Counting publications from ORCID: separate works, versions and sources

A research biography says “six journal articles,” and links to an ORCID record with six entries worth inspecting. Before copying that number, decide what counts as a journal article and check whether the entries represent six different works. A second source for the same paper adds information about that paper; it does not add another publication.

This is a method for checking a particular publication-count claim. If you have not established whose ORCID record it is, first use the namesake comparison workflow. Counting carefully on the wrong person’s record still produces the wrong answer.

Write the counting rule before opening the list

Replace “publications” with a testable scope: for example, “distinct journal articles published from 2022 through 2025, credited to this researcher.” Decide whether reviews, corrections or other article categories belong in your total. If the original claim says “peer-reviewed,” checking that qualification requires evidence about the relevant publication and its review process; a row appearing on ORCID does not settle it.

ORCID’s supported work types include journal articles, preprints, books and other research outputs. A mixed Works list therefore needs classification before it can answer a narrower article-count question. Record the public observation date too: your result describes what you could inspect, not an exhaustive lifetime bibliography.

Keep three separate rows of reasoning

For each candidate, distinguish the source record you read, the underlying work it describes and the decision to include it. These are different bookkeeping jobs. Two records can describe one eligible work; two different works can both be outside your chosen scope.

ORCID explains that a work can arrive from several sources and be grouped using identifiers and their relationships. A group’s preferred source determines the displayed version. Inspect the available versions instead of treating that display choice as an independent quality assessment.

Read what an identifier identifies. A DOI for the article itself is useful for matching that article. A shared identifier marked “part of” may refer to a larger container; it is not a reason to collapse every contribution in that container. For an entry without a usable identifier, compare the original publication’s title, author list, venue and date. Similar titles alone do not justify merging records.

Six rows to review, one confirmed journal article

Invented exercise: all titles, identifiers and records below are teaching material. “Work A” is a label, not a real DOI. Assume the source pages described in the table have been checked for this exercise, and the counting rule includes only journal articles in 2022–2025.

The supported result is “one confirmed distinct journal article in the reviewed material, with one unresolved entry.” It is not “six journal articles,” and it is not “the researcher has only one article.” The table describes a limited set of records. Changing the question to include datasets or chapters would change the inclusion decisions, without changing the underlying evidence.

Investigate the unresolved row without fixing the source

For row E, open the linked publication if one exists and compare it with Work A. ORCID’s metadata documentation identifies fields such as work type, title, publication date and identifiers; manually added entries need not have an identifier. An absent DOI is a matching problem, not a finding that the output is invented.

Save the evidence that resolves the row. If a publisher page connects E to Work A, merge it in your analysis and keep both source references. If it identifies a different eligible article, count that work separately. If it remains ambiguous, preserve the uncertainty. Reviewing a public record does not require editing someone else’s profile.

Also note who supplied each assertion. ORCID defines Source as the person or authorized system that added it. That provenance is useful context, but three imports of one paper still describe one paper.

Attach a count that another editor can reproduce

Use these fields in the evidence section of your report: record URL and observation date; exact counting rule; source row; verified work identifier and relationship; output type and publication date; inclusion decision; unresolved match. Keep the original list alongside your reconciled table.

A useful final sentence names both the result and its boundary: “Within the public records reviewed on the stated date, we confirmed N distinct works meeting this rule; M entries remain unresolved.” That gives the next reviewer something to check, instead of a number whose definition disappeared during copying.