How to Cite a Dataset in Harvard Referencing

This guide explains how to format a dataset citation in Harvard Referencing. Follow the format template below, refer to the real example, and use the step-by-step instructions to build your citation correctly.

Format Template

The general pattern for citing a dataset in Harvard Referencing:

Template

Author/Creator (Year) *Title of dataset* [Dataset]. Repository/Publisher. Available at: DOI or URL.

In-Text Citation Example

How the citation appears within the body of your paper:

In-text citation

(Office for National Statistics, 2024)

Reference List Entry Example

The full citation as it appears in your reference list:

Reference list entry

Office for National Statistics (2024) *Annual population survey: regional labour market statistics* [Dataset]. UK Data Service. Available at: https://doi.org/10.5255/UKDA-SN-7656-20.

Step-by-Step Guide

1
Identify the creator
List the person(s) or organisation that created or compiled the dataset. Use the corporate author name if the dataset was produced by an organisation.
2
Add the year
Place the year the dataset was published or last updated in parentheses.
3
Write the dataset title
Italicise the title using sentence case. Follow with '[Dataset]' in square brackets.
4
Name the repository
Identify the repository or publisher that hosts the dataset (e.g., UK Data Service, Zenodo, Dryad).
5
Include the DOI or URL
Write 'Available at:' followed by the DOI as a URL if available. If no DOI exists, provide the direct URL to the dataset.

Common Mistakes to Avoid

Forgetting to include '[Dataset]' after the title.
Not italicising the dataset title.
Confusing the data creator with the repository — the creator is the author, the repository is the publisher.
Omitting the DOI when one is available.
Using the repository home page URL instead of the direct link to the dataset.

Frequently Asked Questions

How do I cite a specific version of a dataset in Harvard?

Include the version number after the title: *Title of dataset* (Version 2.0) [Dataset]. This helps readers identify the exact version you used.

Should I cite the dataset or the paper describing it?

Cite the dataset directly if you used the data itself. Cite the accompanying paper if you are referencing the methodology. You may cite both if both are relevant.

What if the dataset has no identifiable author?

Use the organisation or repository that published the dataset as the corporate author. If neither is available, begin the reference with the title.

Information You Will Need

Before creating your citation, gather the following information from your dataset:

Author/Creator
Title
Version
Publisher/Repository
Year
DOI/URL

Generate citations automatically with CiteDash

Stop formatting citations by hand. CiteDash automatically generates accurate Harvard Referencing citations for all your sources while you research and write.

Start Free Trial More Harvard Referencing Guides

Step-by-Step Guide

Identify the creator

List the person(s) or organisation that created or compiled the dataset. Use the corporate author name if the dataset was produced by an organisation.

Add the year

Place the year the dataset was published or last updated in parentheses.

Write the dataset title

Italicise the title using sentence case. Follow with '[Dataset]' in square brackets.

Name the repository

Identify the repository or publisher that hosts the dataset (e.g., UK Data Service, Zenodo, Dryad).

Include the DOI or URL

Write 'Available at:' followed by the DOI as a URL if available. If no DOI exists, provide the direct URL to the dataset.

Common Mistakes to Avoid

Forgetting to include '[Dataset]' after the title.

Not italicising the dataset title.

Confusing the data creator with the repository — the creator is the author, the repository is the publisher.

Omitting the DOI when one is available.

Using the repository home page URL instead of the direct link to the dataset.