KONSİL

Glossary · CRediT contributor roles

What does “Data curation” mean in CRediT?

Data curation is the CRediT role for managing research data so that it can be used now and re-used later: annotating it with metadata, scrubbing it and maintaining it. It starts from data that already exist; collecting the data or analyzing it falls under other roles.

Data curation is one of the 14 roles of CRediT (Contributor Roles Taxonomy), published as the ANSI/NISO Z39.104-2022 standard. NISO (National Information Standards Organization) defines it as the management work of annotating research data (producing metadata), scrubbing it and maintaining it so that it serves both its first use and later re-use, including software code when that code is needed to interpret the data itself. The standard’s foreword names data curation among the contributions CRediT is meant to make visible in multi-author papers.

NISO’s examples include cleaning, cataloging, archiving and retaining data; integrating data from different formats and sources; keeping descriptions, metadata, version control and documentation up to date; and developing preservation strategies so that data stay findable, accessible, interoperable and reusable. Those four qualities are the FAIR principles, a published set of guidelines for making data reusable.

Collecting or generating the data is Investigation, and analyzing it is Formal analysis. Writing analysis code is Software; code belongs to Data curation only when it is needed to interpret the data itself. Because cleaning counts as curation, a person who cleaned and then analyzed the data can list both roles.

Example

In a hypothetical multi-center registry study, the team member who merges spreadsheets from three hospitals into one format, writes a definition for each variable and keeps dated versions of the dataset would appear in the statement as: E.F.: Data curation.

A point to watch

Data curation is not data collection. Taking the measurements or running the experiment is Investigation; curation starts from data that already exist and prepares, describes and preserves them.

Related terms

Guides and tools

Sources

  1. Data curation – CRediT — NISO (CRediT)
  2. ANSI/NISO Z39.104-2022, CRediT, Contributor Roles Taxonomy — NISO, 2022
  3. CRediT role descriptors — NISO (CRediT)
  4. The FAIR Guiding Principles for scientific data management and stewardship — Scientific Data (Wilkinson et al.), 2016