> I cannot accurately base those decisions on a methodology that assumes a cost of zero ("no direct schooling costs").

The study measures benefit of education. If you wish to determine whether cost>benefit, join it with research on the cost of education.