How to cite a dataset

TL;DR

• Citing a dataset is required, not optional, in peer-reviewed work.

• APA, MLA, and Chicago each have distinct dataset citation formats.

• Always include the dataset version, access date, and persistent identifier.

• Missing a DOI or repository link can get your paper rejected.

• Proper citation protects you from academic integrity violations.

You found the perfect dataset. It supports your argument, the sample size is solid, and the source is credible. Then comes the question that stops most student researchers cold: how do you actually cite it? A dataset is not a journal article. It is not a book. Most citation guides treat it as an afterthought, buried in a footnote with no clear example.

This matters more than it might seem. Journals like Scientific Data (published by Nature Portfolio) and PLOS ONE have explicit data citation policies. Reviewers check them. If your dataset citation is missing a version number, a persistent identifier, or a repository name, your submission can be returned before peer review even begins.

The good news is that once you understand the logic behind dataset citations, the format becomes predictable. This post walks through exactly how to cite a dataset in the three most common academic styles, and what to do when information is missing.

Why You Must Cite a Dataset (Not Just the Paper That Used It)

Citing a dataset separately from the paper that references it gives original data creators proper academic credit, allows other researchers to verify your findings, and satisfies the data transparency requirements set by most peer-reviewed journals. The Joint Declaration of Data Citation Principles, published by FORCE11, establishes that datasets are legitimate, citable research outputs in their own right.

There is a common mistake here. Many student researchers cite the paper that used a dataset rather than the dataset itself. If you downloaded climate data from NOAA or survey data from the UK Data Service, you cite the dataset directly. Citing only the secondary paper that analysed it is incomplete and can misattribute the original work.

Journals that follow open science standards, including those indexed in the Directory of Open Access Journals (DOAJ), increasingly require a data availability statement alongside a formal dataset citation. This is not a stylistic preference. It is a submission requirement.

Understanding journal-specific expectations before you submit can save weeks of revision time. A guide to impact factor vs CiteScore for student researchers can help you evaluate which journals align with your work before you format a single citation.

How to Cite a Dataset in APA Format

In APA 7th edition, a dataset citation follows this structure: Author(s) last name, Initials. (Year). Title of dataset (Version number) [Dataset]. Repository Name. https://doi.org/xxxxx. The key elements are the dataset label in square brackets, the version number where available, and a persistent identifier such as a Digital Object Identifier (DOI) or stable URL.

Here is a concrete example. If you used data from the General Social Survey, the citation would look like this:

Smith, T., Davern, M., Freese, J., & Morgan, S. (2019). General Social Surveys, 1972-2018 [Dataset]. NORC at the University of Chicago. https://doi.org/10.3886/ICPSR34802.v1

Three things to check before finalising an APA dataset citation:

1. Confirm the version number in the repository record. Datasets are updated, and citing the wrong version undermines reproducibility.

2. Use a DOI rather than a general website URL wherever one exists. DOIs are persistent. Regular URLs break.

3. Include the [Dataset] label in square brackets immediately after the title. This tells the reader and any automated system what type of source it is.

For a full breakdown of APA citation rules across source types, the guide on how to cite sources in APA format covers the complete framework.

If you are working on your first academic submission and want structured guidance on how to cite a dataset and format your references correctly before you submit, joining the Publication Compass waitlist gives you early access to an AI platform built specifically for student researchers navigating this process.

How to Cite a Dataset in MLA Format

MLA 9th edition treats datasets as digital sources. The citation structure is: Last Name, First Name. Title of Dataset. Version, Repository Name, Year, DOI or URL. MLA does not use a [Dataset] label, but it does require the version and the name of the hosting repository as a container field.

The MLA format places the repository name in the position normally used for a website or database title. For example:

Ruggles, Steven, et al. IPUMS USA: Version 11.0. Minnesota Population Center, 2021, https://doi.org/10.18128/D010.V11.0.

MLA also asks for an access date when the source does not have a fixed publication date. For most archived datasets with a DOI, the publication year is sufficient. For datasets without a stable identifier, add the date you accessed the repository at the end of the citation.

Two common errors in MLA dataset citations:

1. Omitting the version number when one exists. MLA 9 treats version information as part of the edition field.

2. Using the landing page URL instead of the direct DOI. Always use the most stable, specific identifier available.

The full rules for MLA source formatting are covered in the guide on how to cite sources in MLA format.

How to Cite a Dataset in Chicago Style

Chicago 17th edition handles dataset citations differently depending on whether you use Notes-Bibliography style or Author-Date style. In Author-Date format, the citation reads: Last Name, First Name. Year. Title of Dataset. Version. Repository Name. Accessed Month Day, Year. https://doi.org/xxxxx. In Notes-Bibliography format, the same elements appear but the year moves to the end of the entry.

Chicago is the preferred style in history and some humanities disciplines, but it also appears in economics and social science journals. If your target journal uses Chicago, check whether it requires Author-Date or Notes-Bibliography before you format your references. The two styles look significantly different even for the same source.

A Chicago Author-Date dataset citation looks like this:

Feenstra, Robert C., Robert Inklaar, and Marcel P. Timmer. 2015. Penn World Table Version 9.0. Groningen Growth and Development Centre. Accessed June 10, 2024. https://doi.org/10.15141/S5J01T.

The access date is more consistently required in Chicago than in APA or MLA, even when a DOI is present. Include it.

For Chicago-specific citation rules across all source types, the guide on how to cite sources in Chicago style is a useful reference to keep open while formatting.

What to Do When Dataset Information Is Missing

When a dataset lacks a named author, a version number, or a DOI, use the closest available equivalent and note the gap. Replace a missing author with the organisation or repository name. Replace a missing DOI with the most stable URL available, such as a repository permalink. If no publication date exists, use the access date.

This situation comes up more often than most guides acknowledge. Government portals, institutional repositories, and older archived datasets frequently lack complete metadata. The principle across APA, MLA, and Chicago is the same: provide enough information for another researcher to locate the exact dataset you used. If that is not possible with the available information, consider whether the dataset is citable at all, or whether you need a more formally archived version.

Steps to follow when information is incomplete:

1. Search the repository for an updated record. Many datasets have been re-archived with full metadata after initial upload.

2. Check whether the dataset has a companion data paper. Journals like Scientific Data publish data descriptor articles that provide citable, peer-reviewed records of datasets.

3. Contact the repository directly. Librarians and data managers at institutions like the UK Data Service or ICPSR can often provide the correct citation format for their holdings.

Getting your citation right also affects how your own work gets cited later. The principles in the guide on how to get your paper cited apply to data outputs as much as written papers.

In-Text Citations for Datasets

In-text dataset citations follow the same rules as other source types within each style. In APA, use the author's last name and year in parentheses: (Ruggles et al., 2021). In MLA, use the author's last name in parentheses: (Ruggles et al.). In Chicago Author-Date, use last name and year: (Feenstra, Inklaar, and Timmer 2015). The dataset label does not appear in the in-text citation, only in the reference list.

One specific consideration for datasets: if you cite the same dataset at multiple points in your paper, you do not need to repeat the full citation each time. The in-text reference covers it. However, if you use two different versions of the same dataset, treat them as separate sources with separate reference list entries and distinguish them clearly in-text.

Frequently Asked Questions

Do I need to cite a dataset if it is publicly available?

Yes. Public availability does not remove the citation requirement. Citing a dataset credits the creators, supports reproducibility, and satisfies journal data policies. The FORCE11 Joint Declaration of Data Citation Principles states that datasets should be cited with the same rigour as any other research output, regardless of access status.

What is a DOI and why does it matter for dataset citations?

A Digital Object Identifier (DOI) is a permanent link assigned to a digital object such as a dataset, paper, or report. It does not break when websites are reorganised. Most citation styles prefer a DOI over a regular URL because it reliably points to the same resource over time. Repositories like Zenodo, Dryad, and ICPSR assign DOIs to datasets on deposit.

How do I cite a dataset with no named author?

Use the organisation or repository that created or hosts the dataset in the author position. For example, if the dataset comes from the United States Census Bureau with no individual author listed, begin the citation with United States Census Bureau as the author. This approach is consistent across APA, MLA, and Chicago guidelines.

Should I cite a dataset and the paper that describes it separately?

Yes, if both exist. A data descriptor paper published in a journal like Scientific Data and the dataset itself are two separate citable outputs. Cite both when your work draws on both. The paper describes the methodology; the dataset is the actual data you used. Conflating them can create ambiguity about what you accessed.

How do I cite a dataset in a school or undergraduate paper if my teacher has not given instructions?

Default to the citation style your institution or course uses for all other sources, typically APA, MLA, or Chicago. Apply the dataset-specific format for that style as described in this post. If your institution has a library guide, check it first. Many university libraries publish citation guides that include dataset examples specific to their preferred style.

Getting Dataset Citations Right

Citing a dataset is a small step with significant consequences. A correct citation tells reviewers you understand data provenance. It protects the researchers who built the dataset. It makes your findings reproducible. And it keeps your submission from being returned on a technicality before it reaches peer review.

The format varies by style, but the underlying logic does not. Identify the creator, the dataset title, the version, the repository, and the persistent identifier. Fill in what you have, source what you are missing, and format it according to the style your target journal requires. For more guidance on the full academic publishing process, the Publication Compass blog covers research skills, citation formats, and journal selection for student researchers at every stage.