Good Data Management Practice extends the notion of supporting research findings with good data to supporting research data with practical approaches and activities.
The following ten guiding principles are not a set of requirements or expectations, but rather ways of thinking about supporting research data in the context of twenty-first century science.
| Principle | Description |
|---|---|
| High quality scientific data requires effective data management | Though some of the related practices and strategies may seem relatively simple, even the most advanced science requires data to be properly stored, organized, and documented. |
| Effective data management requires scientific data to be defined broadly. | Scientific data does not exist in a vacuum. Proper management of data requires taking a workflow-based view of data. All materials needed to trace the provenance of a research finding “count” as data. |
| Effective data management requires research data to be understood thoroughly. | Related practices and strategies are, partially, determined by the characteristics of the data being managed. For example, sensitive data necessitates limiting access. |
| Data management is a set of practices and strategies, not a set of tools. | Though there are a myriad of tools and resources to make the process more efficient, the core of effective data management are practices and strategies implemented by people. |
| Data management practices and strategies should be routine and standardized. | Data management practices and strategies should be a part of the day-to-day conduct of research. Realizing the benefits of data management requires agreement about how data should be organized, stored, described, and representation. |
| It takes planning and communication to manage data well. | Like any workflow, the data management-related activities implemented at an early stage of a project substantially affect what can be done later. Though plans may change, strategizing prospectively helps things proceed efficiently. |
| Managing data requires managing documentation and metadata. | Effective data management often necessitates documenting more than seems necessary in greater detail than may seem necessary. Information about data “counts” as data. |
| All data must be managed, but not all data necessarily needs to be shared (openly). | Releasing scientific data for use by others has numerous benefits, but the protection of human subjects must be centered. Typically, only materials that underlie research findings need to be shared. |
| Data management requires commitment and a range of skills. | Supporting data management is a community effort, involving both researchers and other data stakeholders (e.g. IT professionals, librarians, leadership). To be effective, related practices and strategies need to be verified, audited, and revisited over time. |
| Sometimes “good enough” has to be good enough | Best practices in data management is a moving target. In settings where formal data management standards and expectations are still under development, it may be necessary to consider and adopt a set of minimum guidelines. |