Homework Isn’t the Problem. Bad Homework—and District Silence—Are.

The homework debate is usually presented as a choice between two extremes. Either homework is an unquestionable academic tradition, or it is an outdated burden that should disappear. Neither position is especially useful.

Homework is not automatically valuable simply because a teacher assigns it. Hours of repetitive questions, unfinished classwork pushed into the evening, and copying completed for compliance deserve criticism. But the existence of bad homework is not an argument against every form of work outside class. It is an argument for defining what worthwhile homework should accomplish.

I teach seventh-grade science. My own system combines two scales of homework: micro homework that asks students to revisit an individual lesson, and macro homework that helps them manage learning across an entire unit. Neither piece is perfect, and neither should be copied without adaptation. Together, however, they illustrate what homework can become when its purpose is learning rather than workload.

Start by naming the purpose

Harris Cooper and his colleagues found that the relationship between homework and academic achievement generally becomes more positive as students grow older, while also cautioning that age, design, and amount matter. That is a far more useful conclusion than either “more is better” or “homework does not work.”

Before assigning anything, a teacher should be able to identify the learning function it serves. “Students need practice” is not yet a complete answer. Practice at what?

An assignment does not have to accomplish all six purposes. It should accomplish at least one clearly and efficiently.

Micro homework: ten minutes of thinking

For an individual middle-school class, approximately ten minutes can be enough. Short does not have to mean shallow. A concise assignment can require far more judgment than a page of routine questions.

My recurring lesson-summary structure asks students to identify an answer question or main idea, construct a simple model with labels and captions, select three concise key details, and record at least two questions for the next class. The format remains predictable, but the thinking changes with every lesson.

Each part creates a different kind of productive constraint. Naming the main idea requires prioritization. A model asks students to translate verbal learning into a visual form. Limiting the details forces students to select rather than transcribe. Generating questions asks them to notice what remains unresolved.

The point is not to produce a polished artifact every night. The point is to leave visible evidence that a student returned to the lesson and thought about it again.

Why questions belong in the summary

The final column may be the most useful. Student questions can reveal misconceptions, missing background knowledge, curiosity, or a place where instruction needs to slow down. A question does not have to be profound to be valuable. “Are platelets cells?” and “Why does a bruise change color?” provide different information, but both tell me something worth knowing.

Questions also shift homework away from the pretense that learning should end in certainty. Students are allowed to finish a lesson with unfinished thinking. Their uncertainty becomes part of what the class can use next.

Macro homework: make the whole unit visible

Micro homework helps students revisit one lesson. Longer-term assignments serve a different purpose. At the beginning of a unit, I assign work that remains open for several weeks and is due on the day of the test. This typically includes a study guide and a vocabulary resource.

The Climate Change and Barnegat Bay study guide is organized around six large conceptual questions: human impacts on the bay, salt marshes, sea-level rise, energy and matter in the ecosystem, limiting factors, and sustainable solutions. Each is broken into smaller questions and connected to relevant vocabulary.

This is not merely a test review handed out at the end. Giving students the guide at the beginning tells them where the learning is heading. It also allows them to return to earlier sections as their understanding develops.

The local context matters. Questions about sea-level rise do not remain abstract when students consider Point Pleasant and Barnegat Bay. Ecosystem services, stormwater, salt-marsh loss, and living shorelines become connected pieces of a place students recognize.

Long-term work requires long-term support

Middle school students will not manage a three- or four-week assignment perfectly. Some will work steadily. Some will complete parts during available class time. Some will wait far too long. That is not proof that long-term work is inappropriate. It is proof that executive-functioning skills are still developing.

Teachers can support the process without converting the assignment into a series of rigid daily worksheets:

• Introduce the entire resource at the beginning of the unit.
• Explicitly connect new lessons to the appropriate sections.
• Suggest progress points and provide occasional work time.
• Allow students to revise early answers as their models improve.
• Keep the final deadline meaningful while scaffolding the route toward it.

We would not stop asking students to write because some struggle with writing. We should not stop asking them to manage time simply because they are not yet good at it.

Vocabulary: worth teaching, not worth copying

The vocabulary dictionary is the part of my system I question most. Its structure is simple: a word, a picture, and a definition. In Unit 6.6, students encounter terms such as system, subsystem, structure, function, cross-section, tissue, nucleus, nerve cell, plasma, cell division, development, and growth.

At its worst, vocabulary homework becomes copying for compliance. A student can reproduce a definition without understanding it, and generative AI makes that kind of task even easier to outsource. If the dictionary functions only as filler, then it deserves to be revised.

Still, scientific vocabulary matters. A single unfamiliar term can block a student from understanding an entire explanation. Pairing a student-friendly definition with a visual can help students connect language to a concrete idea. A classroom word wall and opportunities to work during class can also make the assignment more accessible.

The most important test comes later: Do students use the words in their models, explanations, conversations, and assessments? Do they return to the dictionary as a reference?

Homework in the age of AI

A 2026 RAND report found that AI use for homework among students from middle school through college rose from 48 percent to 62 percent between May and December 2025. At the same time, 67 percent of surveyed students agreed that increased AI use for schoolwork could harm critical-thinking skills.

Students themselves recognize the danger of outsourcing too much of their thinking.

The answer is not to abandon homework because AI can complete it. The answer is to reconsider what we ask students to produce. If an assignment consists of generic questions that can be pasted into a chatbot, the problem may not be only student dishonesty. The assignment may not offer enough reason to do the thinking personally.

Almost no take-home assignment is completely AI-proof. Brief tasks tied closely to a particular classroom experience are simply less rewarding to outsource. Ongoing notebooks and study resources also matter because today’s work becomes part of tomorrow’s conversation. The goal is not a flawless final product. It is a visible trail of developing thought.

The second problem: district silence

My district limits homework to 10 percent of a student’s grade. That is a reasonable safeguard against homework disproportionately damaging an average. But it is a grading rule, not a homework policy. It says nothing about how much work students should receive, what purposes it should serve, or how expectations should change as children mature.

When districts remain silent, one teacher may assign manageable practice, another may assign lengthy work with little instructional value, and a third may assign nothing. Students experience dramatically different expectations from classroom to classroom, even within the same school.

An Education Week Research Center survey published in 2026 found that 41 percent of teachers had reduced the amount of homework they assigned over the previous two years. Homework was also less common in higher-poverty districts: 28 percent of teachers in those systems reported assigning none, compared with 17 percent in more-affluent systems.

That disparity should concern us. Eliminating homework does not eliminate out-of-school learning. Families with money and time can provide tutoring, enrichment, test preparation, and academic supervision. Students without those advantages may depend more heavily on schools for structured practice and clear expectations.

Build a vertically aligned K–12 policy

Every district should adopt a K–12 homework policy that gradually increases expectations for time, independence, and complexity. Such a policy should preserve teacher discretion. It should not dictate identical assignments or impose arbitrary nightly quotas. It should establish a common floor for quality and a common ceiling for burden.

A district working group should include elementary, middle, and high school teachers; special educators; multilingual-learner specialists; representatives from major academic disciplines; and curriculum leaders. Together, they could recommend reasonable time ranges, accessibility principles, feedback expectations, long-term assignment supports, assessment-week coordination, and clear rules for AI use.

Subject-area teams could also create choice menus that give teachers practical options without prescribing one method:

Teachers would remain free to select, adapt, or design assignments appropriate to their students. The difference is that those choices would exist inside a coherent district philosophy.

Ultimately, a superintendent or curriculum leader must be willing to say what the district believes homework is for. A grading cap alone is not a homework policy. Neither is leaving every individual teacher to reinvent the answer.

A system, not a stack of worksheets

Teachers are right to reject hours of worksheets, meaningless copying, and assignments that punish students for circumstances outside their control. Bad homework should be challenged. District silence should be challenged, too.

Purposeful homework is brief when it supports daily reflection, sustained when it develops long-term planning, closely connected to classroom learning, and accessible to different students. It should be designed to reveal thinking rather than manufacture points.

Homework earns its place in students’ lives when teachers can explain why it matters—and when districts ensure that answer becomes more thoughtful, demanding, and developmentally appropriate from kindergarten through graduation.