I have data about a few (subnational) geographical units. For each of these units, I know:
1) the proportion of workers by level of education
2) the average income by level of education
3) the average income for the all the population (across all levels of education)
I do not have access to micro-data, and I realise that this prevents me from taking within-group inequality into account.
That being said, what would be a way to calculate between-group (level of education) inequality, so I can compare its values across units?