Home > FAQ > How should a team maintain a source-recovery log for incomplete coauthor references?

How should a team maintain a source-recovery log for incomplete coauthor references?

October 26, 2026
source recovery logincomplete referencescoauthor collaborationcitation verificationreference tracking

A research team should keep one shared source-recovery log in which every incomplete reference gets a row with fixed fields, a named owner, and a status that ends in resolved, rejected with a reason, or assigned for follow-up. The log turns scattered coauthor fragments into a single queue, so two people do not chase the same citation.

Fields to capture for each fragment

  • Fragment as supplied, verbatim, plus who supplied it and when.
  • Search routes tried: which database or index, which query terms, and which filters were applied.
  • Candidate identifiers: DOI, PMID, ISBN, or other record numbers for each candidate match.
  • Field mismatches: which supplied details disagree with the candidate record, such as year, author order, title wording, or journal.
  • Verification status: unresolved, candidate found, verified against the original source, or rejected.
  • Owner and next action with a date.

Working the queue

  1. Enter the fragment verbatim and assign one owner; the expected result is a single row that no other coauthor needs to duplicate.
  2. Run a bibliographic query and record the route and filters used; Crossref documents queries, filters, result selection, and cursor paging for narrowing scholarly metadata, so note which combination you tried before changing it.
  3. Record each candidate identifier and the specific field mismatches; the decision point is whether a candidate matches enough supplied details to verify or must stay unresolved.
  4. Verify the match against the original source, as ICMJE advises authors to verify references against original sources and cite direct sources where possible; the expected result is a status change to verified or a documented rejection reason.
  5. Close the row as resolved, rejected with a reason, or assigned with a next action and owner.

Preventing duplicate work

Check the log for an existing row before starting a search, and merge duplicate candidate records where your reference manager supports it; Zotero identifies potential duplicate items and provides a merge interface that preserves collections and tags while combining records. When a candidate remains unresolved, keep the row open with the routes already tried so the next searcher starts from the recorded evidence rather than repeating it.

←
PreviousHow should a systematic review report a search that was rerun after the original cutoff date?
NextHow should authors cite a conference abstract when no full paper was ever published?
→