And from that, I don't even know why I'm talking to you...
Anyway, what you've been harping on so far is "Let's do more statistical analysis!! That should make it more scientific-y". Well we now know that isn't the case.
So your "hypothesis" is now "OPPOSITES ATTRACT!!!". Well, sure. But you know, sometimes they don't. Sometimes they conflict. Sometimes they annoy each other. Sometimes they lose interest. Sometimes they hate each other.
So it seems more likely that: Opposites attract, if they have something to offer to each other, that the other person is seeking and/or is lacking.
And you might say that's the basis for why F/T, S/N attract.
But people aren't 100% feeling or 100% thinking. Some people are more 50%/50%. Also it's not just limited to areas of F/T/S/N. It could be about a bunch of other stuff. It could simply be gender (biological or psychological).
Or you could also say similarities attract. People aren't necessarily attracted to things that are the complete opposite of them. There'd also have to be some similarities.
Then there are certain things that people are more or less universally attracted to. Then there are cultural and social influences... environmental influences... cognitive influences... personal values and philosophies... upbringings...
Needless to say, it's complicated. And so you're going to be needing bigger frameworks than what Socionics is offering, in order to explain them.