To provide data producers with proper credit for their input in data analysis outputs, and ensure the necessary transparency required to guarantee the "replicability" of analytical work, it is essential that data used by researchers be properly cited. This also allows data repositories to maintain catalogs of citations (such as the one maintained by the IHSN), which are useful to measure and assess the uses of existing data, which in turn can be used by data producers to justify their data collection investments.
The citations should provide information on the source and version of the dataset. The "citation requirement" is one of the elements of the Data Documentation Initiative metadata standard. The citation should include, at a minimum, information on the primary investigator, title and abbreviation of the dataset, reference year(s), version number, and dataset ID number. For data obtained on-line, it can also include the URL of the website from which the data were downloaded and the date of the download.
For the dataset ID, using a Digital Object Identifier (DOI) is recommended.