This page lists every piece of original research this site publishes. Each entry names its dataset, its method, and its limits. Anything not published yet is listed with the reason it is withheld.
Research here means data this site collected, coded and published itself, not summaries of other people's studies. The rules the work follows are on the editorial policy page, and the instrument behind the site is documented in full on the methodology page.
Published Studies
How the Data Is Collected
- Audit studies code public pages only: every claim is taken from a tool's own published site, and the full coding table ships with the study.
- The free scan runs on your device and sends nothing, so this site holds no database of faces, photos or scores to mine.
- No study on this page uses simulated users, purchased panels or invented figures. A number we cannot source does not get published.
What Is Not Published Yet
A score-distribution study, the one that would let this site print percentiles, needs a large set of real, consented scans. That data does not exist yet, so no percentile or confidence band appears anywhere on this site until it does. The same rule holds for a test-retest study of score stability. When either study runs, it will be listed above with its full method.