Same answer, different reasons: AI-moderated interviews uncover topics and reasoning that surveys miss.

AI-moderated interviews with 1,953 US adults extract inductive topics that 94% of participants endorse as fair, revealing reasoning differences even when closed-ended answers match.

• Why it matters: Surveys scale but force pre-set responses, so identical answers can hide distinct reasoning; the study tests whether AI interviews recover that missing rationale at population scale.
• What they did: In a pre-registered randomized counterbalanced design, participants completed three AI-moderated interviews plus closed-ended and open-text survey items; topics were inductively extracted and validated with a confirmatory follow-up.
• The result: For AI interview topics, 21–51% had no closed-ended counterpart and coverage was 17–37 percentage points higher than open text; among identical closed-ended responses, reasoning diverged widely in valence and content.

AI summary of the abstract · may contain errors · not medical advice · read the paper.