[x-factor]

It seems to me like you're not looking at the data in the right way. Looking at simple factors like total unemployement or total per capita GDP isn't going to get it done. I'd suggest taking the analysis deeper. Try looking at economic factors (unemployment, per capita GDP, resources, water, industrialization, etc) cross-referenced by socio-political factors (ethnic group, tribe, region, etc).

As Peregrino suggested, you may find that intrastate conflict is more prevalent where there's a disparity there. Just one example off the top of my head would be the current state of affairs in Iraq which is largely about which ethnic/religious group is in a position to control the countries oil wealth.