Submitting Corrections

posted
08-Oct-16, 13:25
edited about 24 seconds later
by Caro 2 star member
Avatar for Caro
posted about 8 months ago
Hello everyone. My corrections are due within the next couple of weeks and I'm getting worried about how to go about submitting them.

A lot of what they had asked me to do ended up being impossible as they wanted me to change a lot of statistics to complex ones but my data wouldn't fit the assumptions of these tests. So a big chunk of what they wanted me to do can't be done. So my first question is how do I explain this in my list of corrections?

Secondly, I asked for an extension for my corrections as I work about 60 hours a week and have in reality only had about a month all together to do my major corrections, but the internal examiner said no. So...I have done most of the structural changes, spend the majority of the time trying to do statistical tests that ended up being impossible and now have got to the point where I was meant to re-write my discussion and have no time to do so! I only have this weekend left and I need to work through my data chapters to get them smoothed out which will take all the time I have. I'll make an attempt to do what they asked for my discussion but I would have liked another 3 or 4 weekends to do it justice. So second question: am I allowed to explain that I would have done more if I could in correction document? I'm worried it will sound like I'm whinging about not getting the extra time!
posted
08-Oct-16, 16:34
by CR1980 1 star member
Avatar for CR1980
posted about 8 months ago
I wrote a letter to accompany my corrections which responded in detail to what had been asked of me.

If you were provided with a written list of corrections, you should use this as a basis to respond, detailing how you have been able to do x and y but for reasons have not been able to do a and b. Has your supervisor given any guidance on this?
posted
08-Oct-16, 18:27
edited about 25 seconds later
Avatar for allyballybee
posted about 8 months ago
Sounds a difficult situation. Are there any equivalent non-parametric tests you can do?

I know some people who wrote up their statistical tests even though assumptions were violated, but explained that there were violations and that care had to be taken in interpreting the results. Can you use robust standard errors so you don't need to assume normality/independence, depending on the test and the violations?

Alternatively, would it work if you outlined the more complex tests, talk about why they may not be reliable and then discuss your original analysis and contrast your results?
posted
08-Oct-16, 19:07
edited about 29 seconds later
by Caro 2 star member
Avatar for Caro
posted about 8 months ago
Thanks CR1980 and allyballybee. My supervisor has only said to do what is in the corrections and list it in the letter. He didn't have any more advice when I asked.

Ally - my internal supervisor was incredibly picky about stats- hence why I had such specific statistical corrections including putting a qq plot, histogram and boxplots for every test I ran into the thesis which seems very extreme and takes up a lot of space but I did it anyway!

The tests they wanted were involving repeated measures for time series analysis but a lot of my data is count data and it was impossible to get it to be normally distributed no matter what transformation I tried. So the tests I ended up with were non-parametric and I couldn't get it to fit the models they wanted, so apart from adding a ridiculous amount of graphs for 'data exploration' I've ended up with the basically the same statistics they weren't happy with in the first place! I'm just going to add explanation of the drawbacks of my approach in the discussion as you suggest!
posted
08-Oct-16, 19:15
edited about 7 seconds later
Avatar for allyballybee
posted about 8 months ago
Yeah I put these plots into my thesis too - though I ran LOTS of tests, so I ended up putting in an example of one that was pretty perfect in terms of normaliy and one that was not so great as illustrations.

The data are what the data are - you can't change it. I am wondering though if you can assume normality if the CLT applies - even if your actual data points are not normally distributed e.g in a t test as long as the underlying distribution is normal, and you have more than 30 observations then the CLT applies and you can assume normality. if you have less than 30 obs then it is best that the actual data points are normally distributed.I am not sure about time series data though so this may not be relevant. Also, it is multivariate normality that the test assumes?

I think that as long as you address what they say and explain why it doesn't work then as long as your explanation is justified and referenced then you can't do anything else.

Postgraduate
Forum

Copyright ©2011
All rights reserved

Postgraduate Forum

Masters Degrees

PhD Opportunities

FindA University Ltd, Sellers Wheel, 151 Arundel Street, Sheffield, S1 2NU, United Kingdom. Tel +44 (0) 114 268 4940 Fax: +44 (0) 114 268 5766