It's not so much your argument, as being the implication of what you are saying.

There was some hint of condescension in your language as to this being a lack of ability in one side to (paraphrasing) "get the obvious context", and at the same time attribute this to (I'm assuming) social intelligence, or rather, a lack thereof.

What I'm saying, is that you cannot have it both ways here. If the questionnaire aims to get accurate responses, from everyone, you need accurate questions.

Many people you might think this applies to, are perfectly fine understanding the literal meaning, and also any number of "let's assume the question is asking something else instead"-variations. Not that this even matters, as just by accepting the possible existence of variability in how different groups might "be able to understand the obvious context clues", the way you unify responses in the sense of "answering the same question", is by making questions less ambiguous.

Which brings me back to my comment as to how communication works. Concept - symbols - concept. This is always dependant on overlapping agreement in translations at either end, which also depends on context, explicit and implicit. My only argument, the one that you considered might have been tongue in cheek, is that if you want coherent responses to a question, you are better served by a wording that minimises the need for a shared implicit context.

The specifics of my example, I'm guessing, is what you confuse with the more general point. I'm sure that even tho we disagree as to where to draw the line, the general point is still valid.