I don't think the answer is clear since it hasn't been studied to my knowledge. At this low concentration, perhaps the interaction of the alcohol with water is stronger than that with the quat and the interaction of the quat with the record is not disrupted. This assumes no other ions (like salts) in the mixture, which would change everything since water would rather hold hands with the salt ion then the alcohol (this is how one can separate phases). It's a good question. I believe folks add the alcohol to better disperse the rinse solution (lower surface tension) and perhaps aid in drying. I don't know how this effects the bound quat under dilute conditions, but I use just dH2O since I, like you, rinse using a vacuum (VPI RCM in my case) which leaves the surface fairly dry.
EDIT: That said, you can strip the surfactant off a record surface with 95% IPA. You can test this as the surfactant surface will normally hold water and the stripped surface will not. This is why some of us also worry about endogenous lubricants. One study on cleaning artifacts (previously referenced in this thread) shows the threshold is above 50% alcohol before extraction takes place. The ability to hold molecular water is what gives quats their antistatic property so stripping them defeats the purpose of their inclusion in wash solutions. I suppose it would be pretty easy to test what effect 3% IPA has on this. Consider though, that the quat bound to the record in a solution that contained 5-15% IPA (in many recipes) to begin with.