Data Management

Even if one is not required by your funding agency, developing a data management plan (DMP) at the beginning of a new project will inform good practice throughout the data cycle. The following practices are fundamental to effective data management and can be applied to all disciplines:

The Data Management Plan:

• Adhere to the guidelines set by any funding agencies and institutions that are sponsoring the research.
  • For examples of such guidelines, go here.
  • For examples of completed DMPs, go here.
• Complete your DMP early so that it will not be sidelined when you start collecting data.
• The minimum expenses to include when calculating your data management costs are: data creation, processing, analysis, storage, sharing, and preservation.

Data Storage:

• To prevent data from being lost to incompatibility, store it as formats and on hardware that are open standard, not proprietary.

• When naming files in your data set, be consistent and descriptive. 
• Backup your data on local and remote external servers and have a contingency plan for restoring lost data.
• Best Practices Primer from DataONE

Data Documentation:

• Use metadata to record details about a study such as
  • its context
  • the dates of data collections
  • data collection methods, etc.
• If possible, use an established metadata standard appropriate to your field.
  • Producers of ecology data are encouraged to use Ecological Metadata Language (EML) to document their data.
  • The Federal Geographic Data Committee has a metadata schema for digital geospatial data.

Ethical Issues:

• All sensitive information in your data should be redacted before depositing in a public archive or repository.(Also see ICPSR Confidentiality Policy
• Access to the data may need to be embargoed (limited for a certain amount of time) in order to ensure privacy.
• The Boston College Institutional Review Board uses federal policies when approving research on human subjects.
• Develop a data retention policy in order to practice responsible conduct of research. 
• Be aware of the ownership and intellectual property rights concerning your data.
  • Data is not copyrightable, but it can be licensed under Creative Commons.

Sharing Data:

• Archived data sets should include the metadata in a readme file that contains all pertinent details of the study.
• When you are ready to deposit your data, do so in an archive that is public and appropriate for your discipline. Consult your subject bibliographer if you need assistance locating or preparing a submission to an archive.

A data management horror story by Karen Hanson, Alisa Surkis and Karen Yacobucci.