Comparing improvement between two groups. (pre vs post)

I have two groups. Each consists of an equal number of subjects. They are given a test. Then they are 'trained' differently. Then the test is repeated. I want to statistically compare the improvement between groups.

My data is not normal (the scores on the tests). I initially used a rank analysis and simply examined each group. Using the paired Wilcoxon signed rank test (nonparametric paired t-test) I assessed the improvement in each group. But now I want to compare the groups.

At first I was super simple and simply calculated the % change for each subject and classified that as a new variable called 'improvement'. Then performed a Mann Whitney U test to compare this %-change between my groups (nonparametric independent samples t-test). But I thought perhaps I shouldn't do this and instead consider a repeated measures type analysis. Granted my nonparametric alternatives weren't vast, but I was familiar with nparLD in R. I set it up as an F1-LD-F1 analysis (with randomization group as my whole plot factor, and time (pre v post) as my subplot factor).

Both ways give me results that I can use :) Is one method preferable over another? Or is there a simpler third option I'm simply overlooking?
Many thanks!