This is a working draft for Leslie's review. The two competence bars and the calibration protocol are in place; edit freely. It mirrors the Biology certification standard, retuned for the physics bench — once approved, it goes live alongside it.
A demonstration is only worth as much as the judgment behind it. If two guides watch the same student defend the same notebook and reach different verdicts, the credential means nothing. Certification exists to make the verdict reliable — so that "mastered" means the same thing in every room, on every attempt, in front of every certified guide.
Two bars, both required
A certified guide clears two independent competence bars. Neither substitutes for the other.
You can clear, yourself, every demonstration you will assess. A guide who cannot pass the apparatus build-and-defense cannot judge it. You sit the demonstrations before you grade them.
You can run the two-day model, cultivate genuine wonder, score integration on its own line, and apply a rubric as a binary "mastered / not yet" judgment without point-drift.
The certification syllabus
Before a prospective guide may assess this pack, they study and are tested on:
- The eight-unit scope and sequence — the concept spine, each unit's anchor lab, and the integration spokes. You can read the course map as a student would and as an assessor must.
- Every mastery rubric — not just read, but applied to sample work until your verdicts match the cohort standard (see calibration, below).
- The three demonstrations — you sit each one yourself and pass it, then study the demonstration-moment script for how to run it as the assessor.
- The two-day rhythm and pacing model — hold-vs-advance judgment, split-cohort handling, and the concept dependency graph for diagnosing where a student is actually stuck.
- Lab and apparatus safety — secured-mass and projectile safety, apparatus setup and calibration, stable-clamp discipline, and the clear-launch-path baseline for the anchor labs.
Rubric calibration: how two guides reach the same verdict
This is the heart of certification. A rubric on paper does not produce consistent grading — calibrated guides do. The protocol mirrors the proven four-pillar model used in the college program's TA calibration protocol, scaled to a single guide or a small cohort of guides.
The four pillars are a loop, not a one-time checklist. Spot-checks feed drift back into re-calibration, so the standard holds across a whole term instead of slipping a little with every verdict.
| Pillar | What it is |
|---|---|
| ① Calibration session | Before certifying, you grade 20–30 sample student responses per demonstration — a mix of clear passes, clear not-yets, and edge cases — and your verdicts are compared against the standard. Disagreements are surfaced and resolved before you assess a real student. |
| ② Anchor examples | For each demonstration you hold a reference set: one verbatim pass, one verbatim not-yet, and the two or three recurring edge cases with the ruling. You consult it when you hesitate, not when you're certain. |
| ③ Spot-check | Where more than one guide assesses, a coordinator periodically re-judges a sample of each guide's verdicts. Drift — standards quietly relaxing or tightening over a term — surfaces early, while it can still be corrected. |
| ④ Escalation | A genuinely ambiguous response is not adjudicated alone. It goes to a queue decided once, consistently, rather than five different ways by five different guides (or the same guide on five different days). |
Mastery verdicts are binary and atomic: mastered or not yet. There is no partial credit, no "mostly," no splitting the difference. Calibration is what makes that binary honest when applied by a real person under time pressure — without it, "not yet" quietly becomes a 7-out-of-10 and the whole standard collapses into ordinary point-grading with new vocabulary.
Certified guide vs. aide
The line is assessment authority. An aide can supervise a lab, manage materials, support students, and run activities. Only a certified guide renders a mastery verdict — passes or fails a demonstration, signs off that a unit is complete. The credential is specifically the authority to say "mastered," because that word is the entire value of the course.
Re-certification
- Annually, and whenever the rubrics for a unit are revised.
- On drift — if spot-checks show a guide's verdicts have moved away from the cohort standard, re-calibration is required before they resume independent assessment.
- Per pack — certification is subject-specific. A guide certified on Physics re-certifies on the subject competence bar before assessing a different pack, even though the methodology bar carries over.