Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
CHAPTER 8. Data Management Plans Developing and Maintaining Data Management Plans Over the past 20 years or so, there has been increasing interest in making research data more accessible outside of the research project in which they were generated. Recognizing this interest, the National Science Foundation (NSF) in 2010 announced that, beginning in January 2011, all grant applications would have to include a two-page data management plan (DMP). The U.S. DOT followed suit and described its expectations for researchers to create and submit a DMP for any funded research in the U.S. DOT Public Access Plan. As described by funding agencies, DMPs are the means through which researchers (1) describe the data they will generate as a part of their research process and (2) detail how they will make their data accessible to those outside the project team. However, DMPs can be more than just this. They are a useful exercise that can help guarantee that all stakeholders have a shared understanding of how data are to be developed and handled throughout their life cycle. Having a DMP facilitates connections and communication between project stakeholders. It can also make sharing data easier, not only with peers and potential collaborators, but with research partners and current stakeholders as well. Finally, a DMP can help ensure that research data retain their value to the organization over time. A Significant Shift While DMPs are only one piece of the effort to enable research data to be shared in ways that retain or enhance value, they are a critical component. Reviewing DMPs can reveal how well researchers understand what they are being asked to do and how prepared they are to do it. Recent studies indicate that there is work to be done on both fronts. An early study of researchersâ responses to the DMP requirements done in 2012 surveyed principal investigators 8. D A TA M A N A G E M E N T P LA N S In This Section » Developing and Maintaining Data Management Plans » A Significant Shift » Tools That May Help » Essential Requirements for Developing Sections of Data Management Plans » Data Section of the DMP » Standards Section of the DMP » Access Section of the DMP » Use Section of the DMP » Archiving/Preserving Section of the DMP » Going Beyond the Minimum for Data Management Plans » Chapter Checklist 67
68 8. Data ManageMent Plans (PIs) of NSF grant submissions. These PIs reported widespread uncertainty about the new DMP requirements and actions needed to respond adequately. Later studies have provided some additional details. The results show that ⢠There is a lack of understanding of what constitutes public access to data; ⢠Researchers only have a moderate level of awareness of the regulatory environment around issues of data management and sharing; ⢠Data type and format generated vary widely by discipline; ⢠Researchers desire support for their data management needs; ⢠Strategies for sharing data vary dramatically among researchers; ⢠Research assistants handle the majority of tasks related to data management (with the exception of data sharing, which is addressed by the PI); and ⢠When planning for training and outreach, consider the needs of all roles and responsibilities. Changing the culture of how research is practiced in any field will take time as well as dedicated resources, support, and education. Efforts to promote culture change will need to consider how research is currently practiced, articulate measurable benefits for individual researchers and the field (while recognizing potential near- and long-term costs or drawbacks), and identify the gaps between current practices and desired outcomes with actionable plans to close these gaps. Tools that May Help Various tools have been developed in an effort to educate and assist researchers as they navigate the regulatory environment. For instance, the DMPTool walks researchers through the process of developing a DMP for their research. The Massachusetts Institute of Technology (MIT) Libraries, along with other academic libraries, have produced extensive guides on managing, sharing, and preserving data. Data repositories, such as the Inter-university Consortium for Political and Social Research (ICPSR), the UK Data Service, and the Australian National Data Service, have developed guides not only on their data services, but on funding agency requirements and expectations. Finally, the U.S. Geological Survey (USGS) has created a useful webpage that includes information on and examples of DMPs. Although the effect of these tools on researcher compliance has not been studied, state DOTs and other research organizations should emphasize training and support for researchers to meet the DMP requirement. Recommending useful tools, providing DMP support, DMPTool: https://dmptool.org ICPSR: www.icpsr.umich.edu UK Data Service: www.ukdataservice. ac.uk Australian National Data Service: www.ands.org.au ? ? ?
69 8. Data ManageMent Plans and emphasizing the role of DMPs in training are all good ways to help researchers incorporate DMPs in their standard research practices and initiate culture change around data sharing. In addition, researchers who have access to an academic library with data services should contact them for help. Essential Requirements for Developing Sections of Data Management Plans At a minimum, researchers need to generate a DMP for their projects that will meet the expectations of the funding agency to which they are applying. Funding agencies have different expectations of what they want to see in a DMP, and so it is important for researchers to carefully review the call for proposals to be sure they understand what information needs to be included. However, many DMP requirements, including the U.S. DOTâs requirement, are based on the criteria developed by the National Science Foundation and often include five sections: 1. Data: A description of the data to be generated in the project. 2. Standards: A description of the standards that would be used in developing and structuring the data. 3. Access: A policy statement about how others would be able to gain access to the data. 4. Use: A policy statement about what others would be permitted to do with the data. 5. Archiving/Preserving: A statement on how the data would be archived. Some considerations to keep in mind while addressing these five basic sections of a DMP are discussed below. Data Section of the DMP In describing the data that will be generated over the course of the project, the primary focus should be on identifying the parts or portions of the data that will be shared with others outside of the project or organization that collected (and owns) the data. Not all data necessarily have to be shared, but the DMP should include a statement as to why not (including any ethical or privacy issues). Include information about how the data will be stored as they are being developed. It is also advisable to include as much information about the characteristics of the data as you can, including formats, volume, size, number of files, and so forth. Rough estimations are sufficient. ? ? ? Essential DMPs
70 8. Data ManageMent Plans Standards Section of the DMP This section contains information about how the data will be prepared according to community expectations. The standards section primarily focuses on data formats and the metadata that will accompany the data. Formats Consider what formats would make the data most useful or more accessible for others in your field or the general public. Note that open formats are usually better options than proprietary formats. Metadata Consider what information others in your field or the public would need in order to find, understand, and make use of the data set. If there is a metadata standard used in your discipline or for the kind of data you are generating, consider using it for your data. If there is not, consider adapting a general standard such as Project Open Metadata or Dublin Core and applying it to the data. See Chapter 7 for a more detailed discussion on formats and metadata. Although it is not mentioned by NSF or the U.S. DOT directly, including information as to how the data will be documented in your DMP is a good idea. Access Section of the DMP This section describes how the data will be made accessible to people outside of the organization that collected the data. If a repository for the data has been identified, it should be described in this section. If the organization intends to host the data and make it accessible themselves, a description of how this will be accomplished and for how long will be needed. More information about data repositories can be found in Chapter 7. The DMP should also clearly state the intended audience for the data. What research communities are expected to be interested in the data? Is there an expectation that the general public will be making use of the data? Including information as to how the data will be documented in your DMP is a good idea. ? ? ? ? ? ? Data citation and attribution is not only a best practice, but also a good rule of thumb.
71 8. Data ManageMent Plans Use Section of the DMP Once someone is able to access and acquire the data, what will they be allowed to do with them? This is often a difficult section to write, as many researchers and research organizations have not thought critically about how they would like their data to be used once the data have been made available. Following are some elements to consider in writing this section: þ Reuse: Can people reuse the data for research, commercial, or other purposes? Can people create derivatives of your data set or use it to develop new data sets? þ Attribution: Is citation or another form of attribution required if the data are used by others? þ Licensing: What license will be applied to the data to govern how they can be used? Creative Commons licenses are often applied to content in repositories, but they may be somewhat problematic to apply to data, as Creative Commons licenses depend on copyright law, and data sets may not be covered. Other licenses to consider for data are those provided by the Open Data Commons. Archiving/Preserving Section of the DMP Agencies want to be sure that your data are not just available to others for a limited amount of time, but that steps are taken to ensure that the data are available for the long term. Following are some questions to address in the DMP: þ What is the long-term value of the data and how long should they be made available? Some agencies have expectations that the data will be available for a certain amount of time, so be sure to read the requirements carefully. þ What related information (e.g., documentation, software, reports) needs to be preserved alongside the data? þ What steps will be taken to preserve the data by the repository selected to host and make the data accessible? If a repository is not being used to host the data, what steps will the organization take to preserve it? Chapter 7 provides additional information about preserving research data. Checklist: Questions to address in the DMP Archive/Preserve Section ? ? ? Checklist: Elements to consider in the DMP Use Section? ? ?
72 8. Data ManageMent Plans Going Beyond the Minimum for Data Management Plans DMPs submitted to funding agencies are often limited in how long they should be. NSF, for example, limits the DMP to two pages. Therefore, it is important to be succinct and direct in drafting the DMP that will be submitted as a part of the grant application. However, as DMPs are meant to be actionable documents, the version of the DMP used by project personnel may need to expand beyond the initial two page limitation to be useful. More importantly, the DMPs being created should go beyond the considerations of federal and private funding agencies to consider and address local needs. Ideally, DMPs ought to connect with the policies of the state DOT or research institution, as they describe the organization as a whole. In contrast to the high-level focus of policies, DMPs are useful in describing and informing the actual and specific data management, sharing, curation, and preservation practices taking place for individual data sets being developed within the organization. A fully developed DMP can be an extremely useful tool in helping research teams come to a consensus on the work to be done in developing rich data sets, defining expectations for how the work will be done and by whom, and in holding researchers and their research organizations accountable for providing needed resources and completing the work. Moreover, well-crafted DMPs can help change cultures of practice in organizations to help them accomplish more and build their reputation. As more and more DMPs are developed, patterns will begin to emerge that can be used to help state DOTs and other research institutions better understand current practices with data and that can inform where additional support and resources may be needed to ensure that organizational goals are being met. Following are some additional questions to consider including in DMPs that are locally focused: ⢠Beyond funding agency requirements, what additional details are needed to make this DMP fully actionable and useful to the organization? ⢠Who are the stakeholders for the research project and/or the data set(s) that are covered by this DMP? What roles and responsibilities will these stakeholders have in managing, sharing, curating, and/or preserving the data? (See Chapter 5 for more discussion on roles and responsibilities.) ⢠What quality control or quality assurance measures will be taken to ensure the data are reliable and trusted? ⢠What resources within the organization will be needed for the data set(s) being developed? Are there any special needs for the ? ? ? Beyond minimum DMPs
73 8. Data ManageMent Plans data set(s) requiring resources beyond what the organization currently provides? If needed resources are not available currently, how will they be acquired or what alternatives can be put into place? ⢠What costs might be incurred in managing, sharing, curating, or preserving these specific data set(s)? If these costs are outside of the norm, how will they be covered? ⢠How do the DMPs link together or relate to each other? Are there elements of one research project that can be used to inform or improve upon another? Creating DMPs will become easier over time as they become more commonplace. It is expected that some elements of the DMP will be similar or the same for every DMP and, as such, templates can be created with defined options to choose from as a starting point. Chapter Checklist From this chapter, you should be able to þ Know the U.S. DOT requirements for DMPs. þ Understand that other agencies may have their own specific requirements. þ Know the five common elements of DMPs and what goes into them: (1) a description of the data, (2) standards to be used, (3) making the data accessible, (4) provisions for use of the data, and (5) preservation of the data. þ Identify tools and templates that can help with writing a DMP. þ Know how to go beyond with DMPs by adding information related to local research communities and interests. Creating DMPs will become easier over time as they become more commonplace. ? ? ?
