Highly dependent on the person and appears to be dictated by microbiome composition. Which I found more surprising then perhaps I should have.
Studies involving people are always small. There just isn't anyway to make them huge (because it's so expensive) even for pharma with deep deep pockets, clinical trials involve maybe 200 patients in a round and cost tens of millions of dollars. Each. This is one of the drawbacks to actual human study. Now using an animal model you can get larger sample sizes.
Keep in mind this is why proving causation is so exceedingly tough in biological whole organism studies, just not enough replicates to see through the noise.
Yep... But to me. this is a kind of no-go. Without statistics (together with an appropriate way to determine the error, by the way), we are doomed. But I agree that it is good to start somewhere, and for qualititative statement this may be enough
Not everything can be studied to 5 sigma certainty my friend :)
I am too physics-biased :D
:D a high burden of proof is a good bias to have.