My main supervisory is a relatively prominent figure in the field, but as I've gotten to know him better over time, I've discovered some rather critical issues. It's common knowledge that articles often have co-authors, and my supervisor publishes both quantitative and qualitative research. However, to my surprise, I've found from some of our conversations that his knowledge of statistics is actually less than mine... and at times, he even provides incorrect advice on fundamental statistical matters. I'd like to ask how I should handle this situation, as it involves some political aspects. If I were to address this directly, for example, in an email, I'm concerned that my main supervisor might feel embarrassed in front of my other supervisor (who is relatively new). But not addressing this issue could have serious consequences, impacting our research publications. I'm unsure about how my main supervisor managed to publish in those top journals.
-
1please add the "because all emails need to go through supervisory panels" you commented somewhere in an answer. It is very peculiar, and very important to know. Where are you based?– EarlGreySep 25 at 20:30
-
hi Earlgrey i'm in the US. I have three supervisors ,in most cases we made any decision based on the agreement of all supervisors.– roroccSep 26 at 3:20
-
1I don't know what field you and your supervisor are in, but unfortunately, top journals (especially in some fields) are not immune (and especially were not in the past) from elementary 1st year undergrad statistics errors.– kejtosSep 26 at 6:32
-
2Have you tried just having a proper discussion on specific issues with them? Like you get some advice from them, and if you think this is problematic and should be done differently (with arguments and references), you just tell them this, maybe with alternative suggestion? Without mentioning any general concern regarding their competence? So that you can keep your own work correct without getting into more general issues with anyone. How much independence do you have to make methodological decisions?– Christian HennigSep 26 at 10:32
-
Am I correct in assuming that your supervisor practices in a field where statistics is used, but the field is not statistics itself? (Because that makes a great deal of difference to the expectation of what knowledge he would have and whether gaps in knowledge relative to a student would be embarrassing.) Can you please add this information (what field are you and your supervisor in) to your question.– BenSep 29 at 0:55
3 Answers
In no particular order:
This is when you should talk to a co-superviser. That's why they're there!
Distinguished academics often have a "bird's-eye", wide-ranging view of the field, while not remembering the specific details of a topic unless they're personally researching it at the moment. Their perspective is invaluable mainly to draw unexpected connections, which is vital for innovative research.
By contrast a PhD student will often develop a deep understanding of their particular topic and methodology, since they focus on it for several years at the very start of their academic career.
A good supervisor will know the above two points well, and will be open to your feedback. Furthermore, you need to establish an open and strong communicative relationship with your supervisor, so you need to know early on whether they can handle this sort of feedback and how you should deliver it.
Especially the first few times, it can be helpful to show a bit of deference. I have found the "A/B" technique useful: do it their way, do it your way, and show the difference. (If there isn't a clear difference, then maybe it's not as big a deal as you're making it!) Showing them that you didn't outright dismiss their advice can help them be receptive to you.
It would seem to me to be that his assets are elsewhere and you should look to him for that other advice. Some people are much better at the big ideas than at detail. Einstein, for example, was said to be a terrible (and dangerous) experimenter. Bright enough overall, of course.
But, you have learned something valuable and can look elsewhere, or within your own knowledge for the statistical detail and guidance you need.
I don't see a need to "address" the issue directly. He can probably speak for himself. If others find collaboration with him valuable, you can too, within the limits you have noticed. Everyone has strengths and weaknesses.
-
Hi Buffy, the issue is that he's my PhD supervisor, and we're working on a dataset related to my PhD dissertation. Without his approval, I cannot proceed to the next step. The advice he gave me is fundamentally wrong, and I must point it out. However, I don't know how to deal with it properly, and I'm not sure if he will 'feel' embarrassed if I point out such basic mistakes...because all emails need to go through supervisory panels which means there’s a possibility he might feel embarrassed in front of his colleagues (who’re much younger than him). ( perhaps I complex things too much?)– roroccSep 25 at 19:21
-
@rorocc we don't know about the context of the question for which you see he is wrong, but I think it is possible to address this issue even within the email to the supervisory panel. Not all the arguments need to be contrastive; you can explain your approach as a potential solution, explain why you think it will yield the results you are all looking for, and still end the argument with a request for feedback. This is a part of training during PhD, to discuss your opinion, and I think your supervisors will eventually appreciate seeing you grow during your PhD. Sep 29 at 8:17
Don't think about it as an issue over the competency of your supervisor. Think about it as a scientific issue. If the methods your supervisor is suggesting are not adequate, you should be able to have a scientific debate about it with your supervisor, preferable in person.
I've had really good and really bad supervisors in the past, but I've never had one that I couldn't debates science and methodology with.
The debate over scientific methodology should never feel like a judgment on someone's competency; it should be over finding the methods that suits the problem the best. If it's done right, everybody would learn from the discussion.
I learn new things from my PhD students all the time. There is nothing wrong with that.
-
Hi, I don’t think it’s a scientific issue. It’s a basic logic issue about statistics. It’s more like an issue: variable X (combined with item A & B ) is positively correlated to Y, but we found A has no correlation with Y when we solely tested it. My supervisor believed A should be deleted from the data set as it has no correlation with Y. First, A has no correction with Y doesn’t tell the truth, it may due to sample size etc.; second, I don’t think it’s ethical to delete data just bc they’re non-significant.– roroccOct 3 at 9:54