One common question we receive about ROR registry records is whether they reflect organizational hierarchy and relationships between organizations – yes, they do! ROR does indeed support parent-child organizational hierarchies as well as other types of relationships between organizations. In this blog post, we go into more detail about how ROR supports relationships and hierarchies, how these are expressed in the metadata, how we curate them, and how users like you can explore and leverage this metadata.
Relationships in ROR at a glance
The ROR registry currently includes records for nearly 105,000 organizations (to be precise, 104,594 at the time of this writing). Across the entire ROR dataset, there are more than 22,000 total relationships between organizations (22,057), which means that approximately 21% of ROR records have at least one relationship to another record in the registry.
Parent-child hierarchies are the most common relationship type, with more than 17,000 instances. The most common scenario is that an organization has only one parent, but in some cases an organization has multiple parents (1,891 organizations, or about 1.8% of total records, have multiple parents). Several national-level and large-scale research organizations, such as the Czech Academy of Sciences (https://ror.org/053avzc18), have more than 100 of their child organizations enumerated in ROR.
Since its first independent release in March 2022 as part of the planned divergence from GRID, ROR has created over 1,000 new relationships between research organizations in the registry. We do this in collaboration with the global research community through our transparent curation process, and this service is entirely free of charge.
Relationships in the ROR data structure
ROR records store both structural and temporal connections in the relationships field. Connections and relationships between records can therefore be mapped and graphed. The relationship types supported are Parent, Child, Related, Successor, and Predecessor. An organization can have multiple relationships, but each relationship must be classified as only one of the above types.
The United States Department of Energy, for instance (https://ror.org/01bj3aw27), has many “child” organizations such as national research laboratories. Those child organizations themselves often have their own children representing different research units and funding bodies. Pictured below is a vertical “family tree” of the Department of Energy created from ROR records with an organization tree script showing DOE’s children and grandchildren (laterally related organizations are not shown in this view). This structure aligns with how the Department of Energy tracks both its research outputs and the research projects it has supported using ROR IDs and Crossref Funder Registry IDs.
