writersoul,
I am not talking about some fictional scenarios like the one you painted. I am talking about in the real world today. Statistics are a useful way to analyzing evidence, but if they are contrary to widespread well known anecdotal evidence (even if that evidence has never been formally subject to a study) then the statistics should be rejected. My point is that while generally statistical evidence is superior to anecdotal evidence, that is not always the case.
I am not saying NASI is right about the cause of boy/girl discrepancy in our communities but that such a discrepancy exists is undeniable.