1 of 10

Deposit Guidelines for UVic Dataverse

SP Dataverse Contacts Meeting

30 April 2020

​

​

Shahira Khair

Data Curation Librarian, University of Victoria Libraries

​

2 of 10

Closed Deposit

  • Old service model used under�UBC Abacus Dataverse�
  • Users request account setup
    • Fewer users, fewer datasets�
  • Direct communication with users
    • Understand expectations
    • Support curation�
  • Time consuming, hard to scale

3 of 10

Mediated Deposit

  • Service model used under �Scholars Portal Dataverse�
  • UVic-affiliated users create own�Accounts, manage dataset process�
  • Scalable, in line with more user�controlled platforms (Zenodo, OSF)�
  • Curation-intervention before �publishing for reassurance

4 of 10

Our Experience

  • Many self-deposited datasets require curation in some form to help them stand on their own

​

  • Intervention at the end of deposit can lead to miscommunication
    • Users unfamiliar with good data documentation and sharing practices
    • Users unaware of library’s expectations are for assuming responsibility of dataset
    • Users may be reluctant to undo/redo work at end stage of deposit

5 of 10

Deposit Guidelines for UVic Dataverse

  • Communicates minimum expectations for datasets deposited to Dataverse�
  • Clarifies roles and responsibilities for depositor and curator�
  • Hopefully read before deposit process, but helpful to point to at intervention stage

https://libguides.uvic.ca/ld.php?content_id=35154390

6 of 10

Expectations for Depositors

  • Datasets must be derived from or intended for research purposes �
  • Selection and appraisal of files �
  • De Identification of sensitive data�
  • Permission to share files

7 of 10

Expectations for the Dataset

  • Use consistent and comprehensible file names and file structures�
  • Deposit files in preferred file formats to support preservation and reuse�
  • Describe your dataset with rich metadata to support discovery�
  • Include a ReadMe file to support interpretation and reuse

8 of 10

Expectations for the Curator

  • We will review datasets before publishing and will flag any issues that need improvement�
  • DOIs only become activated once a dataset is published�
  • Committed to preserving datasets for a set timeframe - min. 10 years�
  • Reviews do not attempt to judge the research quality of a dataset

9 of 10

Feedback

  • Four datasets started/published since posting this policy�
  • Seems to have helped with dataset quality and in one instance made returning one dataset to author a simpler process

10 of 10

Questions for Discussion

Should libraries be setting minimum standards for research data they will accept? (especially as more sharing is being expected by publishers and funders…)�

How much of the curatorial work should we load onto researchers? Can we go beyond this “minimum” list?�

Are there ideas for tools that can be integrated into Dataverse to promote adherence to best practices, like the Curation Tool?

​