# Results-blind test-contract review

Study ID / protocol digest:  
Review date / timezone:  
Coordinator and conflicts:  
Account reviewers, relationship, and conflicts:  
Role overlap and limitation:

| Review decision | Required readback |
| --- | --- |
| Level of claim | What phenomenon and implemented mechanism are tested? Is this a benchmark or a theory-discrimination study? What consciousness claim is explicitly excluded? |
| Bridges | Which assumptions connect the theory/account to the implementation and the measurement? What follows if one fails? |
| Fair alternative | What is the strongest competing account? Are observations, tuning, capacity, and training resources appropriately matched? What residual difference remains? |
| Diagnostic prediction | Derive both accounts' quantitative regions on the same estimand. Does each permit an exclusive result? If a theory is silent, record no prediction. |
| Six outcomes | Read back target-only, rival-only, both, neither, inconclusive, and invalid interpretations. Preserve disagreement rather than manufacturing consensus. |
| Sensitivity | Show null, minimum-meaningful-target, and rival calibration, assumptions/ranges, independent units, run count, error/precision targets, Monte Carlo intervals, and report pins. |
| Measurement validity | Inspect known identity swaps, measurement failures, implementation controls, negative controls, and how each failure blocks interpretation. |
| Access boundary | Who can see which outputs? When is analysis code locked and registration archived? Where is access logged? What prevents tuning on reserved data? |
| Replication | Identify reserved runs or an independent implementation, independent inputs, pooling rules, and failure interpretation. Otherwise limit the claim explicitly. |
| Reporting | Commit to all hypotheses, attempts, invalid/inconclusive outcomes, and dated amendments with exposure. Preserve the original interpretation matrix. |

## Perspective acceptance

| Account | Actual reviewer / relationship | Accepts design and all outcome interpretations? | Written evidence / locator / time | Remaining dispute / owner |
| --- | --- | --- | --- | --- |
| | proponent / informed proxy / internal critic | | | |

Acceptance does not mean endorsing a theory. Do not claim Melloni, COGITATE, or a theory proponent reviewed this study unless they actually did. Resolve consequential design/interpretation disputes or narrow the registered claim before freeze.

## Close

Decisions accepted:  
Decisions still open / owner / resolution method:  
Artifacts still missing:  
Scope narrowed or postponed:  
Calibration that must be rerun after changes:
