KX-11 · Analysis
Weak-Signal Detection Layer
Characterise low-amplitude signals without inflating them into findings.
Data-SupportedContradicted
Applies pre-registered detection criteria, blinded analysis, and false-positive budgeting to signals near the noise floor.
Institute module developed within the KRYOS v6 framework ecosystem. Established scientific methods retain their conventional attribution.
When to use
When effect size approaches instrument noise or sampling limits.
When not to use
As a way to recover a result from data that failed its pre-registered analysis.
Required inputs
- Raw data
- Noise characterisation
- Pre-registered detection criteria
Optional inputs
- Blinded control datasets
- Environmental monitoring
Assumptions and constraints
- Multiple-comparison exposure is counted and reported.
Process steps
- 01Characterise the noise floor and its drift.
- 02Apply the pre-registered detection criterion.
- 03Run blinded and shuffled controls.
- 04Budget and report false-positive exposure.
- 05State the detection outcome including negative results.
Human review gates
- Statistical reviewer confirms the analysis matches pre-registration.
Primary output — Detection report with false-positive budget and negative results retained.
- noise_characterisation, detection_criterion
- blinded_control_results[], shuffle_test_results[]
- false_positive_budget, outcome
Failure signals
- Signal survives only under post-hoc analysis choices.
Quarantine triggers
- Analysis criterion changed after seeing the data without disclosure.