand the words has little capacity left for the sums, so even easy arithmetic collapses.
Comparison language is genuinely counter-intuitive. Phrases like "3 more than", "2 fewer than", or "twice as many as" describe relationships, and the maths often runs opposite to the surface word. "Mary has 2 fewer sweets than John; Mary has 5; how many has John?" needs addition — despite the word "fewer." Children taught to trust keywords walk straight into these. Researchers call these "inconsistent" problems, and they trip up even strong readers.
Stories hide their structure. Real word problems bury the mathematical skeleton inside irrelevant detail, unfamiliar contexts, and sometimes deliberately distracting numbers. The skill of stripping a story down to "what's being combined, separated, shared or compared?" is exactly the skill schools assume children have and rarely teach directly.
Word-problem difficulty is universal, but different systems have found revealingly different responses. Tap through five.
In England, the reformed maths curriculum and SATs lean heavily on "reasoning" — worded, multi-step problems. This has sharpened a long-standing concern: children with weaker reading comprehension, and the large number of pupils learning in English as an additional language, can be held back in maths by the reading demand rather than the maths. Schools increasingly recognise that a maths intervention sometimes needs to be a reading intervention.
The United States, with a very large population of English-language learners, sees the reading-load issue at scale. National assessments repeatedly show that questions phrased in dense or unfamiliar language depress scores independently of the underlying maths — which is why "plain-language" item design and explicit comprehension strategies have become a major focus of American maths-education research.
Singapore's answer is famous and worth borrowing: the bar model. Children are taught, from early on, to draw the situation as labelled bars before calculating — literally turning the words into a picture of the relationships. It is, in effect, a national programme for building the "situation model" that research says successful solvers use. Many teachers worldwide now adopt it for exactly this reason.
The Netherlands has produced some of the sharpest research on how children represent word problems — distinguishing useful "visual-schematic" pictures of the structure from unhelpful, picture-the-cupcakes mental images. Dutch researchers have shown that teaching children to build the right kind of representation directly improves their problem-solving, lending hard evidence to the "draw the relationships" approach.
Finland, a perennial reading powerhouse, is where several of the key longitudinal studies linking reading comprehension to word-problem skill were carried out. The Finnish data make the long-range case vividly: strong early text comprehension shows up as stronger maths problem-solving years later — a reminder that the reading and the maths are deeply entwined.
The throughline across the strongest responses is the same idea in different clothes: stop treating word problems as a reading test the child must survive, and start explicitly teaching the bridge — turning words into a picture of the situation, whether that picture is a Singapore bar model, a Dutch schematic, or a quick sketch at the kitchen table.
This is one of those happy cases where the most powerful help requires no advanced maths at all. It requires slowing down the reading and making the story visible. Here's how.
Teach your child a fixed routine for every word problem, and practise it until it's automatic: (1) Read it twice, then say it back in your own words. (2) Draw it — picture the situation. (3) Only now decide the maths. The magic is in the order. Rushing children skip straight to step three, grab the numbers, and guess. Forcing steps one and two — the comprehension steps — is what turns a "I can't do word problems" child into one who can. It costs you nothing but a little patience at the start.
The classroom version of this is to teach comprehension as deliberately as we teach calculation, rather than assuming it arrives by osmosis.
Teach representation explicitly. Don't just model the answer; model the reading. Think aloud: "Before I touch a number, I'm going to picture this. Who's involved? What's changing? Let me draw it." Bar models and schematic diagrams give children a concrete tool for building the situation model the research prizes.
Kill the keyword crutch. If your scheme teaches "altogether means add," counter it with inconsistent problems where the keyword lies. Children need to learn that the words point to a relationship to be pictured, not an operation to be triggered. A few well-chosen counter-examples are worth a hundred consistent ones.
Separate the two loads when diagnosing. When a child fails a word problem, find out which task broke: read it to them and simplify the language — if they can now solve it, the gap is comprehension, not maths. This tells you whether to reach for a reading strategy or a maths one. It connects directly to the working-memory issues we cover in our piece on blanking and the foundations point in understanding versus memorising.
Our practice portal includes worded, reasoning-style questions with step-by-step worked solutions that model the situation, not just the answer — so a child learns the read-it, draw-it, solve-it habit rather than guessing from keywords.
Open the practice portal →"If they're bad at word problems, they need more maths practice."
Usually the maths is fine and the comprehension is the bottleneck. Drilling sums reinforces the part that wasn't broken. The fix is reading-and-representation practice, not more arithmetic.
"Teach them the keywords — 'altogether' means add, 'left' means subtract."
Keywords work on easy problems and fail on hard ones, especially comparison problems where the word lies. Successful solvers picture the situation; that's the habit to build.
"Reading is for English; it has nothing to do with maths."
Reading comprehension predicts word-problem performance years into the future. Strengthening comprehension is one of the most effective maths investments a family can make.
The next time your child stalls on a worded question, try resisting the instinct to explain the maths. Instead, ask them to tell you the story, then to draw it. Watch how often the right calculation simply appears once the situation is visible — because the maths was never the missing piece. Word problems feel like a wall, but they're really a bridge between reading and arithmetic, and bridges can be built. Teach the picture, and a great many "I can't do word problems" children quietly discover that, all along, they could.
Because a word problem is two tasks, not one: first understand the situation, then calculate. The reading half is where most children stall. If your child calculates fluently but freezes on worded questions, the bottleneck is comprehension — turning words into a clear mental picture — not the arithmetic, which is why more sums rarely helps.
It's one of the most common traps in maths teaching. Keyword-spotting works on easy problems but fails on anything needing real understanding — "2 fewer than" can require addition, despite the word "fewer." Research finds successful solvers build a model of the situation while unsuccessful ones grab numbers and keywords. Teach the picture, not the keyword.
For word problems, reading comprehension matters a great deal. Studies show comprehension predicts word-problem performance even after accounting for reading speed and arithmetic skill, and early comprehension predicts word-problem skill years later. Strengthening general reading comprehension is one of the most useful things you can do for a child's maths.
Slow the start. Before any calculation, have your child retell the problem in their own words and draw what's happening — without touching the numbers. This forces the comprehension step that rushing skips. A simple bar model or sketch turns an abstract story into something visible, and the right operation usually becomes obvious.
Often, yes. A word problem already loads reading comprehension and working memory heavily; doing it in an additional language adds another layer. This doesn't reflect lower maths ability — it reflects a heavier reading load. Visual strategies like bar models, and pre-teaching the vocabulary of comparison and change, help considerably.
The free assessment pinpoints whether the barrier is the reading or the maths — and shows you the read-it, draw-it, solve-it method that fixes most worded-question struggles. Calm, unhurried, no obligation.
Book the free assessment