ICPSR

1
ICPSRs Approach to Data Citation and Persistent
Identifiers

• Mary VardiganAssistant Director, ICPSR
• Workshop on Persistent Identifiers in the Social
  Sciences -- Bonn, GermanyFebruary 1, 2011

2
Todays Presentation

• ICPSRs use of data citations and persistent
  identifiers
• Ways that ICPSR encourages good practice
• Issues to be resolved
• Future directions

3
ICPSRs Use of Citations

• ICPSR has been providing citations to its data
  since 1990
• Citations based on Cataloging Machine-Readable
  Data Files by Sue Dodd, American Library
  Association, 1982

4
What Makes Up an ICPSR Citation?

• Content Creator/Principal Investigator
• Title
• Distributor ICPSR
• Distribution place and date
• ICPSR study number
• Version number
• Materials designation Computer file
• DOI

5
Example

• Schneider, Barbara, and Linda J Waite. The 500
  Family Study 1998-2000 United States Computer
  file. ICPSR04549-v1. Ann Arbor, MI
  Inter-university Consortium for Political and
  Social Research distributor, 2008-05-30.
  doi10.3886/ICPSR04549

6
ICPSRs Use of DOIs

• ICPSR started assigning DOIs in 2008
• DOIs apply at the study or collection level (a
  study can have multiple datasets)
• DOIs are of the form doi10.3886/ICPSR04549
• DOIs resolve to the study homepage (metadata
  record)

7
How ICPSR Obtains DOIs

• ICPSR uses the CrossRef service, the official
  DOI link registration agency for scholarly and
  professional publications
• ICPSR pays a modest annual Publisher Fee (based
  on publishing revenues) and pays 6 cents per DOI
• To begin assigning DOIs, in 2008 sent CrossRef an
  XML file containing metadata on all ICPSR 7000
  studies
• Now get DOIs weekly

8
Weekly Process

• ICPSR runs script to create XML metadata in
  CrossRef format
• Contributors and their roles
• Title
• Publication date
• Update date
• Study number
• DOI
• URL

9
Weekly Process, continued

• ICPSR submits XML file to register new DOIs
• CrossRef sends email confirming the file is
  correct
• At that point, the DOI has an associated URL on
  the ICPSR Web site

10
Alternative Process

• Registration could happen in a script-driven
  manner through an API
• This would happen without human intervention
• ICPSR database could communicate with the
  CrossRef database with DOIs registered
  automatically

11
Requests for DOIs

• Journals are requiring that authors provide PIDs
  to data they analyzed for their articles
• Authors are coming to ICPSR for DOIs
  pre-publication, generally depositing data into
  the Publication-Related Archive

12
Encouraging Good Practice

• Bibliography of Data-Related Literature includes
  60,000 citations to publications based on ICPSR
  data
• Two-way linking Studies link to publications,
  Bibliography links back to studies
• Widely used DOIs for data would make searching
  for and harvesting related publications much
  easier

13
(No Transcript)
14
(No Transcript)
15
(No Transcript)
16
Making Citations and DOIs More Prominent

• ICPSR provides RIS export for data citations into
  bibliographic citation software
• ICPSR highlights the data citation and DOI in
  several places

17
(No Transcript)
18
For each study
19
Working with Vendors to Promote Links to Data

• ICPSR has a project with Thomson Reuters to
  display data linkages in Web of Knowledge
• Full and summary records in Web of Knowledge will
  link to related data when appropriate
• ICPSR is providing a periodic data feed of
  datasets and related publications to TR
• TR is integrating data feeds from others
  including UK Data Archive

20
Influencing Journals

• On behalf of the Data-PASS partners, ICPSR wrote
  to professional associations in sociology,
  political science, and economics
• Letters urged them to raise the standards for
  data citations in their journals
• Professional associations are in a position to
  set standards for their members and for journal
  editors (including copy editors)

21
More on Influencing Journals

• Approach was to point to the variety of ways that
  data were cited in specific journal issues
• The letter stressed the importance of citing data
  the same way that publications are cited and the
  value of persistent identifiers
• Organizations discussed the letters at recent
  national meetings
• American Sociological Review just revised its
  Notice to Contributors to reflect the importance
  of data citations and DOIs

22
Updating Citation Software

• ICPSR worked with EndNote (owned by Thomson
  Reuters) to ensure that data citations display
  correctly
• The result is that Dataset is now a Reference
  Type in EndNote.
• Zotero also needs adjustment for datasets

23
Working with the Community

• ICPSR has joined DataCite as an associate member
• ICPSR has joined ORCID Open Researcher and
  Contributor ID. ORCID aims to create a central
  registry of unique identifiers for individual
  researchers
• ICPSR is heading up an IASSIST special interest
  group on data citation (SIGDC)

24
IASSIST Session

• IASSIST SIGDC has proposed a session as part of a
  data citation track including DataCite
• Tracking Data Reuse Motivations, Methods, and
  Obstacles -- Heather Piwowar, NESCent, University
  of British Columbia
• Building Data Citations for Discovery  Hailey
  Mooney, Michigan State University, and Mark
  Newton, Purdue University
• ICPSRs Efforts to Encourage Data Citation --
  Elizabeth Moss, Inter-university Consortium for
  Political and Social Research (ICPSR)
• Reactor Panel from SIGDC

25
Issues to Resolve

• With the community, address situations when data
  resources have multiple distributors (and
  multiple DOIs)
• Implement versioning in DOIs
• Address level of granularity for DOIs
• Move to DataCite

26
Multiple DOIs for Same Data

• Eurobarometer 72.2 (Nuclear Energy, Corruption,
  Gender Equality, Healthcare, and Civil
  Protection)   DOI doi10.4232/1.10009
  Principal Investigator Antonis Papacostas
  Publication Agent GESIS - Leibniz-Institut für
  Sozialwissenschaften
• Papacostas, Antonis. Eurobarometer 72.2 Nuclear
  Energy, Corruption, Gender Equality, Healthcare,
  and Civil Protection, September-October 2009
  Computer file. ICPSR28186-v1. Cologne, Germany
  GESIS/Ann Arbor, MI Inter-university Consortium
  for Political and Social Research distributors,
  2010-07-19. doi10.3886/ICPSR28186

27
From CrossRefs Publisher Rules

• CrossRef only registers DOIs for Definitive
  Works but not for Duplicative Works, as defined
  in the CrossRef Glossary.  CrossRef does not
  permit multiple DOIs to be assigned to certain
  closely related versions of a work Where a
  CrossRef member has content which is
  substantially Duplicative of Definitive Works,
  the member must retrieve the DOIs of Definitive
  Works for display in such substantially
  Duplicative Works and must link from the
  substantially Duplicative Works to the Definitive
  Works.

28
More on Multiple DOIs

• CrossRef policy oriented toward publications not
  data
• Arrangement between ICPSR and GESIS is clear, but
  there are other co-distributor relationships
• How much of a problem is this and can we develop
  a community solution?
• Can we use the DataCite metadata kernel
  (relationType) to specify relationships?
• Would providing explanatory text and
  cross-referencing DOIs in archives metadata
  records be useful?

29
Versioning and DOIs

• ICPSR has decided to add version numbers to its
  DOIs
• ICPSR may not have previous versions online
• User will have to contact ICPSR for access
• So far the number of users requesting older
  versions has been very small

30
Level of Granularity for DOIs

• ICPSRs current practice is to assign the DOI at
  the study level
• DOI resolves to the study homepage, which
  includes Version History detailing changes to all
  files in the collection
• Assigning dataset-level DOIs is a challenge
  because ICPSR has over 65,000 datasets
• ICPSR is undertaking a large project to revamp
  archival management and dataset-level DOIs will
  be integrated in the new infrastructure

31
Moving to DataCite for DOIs

• DataCite offers several advantages because of its
  focus on data
• Metadata kernel more robust and intended to
  describe data
• Community of trusted data centers is a shared
  goal

32
Future Directions

• Address situations when data resources have
  multiple distributors and multiple DOIs
• Approach other vendors including Google Scholar
  after TR service deployed
• Contact other professional associations and
  journals
• Work with other data producers on providing
  visible citations and DOIs and encouraging their
  use
• Continue spreading the word about data citation
  and persistent identifiers!

33
Thank you

• Questions?