How to let a PhD student know that their research is not as good as they think it is

Question

I am a postdoc and I have been talking to a PhD student about his research, as his interests are closer to mine than to his advisor's. He is generally nice, smart, and hard working. However, I think he has an unrealistic opinion of his own work.

He has written a paper and wants to submit it to the top conference in my field. (I'm in computer science, so we publish in conferences, not journals.) Unfortunately, I don't think his paper has a chance of getting in. His paper would be a long shot even for a second- or third-tier conference. His paper is also dismissive of prior work, which may offend the authors of that prior work (who are likely to be reviewing his paper).

What should I do about this?

It would be helpful to him if someone gave him a reality check and told him to rewrite much of his paper. However, it is not really my responsibility (or place) to do this. (I'm only two years more senior than him.) If he does submit, he will get a reality check from the reviews, but it would be good if he got some more "gentle" or "constructive" feedback.

I have limited experience with advising. Perhaps someone more experienced can tell me what would be appropriate in this situation. I feel like more than a subtle hint is needed.

Answer 1

Let's start with the obvious:

I'm only two years more senior than him.

You have a PhD and he has not. You have written (many) papers before and he has not. You know your craft. He does not (yet). And he can learn faster by listening to well-meaning advice from more senior and experienced people. The age-difference does not mean anything. And it seems that you care for his well-being (and his success) and that probably makes your advice even more valuable than the advice of a more experienced, yet indifferent person.

Given that you are qualified to give this advice, you should absolutely tell him the honest truth. Start by:

Would you like to hear my true opinion of your paper? You might not like it but I think you should hear it. I think it is better to hear it from me than some random reviewer.

And then wait for his answer. If the answer is "No, I do not want to hear your opinion and my paper is so much better than most of the crap they publish in SODA, VLDB or..." then let him know the hard way. Nothing beats delusions better than three strong rejects and merciless bashing from the reviewers. You should also probably let his advisor know what you think of the paper and then let the advisor decide on how to proceed.

Otherwise, if the student is genuinely interested in your opinion then take the time to explain not only what is wrong with his paper but how he can make it better. Explain why he should be more diplomatic in his view of previous work. How he can present his experiments better. How to further clarify his contribution in comparison to the previous works, and so-on. But also state that even if he implements all that, his paper might still not have a chance in first-tier conference.

Of course, my answer is based on the fact that you do not directly supervise this student. If you do, he has no choice but to hear your opinion. And in this case, crap or half-baked papers should not go to prominent conferences. That makes everyone (co-authors, advisors) look bad. Of course rejection is something normal but only when you submit something that actually stands a chance to be accepted.

Answer 2

He has written a paper and wants to submit it to the top conference in my field...Unfortunately, I don't think his paper has a chance of getting in.

This is not a big deal. I would offer a mild suggestion that a lower tier conference might be better, but ultimately it is fine for him to learn this lesson by experience.

Also, in my field anyway, long shot submissions are very common, even by experienced researchers. Quite often the professors I work with want to try submitting (not very exciting) results to the top journal in our field, then gradually go down the list until someone accepts the work. I kind of think "here we go again; this is a waste of time" but it is a very common approach.

And you never know what a top conference or journal is going to accept; the review process can be somewhat arbitrary. So, it can pay off occasionally.

His paper is also dismissive of prior work, which may offend the authors of that prior work (who are likely to be reviewing his paper).

This is a much bigger deal, and I think you ought to point this out clearly and directly. This endangers the success of any submission.

Presumably, you have been given the opportunity to read the paper and comment on it, and therefore you ought to offer any constructive feedback to help him. It doesn't need to be based on having more experience than him or a higher level position; this is what reviewing papers is about. Being able to give and receive critical feedback in a direct, constructive way is an important skill for a researcher, and this will be part of the PhD student's learning (and perhaps yours as well).

Answer 3

It's his advisor's job to set him straight on this, but if his advisor won't do that for whatever reason, you could always bring it up once as friendly advice. Something simple like "By the way, about your paper, you know you shouldn't diss other people's work without a really good reason, right? You'll make enemies. Also, it's a good paper, but X conference is really hard to get into. It might be better suited to Y or Z." Then drop the issue and let him decide for himself.

Answer 4

Let him submit it and let the rejection be the reality check you are talking about. If its as sub-par as you say he will get the reality check, if not, then he gets to publish. win-win?

Answer 5

This is quite a common scenario. The air of confidence (or maybe arrogance) in the first year of PhD is warranted. You can only advise if he is willing to listen or accept criticisms. However, I doubt that he will at this stage.

My approach would be to provide advice only when he/she asks for it. Otherwise, stay back and let the paper go up in flames; maybe it'll be accepted if the planets are lined up?! He/she will then learn from it or not, in which case only he can save himself. It is worth noting that there are many people who learn by trial-and-error. So this student might be in that category.

Answer 6

As others have clearly displayed, you have a host of options before you. In my opinion, there is only one choice (a step, if you will) that absolutely cannot be wrong:

Talk to the student's adviser.

My heart grows warm by reading all of the answers/responses that heavily support you and your opinion; they encourage the notion that you know what you're talking about as opposed to blatantly discrediting you. I appreciate that.

But the fact of the matter is, just as you wrote, "I think he has an unrealistic opinion of his own work" there may exist a possibility that you have "an unrealistic opinion of his" work. You wouldn't be the first person who had seniority over another and felt that the other person's work was garbage yet turned out to be wrong. In fact, there have been collective groups of people who have shown dissatisfaction or worse towards a singular person's work yet that group ended up being on the wrong side of history. That's not to say that you should lose confidence in your ability to judge another person's research; it's simply a wake-up call to the fact that we can all be wrong sometimes. I am suspicious of your authority to call out a PhD student's work when taking into account that it doesn't seem to be likely that you're that much more experienced than him (because here you are, getting the guidance of others which is an admirable decision but reveals your lack of experience in judging others' work and having to give feedback on that work).

If I were you, take the safer route. Instead of needlessly risking not only the health of your relationship with this student but also risking the confidence of said student, consider approaching his adviser. At the very least, talk this over with more experienced individuals than yourself before you take it upon yourself to be the bearer of bad news. Like I said, there is the possibility that his work is good and you're simply wrong. There's no shame in that but there is shame to be had if you don't consider that possibility.

I also can't help but feel that you really shouldn't say anything because it's simply not your responsibility. Unless he directly asked you for your opinion or you're put in a position to give it (like being an adviser or experienced and respected postdoc), you probably should just talk to the student's adviser. Pointing out gross research errors is always welcome. Correcting flow or grammar of a research paper is no big deal either. But essentially calling his entire paper garbage is a big deal that warrants the experienced eye of someone at the adviser level.

Answer 7

The age difference does not matter. You can supervise students older than you. When you are the supervisor, you are the person entitled to make a judgement.

They do not believe you? Let them submit, but make clear your name does not go on the paper, you may need to become emphatic on that (lest they try to put your name on the paper clandestinely); often that already sends a message. If it doesn't, they will get the feedback they are asking for from the reviewers.

Overconfident students can become real trouble down the line. This is best handled early on if you feel they fall into this category. As they develop the work, you make clear what the level is that you expect for a mark/publication/degree and give top-quality examples. Most overconfident students are quickly grounded by that; if not, the students may either be actually really capable (and their self-confidence justified, thus not "over"confidence; lucky you), or they are really dangerously far from the ability to judge themselves. The latter case is fortunately rare, but should be strongly avoided where possible.

Answer 8

Try not to overthink this. The best education anyone can get is experience, and failure is often the best teacher. As you rightly mentioned it is not your responsibility to take care of this person (it's called co-dependency when you do) and his success or failure is squarely on his/her shoulders. As a friend/colleague you can simply ask if he is interested in a critique of the research and findings. If yes, then be honest. If no, then let it rest. "Gentle" feedback is what a mother gives to 5 year old child.

Answer 9

Let us start from the assumption that a paper, whether presented to a conference or submitted to a journal, must be about new research that has not been published before.

Unfortunately, I don't think his paper has a chance of getting in

If you think so because of mistakes in the paper then address such mistakes and correct them, or have the student correct them by himself. Once those mistakes are solved, then no issue about the submission and the approval should arise.

His paper is also dismissive of prior work, which may offend the authors of that prior work

As long as the paper is correct, that it dismisses prior authors is of not much interest for science.

What should I do about this? It would be helpful to him if someone gave him a reality check

I wonder where his supervisor and perhaps you (no address intended though) have been throughout the entire time the student was writing the paper. Articles do not write themselves alone and if you believe the entire area or topic the paper is dealing with is erroneous, then you (the supervisor) should have warned the student long ago, not to pursue a false line of research.

If the initial assumption, namely that the paper discovers new lines of research, is wrong, then that is not a paper in the first place, by definition, to be presented anywhere, whether good or bad it may be.