Khanh:
The regression result does not show the good result because the data-set was unreliably collected. The numbers of children mortality are identical for many districts, especially the neighboring districts. We suppose that people only measured at some villages and assigned this value for the whole district or even for the neighboring districts. The other hypothesis is down scaling from the total or the average value of a large area to the specific values at each district. Thus, it makes the data biased and there is not any correlation with other factors such as the electricity accessibility, the healthcare center distance,...Consequently, the negative correlation of chosen factors cannot explain the children death. It is illogical that enhancing the adverse conditions like the violence, the healthcare center distance, the family size increases evoke the children death decrease.