Trust through method

Methods, standards & corrections

This project is designed to make the path from question to claim inspectable. It separates source evidence, interpretation, lived observation and hypotheses so readers can see what is known—and what is not.

Evidence standards

How an evidence map is built

01

Question first

Each synthesis begins with a defined research question, population, context and scope.

02

Traceable search

Search terms, source classes, date ranges and exclusions are recorded for substantive evidence maps.

03

Claims are typed

Established evidence, support, emerging findings, hypotheses and community observations are never presented as interchangeable.

04

Uncertainty stays visible

Conflicting findings, missing populations and methodological limits remain part of the public summary.

05

Living documents

Material changes receive a new update date and a concise change-history entry.

Claim labels

Evidence strength is not publication stage

A paper can be published while its underlying claim remains preliminary. An evidence map can be in active analysis while drawing on well-established findings. For that reason, this project displays two labels: evidence status and research stage.

LabelMeaning
EstablishedConsistent evidence or authoritative consensus in the relevant context.
SupportedMultiple credible sources point in the same direction; meaningful limits remain.
EmergingEarly, mixed, narrow or rapidly developing evidence.
HypothesisA proposed explanatory model that requires testing.
Community observationRecurring lived reports used to generate questions, not to establish causality.

Lived experience

Observation generates questions; it does not replace evidence

Lived experience can reveal patterns that formal research has not yet measured. Community reports will be grouped, moderated and clearly labelled. They may guide prioritisation or survey design, but will not be converted into clinical claims without appropriate evidence.

Use of AI

AI assists retrieval and structure—not final authority

AI may support literature discovery, source organisation, translation drafts and retrieval from a curated collection. Substantive published claims require human review against the underlying sources. Future RAG responses will display citations, collection-update dates and a visible separation between retrieved evidence and model inference.

Ethics & participation

Participation will match the level of research activity

Informal audience polls, community observations and formal human-participant research are different activities. Any formal study will display its purpose, consent process, data use, governance and relevant ethics status before participation begins.

Corrections

A public correction trail

Material corrections will retain the original publication date, add a modification date and explain what changed. The initial correction log is empty because the library is launching now.

No material corrections loggedLast checked 2 August 2026

Privacy

Collect only what the project needs

The public MVP does not collect sensitive research or health data. Substack manages newsletter subscriptions and payments under its own terms. Before member submissions or surveys launch, this page will describe the exact data collected, retention period, access controls and deletion route.