Few things cause more household friction than maths homework — and few topics are argued about with less attention to the evidence. The research won't comfort the true believers or the abolitionists: homework does little for younger children, helps modestly in secondary school, and at every age its value depends almost entirely on the type. Here's what that means for your evenings.
It's half past six. Dinner's going cold, a worksheet of thirty near-identical long-division questions sits half-finished, and a nine-year-old is in tears insisting they "can't do it" while a parent, who can do it but can't explain it the way the teacher did, feels their own patience fraying. Somewhere in this scene is a reasonable question that almost nobody stops to ask: is this actually helping? Or are we sacrificing a peaceful evening — and a child's feelings about maths — for a ritual whose benefits we've simply assumed?
It turns out the question has been studied carefully, and the answer is more interesting than either "homework is essential" or "homework is pointless." The evidence draws a clear, nuanced picture — one that should change how families and schools handle those evenings, and that frees a lot of parents from a guilt they didn't need to carry.
Homework occupies a strange place: near-universal, rarely questioned, and quietly dreaded in a great many homes. For some children it's a brief, useful habit; for others it's the daily flashpoint that turns maths from a subject into a source of dread. Parents are told it builds discipline and reinforces learning, so they enforce it — often against a child's resistance, often without knowing whether the specific homework in front of them is worth the fight. The first useful move is to stop treating "homework" as one thing with one verdict, and start asking which homework, for which child, at which age.
The public argument is almost always about quantity — more homework versus less, or none. But the research points somewhere else entirely: to quality and age. A pile of repetitive busywork and a short, well-designed spaced-practice task are both "homework," yet they do completely different things. Asking "does homework work?" is like asking "does food help?" — it depends enormously on which food, how much, and for whom.
Imagine judging exercise purely by hours spent in the gym, ignoring what you actually did there. Two hours of wandering between machines you don't know how to use isn't worth twenty focused minutes with a plan. Homework is the same: ten thoughtful minutes that revisit the right things beat an hour of copying, and an hour of copying can be worse than nothing if it convinces a child they hate maths. Counting minutes measures the wrong thing.
Once you see that, the emotional cost comes into focus too. A homework task that produces a nightly meltdown isn't neutral even if the maths is fine — because, as our piece on maths anxiety shows, the feelings around maths transmit and stick. Bad homework can teach a child to fear the subject faster than good homework can teach them the content.
The most thorough synthesis, by Cooper, Robinson and Patall (2006), found a positive link between homework and achievement overall — but a much stronger one in secondary school (roughly Years 7–13) than in primary (Years K–6), where it was weak. In other words, homework tends to help older students and does little for younger ones. The popular assumption that early homework builds a foundation isn't well supported.
The Education Endowment Foundation reaches the same conclusion: homework shows a positive impact on average at secondary level (worth several additional months' progress) but a much smaller or negligible effect in primary. Crucially, the EEF stresses that the quality of the task and whether it's linked to classroom learning matter more than the sheer amount.
Across countries, the picture gets sharper still. International comparisons (including OECD PISA analyses) find that beyond a modest amount, extra homework time buys little or no extra performance — and some of the highest-performing systems set relatively little. The relationship isn't a straight line where more equals better; it flattens, and past a point the costs (fatigue, lost downtime, resentment) start to outweigh the gains.
If quantity isn't the lever, what is? The homework that helps shares the features of effective practice generally. It's spaced and retrieval-based — revisiting earlier topics from memory rather than re-copying today's (see spaced and retrieval practice). It's pitched right — hard enough to require thought, not so hard it causes overload and tears. It's short and regular rather than long and sporadic. And it comes with feedback, so errors are corrected rather than rehearsed.
The homework that fails is the mirror image: long, undifferentiated worksheets; copying and "finish the questions you didn't do in class"; tasks too hard to do without help that isn't available; and work that's collected but never marked or discussed. There's also a self-regulation angle that explains the age effect — older students can plan, persist and work independently, so they can extract value from homework; younger children often can't yet, so unsupervised homework is easily done badly. The benefit of homework is really the benefit of good independent practice, and that's a capability that grows with age.
Finland is the famous counter-example: comparatively little homework, later school starts, and yet strong international results for years. It punctures the assumption that more homework is the route to higher achievement — though Finland's recent score declines are a reminder that homework is only one variable among many.
Japan and South Korea combine high achievement with heavy out-of-school study, much of it in private tutoring ("juku"/"hagwon"). It shows homework and extra practice can coexist with top results — but also at a real cost in student wellbeing and pressure, which both countries are actively trying to address.
Singapore tops maths tables and has a strong practice culture, but has been deliberately reducing excessive homework and exam pressure in recent years — an implicit acknowledgement that beyond a point, more isn't better and the wellbeing cost is real.
The US produced the foundational homework research (Cooper) and shows the familiar pattern: modest secondary benefit, little primary benefit, and a "10-minute rule" rule of thumb (roughly 10 minutes per night per school grade) that tries to keep amounts age-appropriate.
England sets homework widely, including in primary where the evidence for benefit is weakest. The EEF's guidance — modest amounts, high quality, linked to classwork, with feedback — is the sensible reading of the evidence, even if practice varies a lot between schools.
If your primary-age child finds maths homework miserable and you've wondered whether it's really achieving much — the evidence largely agrees with your instinct. The achievement benefit at that age is small. Reading together, playing maths games, cooking and counting in real life will often do more for their long-term relationship with numbers than a fraught worksheet. You're not failing them by keeping it light.
Set less, but better. The evidence rewards quality over quantity. Short, focused tasks that revisit prior learning (spacing and retrieval), pitched at the right level and tied explicitly to classwork, beat long undifferentiated worksheets — and they're more likely to actually get done.
Close the loop with feedback. Homework that's collected but never discussed mostly rehearses errors. If it's worth setting, it's worth a quick check and correction — otherwise the practice can entrench misconceptions.
Mind the age and the equity. Be especially sparing with primary homework, where the benefit is weakest, and remember that home support varies hugely between families — homework that assumes a quiet space and a knowledgeable adult can widen gaps rather than close them.
If you want short, high-quality practice that revisits the right topics with instant feedback, the practice portal is built for exactly that — the kind of homework the evidence actually supports.
Open the practice portal →"More homework means better results."
The benefit flattens and can reverse past a modest amount. Quality and age matter far more than quantity.
"Early homework builds a strong foundation."
At primary age the homework–achievement link is weak. Real-life maths, games and reading often do more for young children.
"Finishing the worksheet matters more than how the evening felt."
Maths feelings transmit and persist. A meltdown over busywork can cost more than the worksheet could ever gain.
It's age-dependent. In secondary school there's a modest positive link; in primary it's weak to negligible. So homework can help older students, but more isn't automatically better, and quality matters more than quantity.
Keep it modest and age-appropriate — the benefit levels off and can reverse beyond a point, especially for younger children. For maths, short, focused, regular practice beats long marathons. A common rule of thumb is around ten minutes per night per school grade.
Younger children have less developed self-regulation and study skills, so unsupervised homework is easily done badly or abandoned. The benefit grows with age as students become more able to work independently and the tasks become clearer practice.
Short, focused practice that revisits earlier material from memory (spacing and retrieval), is pitched at the right difficulty, and comes with feedback. Busywork, copying and huge undifferentiated worksheets do little and can damage motivation.
Often not, if the homework is low-quality and the conflict is high. Nightly battles can transmit maths anxiety and damage your relationship, which costs more than poor homework gains. Protecting calm and goodwill usually matters more than completing a mediocre worksheet.
Every Insight Bay programme sets short, high-quality, feedback-rich practice — the kind the evidence supports — and keeps the evenings calm. We'll design it around your child in the free assessment.
Book the free assessment