Supervisors & feedback

Lessons from being a supervisor

Why feedback comes out vague or harsh, and what a fairer version looks like — from the other chair.

I have sat in both chairs: the student waiting weeks for comments that then stung, and the supervisor who has, on a bad day, written the kind of note I once dreaded. This is what I learned crossing the desk — mostly about the harm a careless comment does, and how little it costs to do better.

By Heidi Andersén, MD, PhD · Updated 18 July 2026 · 6 min read

“Do you ever feel like you’re a disappointment to your supervisor?” — and, from another thread, “losing confidence due to my supervisor’s harsh criticism.” — two students, describing the same wound

The problemA comment costs the supervisor a minute and the student a fortnight.

When students write about supervision, the recurring word is not confused — it is demoralised. “A disappointment.” “Harsh criticism.” “Even an undergrad would know this.” I used to read those as rare, unlucky pairings. Then I supervised, and understood how a comment I dashed off between clinics could land as a verdict on someone’s worth. The asymmetry is the whole problem: what takes me thirty seconds to write can take a student two weeks to recover from, and sometimes longer to act on.

Why it mattersFeedback is a research instrument. It can be miscalibrated.

I think about feedback the way I think about any instrument in a study: if it produces a signal the receiver can’t use, the reading is noise, however true it is. A comment that is correct but unactionable — “be more critical”, “this isn’t PhD level” — has failed at its one job, which is to change the next draft. And a comment delivered so bluntly that the student can’t work for half a day has actively cost the project time. Harshness is not rigour. It is usually just haste with the safety off.

Common mistakesFour things I’ve done wrong, and see done

Judging the person, not the page. “An undergrad would know this” grades the student. “This paragraph states a claim the design can’t support” grades the work — and can be fixed.
Sitting on a draft for months. Silence is its own message, and the one it sends is “you don’t matter.” A slow answer is rarely worth the wait it cost.
Compressing a real point into a vague one. Expertise makes us abbreviate. “More depth” is a whole diagnosis crushed into two words the student can’t unpack.
Forgetting the power gap. A throwaway line from someone who controls your progression is not thrown away by the person who receives it.

What I actually learnedThree habits that cost nothing

Name the fix, not just the fault. Every criticism I write now has to carry the next action. Not “this is descriptive” but “this is descriptive — add one sentence saying what it means for your question.” If I can’t name the fix, I haven’t finished thinking, and I’ve no business sending the comment yet.

Separate the two feedbacks. There is feedback on the work and reassurance about the person, and students need both, clearly labelled. Mixing them — “this is weak” with no “and you are not” — lets the student read a page-level problem as a self-level one.

Protect the turnaround. A rough comment this week beats a polished one in three months. I would rather send “here are my three biggest concerns, more to follow” on time than a perfect memo the student has stopped waiting for.

Worked exampleThe same note, twice

A comment rewritten

As I might once have written it: “This literature review is superficial. Not PhD level. Be more critical.”

As I write it now: “The review summarises each study well but doesn’t yet weigh them against each other — that’s what pushes it to doctoral level. Pick the three most relevant studies, say where they disagree, and tell me which you find more convincing and why. That paragraph is the ‘critical’ move I’m asking for. The writing itself is clear — this is about the argument, not your ability.”

Same judgement. Same standard. The second one is a task the student can start tomorrow, and it does not cost them a day’s confidence first.

ChecklistBefore I send feedback

Downloadable · supervisor feedback check

The pre-send checklist

Does every criticism name a concrete next action?
Have I graded the work, not the person?
Have I separated “the draft has a problem” from “you are the problem” (you are not)?
Is anything vague (“more depth”) unpacked into what I actually mean?
Am I on time, or is a rough-but-prompt version better than a late perfect one?
If I read this cold, as the student, could I start work in the morning?

Supervisor feedback — before I send [ ] Every criticism names a concrete next action [ ] Graded the WORK, not the person [ ] Separated "the draft has a problem" from "you are the problem" (they are not) [ ] Vague words ("more depth") unpacked [ ] On time — or rough-but-prompt beats late-perfect [ ] Read cold as the student: could they start work in the morning? One line of genuine reassurance included? Y / N Turnaround from receipt to reply: ____ days

Download the checklist (Markdown)

The honest limit. I am one supervisor, in one field, still getting this wrong sometimes — this is what I’ve learned, not a rule for how supervision must be done. If you are the student on the receiving end, the companion guide is the practical one: decode vague supervisor feedback.

Related guidesRead next

There is no product to sell on this page. It is here because the corpus behind these guides is full of students describing supervision as their hardest wound, and the fix is mostly on my side of the desk, not theirs.