AMA Dataset Citation with Repositories, Versions, and Persistent IDs

Research software and methods publications

Record the software version separately from a dataset release or an associated methods article. A recommended book or paper can give scholarly credit while your Methods section identifies the program and packages actually used.

Distinguish a dataset from its associated paper

A dataset reference takes readers to the archived data. A journal reference takes them to the authors’ published analysis. Cite each source where it supports your work. If you downloaded data for a secondary analysis and used the paper to explain the study design, the two references serve different parts of your methods.

AMA Style Insider’s December 2020 post demonstrates this with a melanoma study and its Dryad deposit. The post recommends citing the dataset used and the original source. Its repository pattern includes the source information, repository name, deposit date, and dataset DOI. Use that structure while checking the record currently in front of you.

The linked PLOS One article is a research article, rather than a publication specifically labeled as a data paper. A dedicated data paper would still be a separate publication. In this example, the journal’s Data Availability statement points to the Dryad DOI, and the repository links back to the journal article.

Read the repository metadata before formatting

Open the dataset landing page and inspect its Citation section. The current Dryad record starts with Daniel Almquist, Mahesh Seetharam, and Lisa Kottschade. Its recommended citation differs from the older AMA blog illustration, which starts with DeLeon. The worked dataset reference below follows the current repository order and the initials available there.

Check the title independently, even when it matches the paper. Then record the repository and DOI. This release lists gene_data.xls and shows March 12, 2020, beside its version files. Those details help connect the reference to the downloaded file without turning a spreadsheet filename into the dataset title.

Dryad’s reuse guide recommends checking the metadata and README for collection methods, variable explanations, and limitations. Keep that documentation beside your analysis notes so you can interpret the columns you intend to use.

Write the AMA data citation from checked fields

The example below adapts the AMA repository pattern to today’s Dryad metadata. It uses the date label Published because the current landing page labels March 12 as its publication date. A submission date, a repository publication date, and the day you downloaded a file describe different events. Preserve the label your evidence supports.

Keep the data DOI in the dataset entry and the journal DOI in the article entry. Both records list more than six names, so these AMA examples use the first three followed by et al. The dataset creator initials reflect its metadata; the journal authors retain their published initials.

Check what a DOI identifies across versions

Dryad’s publication guide says the dataset DOI stays the same when a dataset is updated. Its main Download dataset button provides the latest version; earlier releases, when present, are available in Data files by publication date. A Dryad DOI therefore needs to be read alongside the release information when reproducibility depends on a particular version.

For this example, the public page shows March 12, 2020, version files. Dryad’s public API identifies the corresponding release as version 2, resource 56606, and lists gene_data.xls at 164,352 bytes. These are source-identification details. They do not establish that the data have been reanalyzed or that the clinical findings are correct.

Keep the release date and any supplied version number in your analysis notes. Follow the target journal’s requirements for including version information in the finished entry or methods. Avoid constructing a version DOI by adding a suffix yourself. Other repositories may assign identifiers differently; check their own documentation before changing a reference.

Connect the data reference to your analysis

When you use data, make the connection visible where you describe the analysis. Record which file supplied the observations and which variables or subset you used. Explain exclusions or transformations in your methods, alongside the dataset reference. A citation locates the source; your methods explain what you did to it.

Keep your original download separate from a working copy. If you later replace a file, note the new release and review whether the reference or methods need updating. A visit date can document when you obtained data, but it cannot by itself identify which release you analyzed.

When the source is a report or figure

Identify the object you used before choosing a format. A report’s PDF and a figure reproduced from a paper each require the reference and source checks appropriate to that document. An independently archived dataset can have a separate reference even when all three objects discuss the same study.

Review DOI imports before copying the result

AMA Cite has no Dataset form. Its DOI lookup currently requests Crossref metadata and maps returned records into Journal fields. A failed lookup may simply leave you needing manual entry, and a successful lookup still needs a source-type check. Use the manual dataset example here to review the finished reference.

For a report distributed as a PDF, use the report or PDF guide. When checking a dataset DOI, reopen its landing page and confirm that the reference points to the release documented in your methods.