Research records · Checking a publication timeline

Which date tells you when a research paper was published?

A research article can have a recent Crossref timestamp and still be years old. To check a claim such as “new research published this month,” match the exact article, read the publisher’s online and print dates, and keep those dates separate from the registry’s maintenance history.

Start by writing the event you need to date. “First journal publication online,” “appearance in a printed issue,” and “latest change to the metadata” are three answerable questions. The unqualified question “What is the date?” can quietly mix them.

Match the work before comparing its dates

Copy the article’s title, journal, authors and DOI, the persistent identifier often shown beside the citation. Follow the DOI to the publisher and inspect the article history. A similarly titled preprint, correction or dataset may be a different record. Save the page address and the day you checked it.

For a DOI registered with Crossref, its public record adds a second place to examine the dates. Crossref’s REST API has a single-work route, https://api.crossref.org/works/{DOI}; replace the placeholder with the complete DOI. In the returned record, search for the field names below. This is a reading method, not a live test of a particular paper. A DOI registered elsewhere may need that registration agency’s own record.

Read the label attached to each timestamp

The official field definitions distinguish these events. A registry timestamp is not a publication timestamp simply because it includes a precise time of day.

Crossref’s explanation of date types also makes the source of the assertion clear: publication dates come from the submitting member. Treat them as supplied bibliographic information to reconcile with the publisher’s page, rather than an independent observation of when a reader first accessed the article.

A paper that looks new only in the last row

Invented exercise. The following records describe fictional Paper A. No real journal, author or product was tested.

• The publisher lists acceptance on 20 June 2021.
• Its article page and Crossref record both give online publication as 12 July 2021.
• The print date is September 2021, with no day supplied.
• Crossref registration occurred on 4 May 2024.
• The latest deposit is dated 21 October 2026, and indexing 22 October 2026.

A post calls this “research first published after 1 October 2026.” The online publication evidence contradicts that sentence. The October 2026 timestamps concern the record. They do not move the study’s publication into 2026, and the later registration does not establish that the paper was unavailable before 2024.

Nor can you infer that the study’s results changed in October. Crossref notes that index updates can involve citation counts or relationships. To investigate a substantive change, locate the actual notice or compare identified versions; our PDF version comparison explains how to preserve that comparison.

Keep an incomplete date incomplete

Crossref permits publication dates with only a year, or a year and month; see its bibliographic guidance. In a date-parts array, [[2021, 9]] therefore supports “September 2021.” It does not supply 1 September.

Try changing the exercise’s question to “Was the print issue published before 15 September 2021?” That month-only value cannot answer it. Record the month and mark the day unresolved. A spreadsheet that fills missing days with “01” has made a formatting choice, not discovered evidence.

If the publisher and registry disagree about the same event, retain both values and their labels. Recheck that they describe the same article and format. Until a correction or clearer source resolves the conflict, avoid silently selecting whichever date fits the claim.

Write a date finding someone else can inspect

For Paper A: “The publisher page and the Crossref record reviewed in this exercise state online publication on 12 July 2021. The print date is September 2021, without a specified day. The October 2026 registry activity does not support the claim of first publication that month.”

Attach the exact identifier, source links, field labels, observed values, retrieval date and unresolved differences. Put that finding beside the original claim in the OSINT Jet report template. For one article, this small source comparison is often the complete task; it does not require a paid investigation or a claim about the study’s scientific quality.