Need help determining which statistical test to use

#1
I've been struggling to answer the results part for our experimental paper for weeks. I'm working part-time and none of my groupmates have enough experience to switch parts with me. Sadly, I've tried using Mcnemar's test and two-way anova for dependent groups but I'm uncertain that I'm correct. This table attached in the image represents data taken from one group tested at two conditions for touch point threshold, with JND starting from 0-50mm and JUD from 50-0mm. Our purpose for the experiment is to determine which part is most sensitive to touch. I'm guessing the palm is, however I still need to present a statistical test for my experiment paper.
 

hlsmith

Omega Contributor
#3
We can see what others think, but can you have a small sample so the following would be a pretty saturated model:


Y (red numbers) = location (dummy coded), given you don't care about group (JUD or JND). Otherwise you could incorporate group as well as an interaction term between group and location, if you think there is an antagonistic or synergistic effect based on interaction between the two. So try out linear regression and test its assumptions.
 

Miner

TS Contributor
#4
You can use a General Linear Model with Participant as a random factor and Location as a fixed factor. If you use the multiple measurements as replicates rather than using the average, you will have additional degrees of freedom.

Note: there were several errors in the calculation of the mean in your data sheet.
 
#5
Thank you! Will try to solve them using your suggestions as soon as I have the time to do them. Will update and tell you guys how it goes.

Note: yes I was aware of the mean computation error, but I forgot to edit that one when I sent the picture here. Thanks for the input though :D