I need help

#23
1 1 88.4 87.6
2 1 104.3 143.6
3 1 103.7 140.3
4 1 164.5 165.5
5 1 88.6 113.9
10 1 136.2 119.5
11 1 138.1 143.0
16 1 96.8 80.5
6 2 77.6 68.3
7 2 89.8 86.3
8 2 100.9 109.3
9 2 163.8 134.7
12 2 100.8 83.1
13 2 109.3 87.3
14 2 101.9 122.0
15 2 112.0 74.3
 

Miner

TS Contributor
#27
I performed several different analyses, but do not see a statistically significant difference in any or them. One area that I would recommend investigating because it could hide real effects is the measurement variation. How repeatable are the results when repeated on the same subject? By a different user? In different areas of the same leg?

1612470903034.png
 
Last edited:
#29
Thanks for doing this. Apart from the box plot, I've never come across the other graphs. I did manage to get a p value of .051 when doing an independent t test between means of treatment 1 and 2, so I might just leave it at that.
The thing is, if I find nothing significant it will just have to be reported and discussed.
 
#31
Minimal face to face contact at the moment due to covid so it all had to be done via email. The guy is super busy and takes ages to reply to his emails. I messaged him yesterday but I'm not expecting a reply anytime soon. It's basically a nightmare.
 

noetsi

Fortran must die
#32
Yeah that is a nightmare. My professor was not experienced in or really interested in my method either. But not keeping in contact is not ideal. If you do something and he knows you did is likely better than if you do something and you don't tell him. It sounds like he won't care which method you choose. My concern was that he preferred a specific approach. That could be bad if you choose another. Some professors are really interested in methods and others not.

Given that you collected your data and he did not ask about your method is telling. It means he probably will let you decide on the method. But better to have that in writing.

I lack the expertise of Miner. I do time series and regression only :)
 
#33
Well I appreciate the help guys. He is aware of my method as I had to submit a research proposal which he should have read. This is only undergrad level so ultimately it doesn't matter about the method or whether I get statistically significant results or not. This is more about our ability to see flaws in studies and being able to discuss them as well as being able to critically appraise journal articles so we can make informed decisions.
 

Miner

TS Contributor
#34
This is more about our ability to see flaws in studies and being able to discuss them as well as being able to critically appraise journal articles so we can make informed decisions.
Then I would bring up the issues about a potential placebo effect and measurement variation that I mentioned previously, as well as a potentially under powered study (insufficient sample sizes).
 
Last edited:

noetsi

Fortran must die
#35
Well I appreciate the help guys. He is aware of my method as I had to submit a research proposal which he should have read. This is only undergrad level so ultimately it doesn't matter about the method or whether I get statistically significant results or not. This is more about our ability to see flaws in studies and being able to discuss them as well as being able to critically appraise journal articles so we can make informed decisions.
lol when you said dissertation I thought you were working on your PHD.
 

noetsi

Fortran must die
#37
Generally, at least in the US, a dissertation is used only for a PHD. You may write qualifying exams for other degrees but they are not called PHD's. What school do you go to?

A lot of people on this board have struggled through PHD's thus the assumption.
 

noetsi

Fortran must die
#39
I'm in the UK doing a physical therapy degree. I'm in final year and it's quite common over here to do a dissertation at this stage.
I guessed that you were going to school there. In the US the term dissertation is reserved almost exclusively for PHD's. That caused some confusion before on my part.