Which test should i use ?

#1
Hi,

I'm super bad at stats and i rly need your help.
So here's the pb : i want to compare 2 different ingredients (Ingredient A/ ingredient B). Each ingredient is described by 4 independant quantitative variables.
I would like to know if these 2 ingredients are significantly different or not depending on these 4 variables.

How do i check if my data follows a normal distribution? Which test can i use when i have 4 factors ? I know how it works but with 4 factors i dont know which test use and on which data (can i check normal law on all of my data in 1 time or do i need to separate my data depending on the factor or something??)

What should i do? Which test should i use? I'm completly lost. I was thinking about Student test but i dont think its possible w 4 variables. What about ANOVA? Is it ok? ANOVA w 4 factors?

PLEASE PLEASE PLEASE HELP ME
THANKS TO ANY PERSON WHICH WILL HELP ME
 
#3
Probably just run a t-tes on each of the 4 independant vars, one at a time. If you want to test all 4 at once, ie to test the hypotheis of any difference across all 4 vars, then hotelling t^2 i think is the test. Hotelling's T-squared distribution - Wikipedia

Scatter plot matrix is also a must for this sort of thing.Scatter plot - Wikipedia

hope that helps with your indgredients problem.
Thanks for replying! I will do t-test cuz ive never used or heard about hotelling t^2.
I have one more pb : to run t-test i have to check that my data follows normal law ? How can i do that? On which data? Cuz if i do t-test on each independant var, i only have 3 data per variable per ingredient (3 repetitions done for each analysis) and im not sure its enough to run a test to check normal law?!
 
#4
with n=3, you should just plot, there's no real point in testing. You can seen how it is. You have a small n big p problem it usually called. In such cases 'dimension reduction' can be desirable, ie to compute a sum or averge over the 4 independant variables to get one 'very imporatnt variable'. principle components may be an option, but intuition may function better in choosing an appropriate score.