Hello! I need to show that compositions of chemical mixture before and after treatment are significantly different. Composition is given in weight %, so the total of all components' weight % in the mixture is 100%. This also means that if one component weight % in the mixture decreased after treatment all other component weight % should be increased. Can I use "Test Mean" from the Distribution platform on a ratio like this: Mean of Abs(1-Component_Before/Component_After)? Or should I use Matched Pairs Analysis or something else? Should it be a Non-Parametric test as the analysis errors for each component are different?
