Would hugely appreciate a little guidance

So I have a cohort of 35 patients and I have a measurement of chemical at artery (A) and vein (V).

The chemical is present in both serum and red blood cells with different distributions. Normal distribution in RBC and non-gaussian in serum.

If I analyse serum and RBC separately the difference from A to V is significant in the RBC and non-significant in the serum

I want to compare the totals RBC+Serum from A to V. 

My question is as follows, How do I do it!!

If I add RBC and serum for each individual the population has non-gaussian distribution, using a paired analysis such as the wilcoxin test shows they are significantly different.

However, is this the correct approach? Or should I be performing some other multiple comparison test of both serum, RBC and total (for A and V) and assuming non -gaussian distribution for the whole analysis??

As you might be able to tell, I am by no means a statistical expert!!

Thank you for your help

Phillip

More Phillip Freeman's questions See All
Similar questions and discussions