Match Affiliation Text to a Research Organization

Separate the institution from the address

Preserve the original affiliation as printed. In a working field, identify the organization, department, city, and country. For an unstructured affiliation, use the ROR matching workflow appropriate to that input. For a structured organization name and location, compare candidate records with the supplied geographic and institutional context.

Check whether the text refers to a hospital, university, research institute, or a department inside one of them. A valid parent organization and a valid affiliated hospital can both be plausible candidates. Record the granularity required by the research question before deciding which identifier to use. Keep co-affiliations separately when the byline actually supplies both.

Review a hypothetical candidate set

Thirty affiliation strings yield 22 straightforward matches. Five appear to refer to a hospital and its partner university, while three lack enough location information. A reviewer resolves four of the five using the institution named on the paper. The result is 26 accepted assignments and four unresolved strings. The accepted fraction of 86.7% describes this workflow and dataset only.

Keep the raw string, matching date, candidate records, accepted ROR ID, and decision evidence. Audit a sample of apparently straightforward matches as well as ambiguous ones. A high match rate cannot reveal a systematic tendency to replace hospitals with their parent universities.

ROR organization matching guidance

Related worksheet: Database join planning and audit worksheet.

Source checked on 5 November 2026.