In case of a superficial question, which shows little evidence of research or reasons beyond "would have expected b instead", I see these options:
Flag it as low quality. Assessment: low effort, no benefit for others having similar question, educational effect possible
Give a "near duplicate" by referring to a question which is close enough to provide sufficient hints for an interested reader. Assessment: medium effort, medium benefit for other having similar question, educational effect possible
Adding a comment asking to address the deficiencies. Assessment: medium effort, high benefit for others having similar question (but only if question actually gets improved), educational effect possible
Answer it nevertheless. Assessment: high effort, high benefit for others having similar question, no educational effect (or even one in the wrong direction), if option (3) is not chosen in addition
I don't like option (1) due to lack of benefit, and (4) due to bad balance of effort and education. The comments to this question make me wonder, whether (2) is really an option on this sub-site. In contrast to programming questions and in absence of a mathematical equivalence proof, some (possibly minor) differences may always exist. So is (3) the only remaining choice, if not willing to invest the effort of (4)?