Is Written Feedback a Waste of Time?
New study finds 1 in 5 students completely ignore feedback, while nearly half of those who did try to act on the feedback, don't improve at all.
A recent study published in Contemporary Educational Psychology reveals some uncomfortable truths about how secondary school students actually engage with the feedback they receive. Researchers examined 937 German students who were given feedback on their English writing tasks and asked to revise their work.
The results are sobering:
20% of students completely ignored the feedback, making no revisions whatsoever
Of the students who did engage with the feedback, 47% showed no improvement in their work
So in total, that means about two-thirds of students either ignored feedback entirely or failed to benefit from it despite making revisions. This raises serious questions about the efficiency of marking books with written feedback.
There were also several paradoxical findings that challenge conventional understanding of feedback processes. Perhaps most surprising is that higher-performing students were actually less likely to improve after receiving feedback. This counterintuitive result suggests that there’s either a motivation issue where these students may feel less compelled to make substantive changes, or there is some kind of ceiling effect of diminishing returns where further improvement becomes increasingly difficult.
Meanwhile, a troubling disconnect emerged between perception and action: many students reported finding feedback useful but still failed to implement it effectively. In other words, even when they know it’s useful, they still either ignore it or don’t know what to do to improve.
Who Ignores Feedback?
The study identified several patterns regarding which students were most likely to disengage from feedback. Firstly, male students were significantly more likely than female students to ignore feedback completely. This gender gap persisted even after controlling for cognitive abilities, motivation, and prior performance, suggesting deeper socialization factors at play. The researchers noted that this aligns with previous findings showing male students tend to approach evaluative situations more competitively and may be more likely to dismiss external input.
Structural and developmental factors also played significant roles. Students from comprehensive (public) schools struggled more with feedback implementation than their academic-track peers, even when controlling for other variables, pointing to potential systemic inequalities in feedback processing capabilities. Similarly, younger students (Grade 7) were more likely to unsuccessfully engage with feedback compared to older students (Grade 9), suggesting developmental factors beyond mere ability or motivation influence feedback utilisation.
The type of feedback proved crucial for implementation but not for initial engagement—students were equally likely to attempt revisions regardless of feedback design, but more detailed feedback with specific examples significantly increased successful implementation.
Marking: more trouble than it’s worth?
Dylan Wiliam told me once that “if you price teachers’ time appropriately, in England we spend about two and a half billion pounds a year on feedback and it has almost no effect on student achievement”
We’ve known for a long time that not all feedback is effective. Kluger and DeNisi's 1996 meta-analysis found that while feedback interventions overall had a modest positive effect on performance, 38% of the feedback interventions actually decreased performance. This startling finding challenges our fundamental assumptions about feedback's universally beneficial nature. When combined with this recent study that roughly two-thirds of students either ignore feedback entirely or fail to improve despite engaging with it, educators might reasonably question the substantial time investment marking requires.
The issue isn't that feedback cannot be effective, but rather that traditional approaches to marking are done not because they’re effective, but because we think we should do it. Without addressing the mechanisms of student disengagement, from gender differences in feedback receptivity to the gap between recognising feedback value and knowing how to implement it, many educators may be unknowingly performing a laborious, time-intensive ritual that serves institutional expectations rather than actual student learning.
Personally I don’t think marking is a total waste of time (as much as I hated it), indeed there are strong reasons for teachers to engage with student work beyond simply providing feedback. Carefully reading a student's work demonstrates respect for their efforts and helps teachers gain deeper insights into their thinking, misconceptions, and progress. This close analysis allows teachers to better understand individual students' strengths and challenges.
Additionally, the formative dividend of marking provides crucial guidance that can shape instruction. When teachers analyse student work, they can identify patterns across the classroom, spot common misconceptions, and adjust their teaching accordingly. This diagnostic function helps teachers make informed decisions about what to teach next, which concepts need reinforcement, and which students might need additional support. But if teachers are spending a lot of their time marking and the feedback is either not effective or students are not bothering to even read it then where do we go from here?
The Promise of AI and Comparative Judgment
The future of marking and feedback may lie in combining two emerging approaches: artificial intelligence and comparative judgment. AI tools can now analyze student work with increasing sophistication, identifying patterns, providing targeted feedback, and even adapting to individual student needs. Meanwhile, comparative judgment—where teachers make quick relative assessments between pieces of work rather than marking against complex rubrics—has been shown to produce more reliable and efficient evaluations.
has thought about this more than most and as No More Marking's recent work demonstrates, this combination offers genuine potential:Their journey highlights important lessons about AI's limitations, particularly concerning hallucinations and reliability. Rather than fully automating assessment, the most promising approach puts humans and AI in complementary roles with teachers making comparative judgments and leaving brief audio comments that AI then transcribes and synthesises into comprehensive feedback. This hybrid approach preserves the teacher's crucial role in understanding student thinking while eliminating much of the mechanical burden. The most promising aspect isn't just workload reduction but the potential to redirect teacher energy toward high-impact interactions: discussing feedback with students, modelling revision strategies, and creating classroom cultures where feedback is valued and implemented. By addressing both the efficiency and effectiveness dimensions of the marking conundrum, these technologies might finally resolve the tension between what we know students need and what is sustainable for teachers to provide.
I'd like to know if similar research has been done regarding feedback on speaking.
As a student of Catalan, I asked for feedback on my writing from languagetool.org immediately after finishing the writing. This meant the feedback was timely and I used it to Improve my writing, but also to create a database (Excel) of all my errors, the corrections and an explanation.
I didn't feel it was cheating although I imagine it got me better grades. I told my teacher what I was doing and she was supportive. In fact, it probably saved her time and red ink as she didn't have to correct endless surface errors and could comment on my writing in more general terms.
I told her I enjoyed writing in Catalan, and she said, "It shows." I felt very proud of that feedback!
Interesting about the gender gap.... I think that the most significant barrier is always the time between the feedback and when the task was finished. It's hard to juggle in the classroom.