Harmful doctoral feedback can damage confidence and suppress student voice
Updated Jul 02, 2026
student voicefeedbackDoctoral feedback problems are easy to misread as private supervision issues until confidence drops, writing slows, or students stop speaking openly about what is happening. Shulin Yu's Teaching in Higher Education paper, "When feedback is not conducive to learning: a qualitative inquiry into second language doctoral students' experiences of the negative impact of writing feedback", matters because it shows that harmful writing feedback can damage the text, the motivation to keep going, and the relationship around the work itself. For UK universities using PRES, annual review comments, and local postgraduate feedback, that is a useful reminder that feedback quality is part of the student experience, not a narrow supervisory detail.
Context and research question
Higher education research often treats feedback as something designed to improve learning. Yu starts from the harder question: what happens when feedback does the reverse? That matters in doctoral education because writing sits close to progression, confidence, and identity. It also sits inside unequal relationships, where a student may depend heavily on a small number of supervisors, reviewers, or assessors. That concern fits wider evidence on what students say makes feedback genuinely useful, especially when clarity, tone, and actionability matter more than comment volume alone.
The study uses a qualitative design focused on 15 second-language doctoral students. Yu draws on critical incident forms and stimulated recall interviews to examine moments when writing feedback felt harmful rather than developmental. This is not a prevalence study and it does not try to estimate how common these experiences are across the whole sector. Its value lies in surfacing the mechanisms through which feedback becomes counterproductive, which is exactly the kind of detail that broad satisfaction scores often flatten.
Key findings
The paper identifies three connected areas of damage: the text, the student's motivation and emotions, and the feedback relationship itself. That is important because it shifts the discussion away from a simple good-feedback versus bad-feedback binary. Harmful feedback does not just fail to help. It can also make the next stage of writing harder, reduce willingness to engage, and erode trust in the people giving the feedback.
At the level of the text, the problem was not only weak usefulness. Students described feedback that was difficult to interpret, hard to act on, or discouraging enough to disrupt revision decisions. For UK doctoral colleges, that matters because a feedback process can look active on paper while still failing at the point where students need to turn comments into better writing. Usable feedback is not the same thing as abundant feedback.
The emotional and motivational costs were just as serious. Yu reports reduced confidence, negative feelings about writing, and weaker self-efficacy after harmful feedback episodes. In doctoral education, where long projects depend on persistence as much as technical skill, that matters immediately. A student who loses confidence in the writing process is not simply less satisfied. They may also be less willing to draft, revise, ask questions, or test an argument openly.
The third finding is the most relevant for student voice work. Harmful feedback also strained the relationship around the text, making students less likely to challenge, clarify, or discuss what they had received. Yu's summary is direct:
"power dynamics among student writers, supervisors, and journal reviewers" can "suppress students' voices in the feedback exchange process."
That makes feedback a relational issue as well as an instructional one. When students do not feel safe enough to query feedback, silence becomes easy to misread as agreement. For UK teams reading postgraduate comments, that is a serious warning sign because dissatisfaction may surface as vagueness, withdrawal, or softened language long before it appears as a formal complaint.
Practical implications
For UK universities, the first implication is to treat doctoral feedback quality as a student experience issue, not only a supervision craft issue. Staff development should cover specificity, tone, consistency, and opportunities for follow-up, rather than assuming that more comments automatically mean better feedback. The benefit is earlier recognition of the kinds of feedback patterns that affect confidence and progression, not just writing style.
Second, institutions should ask better questions in PRES, annual review, and local postgraduate feedback routes. A generic item on supervision satisfaction is too blunt for the issues this paper surfaces. Students need room to say whether feedback felt clear, actionable, respectful, consistent, and open to discussion. Open-text prompts matter here, and a defensible workflow for analysing open-text feedback makes it easier to separate vague feedback, conflicting advice, and confidence-damaging tone at scale. The benefit is sharper diagnosis before issues become harder to resolve.
Third, doctoral schools should treat strained feedback relationships as early warning evidence. Recent work on why doctoral complaints systems fail when students do not feel recognised points in the same direction: concerns often become harder to read long before they become formal cases. If comments repeatedly point to one-way communication, mixed signals, or fear of questioning feedback, the institution should intervene earlier. The benefit is a better chance of protecting progress while relationships can still be repaired.
Finally, institutions should handle postgraduate comment analysis with stronger governance than a generic survey workflow. Small cohorts, dependency relationships, and sensitive writing-feedback concerns all raise confidentiality risks. A practical student comment analysis governance checklist helps teams decide what can be compared safely, who should see what, and how follow-through is documented. The benefit is evidence that is more actionable for staff and more trustworthy for students.
FAQ
Q: How should a doctoral college respond if local feedback suggests students find writing feedback discouraging or unsafe to challenge?
A: Start by checking whether the issue is isolated to one supervisor, one stage of study, or a wider pattern in the doctoral environment. Then review how feedback is delivered, whether students have a realistic route to clarify or discuss comments, and whether support is available before the issue reaches formal escalation. The practical aim is to restore feedback as a usable dialogue, not just to reduce complaints.
Q: What should teams keep in mind about the methodology of this study?
A: This is a qualitative study of 15 second-language doctoral students, using critical incident forms and stimulated recall interviews. It is designed to explain mechanisms rather than estimate prevalence across all doctoral education. UK institutions should treat it as strong interpretive evidence about how harm can occur, then test for similar patterns in their own PGR surveys, annual reviews, and comment data.
Q: What does this change about student voice in postgraduate research more broadly?
A: It strengthens the case for treating silence, vagueness, and softened criticism as evidence in their own right. In doctoral settings, students may not always name a problem directly if they think the relationship is too unequal or the route to response is unclear. That means student voice work has to look beyond satisfaction averages and ask whether students feel able to speak, be understood, and act on what they hear in return.
References
[Paper Source]: Shulin Yu "When feedback is not conducive to learning: a qualitative inquiry into second language doctoral students' experiences of the negative impact of writing feedback" DOI: 10.1080/13562517.2026.2643830
Request a walkthrough
Book a free Student Voice Analytics demo
See all-comment coverage, sector benchmarks, and reporting designed for OfS quality and NSS requirements.
- All-comment coverage with HE-tuned taxonomy and sentiment.
- Versioned outputs with TEF-ready reporting.
- Benchmarks and BI-ready exports for boards and Senate.
UK-hosted · No public LLM APIs · Same-day turnaround