12-10-2014 11:50 PM
What do you expect the coefficients to test? Are they in relation to another variable? If so, what's the type of variable?
If the dependent variable is continuous then you can look at a regression model and using proc glm or proc reg, this assumes that you have don't want to use a non-parametric model.
If the dependent variable is categorical then you can look at chi-square or fishers tests (from proc freq) for starters.