Every elementary teacher has seen it. A child sails through a page of 34 − 18 problems, gets them all right, and then freezes when the same arithmetic shows up as a word problem. The arithmetic is identical. The difficulty is not the arithmetic.
Word problems are hard because they make three separate demands in sequence. First, the child must *read* and understand the English sentences — a comprehension task. Second, the child must *translate* the English into an arithmetic expression — a modeling task. Third, they have to *compute* the answer — the arithmetic task they already know how to do. Any of the first two can fail independently.
The scaffold we recommend at MathVenture is called "label, then think". When the child sees a word problem, ask them two questions before they touch the arithmetic. First: *what does the question want me to find?* Circle the number you're looking for. Second: *what do I already know?* Underline the numbers that are given. The child is not yet solving anything — they are labeling the problem. That labeling step off-loads the comprehension task so their working memory is free for the modeling step.
Once the problem is labeled, the child can choose an operation. Many early-grade kids will try to guess which operation is needed from keywords: "in all" means add, "how many left" means subtract. Keywords work often enough to be dangerous — they feel reliable right up to the day they fail. A better habit is to ask: "am I putting things together, or taking them apart?" That framing is closer to the math and generalizes to multiplication and division.
The Word Problem Bakery game in MathVenture makes the labeling step visible. The child taps the number they know, then taps the word that tells them what to find, *before* the game lets them choose an operation. The process is slow the first time and fast by the fifth. That is the process we want their brain to learn.
At home, read word problems aloud together. Let the child re-tell the problem in their own words before attempting it. Comprehension comes first, arithmetic second.
A useful piece of structure here is the four-step routine: read, picture, plan, check. Read the problem twice — out loud, slowly. Picture what is happening — a quick sketch, a tally, an arrangement of coins, anything that turns the words into something you can see. Plan the operation that answers the question. Then check by asking whether the answer is reasonable. If a child has bought four apples and ended up with three hundred dollars in change, something has gone wrong before we even check the arithmetic.
For young children — say first or second grade — the picture step is by far the most important. A child who can't draw the problem doesn't yet understand it, no matter how confidently they say a number. A child who draws the problem and then solves it confidently has done real mathematical work. Worksheets that demand only the answer reward the wrong half of this; we get more leverage from a single problem worked carefully than from twenty answered fast.
Older children — say fourth grade and up — start to encounter problems where translation is the actual difficulty. A problem about train speeds is mostly an exercise in turning a paragraph into an equation. Time spent practising that translation, slowly and aloud, pays back disproportionately in middle school. Two short routines are particularly useful at this stage: "what would change if a number went up?" (this builds variable-thinking before algebra arrives), and "what is the question really asking?" (because the actual question often hides behind a paragraph of irrelevant detail). Both are habits more than skills, and habits stick.