Sunil,
First of all, I was not doing a comprehensive analysis of any issue like you attempted, but merely listing the other important issues that Jat people have and should be raised by Jat parties, as Vipin had asked. Nevertheless, thanks for correcting me, the low sex ratio is technically the correct phrase to use as female infanticide is a cause, possibly one of a few.
I'd like to make the following points on your arguments:
1. Your argument that poor have more children than the rich seems correct from observation. But how do you apply that to Haryana? Would you divide this as more children for rural folks versus lesser for urban? If that's the case, then Haryana is only 28% urban. Then a vast majority should have more children. However, these days even rural folks are decreasing the number of children. So it is important to know who are the "poor" in your theory as Haryana doesn't have the poverty to the extent Bimaru states have.
2. What your friend told you was interesting, but was just one anecdote. So we can't derive from there that people don't see girl child as a burden. A boy may be a bigger burden in reality, but how many people think that ways. Personally, I have heard this for the first time. What is beyond debate is that there is an overwhelming desire for male child. Whether that is because of cultural factor X or economic factor Y or social factor Z, is arguable. But the fact is real. Now the question is how many people just desire a boy and how many can live with nothing but a boy. I have no expertise in this field. I don't know many female infanticide cases. But it may be a bit naive to dismiss it as "rare extreme". It can happen at a reasonably early stage of pregnancy and it can always be colored as a mishap. No one would advertise it obviously. So there can be quite a few reasons why we don't see or hear about this.
3. I had almost bought your theory on why the ratio is lower because of the steps you described, till I tested it. By the algorithm you described, one should have exactly same number of boys as girls. Let's divide your algorithm into cases where a couple decides to stop having kids as soon as they have a) 1 boy or 2 kids, b) 1 boy or 3 kids, c) 1 boy or 4 kids, d) 2 boys or 3 kids and e) 2 boys or 4 kids.
The case a) can be split into 3 sub cases with probabilities in bracket- 0 girl 1 boy (50%), 1 girl 1 boy (25%), 2 girl 0 boy (25%). Then you multiply no. of girls in each subcase to the probability of that subcase and sum it across all subcases. Ditto for boys. You get 0.75 for both boys and girls. That is if 1000 families followed this, then you should have got 750 boys and 750 girls.
The case a) can be split into 3 sub cases with probabilities in bracket- 0 girl 1 boy (50%), 1 girl 1 boy (25%), 2 girl 1 boy (12.5%), 3 girl 0 boy (12.5%). Same process as above. For 1000 families, you'd have 875 boys and 875 girls.
Same can be done for other cases.
One reason you may be seeing a different result is because you didn't have any 2 girl or 3 girl situations. If it is strictly algorithmic like above, you should have those situations. In fact, the 1 boy or 3 kids scenario has 12.5% "3 girls" scenarios by probability theory.
As you said, in the real world, you don't see many people with only 2 girls or 3 girls. If it was purely algorithmic (or upto fate/nature/god), you should have seen that. So what gives? That's why I think female infanticide be more of an issue than it seems on the s urface through our "knowledge" of such events.
