I'd like to see the data correlated against religious adherence.
In my experience there is a strong bimodal distribution: one group does it because they don't want their kids exposed to the secular world until they are adults, and the other group is dissatisfied with the quality of the educational facilities available in their area.
In my experience there is a strong bimodal distribution: one group does it because they don't want their kids exposed to the secular world until they are adults, and the other group is dissatisfied with the quality of the educational facilities available in their area.