The Nightly Ritual
The scene plays out in millions of American homes every evening: a child sits at the kitchen table with a worksheet while a parent hovers beside them, correcting answers, coaxing focus, and watching the clock tick past bedtime. The assignment takes forty minutes, and the child doing it is eight years old.
Nobody questions whether this ritual serves its stated purpose, because the assumption that homework helps children learn is so deeply embedded in American education that challenging it feels like arguing against gravity. Teachers assign it nightly, parents demand to see it completed, and school boards codify minimum expectations into district policy.
Harris Cooper, a psychology professor at Duke University, decided to check whether the assumption held up against the research.
What Two Decades of Data Show
Cooper and his colleagues assembled every study they could find on the relationship between homework and academic achievement in the United States between 1987 and 2003, grouping the research into four design types to examine what the collective evidence actually said (Cooper, Robinson, & Patall, 2006).
For high school students, the picture was clear and favorable: more homework correlated with better achievement at r = 0.25, a moderate and statistically meaningful relationship that held across subjects and measurement types.
For elementary school students, the picture was equally clear in the opposite direction. The benefit was zero.
Across 35 correlational studies, the average relationship between time spent on homework and academic achievement for students in grades K through 6 hovered around r = 0, and for third grade and below, the correlation was actually negative at r = −0.04, meaning children who did more homework performed slightly worse than those who did less. When homework was compared directly with in-class supervised study at the elementary level, staying in the classroom and practicing under teacher oversight produced better outcomes than sending the same material home.
Over One Hundred Thousand Students Confirm It
Cooper's synthesis was not the only large-scale analysis to reach this conclusion. John Hattie's Visible Learning project, published in 2009, synthesized 161 studies covering over 100,000 students across virtually every educational intervention that researchers have tested, and his meta-meta-analysis calculated separate homework effect sizes broken down by grade level.
Primary-school homework registered an effect size of d = 0.15, well short of Hattie's d = 0.40 threshold for educational significance, the point below which an intervention produces less than typical yearly growth even when heavily resourced.
"Homework in primary school has an effect of around zero," Hattie told BBC Radio 4 in 2014. "It's one of those lower-hanging fruit that we should be looking at in our primary schools to say, 'Is it really making a difference?'"
International data from the Trends in International Mathematics and Science Study sharpen the point considerably. Güven and Akçay analyzed 2015 TIMSS data across dozens of countries and found no relationship between math homework frequency and fourth-grade achievement in the majority of nations studied, including the United States, while Finland and South Korea routinely top international assessments despite assigning substantially less homework than American schools.
637 Hours With No Return
The National Education Association recommends the "10-minute rule," which prescribes ten minutes of homework per grade level per night, so that a first-grader gets ten minutes and a fifth-grader gets fifty.
If a child follows this guideline from first through sixth grade, the total across six years comes to roughly 637 hours of homework, every one of which the best available evidence suggests produces no measurable academic return.
To put a human scale on that number: 637 hours is 26.5 full days, enough time to read approximately 130 chapter books at the average elementary reading pace, or to spend an entire month of weekday evenings playing, exploring, and getting the sleep that pediatricians say elementary students chronically lack. Every one of those alternatives has stronger empirical support for child development than the homework those hours replaced.
The Strongest Case for Homework
The most serious defense of elementary homework doesn't challenge the achievement data but sidesteps it entirely, arguing that homework builds habits which pay off later: time management, responsibility, and the discipline of completing tasks without a teacher watching. Cooper himself makes this argument, writing that "the homework question cannot be reduced to 'Does homework boost achievement?'" even as he acknowledges that for young children, the answer to that narrow question is no.
This deserves to be stated at full strength, because the habit-building hypothesis is genuinely plausible. But the evidence behind it is far thinner than most advocates realize. Trautwein and Köller (2003) found that homework effort, not homework completion, predicted later academic motivation, and that distinction matters enormously: a child forced through a worksheet is technically completing homework, but the motivational benefit appears to come from self-directed effort that younger children are developmentally less capable of sustaining without parental coercion.
The argument also carries an untested assumption: that these habits can only be built through academic homework specifically, rather than through the sports, chores, music practice, and unstructured play that children already engage in and that also demand responsibility and persistence.
What We Didn't Prove
The near-zero correlation applies to elementary grades K through 6 in aggregate and does not mean that every type of homework is useless for every child. Short, focused reading practice may benefit struggling readers, and some evidence suggests moderate writing homework yields small gains in specific skills. The finding concerns achievement as measured by grades and standardized tests, while non-academic outcomes like family stress and sleep loss represent a separate question on which several studies suggest homework has negative effects for young children. This is correlational evidence rather than experimental, and schools that assign less homework may differ from those that assign more in ways the data cannot fully untangle.
The Bottom Line
Homework works for teenagers, and the high-school evidence for it is moderate, consistent, and real. But the elementary-school evidence, accumulated across decades of research spanning multiple continents and assessed through both correlational studies and international comparisons, points in one direction: the relationship between homework and achievement for children under twelve is so close to zero that no study design has reliably detected a meaningful benefit.
The practice persists for exactly the reason Hattie identified: many parents judge the quality of a school by the presence of homework, and a ritual with no measurable academic return survives because removing it sends the wrong signal to the community it serves.
What You Can Do
Stop fighting the worksheet war. If your elementary-age child melts down over homework, the accumulated evidence says the battle is not buying academic improvement, and the stress it generates may be a real developmental cost with no demonstrated return.
Talk to the teacher. Ask what the homework is designed to accomplish, because if the answer is "reinforce what we learned today" with brief focused practice, that is the one use Hattie considers defensible for young children, while busywork assigned to demonstrate rigor has no evidentiary basis.
Protect the evening. The hours between school and bed are finite, and free play, family conversation, outdoor activity, and sleep all carry stronger evidence for child development than homework at this age.