---
title: "Evidence Library"
group: "Research archives"
order: 1
summary: "Browse Charitology's cited sources and build a private, protocol-bound register of reports and screening decisions."
status: live
conditions: ["measurement-gap", "silence-about-failure"]
help:
  - "Check a source and fix a broken link."
  - "Add a stronger primary source to an entry that cites a secondary one."
related: ["/record/methodology/", "/record/interventions/"]
---

The published bibliography collects sources from Charitology's source-bearing content collections, including journal drafts, and its linked headline figures, country indicators and charity directories. Each reference links back to the pages that cite it. This is the site's bibliography; coverage of the wider literature is unknown.

The study register adds a private workflow for your own question. Select published references or preview a mapped citation CSV, preserve its search provenance, inspect duplicate suggestions and choose which reports to combine. Freeze a local protocol version, record two labeled screening lanes, explain disagreement and keep exclusion or retrieval reasons. Included reports can hold optional extraction inputs with their source, unit and uncertainty. The flow counts come from those records; they do not depend on the active search filter.

The same saved register opens in [Evidence Reviews](/programs/evidence-reviews/). It stays in this browser, with a complete JSON backup, analysis exports and guarded restore controls. Existing narrative extraction notes keep their own storage. A reviewer label or local hash establishes no independent reviewer, registration, performed search or verified result. Nothing is submitted to Charitology automatically.

References start with unknown classification. An organization homepage or a method guide may be useful without being an empirical study. No estimate is pooled, no missing value becomes a zero and no named charity receives a rating.