A systematic review of Zotero-powered evidence auditing, demonstrating how metadata validation and claim tracking reveal gaps between citation frequency and factual support.
Zotero stores rich metadata for every item: author, date, publication type, tags, notes, and attachments. Most researchers use these fields for organization. This case shows how to repurpose them for evidence auditing — tagging each source with a strength tier, linking it to a specific manuscript claim, and flagging citations where the source does not match the claim's scope.
The workflow operates in three stages. First, metadata extraction pulls tags, notes, and extra fields from every Zotero item to build a structured evidence index. Second, claim mapping links each in-text citation to a unique claim identifier stored in Zotero's tags. Third, strength scoring evaluates each source against four criteria: study design, sample size, replication status, and metadata completeness.
The audit revealed that 11% of citations pointed to sources that did not directly support the claim they were attached to. These were not weak sources — many were peer-reviewed and highly cited — but they addressed different research questions. A strong source answering the wrong question is no better than no source at all. Zotero's metadata structure makes this gap visible before publication.
To replicate this audit in your own Zotero library, follow these steps:
t1-rct, t2-cohort, t3-opinion) and apply it to every item.CLAIM-01) to the Zotero item's tags for every citation in your manuscript.This process takes approximately 2 hours for a 200-item library and catches citation mismatches that manual review often misses. The structured metadata remains in Zotero for future audits, creating a cumulative evidence quality record.