Setting an experiment with ordered categories as predictors

#1
I am trying to apply a multi-factorial regression analyses on a model with 4 different predictors each of which can have four different weights as levels: 0.1, 0.2, 0.3, 0.4. I run 24 different samples applying permutations of these levels for the four predictors.
predictor 1 : four levels(0.1,0.2,0.3,0.4)
predictor 2 : four levels(0.1,0.2,0.3,0.4)
predictor 3 : four levels(0.1,0.2,0.3,0.4)
predictor 4 : four levels(0.1,0.2,0.3,0.4)
So, I get this model:
Y = B + B1Predictor1 +B2Predictor2+...
However, my response variable is not normally distributed, it looks more as a factor variable showing only three different results across the 24 samples. Based on this, I am not sure, whether I am using the right model for my study.

Can anyone help?
Thank you!
Ines
 

Miner

TS Contributor
#2
It is not necessary for either the predictors or the response variables to be normally distributed. What is important is for the residuals to be normally distributed.