We have developed a browser-based research instrument that simultaneously captures speech and semantics, voice and prosody, facial behaviour, and standardized self-report (BFI-44) in a single ~7-minute remote session. We observed a consistent pattern: systematic discrepancies between standardized self-report and simultaneously collected behavioural signals. Rather than treating these as measurement error, we propose they may represent a distinct measurable construct — Behavioural Divergence (BD). We are looking for collaborators with expertise in: - Psychometrics and behavioural assessment - Participant recruitment and ethics approval - Independent statistical analysis We provide: operational research instrument, automated data collection, structured CSV/JSON export, and draft validation framework. Our goal is independent evaluation — not confirmation of our hypothesis. Theoretical framework (preprint): DOI 10.5281/zenodo.21706859 Contact: dmitrij.stolyarov@gmail.com