What actually drives speciation...?
I'm sure RAZD will do a better job than I will but I'll have a go at it:
It used to be thought that only if a population was separated geographically could a speciation event occur. This is called "allopatric" speciation. In this case it might be "driven" by lots of things. Great distance might do it if the genes can not flow easily throughout the range. Note the ring-species case: if part of the ring is killed off (by whatever) then there are two (or more) separated populations. Once that is true the the accumulating changes to the now
separate gene pools will eventually (and this appears to take perhaps, millions of years for mammals) will be completely incompatible. (E.g., lions and tigers have been separated for that kind of time and are nearly completely separate).
We now know that many things can drive speciation. E.g., there is a parasite of (I think) fruit flies (don't remember too well) that changes them so only infected animals can breed with each other. It is an immediate separation of populations. There have been cases (many in plants, fewer in animals) where a chromosome number change separates the populations. This can produce nearly "instantaneous" speciation.
A mutation changing something like a birds song can cause a separation of a population into those without is and those with it.
And so on...
Does this mean the species know which members within their group are better suited for survival and they hangout together insuring a better success rate? Like some elitist group of birds weeding out lesser members, eventually leading to a parent/daughter population split?
No. They don't "know". Some survive some don't.
In the specific case of ring species both daughter species may be well suited to the environment. Let's take a case where there are only 3 sub-species in the "ring".
A in the middle and X on the east and Y on the west. We'll take it that A is the parent. The range of the populations is long and narrow (like around a mountain range). The differences in X and Y from A and from each other maybe small. But if the individuals don't travel far relative to the range then there may be enough separation to stop the population from being exactly like one well-mixed gene pool. So there is a possibility of drift occurring.
If the environment where X and Y live is very different then there will be a high selection pressure on them and they may diverge more rapidly.
Now if the remaining gene flow is disrupted from A to X and A to Y and if X and Y are already different enough on the far side of the ring so they don't interbreed (much or at all) we now have room for new species to arise.
It may be that the disruption is by a environmental change (a big city appearing?) in the home range of the parent sub-species and it goes extinct or it maybe that the range is disrupted enough to reduce the gene flow to very, very low levels.
If that is right, my next question would be, wouldn't that almost guarantee the parent populations extinction, since it's a split of better suited species from their lesser suited kin?
In general, no. There are too many variables for a general answer most times
. It is possible to have a speciation event and have both parent and daughter populations do just fine in the same area. There are a lot of niches available.
Cannot mate or choose not to mate due to selective reproduction?
Both. It doesn't matter. Some animals select a mate on song, color or behavior. If any change enough they may "choose" not to mate but it amounts to being unable to mate really. It is entirely possible that we may be interfertile with chimpanzees but we choose not to mate with them (though I have heard rumours ). If we can't get interested then we can't mate.