Why Research Data Citations Need a Version and a Boundary

A dataset name may describe more than one thing

A familiar dataset title can refer to a project that has changed for years. Records may be added, corrected or reprocessed. A citation that names only the project can give proper general credit while leaving another researcher unable to identify the evidence behind a particular calculation. The citation needs a version and, when relevant, a boundary around the subset used.

USGS distinguishes static, versioned and dynamic data products in its citation guidance. That distinction is useful because each changes differently. A fixed release, a sequence of releases and a continuously updated service do not present the same reproducibility problem. Before drafting the citation, determine which type of product supplied the analysis and what information its provider offers for identifying the actual records used.