4 ms·
Take for instance, a diet. It's incredibly difficult to assess compliance with a diet (especially when the diet does not call for precise to-the-gram measuremen
by norswap 6y ago
Take for instance, a diet. It's incredibly difficult to assess compliance with a diet (especially when the diet does not call for precise to-the-gram measurement of ingredients). The best way to make this work is to feed the subjects in the lab (and pray that they don't cheat!). This is expensive, and it's difficult to recruit subject for such constrictive experimental parameters.
Also it goes without saying, but statistical tools won't save you if your data is garbage.
- ehsanu1 6y agoAdd to that the fact that biology is mindbogglingly complicated. Different genetics, epigenetics, microbiome, food supplier, or other environmental factors could cause completely different effects with the same exact intervention. It's hard or impossible to isolate individual effects in such a large multivariate system with very nonlinear interactions.