Scientific domain
Quantum Sensing and Measurement
Programmes pushing sensitivity limits where the measurement chain, not the phenomenon, usually determines the result. Pre-registration and noise characterisation carry more weight here than any single dataset.
Typical research questions
- Is the claimed sensitivity supported by an independent noise characterisation?
- Was the detection criterion fixed before the data were seen?
- How much multiple-comparison exposure does the analysis carry?
- Is the calibration chain traceable end to end?
Common confounders
- Noise floor drift over acquisition windows
- Post-hoc detection thresholds
- Multiple-comparison exposure
- Cross-talk between channels
- Reference and clock instability
- Temperature and magnetic environment coupling
What the framework can help determine
- Whether analysis matched pre-registration
- The false-positive budget carried by the reported result
- Whether blinded and shuffled controls support the detection
- Which sensitivity claims are traceable
What it cannot claim
- A detection that survives only under post-hoc analysis choices
- A sensitivity figure without a traceable calibration chain
Typical inputs
- Raw time-series with acquisition metadata
- Noise characterisation runs
- Pre-registration document if one exists
- Calibration and traceability records
Expected outputs
- Detection report with false-positive budget
- Pre-registration conformance findings
- Model card for any simulation used
- Replication packet