a physical chemistry experiment will have 1000 data points per trend line; I organic chemistry will have 10 data points, and biochem will have 2 data points.

There is an element of truth in this, but that one biochem datapoint probably took more money and (wo)manpower than a hundred phys chem datapoints. Which is sad, because biological systems are usually more complex, and therefore more ‘noisy’, needing more datapoints for a definitive result. Medical studies get a lot of datapoints for obvious reasons, and because they can afford to do it thanks to Merck et al.