Correlation and small sample size

#1
I have data from a research study where we have measured skin temperature of a body part, then used Magnetic Resonance Imaging to measure the thickness of the fat in the layer below the skin of that body part. I only have a small sample as the study was cut short, and wondered if it would be acceptable to use the left and right body parts separately to increase my sample size (then do a simple correlation to explore the relationship between fat thickness and temperature). Our research groups previous studies have shown that there is a significant difference in temperature between left and right, but I guess it is unlikely that they are independent of each other. So would it be wrong to do this? Thank you to anyone who can help.
 

hlsmith

Not a robit
#2
Yes, that would be wrong to do. I would just run a multilevel model with values clustered in patients. A very comparable area of research to look at for reference would by ophthalmology where they cluster eyes in patients.

On a positive note, at least you knew well enough that ignoring dependence would have been suspect.
 
#3
Thank you for your reply.
I knew that this wasn't the right way to proceed.
Opthalmology is a great example, will have a look in the literature.
With my n = 8 I am not sure there is much I can do though.