Use both: search arXiv for preprint candidates and Crossref for registered publication metadata, then compare the results to identify the record and its publication type. A partial title alone often fits several records, so the match depends on the author and date clues you already have.
- Run an arXiv fielded search with the known title words in the title field, adding author or date-range terms if you have them. The result is a set of preprint records; note each candidate's title, authors, and submission date.
- Run a Crossref bibliographic query using the same known terms, then narrow the results with filters such as author or publication date. The result is a set of registered publication records with their own titles, authors, and dates.
- Compare the two sets on the clues you hold. A candidate whose title, author, and date align across both searches gives you a matched record; check whether each source labels it as a preprint or as a registered publication.
- If no candidate aligns on all available clues, record the search as unresolved and state which terms and fields you searched, so the gap is documented rather than assumed.
Preprint labeling matters because the same work can appear in preprint and published form. The NIH Preprint Pilot describes how eligible life-science preprints are included in PMC and PubMed and labeled as preprints so readers can distinguish publication status. Apply the same distinction when you compare an arXiv candidate with a Crossref record: identify the publication type each source reports before you cite the work.
For a manuscript, review, or grant citation, keep the matched record with its publication type and the author and date clues that supported the match. When the search stays unresolved, the documented search terms and fields give you and collaborators a clear starting point for the next check.
