How does SaSame define reproducible MCP research?
Methodology separates measured facts from interpretation and defines what can be reproduced.
- Collection
- research
- Updated
- 2026-07-29
- min read
- 2
- Version
- v1
Research records population, observation window, identity assumptions, instruments, exclusions, transformations, retained evidence and known mismatches.
- Observed
- 2026-07-23 00:00 UTC
- Claim strength
- illustrative
- Dataset scope
- Describes the reproducibility method applied to the Registry-Runtime Gap study — this page is about the method, not a dataset of its own.
- Methodology
- A study binds its dataset version, collection window, instrument and exclusion rules to its published claims. A follow-up reproduction script re-derives each headline figure independently, and any mismatch is recorded on a correction-status page rather than treated as a private engineering task.
- Limitations
- A methodology page describes the process; it does not by itself certify that every current SaSame publication has passed reproduction. See Registry-Runtime Gap's own correction status for the one publication that exists today.
Illustrative — describes a method or schema, not a specific measured result
Purpose
Methodology separates measured facts from interpretation and defines what can be reproduced.
Method
Each study binds methods and dataset versions to publication claims and links them to supporting Evidence records.
Boundary and limitations
A broad scan cannot prove universal behavior; a retained mismatch remains explicit instead of being silently normalized.
"Not currently reproducible" is a recorded status (NOT_REPRODUCIBLE), not a silent omission — see Research Measurements for the one figure in SaSame's current publication that carries this status.
Examples
1- 01
A percentage links to its numerator, denominator, collection time and exclusion rules.