Actually, the disassy and reassy process introduces another variable. In real life experimentation you very often can't control all the variables but one. That's where statistics comes to your assistance. So, if you have (CAUTION MADE-UP NUMBERS HERE!!!) twelve horns of identical build specs and there's a 20% variation amongst these, for whatever the characteristic; and then you take twelve horns in three finishes, four each, and you get 20%, well, it's obvious that the finish isn't doing anything. Even if you get 22% for the different finishes, stat.anal. will certainly tell you that's not significantly different than 20% for those tiny sample sizes and the baseline 20% variation in the population.
The point is, the methods for doing this kind of experimentation and analysis are a hundred years old and well understood. For example, if you want to find out if a new medicine works, you cannot hold constant the variable of the person dosed with it. You can't give the person the med, wait till they die or heal, then turn back time and see what happens without it. Instead you compare two populations one with the med and one without, and you use stat.anal. to tell you if, despite the variability of each population, there's a significant difference.
If there were something there, it would be easy for any number of qualified people to show it. Pseudoscientific experimenty videos make people THINK there's something there, leading to sales revenue, but there isn't.
(bottom of page)
