The Regeneron Science Talent Search is run by the Society for Science and
asks something narrower than most competitions: not a polished poster, but a
complete case that a young scientist has done real, independent,
significant work and could keep doing it. Every submission is read as a whole —
the research paper and the ten-task application together — and scored on
four equally weighted criteria out of 20.
Tian2 has read the complete submissions of confirmed winners across the entire
"On The Table" range and built an AI-evaluation system calibrated against
the official 4-criterion rubric and 463 scored submissions. That evaluator
is used to critique an application against the exact standard the Society applies —
before it is ever submitted.
01/ 5 pts
Entry Form
The ten application tasks — essays, activities, and the record of who the student is beyond a single project. Read for coherence, not decoration.
02/ 5 pts
Scientific Merit
Is the research sound, rigorous, and correctly reasoned? The paper carries this criterion — method, controls, analysis, and honest limitations.
03/ 5 pts
Student Contribution
What did the student actually do? Evaluators separate genuine independent work from a name added to a lab's ongoing programme.
04/ 5 pts
Scientific Potential
The forward-looking read: does the whole application suggest a person who will keep doing significant science? Voice and trajectory matter here.
