Towards a Data Management Protocol for the Social Sciences: The Science Europe Working Group on Research Data and CESSDA Peter Doorn, Director DANS Chair, Science Europe W.G. on Research Data Vice-Chair, CESSDA G.A. @dansknaw @pkdoorn “Widening the European Infrastructure of Social Science Data Archives” Workshop Lisbon, May 3 and 4, 2017

http://www.scienceeurope.org/ Science Europe is an association of European Research Funding Organisations (RFO) and Research Performing Organisations (RPO), based in Brussels. Mission Science Europe promotes the collective interests of the RFOs and RPOs of Europe. It supports its Member Organisations in their efforts to foster European research. It strengthens the European Research Area (ERA), working with other entities such as the European Universities, the European Academies, the European Scientific Intergovernmental Organisations and the European Commission.

Science Europe on Research Data The Science Europe Roadmap states that research data should be permanently, publicly and freely available for re-use. Access to and sharing of research data are central pillars of Open Science, a concept that Science Europe members fully support. Science Europe is committed to supporting data sharing by contributing to the definition and use of consistent data-sharing policies and practices. This includes identifying legitimate reasons for delayed or restricted access when necessary. In addition, it is crucial to enable access to and sharing of data by resolving data management issues.

Science Europe WG Research Data Until 2016, the SEWGRD worked on fundamental aspects of research data, such as: funding of data management and infrastructures legal aspects related to copyright and Text and Data Mining (TDM) common data terminology: http://sedataglossary.shoutwiki.com/wiki/Main_Page Since summer 2016 the Working Group has focused on the topic of Research Data Management Protocols (RDMP)

One size of data management doesn’t fit all  a domain-oriented approach Increasing demands by RFO’s and RPO’s for Research Data Management (RDM) and Data Management Plans (DMP) Specialized domain data protocols address different (sub-) disciplines or communities Will be much more suitable to serve the community needs Will get better acceptance/adoption by the community But: Require a generalized framework (or terms of reference), to ensure minimum requirements, such as mutual compatibilities, standards, exchangeability and other demands not in the direct interest of a specific discipline

Effectively organize involvement of communities Science Europe M.O.’s to set Data Protocols Framework (Terms of Reference for Domain Protocols) Domain Data Protocols (DDPs) to be openly published Report “A Conceptual Approach to Data Stewardship and Software Sustainability”: http://goo.gl/ycj8QH

Domain Data Protocols: the core idea [1] Protocols defined following (sub-) disciplinary guidelines to be formulated and adopted by research communities: Researchers can refer to the data protocol to be followed instead of finding out the DMP wheel individually Protocols will raise quality standard of DMPs and will be regarded as useful, in turn DMPs will be a stronger tool Counter situation that researchers see RDM as yet another bureaucratic requirement Can cover all research outputs relevant for Open Science, including software DMP

Domain Data Protocols: the core idea [2] Framework (“terms of reference”) for Protocols to be defined by RFOs and RPOs Will make life easier for Researchers: Diminish the administrative burden for researchers By having a single generally approved model for RDM (and SoSu) plans across different funders and research organisations: EC/H2020 National funding organisations Universities, Research Performing Organisations Will make life easier for Funders: Instead of checking thousands of individual DMPs, endorse disciplinary/domain/community protocols DMP will not be a paper tiger, impossible to check whether it is obeyed during execution of research project

Definitions for reference: Domain* Data Protocols (DDPs) are defined as generally agreed-upon guidelines, or predefined, written procedural methods. One might also conceive a DDP as a 'model DMP' for a given domain or community sharing common methods. Data Protocols Framework (DPF), agreed upon by the Members of Science Europe, sets a number of minimal requirements for disciplinary/community data protocols. These terms of reference have close resemblance to the requirements of current DMPs and fit perfectly in data policies that have been or are being formulated. Policy: sets out the principles Plan: for individual project, not for many projects Protocol: formal status, officially published, recognized Template: empty frame specifying the subjects Guidelines: help, recommendations * The level of granularity can vary

Authorship of protocols: at which level of granularity? Several ESFRI ERICs are well placed Rely on existing work as much as possible rather than asking potential partner communities to start from scratch:  Think modular: the detail can vary according to need: Even a very generic protocol or ‘model DMP’ will be helpful You don’t have to oblige anything or anybody: Researchers still write their individual DMPs, motivating where they deviate from the norm/protocol in their field Communities will decide on the detail that they find useful (within the margins of the Framework) There may be alternative DDPs for different purposes (depending on size of project, type/volume of data, etc.) within one domain Find volunteer organisations from different domains to kick-off the process (proof of concept), create expert teams, including both content and data specialists

Selection of proof-of-concept communities for domain data protocols Community 1. Humanities (general) DARIAH 2. Humanities – Archaeology PARTHENOS - ARIADNE 3. Linguistics - Language data CLARIN 4. Social Sciences - Survey research CESSDA 5. Social & Behavioural Sciences – Psychology Psychology departments and associations 6. Social Sciences - Ageing Studies SHARE and TILDA 7. Life Sciences - Bio-informatics ELIXIR 8. Plant Science ERA-CAPS (former Working Group on RDM) 9. Climate Research ICOS / ENVRI+

Questions posed to communities Do you consider the domain protocols approach useful and feasible for your domain? 2. Using DCC general DMP template as a starting point https://dmponline.dcc.ac.uk/. Indicate which questions in the DMP template are not answerable for the community you represent? Do you miss particular questions, or would you want some of them to be phrased differently? Which important building blocks do you find superfluous or do you miss in the template? Are you willing on behalf of the community to compile a first draft of a generic protocol for your community?

General reactions from communities Almost all reactions positive, general interest of communities to cooperate with the S.E. initiative Several of them are already working towards this direction: Data policies (e.g. Plant Science, Climate Research) RDM Recommendations and guidelines (e.g. Life Sciences, Bio-informatics) Detailed RDM templates (e.g. Humanities) And even full-fledged Data “Archiving” Protocols! (e.g. Dutch Psychology) Activity seems to fit in with a more general feeling that we have to find the communalities of data management planning, across funders, institutions, and researchers Approach also appears to fit in well with the DMPonline developments of the Digital Curation Centre

CESSDA Reaction I (Ivana & Ron) Existing situation with respect to RDM in the community long-standing tradition among social survey researchers to use standard ways of documenting survey data files. The data archives supporting the domain have always played an important role in setting or supporting standards, such as the DDI (Data Documentation Initiative - see: http://www.ddialliance.org/) and widely accepted practices of making codebooks, preferred formats for data storage and exchange, etc. CESSDA (or its partners) also provide training on RDM and DMPs, online services, documents, webinars and tutorials to support digital preservation, data archiving and data sharing.

CESSDA Reaction II (Ivana & Ron) Interest of the community to participate in the effort to develop domain protocols The protocols approach will be of help for the SSH domain. it is still quite a struggle in some countries to get researchers to submit their DMPs. Domain protocols approach is a way of promoting responsible attitudes towards data management during and after research, including data sharing and re-use for further research.

CESSDA Reaction III (Ivana & Ron) Suggestions/comments of the community on protocol elements to take into consideration All elements of standard DMP template relevant, with a couple of amendments and extensions for social sciences Different criteria may be needed for different data types and volumes (big data) Promote use of “lab journal” for social science research Refer to ICPSR DMP approach Use “Comply or Explain” as principle

Opportunities for CESSDA Formulate standard requirements for Social Science DMP: Higher quality data (and metadata) Easier data delivery to archive Act as intermediary between research community and funding organisations Help researchers in DMP  raise the perceived value of the CESSDA service providers

Summary: the advantages of this approach Prevent situations where scientific domains or scholarly communities find top-down requirements or templates for DMPs not applicable or not useful for their field/research Better DMP acceptance by researchers and better researcher engagement in RDM; Provision to researchers of a learning vehicle on research data management practices in their field, thus raising the general quality level of RDM; and Reduced DMP processing costs and burdens for funders and researchers , and more focus on and better assessment of deviating RDM solutions

Huge support for Domain Protocols approach at 9th Plenary of the Research Data Alliance, Barcelona Draft Report “Framework Document for Discipline Dependent Research Data Management” available at: https://www.rd-alliance.org/ig-domain-repositories-rda-9th-plenary-meeting Or: https://goo.gl/nMTrhI