Empirical validation is always a hard subject to introduce students to. I like to always frame the discussion by discussing the arguments the two dons of philosophy of science, Thomas Kuhn and Karl Popper had half a century ago. I think your position echoes tenets of both of Kuhn and Popper.
Kuhn would agree with the majority of your last paragraph, Science is all about framing, we adopt paradigms that seem to work well, and bitterly defend them until a "Scientific Revolution" forces us to jump off a sinking ship. In social science, the problem is that its hard to tell if the ship is sinking. A great example in Psychology was this horrific period we had in the 90s where witness testimonies featured victims retelling horrific tales of childhood abuse to their "cognitive therapist", who had helped them unearth their "buried memories". We did not know that "false memories" could be so easily implanted by therapists, and lot of terrible court outcomes came from this before the ship was finally abandoned.
Popper hated Kuhn's perspective. Rather, he contended, science is driven by solving real problems, not defending our perspectives until they collapse. His single most important contribution was the concept of falsifiability. The best theories are falsifiable, they can fail. The easier your theory can fail; the better. My favorite example was Einstein's theory of relativity: he made a prediction at some point about a certain visual aspect of an upcoming eclipse that would only be true, if relativity was correct, meaning, it was extremely easy for him to be wrong, the predicted eclipse just had to look a bit off. He was right. A great theory. Frued's theory, like you mentioned, is one of the worst. It fits like a glove, matching any dream to some sort of inner trait. It is not in any way falsifiable, and therefore, an awful theory. Evolution is another great example, thousands of times, it has had the chance to be falsified, to be just darn wrong, and remarkably, it survives.
Personally, I don't think either perspective is at odds. Kuhn calls it like he sees it, and as far as politics are concerned, science is very cliquish, people who tend to agree, hang around each other until their whole theory falls apart (this issue becomes obviously detrimental when you imagine entire departments hiring those like-minded faculty, or are set with the task of reviewing a grant etc.). And like Popper, I think they really are problems in science, worth solving. Consciousness is a great problem that isn't very falsifiable, but still worth solving. We don't have a clue how it works, but its a legitimate problem, and it shouldn't be abandoned.