I am analyzing a psychophysiological experiment.
I ran a model on my response times and accuracy data, which gave 7 parameters [dependent variable] per participant ( n = 12) all depending on one experimental manipulation: Color [Independent variable, 3 categorial levels: green/red/brown].
All 7 parameters depend on the color. I'd like to run a multivariate regression which runs a linear regression for each dependent variable and always the same independent variable color. As I have only 12 observations per color level, I have too few samples and normality cannot be assumed.
I have thought about permutation tests, but do they work with 3 different data sets ( as every parameter has 12 data points per color)?