Laboratory position
Preregister panels, denominators, windows, counterfactuals, falsification tests, and null-result criteria.
Research question
Which conditions support attributing a visibility or restitution change to an intervention rather than general growth, noise, or panel change?
Evaluation axes
- Version models, access regimes, queries, and denominators.
- Define a counterfactual and falsification tests.
- Publish null results and non-identifiable conclusions.
Boundaries and non-claims
- A before-and-after comparison does not demonstrate causality.
- A visibility score does not measure fidelity.
- No GEO result is published in this registry.
Authority allocation
gautierdorval.com carries authorship, the canonical definition, and doctrinal status. InferensLab carries only program scope, its public protocol, and the state of its evidence.