METHODS / EVIDENCE
A claim is only as strong
as its evidence.
Precise language matters. So does a clear record of what was measured, what was assumed and what remains unknown.
Five distinct stages.
- Proposed: a question and possible mechanism.
- Specified: a mathematical model with assumptions and a falsifiable prediction.
- Verified numerically: a computation checked for convergence and benchmark agreement.
- Measured: a calibrated observation with controls and uncertainty.
- Replicated: an independent team reproduces the defined result.
What every experimental claim needs.
Define the observable, operating domain and failure criterion. Preserve raw data and calibration. Test competing mechanisms. Record energy and momentum channels, uncertainty and exclusions. Make sufficient evidence available for independent scrutiny, publicly or under a suitable agreement.
One successful result does not validate a different program or a universal framework. Confidential work remains unverified externally until it can be inspected.
NASA model and simulation credibility standard ↗Public record status — 24 October 2026
The six updated program briefs now identify prior work and proposed decision tests. No new experimental dataset was produced by this website revision. A reported electrical response is not promoted to a calibrated result without its raw record and controls.
| Record | Present status | What releases the next stage |
|---|---|---|
| H1: loaded energy | Proposed test | Lock device inputs, controls, replicate design and acceptance criteria. |
| M1: photon force | Proposed test | Document actual optical and balance calibration and uncertainty. |
| W1/W2: water and visibility | Proposed test | Specify inlet conditions and separate endpoints. |
| C1: cell function | Proposed test | Select model, intervention and primary functional endpoint. |
| P1: preservation | Proposed test | Select specimen and assay-specific quality criteria. |
| O1: chemical response | Proposed test | Specify chemistry, stimulus and analytical assay. |
A DOI provides a persistent identifier. It does not certify a result, establish peer review or replace independent replication. A registered protocol must link to an actual registry record; none is invented here.