Maricianah,
This is a terrific example. Are you going to do this research? Because there is likely a sudden cutoff when school went from being free to not free to free again, you can use the birth cohort information. Seems like you could address a lot of important questions, including questions of intergenerational transmission, since the birth cohorts of women who did not have access to free primary school are now starting to have school age children.
Regarding question 2 above, the point is whether the major scientific questions are: "what is the effect of education on girls' pregnancy rates, psychosocial functioning, etc" or the questions are "what is the effect of free primary school on these outcomes"? I would argue that both ar important but the former question - about education - is more important because there are many other policies we could use to increase educational attainment for girls.
Maria