Metadata is the documentation and context which makes your data re-usable. It includes both file-level information and overarching project information. Metadata is best collected systematically and continually throughout the research process. Examples of metadata include:

• rationale for the research
• descriptions of experiments
• methods, protocols, or other standard practices
• units of measurement, dates, times, 
• analyzes, including justification for missing values, transformations, quality assurance procedures, etc. 
• where your data are stored and backed-up
• definitions of any acronyms, terms, or codes (sometimes called a data dictionary or glossary)

For more information regarding keeping metadata in lab notebooks, check out MIT's Instructions for Using Your Laboratory Notebook. 

There are many different types of documentation that can be created to accompany research data and the type(s) that you choose are dependent on your research context. Some common options are listed below and the rest of this guide goes into each type in more detail.

• File Naming Schema
• Readme File
• Data Dictionary / Codebook
• Laboratory or Research Notebooks
• Standard Operating Procedures
• Protocols
• Project Reports

Metadata standards consist of elements that categorize metadata by function. They often arise out of metadata schemas (a set of elements designed to describe information). Following established standards and schema allow repositories to be inter-operable. This can be very important in the longevity and durability of data. 

Our National Information Standards Organization (NISO) provides a good explanation of metadata standards and schema in their publication Understanding Metadata.

Metadata standards often differ by discipline. Want to find out the international metadata standards for your data? Search the Digital Curation Centre’s list of metadata standards, search metadata specification by research discipline, or contact us at datamanagement@osu.edu.