Research data management compliance: is there a bigger role for university libraries?

ABSTRACT

This article explores how 13 Australian universities are assisting their researchers to manage the growing expectation to make research data more accessible. It identifies which university groups are supporting staff with research data management (RDM) activities and queries whether university libraries might have a bigger role to play in this space. We found that there was not a consistent approach to RDM at the 13 universities and that while there was generally strong encouragement to store research data securely during and after the project, there was overall a lack of practical support in how to undertake this activity. From our findings, a question of whether library staff have the appropriate experience, training and professional development to enable academic libraries in Australia to expand their RDM role arises and warrants further research.

KEYWORDS:

• Academic libraries
• Australian universities
• open research data
• training
• research data
• research data management
• RDM activities

Introduction

For some years now, there has been a move globally to make research data, both research outputs and the underlying research data, more accessible to other researchers. In January 2004, the Organisation for Economic Co-operation and Development (OECD) adopted the Declaration on Access to Research Data from Public Funding (Organisation for Economic Co-operation and Development, 2004). OECD members, including Australia, agreed to ‘work towards the establishment of access regimes for digital research data from public funding’ (Organisation for Economic Co-operation and Development, 2004, Paragraph 11).¹ In 2007, the OECD released the Principles and Guidelines for Access to Research Data from Public Funding (Organisation for Economic Co-operation and Development, 2007).

Originally, the focus was on open access publishing so that research publications could be freed from the constraints of copyright restrictions imposed by commercial publishers of journals. This focus led to an increase in accessibility to research publications through open access journals which do not require a subscription and are available via open access repositories or archives. These repositories are often managed by university libraries. The focus of open access has now moved to open access to research data and many open data repositories (Bird, 2010; Budapest Open Access Initiative, 2002) also aim to provide access to the raw data used for such research. Within Australia, such initiatives are illustrated by the Australian National Data Service (ANDS) (https://www.ands.org.au) and Research Data Australia (http://researchdata.ands.org.au/), which aim to provide collaborative access to research data.² However, apart from this, open access to research data has been slow to develop in Australia compared with developments in the United Kingdom (UK) and Europe (Corti, 2011).

Over the last few years, Australian universities have begun to introduce policies and procedures to address better management of research data, mainly in response to government funding body requirements, but also because of the need to comply with the then 2007 Australian Code for the Responsible Conduct of Research (the Code)³ and the National Statement on Ethical Conduct in Human Research (the Statement).⁴ How these policies and procedures are being implemented and by which university group is not always clear, particularly as there are a number of groups that are involved with research at universities, such as central research offices, e-research groups, information technology (IT) services support and academic libraries, in addition to faculties and schools in which researchers are located.

The realities of research data management project (Bryant, Lavoie, & Malpas, 2017) found that university strategies have elevated research data management (RDM) to an institutional priority and there is a growing demand by researchers for RDM support (Council of Australian University Librarians, 2017). As such, academic libraries is one group attempting to respond by making RDM support a strategic focus (Cox, Kennan, Lyon, & Pinfield, 2017; Cox & Pinfield, 2014; Pryor, Jones & Whyte, 2014). The Society of College, National and University Libraries (SCONUL) report on the future of libraries also found that academic libraries are shifting their strategic emphasis from collections to services to ‘support changing practices in research as well as teaching and learning’ (Pinfield, Cox, & Rutter, 2017, p. 18).

This article seeks to understand how Australian universities are managing and assisting academic staff with RDM activities. It also attempts to assess the extent to which open access to research data is actively encouraged in universities and which university group(s) are involved in the space.

Research Questions and Method

This article attempts to answer a main research question of: How are Australian universities supporting requirements to manage research data to make it more accessible and accountable? Subsidiary research questions explored are: Which groups in the universities reviewed have responsibility for management of and compliance with RDM policies?; Whether open access to research data is being specifically encouraged through the introduction of open access data policies? and Whether there are specific RDM training programmes and support provided to academic staff and by which groups?

RDM is defined as ‘the organisation of data, from its entry to the research cycle through to the dissemination and archiving of valuable results. It aims to ensure reliable verification of results, and permits new and innovative research built on existing information’ (Whyte & Tedds, 2011, Paragraph 4). It covers the managing, sharing, dissemination and reuse of data (Australian Research Council, 2018). It is a broader term than ‘open access to data’ which specifically refers to being able to access research data (Australian Research Council, 2018) such as statistics, survey results or interview recordings.

To understand how Australian universities are managing RDM, desk research reviewing the websites of 13 Australian universities was undertaken. At least two universities were selecting from each of the following groupings:

• Group of 8—Monash University (Monash), the University of New South Wales (UNSW)

• Australian Technology Network (ATN)—RMIT University (RMIT), Queensland University of Technology (QUT)

• Innovative Research Universities (IRU)—Griffith University (Griffith), Flinders University (Flinders)

• Regional Universities Network (RUN)—University of Southern Queensland (USQ), University of New England (UNE)

• Unaligned universities—Australian Catholic University (ACU), the University of Western Sydney (UWS), Deakin University (Deakin), Swinburne University (Swinburne) and the University of Tasmania (UTAS)

A search of each of 13 university’s websites, including the universities’ library website, for publicly accessible information or university policy relating to research, research data management and open access including data storage, was undertaken using the keywords of ‘research’, ‘research data’, ‘open access’ and ‘research data management’.

Using content analysis methods, any publicly accessible relevant university policy or webpage were examined to determine whether the university had policies relating to RDM or open access and which university group (research, e-research, information technology (IT) services, library, faculty or school) was providing assistance to researchers in the areas of RDM guidelines, resources and training, storage options for research data and open access options for research outputs including both data and publications. It is acknowledged that some universities might not have made some or all information about their RDM activities or open access approaches publicly accessible, but the researchers found that a wide range of documents were accessible.

Literature Review

What is open access and RDM?

The European Commission defined open access as ‘the practice of providing on-line access to [scholarly] information that is free of charge to the end-user and that is re-usable … [including] research data (data underlying publications, curated data and/or raw data’ (2013, p. 3). Open access has also been described as an ‘ethical imperative (…) to maximize the value of research data generated from human participants, particularly when using public funds’ (Bobrow, 2013, Paragraph 3).

As referred to earlier, open access was originally associated with access to published research outputs but the term is now understood to include access to research data. Some of the reasons stated for providing open access to research data in addition to the published article describing the research findings are that it will preserve important databases that may otherwise be lost as researchers retire or die, allow better external validation of research findings, enable the exploration of topics not envisioned by the original researchers, permit the creation of new data sets from data from multiple sources, and increase the return of public funding of research by enabling government research funding to benefit a wider group of researchers (Cheshire, Broom, & Emmison, 2009; Organisation for Economic Co-operation and Development, 2007). Quantitative research data sharing has been relatively straightforward but there have been issues around sharing qualitative research data, primarily related to privacy and confidentiality issues arising from the personal data collected in qualitative research (Childs, McLeod, Lomas, Cook, 2014). Anonymising and de-identifying data is viewed as the solution to handling these issues but these practices in turn raise issues about useability and reliability of the resulting data (Childs et al., 2014; Patel, 2016).

Following the release in 2007 of the Principles and Guidelines for Access to Research Data from Public Funding (Organisation for Economic Co-operation and Development, 2007), OECD members including Australia began new initiatives or to strengthen existing ones to implement the Declaration on Access to Research Data from Public Funding. The UK has been a leader in open access to data, archiving quantitative data as early as 1967 and starting archiving qualitative data in the 1990s (Corti, 2011). The United States (US) also started open access repositories for quantitative data as early as 1962 but no progress was made with qualitative data until 2009 (International Federation of Data Organisations, 2014). Similarly, Australia made little progress with open access to qualitative data until the late 2000s (Fitzgerald & Pappalardo, 2007).

A new Australian Code for the Responsible Conduct of Research was released by the National Health Medical Research Council (NHMRC), Australian Research Council (ARC) and Universities Australia in June 2018, with affected institutions such as universities, needing to meet the requirements of the Code by 1 July 2019 (National Health Medical Research Council, 2018a). It is replacing a 2007 Code that stated, inter alia, that ‘research data should be made available for use by other researchers unless this is prevented by ethical, privacy or confidentiality issues’ (National Health Medical Research Council, Australian Research Council, & Universities Australia, 2007, Section 2.5.2).⁵

The new Code contains eight principles which encapsulate responsible research conduct, 13 responsibilities that underpin the responsibilities of research institutions, and 16 principles that relate to the responsible conduct of researchers. Principle three of the Code addresses the need for transparency in reporting research and is satisfied if researchers ‘share and communicate research methodology, data and findings openly, responsibly and accurately’ (National Health Medical Research Council, 2018a, Principle 3). Institutions will be expected to ‘provide facilities for the safe and secure storage and management of research data, records and primary materials and, where possible and appropriate, allow access and reference’ (National Health Medical Research Council, 2018a, Responsibility 8). Researchers will have the following responsibility to ‘[r]etain clear, accurate and complete records of all research including research data and primary materials and, where possible and appropriate, allow access and reference to these by interested parties’ and disseminate research findings responsibly, accurately and broadly (National Health Medical Research Council [NHMRC], 2016, Responsibility 22 and Responsibility 23).

The National Statement on Ethical Conduct in Human Research (the Statement), released in 2007 and updated in 2015, which applies to all researchers conducting research with human participants, provides some guidance about how a researcher should deal with issues of open access should the researcher decide to make the research data available for secondary analysis in a databank such as a repository.

Sections 3.2.1 to 3.2.12 of the Statement provides that a researcher should take steps to consider what information to store and/or omit from the databank to ensure the well-being of the participant. They also detail the information the researcher needs to provide the participant about the potential future use of the data that may be stored. These requirements in the Statement means that researchers must consider in detail the potential of storing information in an open access databank, and how they will go about de-identifying participants. This can be a tedious and cumbersome task to do manually, particularly with the current trend of large data collections. There is the very real potential for researchers to overlook or to forget to omit or change some crucial information.

The ARC, the national funding body for non-medical research, ‘strongly encourages the depositing of data arising from a project in an appropriate publicly accessible subject and/or institutional repository’ (Australian Research Council [ARC], 2017, Funding Rules A12.5.2). However, the ARC does not mandate open access to data. It has an ‘open access policy’ but this only relates to research outputs such as articles and reports and associated metadata.

The ARC’s requirement is designed to encourage researchers to consider the ways in which they can best manage, store, disseminate and reuse data. Researchers, in consultation with institutions, have a responsibility to consider the management and future potential of their research data, taking into account the particular approaches, standards and uses for data that may exist in different institutions, disciplines and research projects. Some institutions may have infrastructure and/or processes in place for storing, managing and sharing data and these are valuable resources that should be utilised.

The NHMRC open access policy which was updated in January 2018 is aligned with the 2017 ARC open access policy (National Health Medical Research Council, 2018b). While it does not mandate open access for research data, it ‘strongly encourages researchers to take reasonable steps to share research data and associated metadata’ (National Health Medical Research Council, 2018b, p. 4). It also ‘supports the overall intent of the F.A.I.R. Access to Australia’s research statement through which [research] data/information are made findable, accessible, interoperable and reusable’ (National Health Medical Research Council, 2018b, p. 4).

The F.A.I.R (Findable, Accessible, Interoperable, Reusable) data principles⁶ developed in 2015, are internationally recognised guidelines relating to research data and metadata being (Wilkinson et al., 2016; Australian National Data Service [ANDS], n.d.; Association of European Research Libraries, 2017):

• easily findable,

• available in a repository and if the research data is not open, ensuring that the metadata is available,

• able to be combined or linked with other research data using international standards related to format, language and vocabularies that follow the FAIR principle,

• well described so that the research data can be replicated and re-used.

Benefits of F.A.I.R. research data include improving the reliability of the research, increasing the citation and impact of the research and alignment with international standards and approaches (ANDS, n.d.).

Attempting to implement the FAIR principles to research data, particularly qualitative data, within an existing repository will be challenging as the repository will have been developed using ‘discipline based practices that have well established protocols for how data should be shared’ (Dunning, de Smaele, & Bohmer, 2017, p. 185). Complying with the FAIR principles may require open guidelines rather than rules, as some of the facets, specifically related to interoperability and reuse, are subjective and open to interpretation (Dunning et al., 2017). However, ‘implementing and publishing policies on identifiers, metadata, licensing and protocol[s] will help all [research data] repositories align with the FAIR principles’ (Dunning et al., 2017, p. 188).

As noted earlier, the term RDM is used to describe the whole process of collecting, managing, using and disseminating data gathered during the research process. It encapsulates the need to provide open access to both publications and data. It requires researchers to consider all stages of data management when undertaking research. RDM is also intended to be used to ensure compliance at an institutional as well as faculty, school and individual researcher levels.

Current roles in research data management

A number of university groups appear to be involved in supporting staff with RDM. It would be expected that a university’s research group, whether centrally and faculty based, would have a major role. With the focus on digital research data, the information technology (IT) group should be involved. One group which has only be established in the last ten or so years, is e-research (Australian Government, 2006). E-research groups have been created to contribute to the use of advanced information and communication technologies to support research (Schroeder, 2007). Australian government initiatives have provided significant impetus for their existence. Their location within a university can vary. Some are located within the research group, such as at RMIT University, while others may be located within the library, as at Flinders University.

The university group that appears to be engaging more and more in RDM is the university library. Universities’ strategic focus on research is requiring academic libraries to review their services and engage in providing research services that go beyond research bibliometrics, but as noted by Groenewegen ‘actual engagement remains a hit and miss affair’ (2017, p. 176). A reason that academic libraries may not be able to fully engage within the activities of researchers is that many of the research services it offers, could be provided by other groups within a university (Pinfield et al., 2017).

A number of studies between 2012 and 2014 found that, at that time, the majority of academic libraries in Australia, Canada, the UK, and the US undertaking RDM were offering information or consultation type services rather than technical support. For example, Tenopir, Birch, and Allard (2012) found that the most common type of RDM service being offered by 221 academic libraries in both Canada and the US was reference support related to finding and citing research data and data sets with 44 percent offering this service and only 14.5 percent providing technical RDM support related to accessing a repository. The Tenopir et al. (2012) study also found that in the next two years, 77 and 66 percent respectively plan to offer technical support in respect to removing research data or data sets from a repository or creating meta data for research data and data sets. Similarly, Corrall, Kennan and Afzal’s 2012 survey of 88 libraries in Australia, Ireland, the UK and New Zealand found that only 8 of the 82 UK libraries in the study were offering technical RDM support in the form of digital curation (for example preserving and archiving research data) but approximately 40 percent planned to offer this type of technical support in the future (Cox & Pinfield, 2014).

Tenopir et al. (2012) concluded that as the requirement for RDM increases and with researchers constrained by time and training in this area, academic libraries are in a unique position to have an active role in RDM activities. Similarly, a 2015 survey by the University of Newcastle in Australia found that researchers need support in the area of RDM (Council of Australian University Librarians, 2017). The Australian National Data Service (ANDS) also recognises that ‘librarians have a wealth of experience and skills to bring to RDM’ (Searle, Wolski, Simons, & Richardson, 2015, p. 442).

Childs et al (2014) study of two RDM projects concluded that there are opportunities for information management professions which would include librarians in RDM and specifically in converting research data to open research data, for example in the anonymisation activity. This conclusion supports the findings of Cox and Pinfield (2014) who concluded that RDM services offered by academic libraries could include working with research data, specifically assisting with storage, security, preservation and access but both Childs et. al (2014) and Cox and Pinfield (2014) noted that new knowledge, skills and training will be needed.

Similarly, Searle et. al’s case study of Griffith University’s management of research services concluded that the University ‘recognised the importance of supporting … librarians in building the knowledge and skills required’ for RDM activities (2015, p. 456). However, Searle et al. (2015) noted that developing these capabilities is challenging as often these skills are acquired by in-house training or via self-training. Some examples of self-training resources include joining information management (IM) discussion groups such as DataCure discussion list (http://bit.ly/22sx7M8) and DataQ (http://researchdataq.org/index.php), following social media such as #datalib (Barbrow, Brush, & Goldman, 2017), participating in webinars and using research data management tools such as the US Research Data Alliance’s ‘23 things: Libraries for Research Data’ (Research Data Alliance, 2015), the University of Edinburgh’s MANTRA tool (https://mantra.edina.ac.uk) and the DataOne modules (https://www.dataone.org/education-modules).

In terms of training for librarians in the area of RDM, a study by Thomas and Urban (2017) found that from 75 respondents of IM professionals from libraries in the US, UK, Europe and Australia, 29 gained their data management knowledge from on the job learning, 19 gained their data management knowledge from a combination of formal training (for example, an information management course) and on the job learning. Only one respondent indicated that as part of their IM course, there was a subject or unit dedicated to data management with 26 responding that no subject or unit as part of their IM course that taught data management. Thomas and Urban (2018) concluded that information management education programs will need to be revised to ensure that data curation relating to the ‘activities required to maintain research data … so that it is available for re-use and preservation’ (p. 4) is included in formal education.

A 2017 study by the Online Computer Library Center (OCLC)⁷ notes that the approach to RDM needs to be a focus of the entire academic institution rather a single department and as such, there is a role for academic libraries in supporting their institution’s RDM goals. With research data as ‘a key [component] of the research and scholarly communication processes, as well as the scholarly record … the acquisition of RDM capabilities will only increase which presents both challenges and opportunities for academic libraries …’ (Bryant et al., 2017, p. 14). to improve and increase this assistance.

Key: Responsibility of

Library

Research

E-Research

Other e.g. I.T.

Results and Findings

The requirement of funders and governments for research outputs to be shared and made public has seen academics institutions in Australia, the UK and the US implement research data management policies and provide resources to assist researchers with research management (Pryor, 2014). Australian universities have begun to introduce policies and procedures to address RDM over the last few years, although a few universities, like the Australian National University (ANU), the University of Queensland (UQ) and Queensland University of Technology (QUT) have had data management policies since at least 2008 (Fitzgerald, Pappalardo, & Austin, 2008). Only a few universities have developed additional policies on open access to research data.

Table 1 sets out a snapshot of our findings from the desk research undertaken of the 13 universities. From our sample, it appears there is no uniform approach employed by Australian universities in relation to RDM with research, e-research, library and IT services all having responsibilities. It was found that the larger the university, the more groups that were involved in managing RDM. As noted earlier, e-research which are relatively recent created groups within universities often have dual reporting lines and quite different responsibilities.

Table 1. Summary of research data management activities of 13 Australian universities.

	Research Policy/Code	Data Management Policy	Open Access Policy	Guidelines		Storage		Open Access Support	Research Repository	Storing Research Data
ACU	Y	Y	N						Y	Y
Deakin	Y	Y	N						Y	Y
Flinders	Y	Y	N						Y	N
Griffith	Y	Y (as part of Code)	N						Y	N
Monash	Y	N	N						Y	Y^a
UNE	Y	Y	Y						Y	Y
UNSW	Y	Y	Y						Y	N
QUT	Y	Y	N						Y	Y^b
RMIT	Y	Y	N						Y	Y^c
Swinburne	Y	N (has broken link)	N						Y	Y
UTAS	Y	Y	N						Y	Y
USQ	Y	Y	N						Y	Y
UWS	Y	Y	Y						Y	Y^d

aVia Monash.figshare.

bVia Research Data Finder.

cVia the RMIT Research Data Catalogue. Formerly known as Research Data Box (RedBox).

dVia Research Data Box (RedBox).

Policies

All the universities we examined have some form of research policy. Some, like Griffith Swinburne, QUT and Monash, have either adopted or adapted the 2007 version of the Code to create their own code of research practice while the remaining 10 universities have drafted their own research policies.⁸ Eleven universities have also introduced a specific RDM policy while Griffith has included RDM in their research code. Swinburne states on its data management webpage that it has a RDM policy but the provided hyperlink did not lead to a policy (Swinburne University, 2017a).

Three universities have also introduced stand-alone open access policies. The UNSW open access policy, like the ARC Open Access Policy, concentrates on research outputs such as publications but it does encourage researchers to make outputs such as research data descriptions and small datasets available via the UNSW institutional repository ‘where publisher permission can be obtained’ (University of NSW, 2017, Section 1.1). The UNE open access to UNE research publications and data policy relates only to research outputs such as publications but it does include metadata for datasets. Similar to UNSW, open access to research outputs is via the University’s research repository known as e-publication@UNE. The UWS policy covers both research publications and data.

Open access to data is addressed in other ways at other universities. For instance, Swinburne provides information via webpages. It has a sharing of research data webpage (http://www.swinburne.edu.au/research/ethics/data-management/sharing-research-data/) which discusses options for making data accessible. It also has another webpage (http://www.swinburne.edu.au/research/ethics/data-management/what-does-open-access-mean/) which encourages researchers to make their research publications available through the University’s open access repository known as the Swinburne research bank. This webpage also draws attention to the fact that the ARC and the NHMRC both require publications to be placed in open access repositories (Swinburne University, 2017b).

Flinders has a management of research data and primary materials policy which states that ‘research data should normally be made available under an open access licence such as a Creative Commons licence or by controlled access …’ (Flinders University, 2016, Section 3.4.3). Similarly, RMIT has a RDM policy process which states that researchers should ‘make research data available to other researchers and support reuse of research data, where possible. When you cannot make the research data available, make the research metadata available’ (RMIT University, 2016, Section 8.1).

The RMIT RDM policy process also refers researchers to the Library for assistance with making research data discoverable. The RMIT IT group is responsible for a webpage on data sharing and collaboration but the content focus on this webpage is on data handling during the project. This webpage provides a hyperlink to the University Library’s RDM guide (http://rmit.libguides.com/researchdata). This guide includes brief sections on data sharing and storing and on data publishing and sharing data.

All universities in the sample except one offered a checklist, guideline or template on research data management. However, that one exception (Flinders) provided a frequently asked questions on research data management webpage (http://www.flinders.edu.au/library/research/data-management/faqs—dmp.cfm).

Data Storage

For data storage options prior to or during a research project, all except two of the universities sampled provided internal storage options for researchers within the university. As shown in Table 1, the group responsible for providing internal storage options varied between the research, e-research, library or IT groups among the universities sampled. In four universities, the responsibility was shared between two groups, including the library and in the case of UNSW between three university groups including the library. As an alternative to the internal option, several universities also recommended the use of external storage options such as Figshare (https://figshare.com) and Cloudstor (https://www.aarnet.edu.au/network-and-services/cloud-services-applications/cloudstor).

At the universities sampled, there was a difference between whether a group only recommended and/or provided a data storage option and whether a group such as the library assisted researchers in using the data storage options. For example, at USQ, the researcher, known as the researcher worker, is responsible for retaining and storing research data collected from a project on a University approved system. Advice is available to research workers via the research group in relation to selecting an appropriate storage option (University of Southern Queensland, 2016). Researchers at USQ appear to upload their own data for retention and storage purposes but can seek assistance from the research group if needed. In regards to uploading their data for sharing on Research Data Australia (http://researchdata.ands.org.au/), USQ researchers can seek assistance from the University’s Library. Similarly, RMIT and UTAS provide internal data storage options for researchers to store their data and recommends external options for sharing data. Researchers at both RMIT and UTAS can seek assistance from their Library in relation to uploading their research data for sharing and re-use by other researchers. In contrast, the ACU Library provides researchers with an RDM toolkit (https://libguides.acu.edu.au/data_management_toolkit/home) and only assists with uploading records to the University’s research repository. Overall, there appears to be a lack of support for researchers in how to use the data storage options.

To enable researchers to share and re-use research data, nine of the universities sampled recommended external data storage options such as Research Data Australia (http://researchdata.ands.org.au/). At three universities, research data is automatically uploaded to Research Data Australia (http://researchdata.ands.org.au/) at the same time as the research data is uploaded by the researcher to their university’s research repository.

All 13 universities sampled host a research repository on their website. Each repository contains records of research outputs authored by the university’s researchers. Due to copyright restrictions, access to the full text of the research output is generally not available via the repository record. Rather, the record, unless a thesis record, contains a description including the metadata of the research output.

As of October 2017, 10 of the universities sampled hosted research data on behalf of their researchers. Six universities included research data records within their research repository. Similar to research output records, the research data records held within the research repositories of the universities sampled, do not generally provide access to the research data. Rather the record provides a description of the data and in some cases, a hyperlink to the data record in an external data storage option such as Figshare or Research Data Australia (http://researchdata.ands.org.au/).

Four universities hosted research data separately from their research repository. For example, Monash uses Monash.figshare and QUT uses Research Data Finder (https://researchdatafinder.qut.edu.au) to manage, store, share and publish research data. Similarly, RMIT and UWS uses Research Data Box (known as ReDBox)⁹ to store and allow the sharing and re-use of research data. However not all research data records in an external data storage option such as ReDBox provide access to the research data. For example, as of October 2017, RMIT holds 323 research data records in ReDBox and has provided 299 records in Research Data Australia (http://researchdata.ands.org.au/).¹⁰ However, when a search for research data in Research Data Australia (http://researchdata.ands.org.au/) is limited to open access, RMIT had provided only one record.¹¹

In terms of universities sampled, as of October 2017, UTAS and UWS provide the most open access research data records in Research Data Australia (http://researchdata.ands.org.au/), with 135 and 120 records respectively.¹² The number of open access research data records provided by the UTAS to Research Data Australia (http://researchdata.ands.org.au/) is not surprising given that all data within the UTAS’s research data discovery service is harvested by Research Data Australia (http://researchdata.ands.org.au/) (University of Tasmania, 2018).

The finding that research data is not always accessible is consistent with a finding from a study by Nuijten et al. (2017) which focused on the availability of open access research data from published psychology articles. The study found that from the 474 articles published in two psychology journals which claimed that the research data was available, only 137 articles (approximately 29 percent) actually provided access to the research data. Nuijten, Borghuis, Veldkamp, Dominguez-Alvarez, van Assen, & Wicherts concluded that their findings were ‘comparable to the findings of Chambers (2017, p. 86), and Kidwell et al. (2016). Kidwell et al. (2016) showed that of the articles from journals … that promised open data, only 40.5 percent actually had data available’ (2017, p. 48).

The reasons given by Nuijten et al. (2017) for a lack of access to research data were that the definition or interpretation of data may be ambiguous, concerns over privacy and the need to anonymise data, limited online data storage options and fear by researchers that other researchers may take advantage of their efforts to undertake the research and collect the data.

Nuijiten et al. (2017) also found that, following the introduction of a data sharing policy by one of the journals in the study, the percentage of articles that provided open access to research data increased from approximately 5 percent to 56 percent. Nuijten et al. (2017, p. 31) concluded that ‘open data should follow the FAIR guiding principles (Wilkinson et al., 2016), which state that data should be findable, accessible, interoperable, and reusable’. As noted above in the literature review, in Australia the NHMRC supports the use of FAIR in relation to research data/information (NHMRC, 2018). Similarly, ANDS, the National eResearch Collaboration Tools and Resources project (NECTAR) and RDS have a FAIR self-assessment tool (https://www.ands-nectar-rds.org.au/fair-tool), that allows users to ‘assess the FAIRness of a [research] dataset [or] determine how to enhance [the] FAIRness [of a research dataset] (where applicable)’ (ANDS, RDS, & National eResearch Collaboration Tools and Resources project [NECTAR], n.d. Paragraph 1) and ANDS provides a list of resources for FAIR training.

Training

Responsibility for RDM training appears to be split between libraries and the research or e-research groups in our sample of universities. However, it was not easy to locate information on training courses or session being offered outside the library, as this information was not available on publicly accessible university webpages. Rather the information was only accessible to researchers of the specific universities on internal only webpages. In addition to some hyperlinks not working, many RDM webpages within the sampled universities hyperlinked to other webpages but it was often too difficult to identify the university group which was responsible for the webpage.

The ACU research data management policy specifically states that the University Librarian is responsible for ‘developing and disseminating procedures and providing training and support for university research and administrative units involved in research data management’ (Australian Catholic University, 2017, Clause 5). Deakin has a manage your data webpage which covers a range of topics including data management planning. It has a useful section on sharing of data which can be via Deakin Research Online (DRO) and Research Data Australia (http://researchdata.ands.org.au/). This information is provided by Deakin e-research, which also provides training on RDM techniques. At UWS, the e-research group provided assistance to researchers on RDM.

The e-research group at Flinders, which is located within its Library, is responsible for support on data sharing and in 2017 sponsored seminars with ANDS on supporting data in the social sciences. Griffith Library appeared to be the department offering RDM training. Monash University Library appears to offer training on RDM, as does the QUT Library. At UTAS, joint seminars on data management were run with ANDS, advertised through the Library website.

For research data management training, three universities link to the University of Edinburgh’s MANTRA tool (https://mantra.edina.ac.uk) which can be used by research students, academics and information management professionals. UNE Library provides a link to the School of Data online courses (https://schoolofdata.org/courses/#DataFundamentals).

Conclusion

There does not appear to be a consistent approach to how the universities in our sample are addressing RDM services. The role of e-research groups appears to add a note of confusion to the overall picture, introducing an extra group to what appears to be an already crowded space—research groups, libraries and IT services. In the smaller universities in our sample, the library had a primary role in managing RDM services. All of the universities in our sample have developed RDM policies as would be expected but only three have a specific open access data policy.

There is strong encouragement to store research data securely during and after the project. Nine universities suggested researchers should submit their data to RDA and ten also hosted research data on behalf of their researchers, although generally this was only a record of the data rather than providing access to the data. Training on open access research data appears to be only offered at three universities through the library and appears to be a gap in ensuring RDM is appropriately supported.

Our sample results found that libraries were quite active in the provision of advice to researchers on RDM, particularly in the areas of support for creating metadata and loading data to repositories. The impression gained though this, is that this role is a specialised one, often undertaken by only one person such as a meta-data librarian.

The extent of training provided to researchers was the one area in which there appeared to be insufficient publicly available data to assess the full range of support offered and would appear to be an important area for further research.

From the literature on RDM and given our findings of an inconsistent approach to RDM services and the fact that it is currently being supported by several groups within a university, academic libraries could take a bigger role within this space. This role may not relate to compliance with funding bodies or the Code and the Statement requirements, but could relate to training, development and implementation of RDM plans or making data more accessible for sharing and re-use, particularly as the FAIR data principles gain prominence.

While our sample of Australian universities found that libraries play a key role in providing guidelines, webpages and hyperlinks to RDM resources, it is not clear that library staff have access to the appropriate training and professional development support, to enable them to expand their support roles in this area.

Whether library staff have the appropriate experience, training and professional development to enable academic libraries in Australia to expand their RDM role warrants further research.

Disclosure statement

No potential conflict of interest was reported by the authors.

Additional information

Notes on contributors

Marita Shelly

Dr Marita Shelly is a reference librarian at RMIT University Library. Prior to becoming a reference librarian, she had worked as a research coordinator. Marita has conducted research in the areas of copyright and the legal issues associated with accessing electronic information. She is the co-author (with Emeritus Professor Jackson) of Electronic Information and the Law, Thomson Reuters 2012.

Margaret Jackson

Margaret Jackson conducts research in the areas of protection of and access to information, privacy and copyright. She is the co-author (with Dr G Hughes) of Private Life in a Digital World, Thomson Reuters 2015; co-author (with Dr M Shelly) of Electronic Information and the Law, Thomson Reuters 2012, author of A Practical Guide to Protecting Confidential Business Information, LawBook Co 2003, and author of Hughes on Data Protection in Australia, LawBook Co 2001. 

Notes

1. OECD Declarations are non-binding on member states but are considered persuasive instruments.

2. Note that ANDS and RDS have now joined with the National eResearch Collaboration Tools and Resources project (NECTAR) to align their projects (https://www.ands-nectar-rds.org.au/) and from July 2018 will be known as Australian Research Data Commons (ARDC).

3. The Australian Code for the Responsible Conduct of Research (the Code) has been reviewed and the new Code was released in June 2018.

4. The National Statement on Ethical Conduct in Human Research (the Statement) was last updated in 2015.

5. Note the ARC also released a Guide to Managing and Investigating Potential Breaches of the Code (the Investigation Guide) (https://www.nhmrc.gov.au/guidelines-publications/r42) in June 2018. It aims to release two more guides to support the Code.

6. Within the four FAIR principles, there are a total of 15 facets relating to the data and metadata. For example, the findable principle requires a unique global identifier to be assigned.

7. Originally known as the Ohio College Library Center, now known as OCLC Research.

8. With the release of the 2018 version of the Code, all Australian universities will need to review and update their research policy.

9. At RMIT, the ReDBox application is now known as the RMIT Research Data Catalogue. Research data records added to the RMIT Research Data Catalogue are now automatically included in Research Data Australia (http://researchdata.ands.org.au/).

10. As of July 2018, this figure has increased to 512 records in Research Data Australia.

11. As of July 2018, RMIT still only has one open access record in Research Data Australia.

12. As of July 2018, these figures have increased to 155 and 131 respectively.