Is it meaningful to do a statistical test for such a question ?

#1
Hello everyone :)

Here is my question : a student of mine is willing to investigate (through electromyography) the strength percentage level of 4 different muscles for 4 different tasks.
So, basically it looks like this :

- Subject A is connected to the electromyography device for muscle 1, muscle 2, muscle 3 and muscle 4 => performs task 1
- Subject A is connected to the electromyography device for muscle 1, muscle 2, muscle 3 and muscle 4 => performs task 2
- Subject A is connected to the electromyography device for muscle 1, muscle 2, muscle 3 and muscle 4 => performs task 3
- Subject A is connected to the electromyography device for muscle 1, muscle 2, muscle 3 and muscle 4 => performs task 4

Then :

- Subject B is connected to the electromyography device for muscle 1, muscle 2, muscle 3 and muscle 4 => performs task 1
- Subject B is connected to the electromyography device for muscle 1, muscle 2, muscle 3 and muscle 4 => performs task 2
- Subject B is connected to the electromyography device for muscle 1, muscle 2, muscle 3 and muscle 4 => performs task 3
- Subject B is connected to the electromyography device for muscle 1, muscle 2, muscle 3 and muscle 4 => performs task 4

Same goes for all 30 subjects, and this allows him to define for example if subject A uses more muscle 1 then muscle 2 for task 1, etc....

He is then willing to check if :
- the different subjects use/do not use the same muscle for a similar task
- each individual tend to mainly use the same muscle, regardless of the task

My question : do you think that this requires a "statistical test" ? Or is it just descriptive statistics ?
If this requires a test...which one would be appropriate ?

Thank you in advance
 

katxt

Well-Known Member
#2
Generally, if there seems to be a conclusion to be drawn just by looking at the descriptive statistics, you can usually do some sort of test to convince yourself (and your supervisor) that what you have seen isn't just a fluke.
What is the sort of data you might see that would lead you to think that "different subjects use/do not use the same muscle for a similar task"?
What is the sort of data you might see that would lead you to think that "each individual tends to mainly use the same muscle, regardless of the task"?
 
#3
Generally, if there seems to be a conclusion to be drawn just by looking at the descriptive statistics, you can usually do some sort of test to convince yourself (and your supervisor) that what you have seen isn't just a fluke.
What is the sort of data you might see that would lead you to think that "different subjects use/do not use the same muscle for a similar task"?
What is the sort of data you might see that would lead you to think that "each individual tends to mainly use the same muscle, regardless of the task"?
Hello,

The only data the student has are the one provided by the electromyography, which are converted into "percentage of maximum strength of each muscle".
Let's say 25 subjects out of 30 mainly use muscle 3 for task 1, then....well....we know that 25/30 subject use that muscle. That's it.

SImilarly, if subject A uses mainly muscle 3 for all tasks, then we know he mainly uses that muscle for all tasks....that's it.

That's not a statistical test tough :)
 

katxt

Well-Known Member
#4
Start slow. I don't know much about physiology or whatever the field is, but this is the sort of question I would start with.
Just looking at muscle1. Do different tasks use muscle1 at the same rate? Is this question of interest? Is the answer obvious? Is the answer important? We can use stats to answer this.
 

AngleWyrm

Active Member
#5
So the measurement of a subject executing a task is four percentages, representing level of use?
For example, Subject 1 performs Task 1 with the outcome measured as {60%, 45%, 11%, 2%}
Untitled.png

If this is an accurate assessment of the situation, then it's possible to calculate the average and standard deviation of each column. So that we have an average muscle activation per subject for each muscle in each task, and how much it differs from subject to subject.
 
Last edited: