20

I am a PhD student, and last year, as a TA I helped grade/supervise a Bachelor's student's thesis. I now notice that I provided her with some wrong suggestions due to which her analysis is also wrong.

The student had approached me with some questions about statistics asking if she should do an ANOVA or a regression. I said that they are the same analysis if you dummy code your variables in your regression analysis (which is true).

However, I only now realize that her specific design (within-subjects) makes the regression analysis not the right choice.

Two other second graders also missed this while grading her thesis, and she got a good grade for it.

I am not sure what to do. This was last year and she has already received her degree. I am very afraid that I will not be allowed to continue my PhD if I inform people about this. However, I am also feeling guilty about the fact that this student learned something wrong and might use this thesis as a writing sample for future admissions.

What would be the ethical/right/pragmatic thing to do here?

Improve this question
7
  • 96
    Just passing by to say that the fear encapsulated in the sentence "I am very afraid that I will not be allowed to continue my PhD if I inform people about this" has absolutely no grounding in reality. Congratulations on having found an issue other people overlooked. Undergraduate theses are regularly rife with errors, and almost none of them are ever caught. It would be awesome of you to bring this up with the primary professor supervising the thesis, but don't be surprised if no one else is particularly interested in addressing this issue and they perceive it as too minor to address.
    – John Madden
    Commented Jan 12 at 16:30
  • 6
    If the thesis will be published as a journal article, this mistake might get fixed there. I have seen PhD dissertations get edits like that before being published (sometimes caught by coauthors and sometimes caught by journal peer reviewers). Thus, another example of similar events occurring and a reason not to worry.
    – Richard Erickson
    Commented Jan 12 at 17:35
  • 3
    "her specific design (within-subjects) makes the regression analysis not the right choice" Do you mean it’s the wrong choice, as in the results are not correct/reliable, or a suboptimal choice, as in the results are slow/imprecise/etc?
    – MisterMiyagi
    Commented Jan 13 at 7:03
  • 1
    At the undergraduate level, you kind of realize some of the professors just teach straight-up false stuff... I recently got a degree after being in the workforce for 15 years with a college diploma. Having obscure database table names is not security at all, but I knew what was expected and didn't raise a stink about it.
    – Nelson
    Commented Jan 13 at 13:52
  • 17
    Statistician here. Believe me, while this modeling approach is indeed wrong, it is absolutely no big deal. I could point you to worse howlers that are regularly published in high ranking journals. Kudos to you for actually thinking about this stuff; too many people just write down the first few numbers that come out of SPSS. I agree with everyone else here that this is absolutely not something to worry about - neither for you, nor for the student.
    – Stephan Kolassa
    Commented Jan 13 at 14:36

6 Answers 6

Reset to default
79

Assuming that her outcome was good ("... she got a good grade...") I don't think there is anything you should do other than, perhaps making a personal apology to her for a flawed analysis.

People make mistakes. It is good that you recognize yours, but you don't need to parade in ashes and sackcloth over something from the past that didn't result in harm to others. Let it go.

Improve this answer
8
  • 20
    +1 for framing this as a personal apology if you send anything at all. Though I wouldn't send anything at all... it's just not going to have any practical effect other than potentially making them unhappy.
    – user541686
    Commented Jan 13 at 8:41
  • 8
    I fail to see what benefit there would be in sending this poor student an apology for not spotting their shoddy work, a year after their graduation.
    – Valorum
    Commented Jan 13 at 11:58
  • 14
    @Valorum you alert them that there's something wrong in the thesis, so that they might not repeat the same analysis again. If nothing else, they learn something about the analysis method.
    – Allure
    Commented Jan 13 at 12:34
  • 17
    @Valorum informing the student will correct their misconception. If I had learned in university that NaCl tastes sweet, I would appreciate someone telling me that no, actually, that was a mistake they had made and in fact NaCl tastes salty. University is (or should be) about knowledge first and degrees second. Correcting flawed knowledge is every (decent) teacher's end goal, so correcting the student's flawed knowledge is good for the student and teacher alike.
    – terdon
    Commented Jan 13 at 15:52
  • 7
    @Valorum didn't deserve their degree!? Who said anything like that? We're only talking about a letter to correct a misunderstanding so the student doesn't spend their life believing something wrong.
    – terdon
    Commented Jan 13 at 16:14
22

This was last year and she has already received her degree

So the problem is over, nobody cares about the contents of a BA anyway.

I am very afraid that I will not be allowed to continue my PhD if I inform people about this

Nobody cares about that either. You made a mistake on a minuscule, microscopic thing whose half-life time is about 1 minute.

However, I am also feeling guilty about the fact that this student learned something wrong and might use this thesis as a writing sample for future admissions.

Yes, she made a mistake. But nobody cares. She will use her BA to get a job where she will not use the regression formula you mention, or if she uses it someone will tell her that this is wrong. She may ponder about a minute about the fact that she did a mistake in her BA and hope that this will be forgotten.

And it will be forgotten.

Really, nobody cares about BAs or MSc thesis. Do not overthink it and move ahead.

Improve this answer
4
  • 10
    "nobody cares about the contents of a BA anyway." <- She cares; that's her understanding of the subject matter; the readers/graderds of her thesis have been influenced somewhat by her choice; any other students she had discussed the matter with are also influenced. So, -1.
    – einpoklum
    Commented Jan 14 at 12:41
  • 3
    @einpoklum she does not care, a BA thesis is the chore you have to do to get your BA. Do you know people who keep a copy of their BA? Discuss it with others? Her understanding of the matter subject is almost nil - she will learn everything in her job. Get real.
    – WoJ
    Commented Jan 14 at 12:54
  • 12
    "she does not care" <- 1. You don't know that and 2. She might care when told about it. "BA thesis is the chore" <- it is an important part of your education as a Bachelor. "Do you know people who keep a copy of their BA?" <- the thesis or the degree? Anyway, of course I do, for both. "Her understanding is almost nil" <- I doubt that's the case.
    – einpoklum
    Commented Jan 14 at 12:57
  • I think it is not very common that a BA thesis is written on a level (content-wise and language-wise) to be comparable to a published paper. In my experience most people who do a PhD thesis are a bit embarrassed about their BA thesis, since it (naturally) has way lower quality than their writing later. This is totally normal. A BA is a chance to practice working on a bigger writing project. It will contain some mistakes.
    – Per Alexandersson
    Commented Feb 7 at 20:14
12

No matter what, a student is fully responsible of their thesis content. You did a bad supervising job, but what the student did is what the student did.

Back to your problem:

I said that they are the same analysis if you dummy code your variables in your regression analysis (which is true).

However, I only now realize that her specific design (within-subjects) makes the regression analysis not the right choice.

Remove the fears about your reputation, ignore the grades she got. What is left? a public text with a methodological error. Libraries are full of them, and still both libraries and sometimes even those texts are useful. Not particularly useful to chase the +1 that just come to the world.

You can write an email congratulating the student about finishing her Bachelor degree and letting her know that you missed completely the fact that her specific design made regression not the right choice and that "insert other method" would have been the accepted state-of-art method.

Apologize again, offer her to collaborate on doing the state-of-art method, conclude with best wishes and good luck for her future endeavours, send the email.

Then make a note to yourself: always reply to questions "should I" from a student you supervise with the golden answer "why should you?", to avoid ending up in the same situation.

Improve this answer
8
  • 35
    I would also recommend OP to be careful when writing the email so they don't traumatise the student by putting forward the same state of mind of this question. The student should not feel like an impostor because of this mistake.
    – The Doctor
    Commented Jan 12 at 17:12
  • 4
    @TheDoctor - Sending her an apology for not spotting the error on her thesis is selfish in the extreme. Given that there's no realistic possibility of fixing it, I'd spend the rest of my life hating the person who sent me that message when they could have just kept their mouth shut.
    – Valorum
    Commented Jan 13 at 11:48
  • 6
    If I were to send an email to a student in a similar situation, I think I would completely avoid formulating as "There was an error in your thesis" and instead I would say something like "To continue our discussion about the choice between ANOVA and regression...". In other words: "Your thesis was great, here is more food for thought on the topic" and certainly not "Your thesis contained a mistake".
    – Stef
    Commented Jan 13 at 12:07
  • 3
    Oooh: ‘a bad supervising job’ is overstating it – the OP missed a technical detail when acting as one of a team of markers. An apology isn't necessary and, as TheDoctor and Valorum mention, could have a negative effect. If the ex-student ever notices this, they'd learn that even their teachers sometimes nod (gasp!) and (a good outcome) end up with a very proper scepticism about the general reliability of assessment grades.
    – Norman Gray
    Commented Jan 14 at 11:30
  • 1
    @Valorum OP is thinking about the consequence on their own reputation, so a selfish email fits with their point of view.
    – EarlGrey
    Commented Jan 15 at 6:29
8

You can always state the facts and leave sentiment out of it.

"Congratulations on your degree! I've recently revised the advice I've given you on your thesis. When doing [work similar to your BA thesis], regression can produce incorrect results and I recommend ANOVA in the future."

A BA degree isn't invalidated by coursework mistakes, no damage there. No one relies on BA theses as a source of truth. And any future damage, should the student over-rely on incorrect, advice is prevented by a simple correction. Science is about search for the truth, not instant perfection.

Improve this answer
3

As others have already said, you are not in danger of being kicked out of your Ph.D. But unlike some have suggested, I believe it is important for you to rectify the situation - even if the grade is given and the student has moved on.

I suggest you get in touch with the student (e.g. via email) and ask for a phone conversation / voice chat with her. In that chat, explain the situation. Don't use a distressed and apologetic tone; only once you've described how your advice was invalid - then make your apology. Tell her it was important for you to correct the wrong impression/conception you conveyed last year.

Improve this answer
1

Main thing is that not a note will grade someone, but a lot of notes and a lot of decisions. Some of them will be unavoidably false, in both directions. Idea is that the final result, based on many decisions like yours, will roughly correlate to the real value of the student; and the law of the large numbers is a strong supporter of this idea.

It does not mean that you would not need to invest all reasonable effort to give a fair grade, but it does mean that single cases are not so bad as they seem.

There is also an asymmetry in the scenario: Regretting a deviation into the positive direction has a clearly hostile favor, while doing the opposite is not. That is because the interests of the University, or the collective interests of the education system of your country, or the collective insterests of the people having a degree, these are rational things, but not persons with feelings. The guy whose note you want to degrade retroactively, yes he is a person with feelings.

Improve this answer
1
  • 1
    Btw, beside all of these: you can be sure that no one will ever check the thesis of that guy. Also no one will ever think about, why you gave that degree. Even if the doubly unthinkable would happen, then no one will be ever sure that you made a mistake or you had a real reason to give that note. The only way for you the get any negative backslash from that, if someone directly above you, checks directly your work, with the explicit goal to find mistakes. Note, if that would happen, this guy would surely find mistakes, had you made that fault or not.
    – peterh
    Commented Jan 14 at 19:47

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .