A researcher wanted to know if there was a difference in depression ratings in patients taking Lopressor. Depression was measured with an instrument that used a 1 - 5 ranking scale with '5' indicating the most depression. The depression instrument was administered to the same patients, every 6 months for 3 years. Which statistical test would be appropriate for these data?

A. Paired t test

B. Kruskal Wallis

C. Phi

D. ANOVA

E. Chi Square

F. Wilcoxon

G. Spearman rho

H. Person r

I. Mann Whitney Utest

J. Friedman test