Data and Information Management, 2019; 3(2): 84–101

Research Article Open Access

Rong Tang*, Zhan Hu Providing Research Data Management (RDM) Services in : Preparedness, Roles, Challenges, and Training for RDM Practice

https://doi.org/10.2478/dim-2019-0009 received April 19, 2019; accepted June 9, 2019. 1 Introduction

Abstract: This paper reports the results of an international Around the world, research data management (RDM) has survey on research data management (RDM) services in been becoming an increasingly important service that libraries. More than 240 practicing responded a variety of information centers and libraries provide. to the survey and outlined their roles and levels of According to Whyte and Tedds (2011), “Research data preparedness in providing RDM services, challenges management concerns the organization of data, from its their libraries face, and knowledge and skills that they entry to the research cycle through to the dissemination deemed essential to advance the RDM practice. Findings and archiving of valuable results” (p. 1). Two years of the study revealed not only a number of location and later, in the ARL (Association of Research Libraries) organizational differences in RDM services and tools SPEC (Systems and Procedures Exchange Center) Kit provided but also the impact of the level of preparedness 334 entitled “Research Data Management Services,” and degree of development in RDM roles on the types of Fearon et al. (2013) defined RDM services as “providing RDM services provided. Respondents’ perceptions on both information, consulting, training or active involvement in the current challenges and future roles of RDM services data management planning, data management guidance were also examined. With a majority of the respondents during research (e.g., advice on data storage or file recognizing the importance of RDM and hoping to security), research documentation and metadata, research receive more training while expressing concerns of lack data sharing and curation (selection, preservation, of bandwidth or capacity in this area, it is clear that, in archiving, citation) of completed projects and published order to grow RDM services, institutional commitment data” (p. 12). Within the US, a variety of federal funding to resources and training opportunities is crucial. As an agencies or foundations including National Institutes emergent profession, data librarians need to be nurtured, of Health (NIH), National Science Foundation (NSF), mentored, and further trained. The study makes a case Department of Energy (DoE), Department of Education for developing a global community of practice where data (DoED), Environmental Protection Agency (EPA), and the librarians work together, exchange information, help one Alfred P. Sloan Foundation (Sloan) have started requiring another grow, and strive to advance RDM practice around the sharing of research outputs (National Institutes of the world. Health, 2005; National Institutes of Health, 2008) and mandating data management plan (National Science Keywords: research data management, RDM services Foundation, 2010a, 2010b). Similar mandates from and tools, locations and organizational types, level of funding agencies of a variety of other countries include preparedness, degree of development, evolving roles the UK (UKRI, n.d.), Canada (Government of Canada, 2016), Australia (ARC, 2017), European Union (Shearer, 2015), Japan (JST, 2013), and India (Department of Science & Technology, Government of India, n.d.). In the UK, the first data management plan (DMP henceforth) requirement *Corresponding author: Rong Tang, School of and was put in place by the Medical Research Council in 2006, Information Science, Simmons University, Boston, Massachusetts, soon followed by the Wellcome Trust in 2007 (Smale, USA, Email: [email protected] Unsworth, Denyer, & Barr, 2018). The NSF in 2011 was the Zhan Hu, School of Library and Information Science, Simmons first funder in the US to implement a DMP requirement University, Boston, Massachusetts, USA

Open Access. © 2019 Rong Tang, Zhan Hu, published by Sciendo. This work is licensed under the Creative Commons Attribution- NonCommercial-NoDerivatives 3.0 License. Providing Research Data Management (RDM) Services in Libraries: Preparedness, Roles, Challenges... 85

(National Science Foundation, 2010a). In 2013, the Office However, even though there is a growing momentum of Science and Technology Policy released a memo to the in RDM, the level of success in implementing and providing heads of executive departments and agencies requiring RDM services at various data-intensive organizations that each agency develop a plan to promote public has not been consistent. According to Perrier, Blondal, access to research. The public access plan included a and MacDonald (2018), “although libraries play a role requirement that “all extramural researchers receiving in RDM at academic institutions, they have experienced Federal grants and contracts for scientific research and varying degrees of success with the development of RDM intramural researchers develop data management plans, support and services given this expanded responsibility” as appropriate” (OSTP, 2013, p. 5). In August 2017, Smale (p. 173). Furthermore, as noted by Faniel and Connaway et al. (2018) surveyed 16 US national funding bodies and (2018), “research into RDM still is in the early stages, found that 10 required DMPs as did six of seven research and few studies focus on the library community’s RDM councils in the UK (p. 10). experiences” (p. 100). In this study, with a focus on the Given these mandates, universities or research libraries RDM practice in terms of the librarians’ role, their levels from different countries have redefined or extended their of preparedness in providing RDM services, and the role in RDM services. Their efforts have made librarians current RDM services and tools provided, we conducted not only an important contributor to the research process an international survey involving RDM service librarians but also an essential partner in the entire research data around the world who worked in a variety of organizational ecosystem. As pointed out by Choudhury (2008), the types. We also probed into the knowledge and skills new role of librarians in “supporting new forms of data- that respondents deem crucial for RDM services, as well intensive scholarship” (p. 215) has transferred librarians as their vision for RDM roles in the future. Our primary into a “data scientist” or “data humanist.” In such a role, research questions (RQs) were as follows: “they act as the human interface between the library and RQ1. What is the state of current practice of RDM the eScience projects. In a fundamental sense, they may services in libraries? represent the future of subject librarianship and help craft RQ2. What current role do librarians play in providing a new relationship between the library and scientists” (p. RDM services? 217). As an example, in health science libraries, the demand RQ3. What specific knowledge and skills do librarians for RDM services has been very strong. As indicated by believe as needed for RDM training? Martin (2013), “As biomedical science becomes more RQ4. How do participants see as the future evolutions data-intensive, researchers are faced with a range of data of the librarians’ role in providing RDM services? management challenges, problems, and needs. Health sciences librarians are ideal partners for offering scientists at their institutions a range of data management services” (p. 1). Data services provided by health science libraries 2 Literature Review may vary from creating “a comprehensive data dictionary” and “a standardized data request form” (Gore, 2013, p. 2.1 Components and Models of RDM 22) to assign metadata to incoming data, naming and Services labeling research variables, and finding a data repository (Hasman, Berryman, & Mcintosh, 2013). As reported in by Empirical research studies on RDM services in libraries Hanson et al. (2013), librarians are now providing “the tend to use a similar set of definitions for RDM. A deepest and most detailed data management support” to frequently cited definition is by Cox and Pinfield (2014), scientific research teams “to date” (p. 25). which states that RDM “consists of a number of different It should be noted that data services are also provided activities and processes associated with the data lifecycle, in social science and humanities disciplines. In the involving the design and creation of data, storage, security, “Special Report: Digital Humanities in Libraries” by preservation, retrieval, sharing, and reuse, all taking into Varner and Hswe (2016), the authors argue that with the account technical capabilities, ethical considerations, rapid evolution of digital humanities and the uncertainty legal issues and governance frameworks” (p. 300). Both of its future trajectory, “it may be more productive—and this definition and the one by Fearon et al. (2013) as cited more honest—to position the library as a research partner earlier define RDM through a series of activities associated that can explore new solutions with researchers rather with data lifecycle. Meanwhile, the conceptualization of than as a service provider that either has what a researcher RDM is enriched through the development of frameworks is looking for or doesn’t” (para 9). or process models of RDM services. For example, in their 86 Rong Tang, Zhan Hu

model of institutional RDM.” The model features factors including drivers (why), components (what), influencing factors (how), and stakeholders (who). Those within the stakeholder section include institutional units such as the library, IT services, academic departments, senior university managers, research support services, and other support services. Researchers from various disciplines are listed as stakeholders. The examples for drivers include storage, security, preservation, and compliance, whereas influencing factors consisted of demand, roles, resources, acceptance, and communications, among others. The authors indicated that this model aligns well with previous models such as the one by Jones et al. (2013) and Whyte (2014). Its uniqueness is that it “deals more with complexity of the underlying drivers and influencing factors that could shape which types of specific services Figure 1. Components of research data management support are developed and what they might look like” (p. 24). services (Source: Jones et al., 2013, p. 5) Another notable model is an RDM maturity model by Cox et al. (2017). In this model, four levels of RDM maturity working level guide for the Digital Curation Centre, Jones, are specified, with Level 0 as “none,” Level 1 as “basic,” Pryor, and Whyte (2013) created a model outlining the Level 2 as “developing,” and Level 3 as “extensive.” These components of an RDM service (see Figure 1). As pointed levels are categorized according to the existence or absence out by Jones et al. (2013), “In order to support effective of services and support, compliance, skills, roles and data management and sharing, an institution needs a structures, embedded practices, and cultural acceptance. coherent strategy and suite of services” (p. 5). In the RDM All the models reviewed in this section provided relevant components diagram, the overarching activities include ideas for our current study. However, even though several “RDM policy and strategies” and “business plan and of these models hold valuable practical implications, sustainability.” Under the RDM service establishment, there has been rather limited empirical research various levels of guidance, training, and support are investigating and confirming the constructs such as required. The core RDM process features the service drivers, components, and influencing factors, which components of data management planning, managing are associated with RDM. Even with the RDM maturity active data, data selection and handover, data repositories, model outlined by Cox et al. (2017), substantial empirical and data catalogs (p. 5). evidences are needed to test, verify, or enhance the model. Building upon Jones et al.’s RDM components model, Whyte (2014) presented a process pathway model to illustrate, among other things, the steps required in 2.2 RDM Services in Academic Libraries developing an RDM service. In Whyte’s model, higher level factors, which are positioned together, include Although information professionals served a role in context, principles, inputs, outputs, and outcomes. data management before the proliferation of DMP Six steps are included within “outputs”: envision, requirements, the adoption of such requirements provided initiate, discover, design, implement, and evaluate a “catalyst” for academic libraries and librarians to (p. 60). In the same “output” section, actors involved expand their data management services (Dietrich, Adams, in RDM are listed as researchers, research users, data Miner, & Steinhart, 2012). According to Cox et al. (2017), managers, research support, IT support, librarians, from 2013 to 2016, the number of academic libraries with archivists, compliance officers, senior managers, funders, an RDM policy in Australia nearly doubled from 29% to and others. Even though Whyte’s pathway model is 56%, while in the same time period, the number in the relevant and comprehensive in listing diverse research UK more than doubled from 17% to 42%. Furthermore, communities, it does not, as pointed out by Pinfield et al. the number of academic libraries providing data literacy (2014), address “any constraints on the service delivery training from 2013 to 2016 also increased dramatically, side” (p. 4). Based on their qualitative investigation, from 26% to 74% in Australia and from 14% to 65% in the Pinfield et al. (2014) developed “a library-oriented UK. Similar comparisons for the US were conducted but Providing Research Data Management (RDM) Services in Libraries: Preparedness, Roles, Challenges... 87 with different results. According to Tenopir et al. (2015) copyright/intellectual property rights (IPR) for data and who conducted the same survey in 2011 and 2015, there on running a DR. was no significant growth in the number of RDM services In a study of services at 128 academic institutions in provided by academic libraries in the US. the US, Tenopir et al. (2015) found that most institutions A number of international surveys shed interesting did not offer or plan to offer a wide range of RDM services. light on the growth of RDM services and tools. In 2013, Consultative services such as providing reference and Fearon et al. (2013) conducted a SPEC survey among support for citing data (29.5%) and publishing web ARL member institutions to understand current RDM guides for data management (21.5%) were offered more activities in ARL libraries. Out of the 125 member libraries, frequently than technical services such as providing 73 responded to the 67-question survey covering topics technical support for RDS systems (approximately 15%) of RDM services, data archiving services, RDM services, and creating or transforming metadata (less than 10%). staffing, education and skills, training needs, and funding Research results concerning institutions in the US RDM services and partnerships. The results showed differ strongly from findings from a similar study based that RDM services and support for data archiving were in Europe. Tenopir et al. (2017) surveyed the prevalence of the emerging services at the member libraries. Around research data services among members of the Association 54 ARL libraries already provided RDM services, with of European Research Libraries (LIBER). The findings varying capacities, and 17 libraries reported planning revealed that 45% of libraries surveyed provided services for RDM service provision in the near future. Of the in the form of consultations with researchers and students RDM activities conducted, the focuses were on DMP regarding DMP, 43.5% provided consultations on metadata for grant applications, data management consultation, standards for RDM, and 36.6% provided reference support. and research data archiving. Another important finding The authors concluded that, while disappointing, this from the survey was the need for collaborations between gap between the provision of RDM services in the US libraries and other units within the institution. and EU might be due to the earlier adoption of DMP Meanwhile, in a survey of 500 science librarians at requirements in the EU. It is also possible that there were ARL-affiliated libraries, Antell et al. (2014) found that regional differences in libraries’ abilities to overcome the 94.9% of librarians were aware of the 2011 NSF requirement challenges of implementing research data services. that researchers needed to submit a DMP. Ninety percent Overall, it is clear that since the emergence of RDM of the respondents reported that their institutions had an services in libraries, there have been research studies institutional repository (IR). Only 23.5% indicated that they surveying the kinds of RDM services that libraries of had a data repository (DR henceforth) while another 27.1% varying geographical locations provide. The comparative indicated that a DR was being planned. Of those surveyed, investigations into RDM services in the US and Europe 60.1% provided some kind of data management services seemed to show that EU was more advanced than the US. to researchers. Such services were being planned in 17.8% A broader comparative study examining the differences or of cases. It was also found that 39.5% of those surveyed gaps among RDM practices in the US and in other non-US had duties related to the IR, DR, or data management. world regions would be valuable. Furthermore, within the The most common duties were “liaise, consult, and refer” US, there has been no study exploring the differences or (39%); “just starting” (16%); and “promote, publicize, or gaps of RDM services among the US regions. Specific focus advocate” (14%). on individual regions and the types of RDM services they In a survey of six European countries and Canada provide would produce helpful information to identify totaling 170 responses, Cox et al. (2017) found that the both the gaps in services and the RDM niche associated most common service across regions was the provision with each geographical region within the US. of web resources or guides for RDM. RDM training or data literacy training was a growing service that was considered as either basic or well developed depending on 2.3 Challenges in Providing RDM Services the institution. Some services were offered by only a few institutions, including promoting awareness of reusable Many empirical studies on RDM services investigated the data , data publication advisory services, data challenges that libraries or librarians face. For example, storage advisory services, and providing access to tools to in the survey by Fearon et al. (2013), it was reported that assist with RDM. The top strategic priorities for services establishing and maintaining effective partnerships and indicated by respondents included giving advice on fostering campus-wide collaboration were mentioned by 88 Rong Tang, Zhan Hu many respondents (38%) as the top challenge, followed in RDM services (e.g., Cox et al., 2017; Tenopir et al., by securing financial resources and funding (35%), 2017), it is necessary to examine the extent to which RDM increasing the engagement with faculty and researchers practices vary by locations. The revelation of the location (31%), and providing technology infrastructure (27%). differences will help to identify the impact of cultural and Limited staffing (25%), marketing services (25%), and staff service infrastructure on location-based practices. The training (23%) were also among the top six challenges. strengths and niche that are identified to be associated The need to provide training for library staff was also with location-based services would provide guidance for recognized in the article by Tenopir et al. (2015), as the institutions lacking such strengths to improve and create authors pointed out that “in order to fully offer technical their unique RDM service structures. In the same vein, RDS, libraries need to have technologically skilled staff examining the differences among organizational types or greatly increase opportunities for technology training such as academia versus non-academic health sectors for their existing staff, which might not be feasible due to will help various types of institutions to learn from one resource constraints” (p. 17). another and to develop their own effective strategies Meanwhile, Cox et al. (2017) indicated that one of for providing RDM services. Meanwhile, probing into the major concerns related to the development of RDM measures of preparedness and degrees of development services across various countries was that technical- would enable us to operationalize RDM maturity levels as oriented services such as providing a data catalog per Cox et al. (2017), based on which, we would be able and curating active data were rather limited. Data to further examine the variations in kinds of services and curation skills and capabilities were not consistently tools provided by librarians working in environments in place. Other challenges identified by Cox et al. with varying levels of RDM maturities. These empirical (2017) “include resourcing, working with other support constructs are important and worthwhile to pursue, not services, and achieving ‘buy-in’ from researchers and only because they have not been investigated much in senior managers” (p. 2182). Similar findings were the existing RDM literature but also due to the valuable reported in the study by Faniel and Connaway (2018). insights and enriched understanding that they afford. The In conducting individual or focus group interviews present study intends to address this gap by developing an that involved 36 library professionals in 2012 and 2013, in-depth understanding of the RDM services around the Faniel and Connaway (2018) found that there were five world and examining various differences and challenges factors contributing to librarians’ RDM support: human surrounding the provision of RDM services. resources, communication, coordination, collaboration, technical resources, leadership support, and researchers’ perceptions of the library. Specifically related to technical resources, 36% of librarians indicated that the solutions 3 Methodology related to data storage and preservation presented serious challenges, particularly with regard to the financial 3.1 Research Variables commitment required to sustain a long-term solution. In terms of human resources, the constraints noted included The current study included a number of quantitative “demands on time” and “limited number of expert staff.” variables. As reviewed earlier, even though there were For 25% of the participants, researchers’ misperception research studies examining RDM practices and policies about the library and librarians’ expertise was a major through a global lens, seldom has any research compared constraint for RDM services. These findings echo with country or location differences in RDM services. In this other studies such as Cox et al. (2017) and Fearon et al. study, because we are interested in investigating the (2013). location and organizational differences in RDM services Despite a great number of studies on RDM services and tools, we have location groupings and organizational in libraries and the perceptions of faculty or librarians types as our independent variables. In terms of locations, in their endeavor of working collaboratively on RDM, in addition to grouping participants by world regions, there have been limited studies investigating into the we also categorized people based on whether they were differences in locations, organizational types, level of from the US or non-US countries. For people who were preparedness, and the degree of development associated located in the US, we were also interested in exploring the with the types of services and kinds of training that differences among the five regions (National Geographic librarians and researchers should provide and receive. Society, 2009). For organizational types, we had four With the evidence that there have been global-level efforts categories including “university/college,” “government Providing Research Data Management (RDM) Services in Libraries: Preparedness, Roles, Challenges... 89

Table 1 Variables Used in This Study

Independent Variables Dependent Variables

Location Data management services ● World regions ● RDM planning ● USA vs non-USA ● Data discovery and access ● US regions ● Data organization and curation Organization type ● Metadata ● Academic versus non-academic ● Protocol documentation Level of preparedness ● Data sharing and dissemination ● Prepared versus unprepared ● Data preservation Role development ● Data visualization ● Five-point scale development level from not developed to very developed RDM tools ● Electronic lab notebooks ● Data citation manager ● DR ● Data search engine ● Data processing software (R, Python, etc.) ● Preprints Challenges

agency,” “non-academic health sector,” and “others.” For in RDM, knowledge and skills needed for RDM, and more. this particular variable, we were interested in comparing The questionnaire was hosted on SurveyMonkey. On participants from academic settings with those who May 23, 2018, a call for participation was sent to various were outside academia. The remaining two independent communities and international conferences’ variables were levels of preparedness in providing RDM attendees through email, blog posts, listservs, and word services and the degree of development that participants of mouth. Library Connect, an Elsevier outreach program perceived that their institutional RDM services belonged. for librarians, was used to distribute the online survey. A These two variables are related to RDM maturity (Cox et total of 241 responses were received. Answers to individual al., 2017), which are good indicators of the differences questions were not mandatory, so the number of responses in RDM services. The dependent variables of this study to each question varied. The two authors processed survey included various forms of RDM services and tools data, coded responses, and analyzed the data. Results are provided in a given institution and respondents’ view of outlined in the next section. the kinds of challenges that they face in delivering RDM services. All the research variables that were included in our inferential statistics analysis are listed in Table 1. 4 Results

3.2 Research Questions 4.1 Respondents

In this study, we pursue the answers to the following RQs, 4.1.1 Locations which include four primary RQs and several sub-RQs under each primary RQ. Table 2 lists all the RQs. The 241 respondents came from five continents and 29 countries. As shown in Figure 2, 198 (82.2%) responses were from North America, 21 (8.7%) were from Europe, 13 3.3 Procedure (5.4%) were from Asia/Pacific/Middle East, five (2.1%) were from Africa, and four (1.7%) were from South America. From late May 2018 to early September 2018, we conducted Figure 3 shows the distributions of the respondents by an international survey featuring an online questionnaire country in the format of a pie chart. Seventy-eight percent containing 19 questions that inquired into the current of the participants were from the US. practice of RDM in libraries, how prepared librarians were 90 Rong Tang, Zhan Hu

Table 2 RQs and Sub-RQs Investigated in the Study

RQ Primary RQ Sub-RQs

RQ1 What is the state of current 1.1 What specific RDM services and tools are currently provided? practice of RDM services in 1.2 What aspects of RDM services and tools provided differ among various locations and libraries? organizational types? 1.3 What challenges do libraries face in providing RDM services?

RQ2 What current role do librarians 2.1 Do participants consider that they are personally prepared in providing RDM services? play in providing RDM services? 2.1.1 What aspects of RDM services and tools provided different among participants who indicated that they are prepared or unprepared to provide RDM services? 2.1.2 Are there statistically significant differences between participants who considered themselves as “prepared” and the kind of challenges they perceived in providing RDM services? 2.2 How developed do participants perceive their current role in RDM services, and do they believe that they should take on a more formal role? 2.2.1 What aspects of RDM services and tools provided different among participants who reported their RDM role at different degrees of development?

RQ3 What specific knowledge and 3.1 What sources do librarians use to identify RDM training opportunities? skills do librarians believe are 3.2 What are the most important and useful knowledge and skills in RDM training? needed for RDM training? 3.3 What content areas and specific skill sets are useful for RDM training targeted to librarians? 3.4 What content areas and specific skill sets are useful for RDM training targeted for researchers?

RQ4 How do participants see as 4.1 What are the most common future RDM roles that respondents identified? the future evolutions of the librarians’ role in providing RDM services?

Figure 2. Locations of survey participants Providing Research Data Management (RDM) Services in Libraries: Preparedness, Roles, Challenges... 91

Figure 3. Distributions of survey participants by country

Table 3 Distribution of Survey Participants by US Regions (n=186, plus 222 respondents who belonged to an institution, more Puerto Rico) than half were from an academic institution (n=152, 63.1%) and seven (2.9%) were from a non-academic US Regions Counts (Percentage) health sector. In our inferential statistics analysis, we compared responses from academic participants with Northeast (DC included) 70 (37.63) those from non-academic participants. Furthermore, West 43 (23.12) for individual questions, there were responses from Midwest 32 (17.20) participants of the same institutions. We compared

Southeast 26 (13.98) responses from participants of the same institutions with regard to factual information such as the kinds of Southwest 15 (8.06) RDM services and tools that a given institution provided; there were variations in these respondents’ answers. We believed that the varied responses were due to either The 187 US respondents came from 40 different states the respondents were from different libraries within and District of Columbia (DC) as well as Puerto Rico (n=1) an institution or the librarians had different levels of (see Figure 4 for distribution). The US respondents were awareness of the services and tools that a given library from all five US regions. The Northeast (n=70, 37.6%) provided. Because our study is based on librarians’ had the highest number of participants and Southwest perceptions and self-reported data, we decided to treat (n=15, 8.1%) had the lowest. Table 3 lists the number of these responses from the same institution as the same as participants from various US regions. they were from different institutions.

4.1.2 Organization Types 4.2 RDM Services and Tools Provided in Institutions The 241 respondents were from four types of institutions, including universities/colleges, government agencies, 4.2.1 RDM Services non-academic health organizations, and other types of organizations (see Table 4 for distribution). The In answering the question of “What RDM services does affiliations of 19 respondents were not available. Of your institution offer?” 63 respondents reported that their 92 Rong Tang, Zhan Hu

Figure 4. Distribution of survey participants by state

Table 4 Distribution of Survey Participants by Organization Types (n=241) most frequently offered services include RDM planning (n=51, 81.0%) and data sharing and dissemination Organizational Types Counts (Percentage) (n=49, 77.8%). Table 5 lists the range of RDM services that participants reported that their institutions provided. Universities/Colleges 152 (63.07) Through a series of Chi-square tests, statistically Government Agencies 22 (9.13) significant differences were found in terms of location, Non-Academic Health Sectors 7 (2.90) organization type, and the kinds of RDM services that librarians provided through their institutions. In the Other Types of Organizations 41 (17.01) following sections, we outline these differences. Not Available 19 (7.88)

Table 5 4.2.1.1 Location Differences – World Regions RDM Services Provided in Participants’ Institutions (n=63) Significant differences among North America, Europe, and Asia-Pacific regions were found in terms of the provision RDM Services Provided Counts (Percentage) of data discovery and access services (χ2(2, n=60)=6.085, 2 RDM planning 51 (80.95) p=0.048, V=0.32) and data visualization services (χ (2, n=61)=17.253, p=0.000, V=0.53). Significantly higher Data sharing and dissemination 49 (77.78) proportions of institutions in North America (75.0%) Data preservation 42 (66.67) provided data discovery and access services than those Data discovery and access 41 (65.08) in Asia-Pacific (66.7%) and Europe (33.3%). In addition, Metadata 41 (66.67) significantly higher proportions of institutions in North America (71.4%) provided data visualization services than Data visualization 37 (58.73) those in Europe (11.1%) and Asia-Pacific (0.0%). Data organization and curation 37 (58.73)

Protocol documentation 20 (31.75) Providing Research Data Management (RDM) Services in Libraries: Preparedness, Roles, Challenges... 93

Table 6 4.2.1.2 Location Differences – USA versus Non-USA RDM Tools Available in Participants’ Institutions (n=57) Significant differences between USA and non-USA countries were found in terms of the provision of data RDM Tools Provided Counts (Percentage) organization and curation services (χ2(1, n=61)=4.736, p=0.030, V=0.28), metadata services (χ2(1, n=61)=4.891, Data repositories 44 (77.19) p=0.027, V=0.28), and data visualization services (χ2(1, Data processing software 38 (66.67) n=61)=11.716, p=0.001, V=0.44). Significantly higher Data citation manager 23 (40.35) proportions of institutions in the USA (66.0%) provided Electronic laboratory notebooks 21 (36.84) data organization/curation services than those in non-USA (35.7%) locations. Significantly higher proportions of Preprints 18 (31.58) institutions in the USA (72.3%) provided metadata services Data search engine 16 (28.07) than those in the non-USA countries (42.9%). Finally, significantly higher proportions of institutions in the USA (70.2%) provided data visualization services than those in 44 indicated that they provided data repositories (n=44, the non-USA countries (21.4%). 77.2%). Other frequently mentioned tools were data processing software such as R, Python (n=38, 66.7%), data citation manager (n=23, 40.4%), and electronic lab 4.2.1.3 Location Differences – US Regions notebooks (n=21, 36.8%). Table 6 lists the RDM tools Significant differences among five US regions were found selected by participants as being provided through their in terms of the provision of protocol documentation institutions. services (χ2(4, n=47)=16.599, p=0.002, V=0.59) and Through a series of Chi-square tests, statistically data preservation services (χ2(4, n=47)=9.598, p=0.048, significant differences were found in terms of location, V=0.45). Significantly higher proportions of institutions organization type, and the kinds of RDM tools that in the Midwest (66.7%) provided protocol documentation institutions provided. In the following sections, we outline services than those in the West (58.3%), Southeast (50.0%), these differences. Northeast (8.7%), and Southwest (0.0%). Meanwhile, significantly higher proportions of institutions in the West (100.0%) provided data preservation services than those 4.2.2.1 Location Differences – World Regions in the Midwest (83.3%), Northeast (52.2%), Southeast Significant differences among North America, Europe, and (50.0%), and Southwest (50.0%). Asia-Pacific regions were found in terms of the provision of data processing software (χ2(2, n=55)=8.054, p=0.018, V=0.38). Significantly higher proportions of respondents 4.2.1.4 Organizational Differences from institutions in North America (75.6%) reported that Significant differences were found between academic their institutions provided data processing software than institutions and non-academic institutions in terms of those from Europe (57.1%) and Asia-Pacific (0.0%). service provision of RDM planning (χ2(1, n=58)=7.344, p=0.007, V=0.36) and data preservation (χ2(1, n=58)=3.888, p=0.049, V=0.26). Significantly higher proportions of 4.2.2.2 Location Differences – USA vs. Non-USA respondents working in universities (90.7%) reported Statistically significant differences between US that their institutions provided RDM planning services institutions and non-US institutions were found in terms than those working in non-academic organizations of providing data processing software (χ2(1, n=55)=5.406, (60.0%). Additionally, significantly higher proportions of p=0.020, V=0.31). Significantly higher proportions of respondents working in universities (74.4%) reported the respondents working in US institutions (86.8%) reported provision of data preservation services than those working that their institutions provided data processing software in non-academic organizations (46.7%). than those working in non-US organizations (13.2%).

4.2.2 RDM Tools 4.2.2.3 Organizational Differences University and non-academic institutions differ With regard to the RDM tools offered by their institutions, significantly in terms of the provision of data processing among the 57 participants who answered the question, software (χ2(1, n=53)=5.487, p=0.019, V=0.32). Significantly 94 Rong Tang, Zhan Hu

Figure 5. Summary of significant location and organizational differences in RDM services and tools provided

Figure 6. Common challenges identified by respondents (n=54) higher proportions of respondents working in universities 4.3 Challenges in Providing RDM Services (76.7%) reported that their institutions provided data processing software than those working in non-academic In responding to the question “What challenges does organizations (41.7%). your institution face in offering those services?”, among In summary, a number of significant location or the 54 participants who provided their answers, the organizational differences were found in terms of the most frequent theme was “capacity/bandwidth; limited provision of the data services or RDM tools. Figure 5 staffing” (n=28, 51.9%), followed by “marketing and presents an aggregated view of various significant outreach of RDM services” (n=16, 29.6%) and “collaborative differences. understanding among campus departments” (n=16, 29.6%). Other common responses included “upskilling Providing Research Data Management (RDM) Services in Libraries: Preparedness, Roles, Challenges... 95 staff,” “providing consistent service in terms of quality n=54)=4.441, p=0.035, V=0.29). A significantly higher and options,” and “handling researcher’s and faculty’s proportion of the respondents who felt they were misconception of RDM and library services.” Figure 6 “unprepared” identified “bandwidth and capacity” shows the categories of the common responses. as a challenge (77.8%) than those in the “prepared” category (38.9%). A significantly higher proportion of the respondents who felt they were “unprepared” also 4.4 The Role of Librarians in RDM Services identified “upskilling staff” as a challenge for them (38.9%) than those in the “prepared” category (2.8%). However, a 4.4.1 Level of Preparedness significantly higher proportion of the respondents who felt they were “prepared” identified “marketing outreach A total of 87 respondents answered the question “Do you and awareness” as a challenge (38.9%) than those in the personally feel prepared to offer RDM services?” Among “unprepared” category (11.1%). them, 61 (70.1%) said “yes,” whereas 26 (29.9%) indicated “no.” For those who said no, in their explanation of the reasons for feeling unprepared, they listed “lack of 4.4.2 Degree of Development in RDM Role training” (n=8, 30.8%), “lack of knowledge and skills” (n=7, 26.9%), “only comfortable with providing basic A total of 239 respondents answered the question “how service” (n=7, 26.9%), and “not prepared for specific/ developed is your role with your institution’s RDM?” using advanced RDM areas” (n=5, 19.2%). Other reasons listed a 5-point scale with 1 being “not developed” and 5 being included “lack of experience,” “job title was not Data “very developed.” Among these respondents, 89 (37.2%) Librarian,” “lack of time and resources,” and “lack of indicated that their role was “not developed” and 19 (8.0%) cooperation of researchers.” stated their role as “very developed.” Figure 7 displays the Through a series of Chi-square tests, statistically distribution of the response categories. Parallel to the fact significant differences were found in terms of the type that less than 8% of respondents felt that their role was of RDM services provided and the challenges that the “very developed” when responding to the question “Do you respondents identified. In the following sections, we wish you had a more formal role with RDM?”, 121 (88.3%) outline these differences. said “yes” and only 16 (11.7%) said “no.” The contrast between the responses on preparedness and these two questions suggests that even though the librarians did not 4.4.1.1 RDM Services feel that their RDM role was fully developed and wished to Significant differences between respondents who indicated have a more formal role, they personally felt more prepared that they were prepared and those who believed that they to provide such services. were unprepared were found in terms of the provision of protocol documentation services (χ2(1, n=61)=7.864, p=0.005, V=0.36) and data preservation services (χ2(1, 4.4.2.1 Organizational differences in role perception n=61)=4.051, p=0.044, V=0.26). A significantly higher A Chi-square test indicated that there is a significant proportion of the respondents who felt they were prepared difference between university/academic institutions and provided protocol documentation services (42.5%) than non-academic organizations in respondents’ perception those in the “unprepared” category (9.5%). A significantly on whether they wished to assume a more formal role in higher proportion of the respondents who felt they were RDM (χ2(1, n=126)=5.449, p=0.020, V=0.21). Significantly prepared provided data preservation services (75.0%) higher proportions of respondents working in non- than those in the “unprepared” category (52.4%). academic organizations (97.6%) wished to assume a more formal role in RDM than those working in universities/ academic institutions (83.3%). 4.4.1.2 Challenges in providing RDM services Significant statistical differences were found in the level of preparedness and the challenges that respondents 4.4.2.2 Level of preparedness and degree of identified in providing RDM services, specifically in development in RDM role “bandwidth and capacity” (χ2(1, n=54)=7.269, p=0.007, Significant differences were found in the level of V=0.37), “upskilling staff” (χ2(1, n=54)=12.399, p=0.000, preparedness with regard to their perception of the degree V=0.48), and “marketing, outreach and awareness”(χ2(1, of development in RDM services (χ2(2, n=126)=9.870, 96 Rong Tang, Zhan Hu

4.5 Essential RDM Skills, Knowledge, and RDM Training

A number of questions in the survey probed the issue of RDM training opportunities, the essential knowledge and skills needed, and what RDM training was needed for librarians and for researchers.

4.5.1 RDM training opportunities

A total of 68 respondents answered the question concerning where they learned about RDM training opportunities. More than 55% of the respondents (n=40, 58.8%) indicated that they identified training opportunities through “email listservs.” Other sources included announcements from associations (n=16, 23.5%), messages from organizations Figure 7. Degree of development in RDM role (n=239) (n=15, 22.1%), exchange from experts or colleagues (n=12, 17.7%), by searching online for workshops (n=12, 17.7%), receiving feeds from social media (n=11, 16.2%), through p=0.007, V=0.34). Significantly higher proportion of the word of mouth (n=8, 11.8%), and via conferences (n=8, respondents in the “developing” category indicated 11.8%). When asked “how likely would you participate that they were “unprepared” (43.5%) than those in in online training on RDM?,” 111 (92.5%) out of 120 the “developed” category, indicating that they were selected “very likely” or “somewhat likely” and three “unprepared” (21.7%), and those in the “very developed” (2.5%) selected “somewhat unlikely” or “very unlikely.” category, indicating that they were “unprepared” (5.6%). Six (5.0%) participants selected “neutral.” In terms of whether getting academic credit would motivate them more in getting the RDM training, the answer was nearly 4.4.2.3 RDM services and degree of development in RDM 50/50 with “yes” as 52.5% and “no” as 47.5%. role Significant differences among respondents of different degrees of development in their RDM role were found 4.5.2 Essential RDM knowledge and skills in terms of the provision of data organization and curation services (χ2(2, n=61)=10.963, p=0.004, V=0.42), Sixty-three respondents answered the question “What metadata services (χ2(2, n=61)=9.666, p=0.008, V=0.40), three things have you learned about RDM that was and protocol documentation services (χ2(2, n=61)=6.318, absolutely mandatory?” The most frequent responses p=0.042, V=0.32). A significantly higher proportion were “data/file documentation” (n=17, 26.9%), “metadata” of the respondents in the “very developed” category (n=13, 20.6%), and “DMPs” (n=13, 20.6%). Figure 8 indicated that they provided data organization and includes the common responses. curation service (81.8%) than those in the “developed” As to what further RDM training was needed, the category (80.0%) and those in the “developing” category most frequent answers were “advanced data management (42.9%). Meanwhile, a significantly higher proportion of skills (e.g., data analysis, preservation, acquisition and the respondents in the “developed” category indicated “de-identification)” (n=15, 22.7%), “learning about DM that they provided metadata service (86.7%) than those or Open Source tools” (n=14, 21.2%), “hands-on projects” in the “very developed” category (81.8%) and those in (n=8, 12.1%), and “RDM related policies and regulations” the “developing” category (51.4%). A significantly higher (n=7, 10.6%). Figure 9 includes all the response categories proportion of the respondents in the “very developed” concerning this question. category indicated that they provided protocol service (63.6%) than those in the “developed” category (26.7%) and those in the “developing” category (22.9%). Providing Research Data Management (RDM) Services in Libraries: Preparedness, Roles, Challenges... 97

Figure 8. Essential RDM knowledge and skills

Figure 9. Common responses to what further RDM training needed 98 Rong Tang, Zhan Hu

Table 7 RDM Training for Librarians and for Researchers

RDM Training for Librarians (n=61) RDM Training for Researchers (n=68)

● Data services/science skills (13, 22.95%) ● Data management/data processing (12, 17.65%) ● Skills regarding data reference interview (11, 15.3%) ● Data storage/data preservation (11, 16.18%) ● Basic training (8, 13.11%) ● DMP (9, 13.24%) ● Connecting with the researcher and faculty (8, 13.11%) ● Data sharing/data dissemination (9, 13.24%) ● Depending on the knowledge that the librarians have and their ● Data file/documentation (8, 11.76%) institutional needs (6, 9.84%) ● All/everything/a lot (7, 10.29%) ● All levels of trainings (5, 8.20%) Education/training (7, 10.29%) Software/systems/tools (7, 10.29%) Metadata (7, 10.29%)

4.5.3 Training for Librarians and Researchers and discoverable. As collections are seen as data and data as collections and as researchers now want to use Sixty-one participants answered the question about RDM natural language processing and machine learning training for information specialists, and 68 answered the on subscription library resources – we are key parts question about RDM training for researchers. As seen in of the ecosystem.” A great number of participants saw Table 7, participants’ top six responses about RDM training “supporting researchers” for their RDM needs as a needed for librarians differed from the top responses continuing role for the future. As commented by P201, “It about RDM training needed for researchers. While depends on the institution, but I see librarians becoming librarians were mostly interested in learning about data more hands-on in helping researchers organize their service skills, the other frequently mentioned categories data, and learning/teaching software and tools.” Another were also service-oriented or outreach activities. On the respondent (P170) described an entire suite of researcher’s other hand, respondents believed that researchers should activities that RDM libraries may support: “I see librarians get rather specific RDM training such as management, in a position of advisor and reviewers. As a librarian, we process, storage, preservation, DMP, data sharing and can give researchers all the keys to start their study with dissemination, and more. a good architecture, support them in the DMP writing, answer questions during the study, help them to identify sensitive data, give information about legislation, curate 4.6 RDM Evolving Role the data, review an article before publication (see if the data given are enough for reproducibility), give advice on Fifty-three participants shared their thoughts on the metadata, the data, share the data.” Meanwhile, “evolving role of librarians in the context of RDM.” Sixteen several respondents believed that it is very important (30.2%) stated that one aspect of librarians’ continuing to connect with other divisions on campus to provide role is to “support researchers with RDM,” while 14 (26.4%) a full range of RDM services. As pointed out by P137, acknowledged the “importance of the role” librarians “partnering with others on campus for a full-fledged/ play in RDM services. Twelve (22.6%) participants all services approach – need to work more with ITS and suggested that librarians need to “connect with others in others providing more data analysis help, and storage the institution” and form “partnership,” and 11 (20.8%) capacity.” The value of extended collaboration across believed librarians should “embed in research/data campus was also advocated by P199 who stated that lifecycle.” Other common responses included “teaching “Working in partnership with other offices, directives, RDM” (n=7, 13.2%), providing “consultation” (n=6, 11.3%), and departments on campus; focus on training and and “collaborating with IT departments” (n=5, 9.4%). keeping up with new tech developments such as ELNs.” Several respondents stressed the value and Another future development seen by several contribution of librarians to RDM. As stated by a respondents was to embed the RDM librarian and the participant (P240), “I am the only librarian by degree services they provide into research data lifecycle. As noted in a data services group of six people; I think data viz, by P189, “Eventually, we will be more embedded in the GIS, DH, etc. are attracting a wide variety of people, but process as consultants and trainers.” P130 also indicated RDM is a place where librarians excel because it speaks that “I’d like to see librarians as partners in research to our strengths for making information documented groups, involved throughout the research data lifecycle.” Providing Research Data Management (RDM) Services in Libraries: Preparedness, Roles, Challenges... 99

Figure 10. Statistically significant differences found in this study

P215 declared a similar sentiment: “I hope that our best AI will be huge and determining significant roles will be and brightest don’t leave the profession to become ‘data important.” scientists’ and that data librarians become integral to Overall, participants’ qualitative comments in their researchers’ research lifecycle.” Parallel to P215’s discussing the evolving role of librarians in providing RDM concern of losing the RDM librarian workforce, P191 also services are all rooted in actual practice and corresponded articulated the challenge of misconception of researchers well with their responses in other questions such as the about librarian role – “Honestly I’m worried about the kinds of services and tools they provided, challenges future of RDM as a librarian role, mainly because in my and the level of preparedness, and essential skills and experience it’s such a challenge to get researchers to think knowledge needed for RDM. of libraries and librarians that way.” Multiple respondents also expressed their vision of incorporating data access and discovery into the 5 Discussion and Conclusion existing library search and discovery systems. While P151 discussed their concerns with the lack of efficient In this study, we investigated extensively current RDM data discovery, data description, and data depository practice offered from institutions around the world. mechanism by stating “I’m guessing librarians are going We identified a number of location and organizational to have to be as facile with data as they are with other differences in the RDM services and tools provided as well products of scholarship. Data discovery will continue to as the impact of the level of preparedness and degree of be a huge challenge, as data are not well described and development in RDM roles on the types of RDM services are deposited all over creation without standardized provided. Figure 10 lists all the statistically significant metadata,” P029 claimed that “It’s [data is] a new content differences found in this study. As it is shown, not only type and we must set up our systems to enable search and locations and organizational types had an impact on discovery like we do for other research output.” Several RDM services and tools offered but also participants’ participants pointed out that RDM role in the future perception of their levels of preparedness and degrees of will be more technologically driven, as commented by development impacted the kinds of services they provided P088, “Library roles become increasingly technical and and the kind of challenges they perceived. technology-focused.” Interestingly, a few participants We also examined respondents’ perception on both suggested paying attention to the advancement of AI and the current challenges and future roles of RDM services. its impact on RDM services. P037 indicated that “I think Although our results provided necessary answers to our AI will have a huge impact. Positioning ourselves to take RQs, there are several limitations in our study. One of advantage of that space is key.” P169 echoed, “Impact of the limitations is related to the composition of our study 100 Rong Tang, Zhan Hu sample. The majority of our sample was from North expressing concerns of lack of bandwidth or capacity, it America (82.2%) and academic institutions (63.1%). When is clear that, in order to grow RDM services, institutional we compared the locations and organizations, it would be commitment to resources and training opportunities more desirable if we had a balanced sample. In addition, is crucial. As an emergent profession, data librarians even though we were able to perform comparative need to be nurtured, mentored and further trained. The analyses among five of the US regions, with the range study makes a case for developing a global community of from Northeast (37.6%) and Southwest (8.1%), it would be practice where data librarians work together, exchange helpful to have an even distribution of participants from information, help one another grow, and strive to advance the five regions. Our study revealed interesting results RDM practice around the world. with regard to the RDM service strength associated with each US region. A potential follow-up study focusing Acknowledgment: This research study was a part of exclusively on comparing regional strength in RDM needs assessment project, funded by Elsevier to establish service provision will need to perform systematic or the Research Data Management Librarian Academy cluster-based multistage random sampling in order to (RDMLA). The authors wish to thank Jean Shipman and achieve a balanced and comparable sample. Another Elaine Martin for developing questions for the online limitation of the study design is that one of the main survey and Alyson Gamble for performing preliminary distribution methods that we used was through Elsevier data processing and analysis of the survey dataset. The Library Connect blog post. A number of participants were authors also wish to thank all the respondents who resistant to providing full responses to questions due to answered our survey. their concern of the involvement of a publisher in the RDM realm. Nevertheless, we were able to obtain a good representative sample both in terms of the locations and in terms of a variety of organizational types. Furthermore, References the findings of this research study are confined to the Antell, K., Foote, J. B., Turner, J., & Shults, B. (2014). Dealing with survey responses; in order to obtain a more in-depth data: Science librarians’ participation in data management contextual understanding of these responses, follow-up at Association of Research Libraries institutions. College & interviews could be conducted. In further studies, a multi- Research Libraries, 75(4), 557-574.doi:10.5860/crl.75.4.557 phased and multimethod design may be used to warrant Australian Research Council (ARC). (2017). ARC open access policy. Retrieved from http://www.arc.gov.au/arc-open-access-policy a comprehensive understanding of the challenges and Choudhury, G. S. (2008). Case study in data curation at Johns practices of RDM services and tools. Hopkins University. Library Trends, 57(2), 211-220. Retrieved As one of the first studies investigating the location from http://hdl.handle.net/2142/10669 and organizational differences in the RDM practice Cox, A. M., Kennan, M. A., Lyon, L., & Pinfield, S. (2017). around the world, the present research revealed gaps Developments in research data management in academic in the global-based effort of developing RDM services. libraries: Towards an understanding of research data service maturity. Journal of the Association for Information Science and As argued earlier, the understanding of location and Technology, 68(9), 2182-2200. doi:10.1002/asi.23781 organizational differences would help us learn more Cox, A. M., & Pinfield, S. (2014). Research data management and about the strengths and weaknesses associated with each libraries: Current activities and future priorities. Journal of location or organizational types. It would also help to Librarianship and Information Science, 46(4), 299–316. identify the unique characteristics, or the niche associated doi:10.1177/0961000613492542 Department of Science & Technology, Government of India. (n.d.). with a given location or institutional type. Moreover, our National data sharing and accessibility policy. Retrieved from investigation into the level of preparedness, the degree of http://www.dst.gov.in/national-data-sharing-and-accessibility- development, as well as the obstacles that RDM librarians policy-0 face helps to paint a realistic picture of the problems Dietrich, D., Adams, T., Miner, A., & Steinhart, G. (2012). that librarians encounter in their collaborative endeavor De-mystifying the data management requirements of funders. of positioning themselves in a larger RDM lifecycle. Issues in Science and Technology Librarianship, 70(1), 1-16. doi: 10.5062/F44M92G2 This study contributes to the advancement of empirical Faniel, I. M. & Connaway, L. S. (2018). Librarians’ on the factors understanding of the RDM phenomena by introducing influencing research data management programs. College & and employing new comparative perspectives such as Research Libraries, 79(1), 100-119. doi:10.5860/crl.79.1.100 location and organizational types as well as exploring the Fearon, D., Gunia, B., Pralle, B.E., Lake, S., & Sallans, A. L., impact of varying levels of RDM maturity. With a majority (2013). Research data management services. SPEC Kit of the respondents hoping to receive more training while Providing Research Data Management (RDM) Services in Libraries: Preparedness, Roles, Challenges... 101

334. Washington, DC: Association of Research Libraries. Perrier, L., Blondal, E., & MacDonald, H. (2018). Exploring the doi:10.29242/spec.334 experiences of academic libraries with research data Federer, L. (2018). Defining data librarianship: a survey of management: A meta-ethnographic analysis of qualitative competencies, skills, and training. Journal of the Medical studies. Library & Information Science Research, 40(3-4), 173- Library Association, 106(3), 294-303. doi:10.5195/ 183. jmla.2018.306 Pinfield, S., Cox, A. M., & Smith, J. (2014) Research data Gore, S. A. (2013). A librarian by any other name: The role of the management and libraries: Relationships, activities, drivers on a clinical research team. Journal of eScience and influences. PLoS ONE, 9(12), e114734. doi:10.1371/journal. Librarianship, 2(1), 20-24. pone.0114734 Government of Canada. (2016, December 21). Tri-Agency Statement Shearer, K. (2015). Comprehensive Brief on Research Data of Principles on Digital Data Management. Retrieved from Management Policies. Retrieved from: https://portagenetwork. http://www.science.gc.ca/eic/site/063.nsf/eng/h_83F7624E. ca/wp-content/uploads/2016/03/Comprehensive-Brief-on- html?OpenDocument Research-Data-Management-Policies-2015.pdf Hanson, K. L., Bakker, T. A., Svirsky, M. A., Neuman, A. C., & Rambo, Smale, N., Unsworth, K. J., Denyer, G., & Barr, D. P. (2018). The N. (2013). Informationist role: clinical data management in History, Advocacy and Efficacy of Data Management Plans. auditory research. Journal of eScience Librarianship, 2(1), bioRxiv, 443499. doi:10.1101/443499 25-29. Tenopir, C., Hughes, D., Allard, S., Frame, M., Birch, B., Baird, L., Hasman, L., Berryman, D., & Mcintosh, S. (2013). NLM Informationist …& Lundeen, A. (2015). Research data services in academic Grant – web assisted tobacco intervention for community libraries: Data intensive roles for the future? Journal of eScience college students. Journal of eScience Librarianship, 2(1), 30-34. Librarianship, 4(2), 1-21. doi:10.7191/jeslib.2015.1085 Japan Science and Technology Agency (JST). (2013). JST Policy on Tenopir, C., Talja, S., Horstmann, W., Late, E., Hughes, D., Pollock, Open Access and Research Publications and Research Data D., ... & Allard, S. (2017). Research data services in European Management. Retrieved from https://www.jst.go.jp/EN/about/ academic research libraries. Liber Quarterly, 27(1), 23-44. DOI: openscience/index.html 10.18352/lq.10180 Jones, S., Pryor, G., & Whyte, A. (2013). How to Develop Research UK Research and Innovation. (n.d.). Data Policy. Retrieved from: Data Management Services - a guide for HEIs. DCC How-to https://www.ukri.org/funding/information-for-award-holders/ Guides. Edinburgh: Digital Curation Centre. data-policy/ Martin, E. R. (2013). Highlighting the informationist as a data Varner, S., & Hswe, P. (2016). Special report: Digital humanities in librarian embedded in a research team. Journal of eScience libraries. Retrieved from American Libraries website: https:// Librarianship, 2(1), 1-2. americanlibrariesmagazine.org/2016/01/04/special-report- National Geographic Society. (2009). United States Regions. digital-humanities-libraries/ Retrieved June 19, 2019, from https://www.nationalgeographic. Whyte, A. (2014). A pathway to sustainable research data services: org/maps/united-states-regions/ From scoping to sustainability. In Pryor, Jones, & Whyte (eds.). National Institutes of Health. (2005). Policy on enhancing public Delivering research data management services (pp. 59-88). access to archived publications resulting from NIH-funded London: Facet. research. [Internet]. Retrieved from http://grants.nih.gov/ Whyte, A., & Tedds, J. (2011). Making the case for research data grants/guide/notice-files/NOT-OD-05-022.html management. Retrieved from Digital Curation Centre website: National Institutes of Health. (2008). Revised policy on enhancing http://www.dcc.ac.uk/resources/briefing-papers public access to archived publications resulting from NIH- funded research. [Internet]. Retrieved from http://grants.nih. gov/grants/guide/notice-files/NOT-OD-08-033.html National Science Foundation. (2010a). Dissemination and sharing of research results. [Internet]. Retrieved from http://www.nsf.gov/ bfa/dias/policy/dmp.jsp National Science Foundation. (2010b). Scientists seeking NSF funding will soon be required to submit data management plans. [Internet]. Retrieved from http://www.nsf.gov/news/ news_summ.jsp?cntn_id=116928&org=NSF&from=news National Science Foundation. (2011). Grant proposal guide: Chapter II - proposal preparation instructions. [Internet]. Retrieved from http://www.nsf.gov/pubs/policydocs/pappguide/nsf11001/ gpg_2.jsp#IIC2j Office of Science and Technology Policy (OSTP). (2013, February 22). Memorandum for the heads of executive departments and agencies. Retrieved from https://rosap.ntl.bts.gov/view/ dot/34953