Transportability
0ConceptWhether metrics and findings from one cohort or context transfer to another. Can results from one population generalize?
Last updated: 2026-09-01Spec: SigRank Standard v1.0 (proposed)
Definition
Whether metrics and findings from one cohort or context transfer to another. Can results from one population generalize?
Inputs
Metric distributions from two or more cohorts/contexts.
Derived variables
Distribution overlap, transfer error rate.
Claim
If metrics are transportable, findings from one cohort should generalize to another (within reasonable bounds).
Test
Compute metrics for cohort A, apply any cohort-A-specific thresholds to cohort B, measure transfer error.
Observable
Transfer error rate, distribution overlap statistics.
Falsifier
High transfer error — findings from one cohort do not generalize to another.
Evidence
No formal transportability studies yet. Proposed validation method.
Limitations
Different cohorts may have fundamentally different operating conditions. Transportability is not all-or-nothing.
Lineage
Psychometric validation methodology, proposed for SigRank. Architecture: mos2es.com/architecture.