How to cite a dataset in a research paper: a complete guide for high school students

This post answers a specific question: how do you cite a dataset correctly in an academic research paper? It is written for high school students who are conducting original research and need to document their data sources in a format that meets publication standards. After reading, you will know which citation elements are required, how format varies by citation style, and what mistakes most commonly cause reviewers to flag a paper's references. When your research is ready for peer review, the Princeton Journal of Pre-Collegiate Research publishes original student work across all academic disciplines.

Why dataset citation matters in student research

Dataset citation is one of the most frequently mishandled elements in student research papers, and it is one of the first things a peer reviewer checks. When a student analyzes data from an external source without citing it correctly, the paper cannot be independently verified. Reproducibility is a foundational requirement of academic research, not an optional standard. A missing or malformed dataset citation tells a reviewer that the work cannot be checked, which is sufficient grounds for rejection at most journals.

The challenge is that most citation guides written for students focus on books, journal articles, and websites. Datasets follow different rules. They have their own identifiers, version numbers, access dates, and repository names, all of which must appear in the citation. This guide addresses that gap directly.

How do you cite a dataset in a research paper?

To cite a dataset, include the author or organization, the year of publication or release, the dataset title, the version number if available, the repository or archive name, and a persistent identifier such as a DOI or URL. The exact order and formatting of these elements depends on the citation style your target journal requires, most commonly APA, MLA, or Chicago.

What elements does a dataset citation require?

Every complete dataset citation contains six core elements: (1) the creator or authoring organization, (2) the year the dataset was published or last updated, (3) the full title of the dataset, (4) the version or edition if one is specified, (5) the name of the repository or data archive, and (6) a persistent identifier, preferably a Digital Object Identifier (DOI). If no DOI exists, a stable URL with an access date is acceptable.

A DOI is a permanent digital address assigned to a dataset by its publisher or repository. Unlike a standard URL, a DOI does not change when a website is restructured. The DOI Foundation, which administers the system, reports that over 200 million DOIs have been registered across academic content. When a dataset has a DOI, always use it rather than a URL.

Version numbers matter more than many students realize. A dataset that is updated annually, such as a government economic survey or a public health registry, may produce different results depending on which version was analyzed. Citing the version ensures that another researcher can access the exact data you used. If no version number is listed, note the access date instead.

How to cite a dataset in APA format

APA 7th edition, the most widely used style in the social sciences, formats a dataset citation as follows:

Author Last Name, First Initial. (Year). Title of dataset (Version number) [Dataset]. Repository Name. https://doi.org/xxxxx

For example, a citation for a publicly available dataset from the U.S. Census Bureau would list the bureau as the author, the year of release, the dataset title in italics, the label [Dataset] in square brackets, and the DOI or stable URL. The bracketed label [Dataset] is an APA 7th edition requirement that distinguishes a dataset from a journal article or report. Many students omit this label, which is a formatting error that reviewers notice.

How to cite a dataset in MLA format

MLA 9th edition treats datasets as a type of digital source. The format is:

Author Last Name, First Name. Title of Dataset. Version, Repository Name, Year, DOI or URL.

MLA does not require the [Dataset] label, but it does require a container field, which is the repository name. If the dataset is hosted on a platform such as Harvard Dataverse, ICPSR, or the Inter-university Consortium for Political and Social Research, that platform name functions as the container and must be included.

How to cite a dataset in Chicago format

Chicago style, used most often in history and the humanities, formats a dataset citation in a footnote and bibliography entry. The bibliography entry follows this structure:

Author Last Name, First Name. Title of Dataset. Version. Repository Name, Year. DOI or URL.

Chicago also permits the use of a short-title footnote for subsequent references to the same dataset. If your paper analyzes the same dataset multiple times, use the abbreviated form after the first full citation.

For guidance on citing other source types in your paper, the post on how to cite sources correctly in a research paper covers books, articles, and web sources in comparable detail.

What happens when a dataset does not have a DOI or stable URL?

Some datasets, particularly those shared directly by researchers or hosted on institutional servers without a formal repository, do not have a DOI. In these cases, the correct approach is to cite the most stable access point available and record the date you accessed it. If the dataset was shared with you directly by a researcher, cite it as personal communication in APA or as an unpublished dataset in Chicago, and describe the data source clearly in your methods section.

If you requested data directly from a researcher, the post on how to ask a researcher for their dataset or materials explains the appropriate protocol for that process, including how to document the exchange for citation purposes.

Datasets hosted on government portals such as data.gov or the World Bank Open Data platform are generally stable and can be cited with a URL and access date when no DOI is assigned. Always check whether the hosting repository has assigned an accession number, as some archives use these instead of DOIs.

What are the most common mistakes students make when citing datasets?

The four most common errors in dataset citation are predictable and preventable. Each one has a direct consequence for the credibility of the paper.

The first mistake is citing the paper that describes the dataset rather than the dataset itself. Many publicly available datasets have an associated methods paper or data descriptor published in a journal. Students often cite that paper and assume the dataset is covered. It is not. The dataset and the paper describing it are separate citable objects and must each be cited independently if both are used.

The second mistake is omitting the version number. When a dataset is updated periodically, citing the dataset without specifying the version makes the citation unverifiable. A reviewer cannot confirm your results if they cannot access the exact version of the data you analyzed.

The third mistake is using a standard URL instead of a DOI. URLs to dataset landing pages frequently change when repositories reorganize their file structures. A DOI resolves permanently. If a DOI is available, the URL is not an acceptable substitute.

The fourth mistake is failing to cite the dataset at all and instead describing the data source only in the methods section. A methods section description is not a citation. The dataset must appear in the reference list with full bibliographic information.

How to cite a dataset correctly, step by step

1. Identify the dataset's creator or authoring organization. This may be an individual researcher, a government agency, or a research institution.

2. Record the year the dataset was published or last updated. Check the repository landing page, not the file itself.

3. Copy the exact title of the dataset as it appears in the repository.

4. Note the version number or edition. If none is listed, record the date you accessed the dataset.

5. Record the full name of the repository or data archive where the dataset is hosted.

6. Locate the DOI. It is usually displayed on the dataset landing page. If no DOI exists, copy the most stable URL available.

7. Format the citation using the style required by your target journal: APA 7th, MLA 9th, or Chicago 17th edition.

8. Place the citation in your reference list and add an in-text citation at every point in the paper where you use the data.

If you are working with publicly available data and need help identifying appropriate sources, the post on free datasets every student researcher should know lists verified, citable repositories across multiple disciplines.

PJPCR publishes original student research across all academic disciplines. If your paper uses external data and is ready for peer review, review the submission guidelines at princeton-jpcr.org/submit.

Frequently asked questions about how to cite a dataset

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

A DOI, or Digital Object Identifier, is a permanent alphanumeric string assigned to a digital object such as a dataset, journal article, or report. It resolves to a stable URL regardless of where the file is hosted. For dataset citation, a DOI is preferred over a standard URL because it does not break when a repository reorganizes its file structure. Most major data repositories, including Harvard Dataverse, Zenodo, and the ICPSR archive, assign DOIs to deposited datasets automatically.

How long does it take to properly document all data sources before submitting a paper?

Documenting data sources thoroughly typically requires two to four hours for a paper using one or two datasets, assuming the datasets are hosted in established repositories with DOIs. Papers using multiple datasets or data from non-standard sources may require longer. The standard review and publication timeline at PJPCR is 2 to 3 months from submission to decision. A fast-track option is available for students who need a quicker turnaround. Completing all citations before submission avoids revision requests that delay the process.

Do I need to cite a dataset if I collected the data myself?

If you collected your own original data, you do not cite it in the reference list as an external source. Instead, describe the data collection process fully in your methods section. However, if you deposit your original dataset in a public repository after collection, that deposited dataset receives its own DOI and can be cited by other researchers. Depositing original data is considered good practice in many disciplines and increases the verifiability of your findings.

What makes a dataset citation complete enough to satisfy a peer reviewer?

A complete dataset citation allows a reviewer to locate the exact data you used without contacting you. It must include the creator, year, title, version or access date, repository name, and a DOI or stable URL. A citation that omits the version number or uses only a general repository homepage rather than the specific dataset landing page is incomplete. Reviewers flag incomplete citations because they prevent independent verification, which is a core requirement of peer-reviewed research.

What kinds of research does PJPCR publish, and does it accept data-driven papers?

The Princeton Journal of Pre-Collegiate Research publishes original, peer-reviewed research by high school students across the sciences, social sciences, humanities, and interdisciplinary fields. Data-driven papers are accepted across all disciplines, provided the methodology is sound and all data sources are fully documented and cited. Submission and peer review are free. A publication fee applies for accepted papers. To understand how the review process evaluates submitted work, visit the peer review process page.

Conclusion

Citing a dataset correctly requires six specific elements: creator, year, title, version, repository, and a persistent identifier. The format of those elements varies by citation style, but the requirement to include all six does not. Omitting the version number, substituting a URL for a DOI, or citing a descriptive paper instead of the dataset itself are the errors most likely to draw a reviewer's attention. Addressing them before submission is more efficient than correcting them during revision.

If your research uses external data and your paper is ready for peer review, submit it to PJPCR at princeton-jpcr.org/submit.