For comparing both levels of IV of different ethnic groups then a Wilcoxon Signed rank test or better still normalize the data structure and apply a Paired- Samples t-Test.
In the caption you said 2DVs, in the detailed info you said DV . In case of 1 DV Well, a three-way ANOVAwill be needed , to see the contribution of each and combination effect of the three.Which ever way, you can regress all the three IVs on each DV or use simple correlation.
In case of 2DVs, then use each IV on the two DVs and work with MANOVA.