You have found a web page that proves exactly what your assignment claims. Half the reference details are then missing: no author, no date, and a URL running across three lines. Most reference lists fail on that gap, not on the style itself.
An internet source is referenced with the author, the year the page was published or last updated, the title, the name of the website, the URL and, in Cite Them Right Harvard, the date you accessed it. The in-text citation carries the surname and year in round brackets.
📌 Key takeaways
- Every web reference needs author, date, title, URL and site.
- Cite Them Right Harvard closes web entries with an accessed date.
- OSCOLA reached its fifth edition, which covers generative AI.
- APA 7 asks for a retrieval date only in exceptions.
- Pew Research Center: 38% of 2013 web pages have gone.
Create a survey for free
With empirio.ai you can create a modern online survey in minutes — with hosting in the EU.
Start for freeWhat counts as an internet source, and when can you cite it?
An internet source is a text, image, video or dataset published online and reachable through a URL, with no printed edition that carries its own page numbers. Finding something on the web does not by itself make it an internet source.
The distinction gets missed constantly, and it produces the first wave of faulty references. A journal article stays a journal article once you download the PDF, so it takes volume, issue, page range and, where one exists, a DOI. An e-book stays a book, as our guide to citing books sets out. Genuine internet sources are the items that exist only online: blog posts, press releases, government pages, online encyclopaedias, statistics portals, videos, forum threads and social media posts.
Whether the source is worth citing at all is the second question. Preferring print on principle no longer holds, because a great deal of reliable material appears online and nowhere else, from official statistics to legislation. Three tests decide it instead: a named person or organisation is responsible, a specialist can check the content, and the page can be found again reliably. A Wikipedia article fails the first test, which makes it a starting point rather than evidence.
The six parts of a reference for a web page
A complete reference for a web page has six parts: author, date, title, name of the website, URL and, depending on the style, the date you accessed it. Order and punctuation differ between styles, the six parts do not.
All six sit on the page itself or on its about page. Collecting them in that order before you format anything saves rework. Anyone who formats first and hunts afterwards misses the last-updated date, which many systems park right at the foot of the page.
| Part | Where to find it | If it is missing |
|---|---|---|
| Author | above or below the article, else the about page | use the organisation, else start with the title |
| Date | article header, page footer, last-updated note | mark the entry as undated |
| Title | heading of the article, else the browser tab | describe the function of the page briefly |
| Website name | masthead, logo, about page | use the domain without the protocol |
| URL | address bar, better the permalink of the page | never optional, no URL means no reference |
| Accessed date | the day of your own last visit | required by Cite Them Right, rare in APA |
For the author, an organisation steps in when no person is named, so a government body or a university takes the slot. Only when that fails does the title move to the front of the entry.
For the date, the last-updated date beats the original publication date, because the version you read is the version you reference. Cite Them Right asks for the year the page was published or last updated, and the accessed date pins down which version you saw.
Accessed dates: what Cite Them Right requires and APA does not
Cite Them Right Harvard puts an accessed date on every online source, and APA 7 puts one only on unarchived pages that are meant to change. Applying the rule of one style inside the other produces a formal error.
The reasoning behind each rule differs rather than contradicting the other. Cite Them Right treats the accessed date as proof of which version you saw, so the closing element appears on web pages, online reports, newspaper articles and journal articles alike. APA treats a publication date as sufficient identification, so a retrieval date adds only a note about your own calendar.
| Style | Accessed date | What the rule says |
|---|---|---|
| Cite Them Right Harvard | yes, on online sources | entry closes with the accessed date in brackets |
| APA 7 | exception only | unarchived pages likely or intended to change |
| OSCOLA 5 | persistent link preferred | section 3.7.1 encourages permalinks or DOIs |
| MHRA and RSC | department decides | check the guidance issued by your school |
Status: August 2026, subject to change.
Newcastle University sets out the pattern for a web page with a named author in its Cite Them Right Harvard style sheet for 2025: surname and initial, the year the page was published or last updated, the title, then Available at: with the URL and the accessed date. A worked example there closes with 26 March 2025.
Purdue OWL states the APA rule the other way round: retrieval dates are required for unarchived sources that are likely, or intended, to change over time. A public wiki qualifies, a dated press release does not.
⚠️ Watch out
An accessed date does not replace a missing publication date. Where both are absent, mark the entry as undated and check whether a firmer source carries the same claim.
Create a survey for free
With empirio.ai you can create a modern online survey in minutes — with hosting in the EU.
Start for freeCiting internet sources: worked examples in three styles
Three referencing systems cover most UK assignments, and each arranges the same six parts differently. All three examples below use one source: the Pew Research Center study on disappearing web content, published on 17 May 2024.
One warning saves a lot of rework, because referencing systems are closed. Author-date systems such as Cite Them Right Harvard and APA put the citation in round brackets in the running text, while footnote systems such as OSCOLA send it to the bottom of the page. Mixing the two inside one assignment is the most common formal error of all. The groundwork is in our guide to correct citing.
Referencing a web page in Cite Them Right Harvard
Cite Them Right Harvard opens with surname and initial, follows with the year in round brackets, italicises the title of the web page and closes with Available at: the URL and the accessed date. The reference list records every author, while the in-text citation shortens to et al. from four authors upwards. More detail sits in our guide to the Harvard referencing style.
Reference list: Chapekis, A., Bestvater, S., Remy, E. and Rivero, G. (2024) When online content disappears. Available at: https://www.pewresearch.org/data-labs/2024/05/17/when-online-content-disappears/ (Accessed: 27 August 2026).
In-text citation: (Chapekis et al., 2024)
Referencing a web page in APA 7
APA 7 puts author and full date at the front, italicises the title of the page, names the site and ends with the bare URL. No retrieval date appears here, because the page carries a fixed publication date. Psychology departments in the UK usually work in APA rather than Harvard, so confirm which of the two your school expects before you build the list.
Reference list: Chapekis, A., Bestvater, S., Remy, E., & Rivero, G. (2024, May 17). When online content disappears. Pew Research Center. https://www.pewresearch.org/data-labs/2024/05/17/when-online-content-disappears/
In-text citation: (Chapekis et al., 2024)
Referencing a website in OSCOLA
OSCOLA works in footnotes and gives the author, the title in single quotation marks, then the site name and date in round brackets, followed by the link in angle brackets. Section 3.1.1 of the fifth edition shortens more than three authors to the first name plus and others, so the Pew study runs under Athena Chapekis and others. Section 3.7.1 prefers a persistent link or a DOI, and with one of those the date of access falls away; where neither exists, the page address goes in angle brackets and the date of access follows it. The free PDF comes from the Oxford Law Faculty.
Footnote pattern: Author and others, ‘Title of Page’ (Site Name, Day Month Year) <URL> accessed Day Month Year.
Applied to the example: Athena Chapekis and others, ‘When Online Content Disappears’ (Pew Research Center, 17 May 2024) <https://www.pewresearch.org/data-labs/2024/05/17/when-online-content-disappears/> accessed 27 August 2026.
Bibliography: Chapekis A and others, ‘When Online Content Disappears’ (Pew Research Center, 17 May 2024) <https://www.pewresearch.org/data-labs/2024/05/17/when-online-content-disappears/> accessed 27 August 2026
Special cases: no author, PDF, video, social media, AI chat
Special cases in web referencing all follow one logic: when a part is missing, the next most important part moves up, and the gap gets marked rather than passed over silently. Nothing may be invented, including a year that looks about right.
Seven cases cover most of what turns up in a dissertation. Departmental guidance overrides each of them where it differs.
- No author: put the organisation in the author slot. Where none is identifiable, the entry starts with the title.
- No date: Cite Them Right marks the entry undated, APA uses n.d. Both markers also appear in the in-text citation.
- PDF on a website: reference by document type, so as a report or a paper, and give the URL of the PDF itself.
- Video: the channel or uploader is the author, the upload date is the date, and a timestamp pins down a specific claim.
- Social media post: profile name, date, the opening words as the title, and the permalink of the single post.
- Legislation: reference by short title and section rather than as a web page. Law schools use OSCOLA for the whole category.
- Secondary referencing: read the original wherever you can. Otherwise reference the work you read and name the original with cited in.
Output from a language model deserves its own sentence, because the rules have moved every year since 2023. OSCOLA added guidance for generative artificial intelligence at section 3.7.13 of the fifth edition and a section for podcasts at 3.7.2, while social media posts are handled inside the website section at 3.7.1. Whether your department adopts that, demands a declaration or bans the tool outright is set out in your assessment regulations. What follows a breach is covered in our guide to the plagiarism check.
Why the standard most UK universities use sits behind a paywall
Cite Them Right is a commercial publication rather than a public standard, and the binding version sits behind an institutional library licence. Students without a university login cannot reach the text their marks depend on.
Manchester Metropolitan University states the position plainly on its referencing page: most students there follow Cite Them Right Harvard, which the library calls an externally published version of Harvard, while some departments use a different style altogether. Every route to the guidance runs through a library sign-in, and the library recommends using the ebook or the print book in conjunction with the website. Calling Harvard a single national standard is therefore wrong on two counts.
The regular case for you is straightforward: the binding rule is whatever your school, department or assessment regulations set out. The exception is the case where no guidance exists at all, and you then pick one established style and apply it consistently. Responsibility for interpreting the rule rests with your supervisor, never with a guide on the web. Two free routes cover the outline in the meantime, since Newcastle University publishes its style sheet as an open PDF and the Oxford Law Faculty publishes OSCOLA under a Creative Commons licence. This article does not replace assessment advice.
Marks follow consistency, not the choice of style.
Create a survey for free
With empirio.ai you can create a modern online survey in minutes — with hosting in the EU.
Start for freeDead links: why the URL alone is not enough
A URL is a state rather than a permanent address. Research by the Pew Research Center from May 2024 found that 38% of the web pages that existed in 2013 were unreachable ten years later.
For a dissertation the decay is a practical problem: if someone checks your sources a year after submission, the page may be gone. Pew also found that 23% of the news pages sampled carried at least one dead link, and that more than half of the Wikipedia articles sampled had at least one dead reference. Three tools help against it, and none of them costs anything.
- DOI: a Digital Object Identifier points at a document permanently, even after the publisher moves the address. Crossref asks for the form https://doi.org/10.xxxx in its display guidelines.
- Persistent links: permalink services mint a stable address for a single page. OSCOLA encourages exactly that at section 3.7.1 of the fifth edition.
- Web archive: the Wayback Machine of the Internet Archive stores a version of the page at its own dated address, and linking to it is expressly allowed.
💡 Tip
Save every internet source on the day you access it with the Save Page Now function of the Wayback Machine, and note the archive address beside your reference. A screenshot in your own files costs nothing and settles arguments later.
Common mistakes when citing internet sources
Most marks lost on internet sources come from haste in the final days before submission rather than from ignorance. Four mistakes recur often enough to be worth a targeted check.
The repairs take minutes each. Running the check once across the whole reference list is faster than working through it entry by entry.
Giving the home page instead of the exact page
A reference has to lead to the page the claim sits on. Giving the home page of the site instead forces the reader to search and defeats the purpose of the entry. Statistics portals and government sites are the worst offenders, because their front pages change daily. Take the address of the specific page, and prefer the permalink wherever the site offers one.
Breaking a URL with a hyphen
A URL is never broken with a hyphen, because the hyphen might belong to the address and makes the link unusable. Break long addresses after a forward slash instead, or switch off automatic hyphenation for the entry. Shortening services solve nothing either, since they can be switched off and they hide where the source came from.
Using the copyright year in the footer as the publication date
The year in the footer belongs to the website as a whole rather than to the individual page, so it is not a publication date. Taking it over dates a text from 2019 to 2026 by accident. Look for the date on the article itself, check for a last-updated note, and mark the entry as undated when nothing at all can be found.
Putting internet sources in a separate list
A separate list for web sources alone appears in none of the major referencing styles and makes looking things up harder, because the reader has to search twice. Where your department demands one, follow the department. Without such a rule, every source belongs in one alphabetical list, which Cite Them Right calls a bibliography once it also holds background reading.
Conclusion
Referencing internet sources rests less on rule knowledge than on care while collecting the six parts. Settle first which style your department requires, gather author, date, title, site name and permalink while you read, and archive the page before it disappears. Add the accessed date wherever your style asks for it, which in Cite Them Right Harvard means almost every online entry.
One habit carries the whole job: build the reference while the page is still open in front of you, not in the week of submission. Cite Them Right Harvard, APA and OSCOLA all reward the same discipline, and each of them punishes a reference list assembled from browser history.
Where to go next
- Starting from the basics? Correct citing: the full guide
- Referencing books and e-books? How to cite books correctly
- Writing the whole dissertation? Bachelor thesis: guide and structure
Do you need your own data instead of somebody else's web pages?
Your own survey is never referenced as an internet source, because it becomes your own dataset. With empirio.ai, an online survey tool from Germany, you can build one free of charge. How to structure the questions is set out in our guide to creating a questionnaire.
Frequently asked questions
Create a survey for free
With empirio.ai you can create a modern online survey in minutes — with hosting in the EU.
Start for free