And you would be wrong if you think this is exclusively how knowledge in physics develops. One clearest example is antimatter that first came out of Dirac's formulation of QM. No such thing was ever observed experimentally, or even thought of, before his formulation.
Secondly, I really don't understand this "analogy" stuff. Analogies are simply a small subset, and application, or simply some resemblance of some THING. It is NOT the thing in itself. It is a way to convey some parts of an idea to someone who does not know, or have no ability, to know the thing. Pop-sci authors do this all the time. As Integral has pointed out in another string, there is a difference between learning physics and learning ABOUT physics. I would not be quick to make any definite statements about a subject matter if all I have acquired about that subject is simply via pop-sci books, or even conversation with people in that area. If understanding physics is that easy, everyone could do it. The essence of physics often lies in the subtleties in many areas - one does not understand it until one graps these subtleties.
Take note that there is a difference in "analogies" and constructing a "model" that represents the mathematics involved. Even the Bohr-Sommerfeld quantization model that was supposed to represent the atomic orbits came out of classical, mathematically-developed mechanics! It didn't come out of thin air with ambiguous hand-waving arguments. There must be both qualitative and quantitative descriptions. Physics just doesn't say that everything that goes up, must come down. It must also say when and where it will come down!
While "analogies" and "visualization" can help to convey an idea, one should never fool oneself into thinking that these are the whole idea and that one has gotten all there is to know about that idea.