Or what about this experiment actually working, this specific species going extinct (or at least driven out), and that 4% nice that's now empty gets taken over

Or what about this experiment actually working, this specific species going extinct (or at least driven out), and that 4% nice that's now empty gets taken over by another type of mosquitoes or other type of insect that has the bad properties you describe? Not too unlikely as we tend to have no clue what will happen when we eco-engineer something away. And if we think we have a clue, it tends to be wrong. Or what about this: they go extinct/away and it turns out there was some predator that now doesn't get food anymore and is decimated while their competition gets now out of hand? Next pest please.