In the majority of physical experiments doing a blind test is not that helpful, since it will not change the physical data. It can happen that blind testing is

In the majority of physical experiments doing a blind test is not that helpful, since it will not change the physical data. It can happen that blind testing is helpful, but most of the time it is a sign that the protocol or equipment needs to improved. Most of the time bias creeps in at the analysis stage. In my experience. Deciding what data to keep and what to reject, deciding which fitting algorithm is working correctly, etc.