The explanation of Socionics is pretty clear, it's not telling you some magic is making stuff happen, but there are differences in how the brain, body as an metabolism processes information and these differences has an effect on the quality of relationship between people.
Imo, types are predictions from the model, not the other way around in socionics. Types are also observed empirically by other studies, such as Jung and MBTI. So in a sense the conclusions of Jung/MBTI can be derived from socionic's model. Althrough the model of socionics would not be compatible with the J/P terminology of MBTI. There are however no such incompatibilities with Jung's observations.
Go make some predictions then and test them. Experiment and what not. Come to your own conclusions. I'm not here to tell you if socionics is true or not, nothing can tell you that. If you think it's false, might as well discard it now. Good luck with that.
When I say measurement I mean measurement and not measurement tool. So it doesn't really matter how you come to socionics typing,
if the socionic typing is accurate and the MBTI is accurate then the socionic typing imo is better than the MBTI typing.