It's easily falsifiable and therefore much harder to definitely call it an observable material concept.

This may be a bit of a nitpick, but you have this backwards. Falsifiability is a prerequisite for any kind of hypothesis to be scientific. If a hypothesis, theory, or model is not falsifiable, what that means is it can never be shown to be wrong (false), and so it is fundamentally not scientific. And in this case, it is the difficulty (the impossibility, even) of falsifying what a person says they're feeling that puts statements like that on shaky ground, scientifically speaking. Having to take someone at their word is not "easily falsifiable," it is unfalsifiable, and that's where the problem lies. If someone says "I feel sad today" then there is virtually no way we can ever prove this statement false: hence it is unfalsifiable. However, given the understanding of that caveat we do scientific studies all the time that involve the subjectivity of a person's experience, even as a focal point. From the efficacy of depression medication to the polling done in order to sell more products/candidates, countless scientific studies still rely on people self reporting their feelings. The subjectivity just has to be recognized and factored in as part of the study.

In short, the unfalsifiability that is inherent in dealing with human experience doesn't suddenly make it impossible to study human experience. We just have to control as best we can for things like bias in self reporting and recognizing and taking measures at eliminating reasons participants may have for saying things that aren't accurate about their experience, and including relevant error margins.