Handing questions out is not the job
Homework gets set for two different reasons and mixing them achieves neither. One is to score, which is summative assessment. The other is to steer your teaching, which is formative assessment. In the second case the output is not the student's mark but the subject of your next lesson. Homework set without deciding which one it is usually manages neither: too small to count for a grade, too late to change anything.
Which question defeated everyone
The most useful thing in thirty submissions is not the average but the spread. If twenty five of the class got one question wrong, that question is measuring itself rather than the students: either the topic was not taught or the wording is ambiguous. Both are things you need to know. Reading papers one by one hides this, because each paper tells its own story. Looked at per question, the errors stacked on one item collapse into a single sentence: this needs teaching again.
When to set it
Homework set the day after the lesson tests a topic before it has faded, which makes it feel easy, and easy homework measures nothing. The spacing effect says the opposite: asking a few days later is both harder and more instructive. The arrangement that works is two short pieces on the same topic, one the next day and one a fortnight later. The second one usually tells you far more than the first.
The students who do not hand it in
Used as a discipline tool, the submission list kills the learning side of homework. A more useful reading is that a student who did not submit is usually not lazy but stuck at the start. When the entry cost is high, when you first have to find the source and then work out what is being asked, the student who needs it most never begins. Lowering that cost, by stating the source and the scope plainly, raises submission rates more reliably than penalties do.
Where grading does harm
The moment homework carries a mark, the student's goal shifts from learning to scoring and being wrong becomes expensive. But the testing effect comes precisely from being wrong and correcting it. So homework meant to steer works better ungraded, or graded on submission alone. Roediger and Karpicke 2006 measured that the act of being tested is what teaches; it teaches without a mark attached too.
Doing all this with less effort
None of the above needs software; it works on paper, it just takes longer. In WeSolve+ you assign a quiz as classroom homework, see who has submitted on one screen, and the question the class found hard surfaces on its own as submissions arrive. Because the questions come from the file you uploaded, the homework never wanders outside what you taught. The same file also becomes an audio recap and a study booklet.
By Eren Kayıkçı (Study planning and habits)
Sources used on this page
- WeSolve+ App Store listing
- Formative assessment
- Summative assessment
- Educational assessment
- Item analysis
- Feedback
- Corrective feedback
- Homework
- Classroom management
- Teaching
- Bloom's taxonomy
- Bloom's 2 sigma problem
- Peer tutoring
- Peer assessment
- Peer instruction
- Learning by teaching
- Testing effect
- Spacing effect
- Forgetting curve
- Active recall
- Roediger and Karpicke, Test-Enhanced Learning (2006)
- Karpicke and Blunt, retrieval practice (2011)
- Cepeda et al., distributed practice (2006)
- Dunlosky et al., improving students' learning (2013)
- Metacognition
- Self-assessment
- Rubric (academic)
- Grading in education
- Curriculum mapping
- Test (assessment)
Should I grade the homework?
If it is meant to steer your teaching, ungraded or graded on submission alone works better. A mark shifts the student's goal from learning to scoring and makes being wrong expensive.
How many questions should I set?
Short and frequent beats long and rare. Five to ten questions get submitted and can actually be read.
Is setting two pieces on the same topic wasteful?
No, that is where most of the information is. Setting the second one a fortnight later shows whether the topic has actually held.
What about students who never submit?
Look at the entry cost first. Homework that states its source and scope plainly gets submitted more often than homework backed by penalties.
Do I have to read all thirty papers individually?
The most useful signal is not in individual papers but in the errors stacked on the same question. Reading per question is both faster and more informative.
Last updated: 2026-08-25 · Written with AI assistance and reviewed before publishing.