An analytic framework that takes multiple comparisons etc. into account and lets you see if any correlations are statistically significant.
Blinding.
Two issues, one of which I did not think of, out of like 20.
EDIT: I suspect, including from my own experience, that many problems can be solved without resorting to advanced statistics. Often by using through experimental procedure instead. Like eliminating a food type for a month then not doing an intervention for a month. Repeat. Trying out medications sounds like it should be done safely. This safety can only be achieved by monitoring vital signs and analyzing them using advanced statistics.
Two issues, one of which I did not think of, out of like 20.
EDIT: I suspect, including from my own experience, that many problems can be solved without resorting to advanced statistics. Often by using through experimental procedure instead. Like eliminating a food type for a month then not doing an intervention for a month. Repeat. Trying out medications sounds like it should be done safely. This safety can only be achieved by monitoring vital signs and analyzing them using advanced statistics.