EVIDENCE CONTROL

Evidence-to-claim map

The map prevents a source from being treated as proof of a proposition it does not directly test.

Current position: the structured literature review is not yet complete. Entries below define the evidence needed; they are not findings.

Claim	Evidence required	Competing explanation	Current assessment
P01 Family-level continuity	Longitudinal, multi-informant and within-family evidence with credible controls	Shared environment, socioeconomic persistence, selective memory	Review required
P02 Context moderation	Comparative evidence across institutions, periods or social conditions	Measurement differences, cohort effects, unobserved confounding	Review required
P03 Interruption and redirection	Prospective evidence around identifiable changes or disruptions	Regression to the mean, maturation, unrelated life events	Review required
P04 AI cognitive environment	Longitudinal and experimental evidence on skill use, delegation and verification	Education, workplace design, novelty and self-selection	Open research question
P05 Societal aggregation	Multi-level evidence linking individual mechanisms to institutional change	Top-down policy, economics, demography and historical contingency	Review required

How entries become assessable

Record a narrow claim and proposition ID.
Add a stable citation and source type.
Extract method, population, result and limitation.
State whether the source supports, challenges or is neutral.
Record an independent confidence judgement with reasons.
Update the proposition assessment only after contrary evidence is also reviewed.