

How to Cite a Dataset in APA 7: Format and 2 Real Examples
An APA 7 data set reference gives the author, the year, the title in italics, any identifier and version number in parentheses, the description [Data set] in square brackets, the publisher, and the DOI (American Psychological Association [APA], 2020):
Author, A. A. (Year). Title of data set (Identifier; Version number) [Data set]. Publisher. DOI
APA's own example is a data set published by ICPSR (APA, 2020):
O'Donohue, W. (2017). Content analysis of undergraduate psychology textbooks (ICPSR 36966; Version V1) [Data set]. ICPSR. https://doi.org/10.3886/ICPSR36966.v1
Filled in with a real data set from Zenodo, which has a version number but no separate identifier:
Wellington, S., & Clayton, J. (2019). Fourteen-channel EEG with imagined speech (FEIS) dataset (Version v1.1) [Data set]. Zenodo. https://doi.org/10.5281/zenodo.3554128
In the text: (Wellington & Clayton, 2019) or Wellington and Clayton (2019). The year is that of the version you used (APA, 2020), which may not match a repository's own suggested citation.
Both real data set records were checked on November 9, 2026, and are written as they appeared that day. The O'Donohue entry and the unpublished-data entries are APA's own examples.
How to format each part of an APA data set reference
Each part of the FEIS entry (APA, 2020, 2025):
| Part | APA 7 rule | In the FEIS entry |
|---|---|---|
| Author | Individual names inverted: surname, comma, initials, with an ampersand before the last name. A group author is spelled out in full, followed by a period | Wellington, S., & Clayton, J. |
| Date | The year of publication for the version of the data you used, in parentheses | (2019). |
| Title | Italics and sentence case, as for other works that stand alone | Fourteen-channel EEG with imagined speech (FEIS) dataset |
| Identifier and version | In parentheses after the title, not italic. With both, separate them with a semicolon | (Version v1.1) |
| Description | In square brackets, before the period, first letter capitalized, not italic | [Data set]. |
| Publisher | The publisher of the data set; for a work that stands alone, that can be a database or archive. APA's example names ICPSR | Zenodo. |
| DOI | A doi.org link at the end | https://doi.org/10.5281/zenodo.3554128 |
Zenodo shows the title as "Fourteen-channel EEG with Imagined Speech (FEIS) dataset." Data sets are on APA's list of works that stand alone, so the title goes in italics and in sentence case (APA, 2025): "Imagined Speech" becomes "imagined speech," while EEG and FEIS keep their capitals. Sentence case vs title case covers which words keep capitals.
Why the [Data set] label is there
APA adds a description in square brackets "to help identify works outside the peer-reviewed academic literature," and data sets are on its list of works that take one (APA, 2025). The wording is flexible: APA gives "[Data set]" and "[Data set and code book]" as examples (APA, 2020).
Which version and year to use
APA ties the reference to one version: the date is "the year of publication for the version of the data used" (APA, 2020). Write the version label the way the repository shows it, after the word Version, as APA's example does with V1. Zenodo and Harvard Dataverse handle versions differently.
A new DOI for each version: Zenodo
Zenodo registers one DOI for the specific version and another representing all versions of the record (Zenodo, 2026). FEIS has two versions:
| The DOI points to | Version | Published | DOI |
|---|---|---|---|
| One version | v1.0 | August 15, 2019 | 10.5281/zenodo.3369179 |
| One version | v1.1 | November 26, 2019 | 10.5281/zenodo.3554128 |
| All versions | 10.5281/zenodo.3369178 |
The entry above names v1.1, so it uses the v1.1 DOI. APA's example pairs Version V1 with a DOI that ends in .v1.
One DOI for all versions: Harvard Dataverse
Dataverse "creates one persistent identifier per dataset and doesn't create new ones for each version," and later versions do not change the citation date (Dataverse Project, 2026). The MIT Election Data and Science Lab's data set U.S. House 1976–2024 is on version 15, released March 9, 2026. The citation Dataverse generates for that version still gives the year as 2017. Under APA's rule, version 15 gets 2026:
MIT Election Data and Science Lab. (2026). U.S. House 1976–2024 (Version V15) [Data set]. Harvard Dataverse. https://doi.org/10.7910/DVN/IG0UN2
In the text: (MIT Election Data and Science Lab, 2026) or MIT Election Data and Science Lab (2026).
The lab is a group author, spelled out in full (APA, 2025); group author abbreviations covers shortening it in the text. With one DOI for every version, the version number is what tells a reader which release you analyzed. Titles change too: versions 12 and 13 were titled "U.S. House 1976–2022," so copy the title from your version.
Data set or the article that reported it?
APA gives two cases for citing a data set: you ran "secondary analyses of publicly archived data," or you archived your own data that are "being presented for the first time in the current work" (APA, 2020).
For numbers someone has already reported, APA points elsewhere: "If you are citing existing data or statistics, cite the publication in which the data were published (e.g., a journal article, report, or webpage) rather than the data set itself" (APA, 2020).
So a percentage you copied from a paper's table is cited to that paper, and files you downloaded and reanalyzed are cited as the data set.
How to cite unpublished raw data in APA 7
Unpublished data can be cited too. An APA Style Blog post sets out what changes (Adams, 2025):
- The date is "the year(s) the data were collected," with an en dash between the first and last year.
- Its examples use the description [Unpublished raw data].
- When the author of a data set is also the source, "do not repeat the author name in the source element," so the reference ends after the bracket.
APA's example, with a company as the source:
Aims, O., Modelo, I., Kapowski, K. C., & Nguyen, Y. (2023–2025). Scoring models and their impact on sales growth [Unpublished raw data]. Hawthorne Advanced Sales Solutions.
In the text: (Aims et al., 2023–2025) or Aims et al. (2023–2025).
APA's example where an organization is both author and source:
Jackson Housing Authority. (2022). Results of affordable housing initiatives on formerly transient persons [Unpublished raw data].
In the text: (Jackson Housing Authority, 2022) or Jackson Housing Authority (2022).
An untitled data set gets a description of "the publication status and focus of the data" in square brackets instead of a title, and an unknown source is left out (Adams, 2025). APA's example:
Carter, W. D. (2024). [Unpublished raw data on the correlations between sleep deprivation from playing video games and failing test scores of high school students].
Once the data are online, use the regular data set format; such references "usually include the URL" (Adams, 2025).
Check each DOI, then the rest of the list
Before you submit, open each DOI in your list and compare the record with your entry: the authors, the title, the version, and the year that version came out. If a database record or the repository's own citation shows an earlier year, keep the year of the version you used (APA, 2020).
The journal articles and books on the same list can be checked against their published records. In PerfectRefs, you paste your reference list; you don't upload the paper. The Existence check looks up each reference in Crossref, OpenAlex, and other scholarly databases and flags any it can't find. It comes with a free daily allowance. The Style check compares authors, title, journal, year, volume, issue, pages, and DOI field by field with the database record, shows the record's values with a link to the source, and lets you accept a fix with one click. A reference with no matching record gets nothing filled in. The Style check uses premium credits (2 per reference), and new accounts get 10 premium credits.
FAQ
Is it "data set" or "dataset" in APA 7?
APA writes two words. Its page is titled "Data Set References," and the description in its example is [Data set] (APA, 2020).
Does a data set reference need a retrieval date?
APA adds one "only if the work is unarchived and designed to change over time" (APA, 2025). APA's data set example has none, and neither do the entries here.
What if the data set has no version number?
APA asks for "any numerical identifier and version number" (APA, 2020). If the record shows neither, the bracketed description follows the title directly, as in the unpublished examples above.
References
- Adams, A. (2025, August 7). How to cite unpublished works. APA Style Blog. https://apastyle.apa.org/blog/cite-unpublished-works
- American Psychological Association. (2020). Data set references. APA Style. https://apastyle.apa.org/style-grammar-guidelines/references/examples/data-set-references
- American Psychological Association. (2025, October). Elements of reference list entries. APA Style. https://apastyle.apa.org/style-grammar-guidelines/references/elements-list-entry
- Dataverse Project. (2026, October 9). Dataset + file management. Dataverse Guides. https://guides.dataverse.org/en/latest/user/dataset-management.html
- Zenodo. (2026). What is DOI versioning? Zenodo FAQ. https://support.zenodo.org/help/en-gb/1-upload-deposit/97-what-is-doi-versioning
Original article by PerfectRefs. To republish it, credit PerfectRefs and link to this page. Do not edit it or use it commercially.
