Wednesday, November 02, 2016

Should We Keep Everything Forever? Determining Long-Term Value of Research Data

Should We Keep Everything Forever? Determining Long-Term Value of Research Data. Bethany Anderson, et al.  iPres 2016. (Proceedings p. 284,5/ PDF p. 143). Poster.
     The poster describes efforts by the institution to launch an institutional data repository called the Illinois Data Bank. The Research Data Service is committed to preserving and providing access
to published research datasets for a minimum of five years after the date of publication in the Data Bank. A preservation review developed preservation review processes and guidelines for datasets that will help promote the discoverability and use of open research data. They offer a preservation and access solution that is trusted by researchers.

The framework includes guidelines and processes for reviewing published datasets after their five-year commitment ends and decide if they should be retained or, deaccessioned. This systematic appraisal approach helps them decide the long-term viability of a dataset, its value to research communities and its preservation viability.

The Preservation review guidelines for the Illinois Data Bank are:

Evaluated by Curators/Librarians/Archivists
  • Cost to Store:  estimated cost of continuing to store
  • Cost to Preserve: estimated cost of continuing or escalating preservation
  • Access: use metrics to determine interest in this dataset
  • Citations:  has the dataset been cited in any publications
  • Restrictions: are there access or re-use restrictions
Evaluated by Domain Experts
  • Possibility of Re-creation
  • Cost of Re-creation
  • Impact of Study: did the study for this dataset significantly impact research
  • Uniqueness of Study
  • Quality of Study
  • Quality of Dataset
  • Current Relevance to contemporary research questions
Evaluated by Curators/Librarians/Archivists and Domain Experts
  • Are other copies available
  • Understandability: is the metadata & documentation for access / reuse sufficient
  • Dependencies: what are the software and environment dependencies
  • Appropriateness of Repository: is there a better repository for the dataset

No comments: