Hmm, I'm glad the FDA doesn't test medications in this manner.
Using a sample size of 1 means it is next to impossible to obtain statistically different results.
I understand there are financial limitations and that according to the MARSH site "... there are ten replicates. Ten batches of salt, one each month. So there is replication for the treatments" ... instead of multiple tanks for each study group.
The idea that a single tank (or person) can be used multiple times to increase the power of the study is simply incorrect. Any source of error (i.e. a rogue/diseased ramet, contamination of a tank, different bacteria levels between tanks, etc., etc.) can potentially be there throughout all 10 months....
For future studies (maybe funding could come from the salt manufacturers) statistically power could be greatly increased by testing 10 instant ocean tanks versus 10 natural sea water tanks, etc.
Another idea is to do a cross-controlled study. Take the best performing salt mix and switch it with the worst... etc. If the worst tank suddenly improves dramatically, its much more likely to be due to the salt, and vice versa.
Kudos for attempting to scientifically study different salt mixes, but this is essentially a case report with n=1 and will have to be taken with a grain of salt.