I am comparing scores from two tests (new test vs gold standard).
The new test is scored as 1=initial; -2=elementary low; +2=elementary high; 3=advanced.
The gold standard test is scored as 0 through 6 (continuous scale).
I am administering both tests at about the same time (concurrent criterion validity)
My question is: how do I go about comparing the scores for both tests? Would be appropriate to simply use correlation methods (Pearson) or is there an alternative (better) method to compare scores in this case?
I think you might not actually be talking about comparing scores here - presumably you're actually interested in the relationshipbetween scores on the two tests? Given that at least one of your tests has responses that seem to be ordered categories rather than continuous it might be worth using a non-parametric correlation coefficients such as Spearman's rho or Kendall's tau.