Datasets should be described with enough detail that another researcher with a level of disciplinary expertise similar to the creator would be able to understand and use the data independently. Documentation may include the context of data collection, variable descriptions and codes, population and sample information, data processing procedures, and any other relevant information.

All datasets must be accompanied by a readme file in either TXT or PDF format. Readme files should contain information about the dataset creator, the context of data collection or dataset creation (who, where, when, what, etc.), and the licensing terms of the dataset.

Datasets based on human research should include the precise terms to which participants agreed. Depositors are encouraged to include a blank consent form with their dataset so that secondary users and research ethics boards can understand the full context of participant consent for the original data collection.