Is this a fail?!!

posted
29-May-19, 14:36
Avatar for Dr_Crabby
posted about 5 months ago
My viva is next week and I am having a bit of a meltdown.

To cut a long story short, my only supervisor has been an absolute waste of time for about the last year (final year), however, she is heavy into stats, written books on the subject and is essentially a statistician so when I came to doing my statistical analysis, I took her word for it that the tests she was advising would be correct.

Only, when I had my mock viva last week they picked up on the fact that not only have I used the wrong tests and explained them wrong, if I had carried out the correct tests, I would have generated a number of significant results and supported ALL of my hypothesis. As it stands, I had to say throughout my entire thesis that my predictions were wrong because the results were not significant.

Fast forward to speaking to my supervisor about this, she said there is only one test that I have done wrong and she did tell me not to do that but I misunderstood her, even though I have the email that says what test she was advising and why. But she swears the rest are correct.

Anyway, after ignoring her advice and carrying out what I now think are the correct tests, I have significant results popping up everywhere, supporting my hypothesis (yey) only I have already submitted and written throughout each chapter how the qualitative findings are the opposite of the statistical findings.

I know that it's better to have picked up this before the viva so that I can go in with the correct results but my worry now is that it is going to be a complete revise and resubmit with a second viva and everything because of the extent of the corrections that will need to be made (i.e. fixing every single findings chapter, there are 5/6). Has anyone had this before??

p.s. any comments about relying on my supervisor for help with the stats are not needed right now.
posted
29-May-19, 16:07
edited about 4 seconds later
by rewt
Avatar for rewt
posted about 5 months ago
Quote From Dr_Crabby:
I would have generated a number of significant results and supported ALL of my hypothesis.


Honestly, I don't think you will get revise and resubmit. Your statistics are wrong but the hypothesis, qualitative analysis and data are right which counts for something. You have just interpreted the data wrong based of the wrong application of statistics. It sounds like major corrections to me.
posted
29-May-19, 17:34
edited about 2 seconds later
by pm133
Avatar for pm133
posted about 5 months ago
I would be re-writing my thesis to make these changes right now.
That way you are already ahead of the curve in terms of fixing things when you have your viva.
Rewt is right that your analysis of dodgy statistics is worthy of note but none of us can ease your mind over how your external will view that. Already being ahead of the curve and being able to demonstrate that during the viva will definitely help you.

As an aside, I always advise against people blaming their supervisor for mistakes like this. At this stage you are expected to know what you are doing and why. You are expected to be an independent researcher now. You shouldn't be relying on your supervisor for core parts of your thesis (and what you are describing is certainly core). That falls squarely on your shoulders I'm afraid. It sounds like you are owning this now though and are making the changes required.
posted
29-May-19, 18:56
edited about 13 seconds later
Avatar for Dr_Crabby
posted about 5 months ago
As I said, comments about relying on her for help aren't helpful!

My understanding of the stats and everything I had learnt about them was questioned when she told me I had got it all wrong and do this instead, being that she is my supervisor and has a lot of experience with stats, I assumed I had got it wrong and "corrected" it. So yes I am placing some of the blame onto her!!

I am in the process of redoing all the tests and rewriting the stats chapter to take in with me but I have focused a lot of my qual analysis around trying to explain why the stats didn't show any significance which is why I'm panicking that they will want a total resubmission.
posted
29-May-19, 23:33
edited about 57 seconds later
by pm133
Avatar for pm133
posted about 5 months ago
Blaming her is misguided.

It's your PhD, your thesis and your viva defence. You are going to be challenged on what you did and why you did it. Saying that you did it because you assumed your supervisor was correct and that you didn't subsequently and independently verify her advice would be a disastrous strategy. You need to think through exactly what you are going to say when asked about this. The viva is there to check that you have personally done the work and made the crucial decisions. A major slice of your thesis has fallen foul of this.

If you don't see that my pointing this out to you in advance of your viva and warning you what to expect is helpful to you then that would be a real shame.
posted
30-May-19, 11:06
Avatar for Dr_Crabby
posted about 5 months ago
I understand what you're saying, yes I know it is a fuck up on my part and I don't need that pointing out to me.

We went through the stats together and she explained to me how I had got it "wrong" and she made sense at the time so I had my argument in the viva without saying she told me to do that but I can't change any of that now!!

I have fixed the stats and went over and over it all again so that I can go into the viva prepared for it coming up, my question isn't around how much of a fuck up I've made, it's around my chances of getting through the viva without a fail and if you haven't got any advice on that I'd appreciate it if you stop making me feel worse.
posted
30-May-19, 12:39
edited about 4 seconds later
Avatar for Tudor_Queen
posted about 5 months ago
Hi DrCrabby, Just to take a step back - have you fully considered the approach the examiners from the mock viva were suggesting you do, and whether indeed this is "correct" and the others "incorrect"? Is it this black and white? Knowing why you did the tests you did is the most important thing here. Maybe the ones your supervisor had you do are OK actually? Maybe the results are the same anyway (e.g., the mean was higher for x group) - but it makes them significant if you use the other approach? Or - does it change your whole interpretation of the results? I'd be wanting to get down into the detail of what the differences are in the results and interpretation of the results depending on the approach I use.

I think the following:
- if you can say why you did the stats the way you did (i.e., other researchers have done it, x assumptions were met, etc), you will be in a much better position, even if they do have you re do them because they think it wasn't the best approach
- if appropriate (I don't know specifics) this could make for an interesting discussion in your viva - stats are not the black and white thing people think they are - papers are often published saying people have been doing x test when actually y is more appropriate and would yield different results... you could also bring in the whole p value thing - and how there is great debate about how much weight is placed on that
- if you DO get major revisions one way of looking at it is that often the big difference here is the time they are allowing you to re do stuff - if they think re doing all the stats will take a lot of time when you might already have other commitments, then giving you major revisions is doing you a favour as you have more time to complete them (I remember a similar thing being raised before on this forum)

In sum, try stay calm and know your stuff.
posted
30-May-19, 14:06
edited about 5 seconds later
Avatar for Dr_Crabby
posted about 5 months ago
Hi Tudor_Queen

The way the mock examiners explained it is the way I had originally done it and the way that I had understood it. My supervisor told me to do non-parametric tests even though some of the data was normally distributed because she said it didn't make any difference and I was doing non-parametric tests on one variable so keep it the same. She also advised I do a chi-square on variables that are categorical, which I admit could have been a misunderstanding on my part because I get Pearson's and chi-square mixed up, so that was likely my error, although she did call it a chi-square but I should maybe have known that's not what she meant. I don't know.

I suppose I'm more annoyed because it literally has been one thing after another with this whole process and she went awol for a good chunk of last year and didn't read any of my work, however, she did say she had read the stats chapter and it was correct. Other than the stats stuff which is not my strong point, I would have managed fine without her help. I have redone the tests and re-written the chapter and had someone else read it and check it for me but I know when I email it to my supervisor she will say it's all wrong and I'm in a catch 22 whether to go in with stats that she says are right but I don't believe they are or whether to go in against the advice of my only supervisor and potentially have it all wrong and shoot myself in the foot.

I have accepted the fact that I'm going to get major revisions and I'm ok with that as long as it's a pass but right now I honestly feel like flinging the whole lot in the bin!!
posted
30-May-19, 14:08
Avatar for Dr_Crabby
posted about 5 months ago
Also, pm133 I apologise for being a bitch, I'm pretty stressed but that's no excuse :)
posted
30-May-19, 14:37
edited about 3 minutes later
Avatar for Tudor_Queen
posted about 5 months ago
OK, well - for the parametic issue - you could argue that the data do not follow a normal distribution (this is rather subjective anyway) or show that there were some outliers, and so you decided to use non parametric tests. The thing is - it shouldn't change the result too much. Yes, non parametric are less powerful, so you might not get the significant result - but the actual pattern of the results (e.g., x was higher for females) should be the same. So it is not as though you have done something so terrible if you conclude a little less strongly than you might had you got the significant results. Actually, it is better than the other way around. For the chi square thing, as far as I know, yes, you would use chi square for categorical variables.

If two sets of academics (i.e., your supervisor and the ones in the mock viva) are in disagreement about how you have done it, then it is possibly one of those slightly fluffy things that could be done either way. I'd take the rewritten chapter to the viva (or be ready to say you're aware of it and have made the changes) but only mention if if they actually raise it. They may be of the same opinion as your supervisor (and for that reason I wouldn't send the rewritten ch to my supervisor at this stage either - you may not end up needing to redo anything).
posted
30-May-19, 14:48
Avatar for Dr_Crabby
posted about 5 months ago
When I highlighted it to my supervisor following the mock she agreed that the chi-square was wrong and it should have been a Pearson's which is where the confusion has probably come in. I'm not so worried about defending it in the viva, I've made some errors, this is why and this is what I've done to fix them. My concern is that they will see that as more than major amendments and want the whole thing rewritten, not just because the stats are potentially wrong but because the rest of the qualitative chapters highlight the original stats e.g. "although the hypothesis have not been supported, the qual findings suggest that xyz".

I'm not entirely sure what counts as major corrections and where the cut off is for that, or does this depend entirely on your examiners?

I just can't get out of my own head right now and I'm already not sleeping worrying about it so by the time the viva actually comes around I'll be half dead lol.
posted
30-May-19, 14:51
edited about 16 minutes later
Avatar for softykitty
posted about 5 months ago
Dr_Crabby I had exactly the same issue with my supervisor, and you cannot only blame everything on yourself. All our research topics are too specific and there is no one else knowing what decision to make than our supervisor. Sometimes we had no choice but believe our supervisors. My supervisor is extremely busy everyday, and he often forgot what advice he gave me a couple of days ago. We had lots of arguments over his advice because most of them were contradicted. He still didn't suggest me to submit my thesis even if I sent him to read for couple of times and changed hundreds of times on his comments. I insisted to submit eventually and he had no right to stop. I am waiting for my viva and I also found lots of mistakes on my thesis. It is hard to define independent research for a PhD student because our supervisor will always more or less steer our wheel. I agree with Tudor_Queen, there is no correct or wrong in the research, it depends on how you defend your results and methodology, I'm sure there must be lots of alternatives to your methodology. As long as the knowledge of your thesis is not fundamentally wrong, it should be OK. Having different opinions is normal in academia, the viva is to test how you react and justify your research topic. No research results are flawless. Original work, clear motivation and objectives, analysis in wider areas of your topic, explaining basis as well as detailed questions, knowing the challenge and discussing future work. These should be the standards on justifying the thesis as doctorate. The viva is also very personal, not all examiners will ask nightmare questions, but do prepare for the worst because that is the pressure to make yourself through the viva.
posted
30-May-19, 14:58
edited about 1 minute later
Avatar for Tudor_Queen
posted about 5 months ago
Quote From Dr_Crabby:
I'm not so worried about defending it in the viva, I've made some errors, this is why and this is what I've done to fix them. My concern is that they will see that as more than major amendments and want the whole thing rewritten, not just because the stats are potentially wrong but because the rest of the qualitative chapters highlight the original stats e.g. "although the hypothesis have not been supported, the qual findings suggest that xyz".

I'm not entirely sure what counts as major corrections and where the cut off is for that, or does this depend entirely on your examiners?

I just can't get out of my own head right now and I'm already not sleeping worrying about it so by the time the viva actually comes around I'll be half dead lol.


I think you just have to wait and see then. It's tough, but you just don't know the outcome until you get there.
posted
30-May-19, 15:02
edited about 2 minutes later
Avatar for Tudor_Queen
posted about 5 months ago
Quote From softykitty:
Dr_Crabby I had exactly the same issue with my supervisor, and you cannot only blame everything on yourself. All our research topics are too specific and there is no one else knowing what decision to make than our supervisor. Sometimes we had no choice but believe our supervisors blindly.


I don't think I've ever believed mine blindly so as to speak. If they're suggesting something and I can't see the rationale for it, then I don't do it (unless it's so petty that it is easier to just go ahead and do it rather than cause a fuss). But maybe that's just me (and quite possibly pm133!).
posted
30-May-19, 15:13
Avatar for softykitty
posted about 5 months ago
Quote From Tudor_Queen:
Quote From softykitty:
Dr_Crabby I had exactly the same issue with my supervisor, and you cannot only blame everything on yourself. All our research topics are too specific and there is no one else knowing what decision to make than our supervisor. Sometimes we had no choice but believe our supervisors blindly.


I don't think I've ever believed mine blindly so as to speak. If they're suggesting something and I can't see the rationale for it, then I don't do it (unless it's so petty that it is easier to just go ahead and do it rather than cause a fuss). But maybe that's just me (and quite possibly pm133!).


Maybe blindly is not appropriate word here

Postgraduate
Forum

Copyright ©2018
All rights reserved

Postgraduate Forum

Masters Degrees

PhD Opportunities

PostgraduateForum is a trading name of FindAUniversity Ltd
FindAUniversity Ltd, 77 Sidney St, Sheffield, S1 4RG, UK. Tel +44 (0) 114 268 4940 Fax: +44 (0) 114 268 5766