CiteOwl

Blog/Getting citations right

How to check if a citation is real (in about five minutes)

The CiteOwl Team· Updated 24 July 2026· 13 min read

To check whether a citation is real, search the exact title in Google Scholar, resolve its DOI at doi.org, confirm the author and journal exist, and make sure the source actually says what it is cited for. Fake references are easy to make and easy to miss, whether one came from a chatbot, a co-author, or your own notes, so this guide turns those checks into a repeatable five-minute routine you can run on any reference before it lands in your bibliography. It also covers the harder cases: how to tell a wholly invented citation from a garbled real one, when a source you can't find is genuinely real, and what to do the moment a reference fails.

Every tool named on this page is free. You need no subscription, no library login, and no browser extension to run the whole routine. The six checks are ordered so the fastest ones catch the most fakes first, and once it's a habit you'll do it almost without thinking.

Search the exact title in quotes

Copy the title and paste it into Google Scholar or plain Google with quotation marks around it. The quotation marks matter: without them the search engine helpfully finds pages that are merely about your topic, which is the noise a fabricated title hides in. A real paper almost always turns up as the first hit. If nothing matches the exact string, that's your first and biggest red flag, and this single check exposes most fabricated references in seconds.

If Scholar comes up empty, try one more index before you conclude anything. OpenAlex is a free, open catalogue of scholarly works, and PubMed covers biomedicine specifically. A paper missing from all three indexes is not a paper that is hard to find. It is a paper that does not exist.

Resolve the DOI

If the citation includes a DOI (a string like 10.1037/abc1234), paste it after https://doi.org/ in your browser. A real DOI takes you straight to the article's official page. A made-up one returns a "DOI Not Found" error. Chatbots are good at generating DOIs that look correctly formatted, so this is one of the most reliable tells: the format can be perfect and the link still dead.

One trap here catches almost everyone. A DOI that resolves is not the same as a DOI that resolves to the paper your citation describes. When researchers examined GPT-4o's fabricated citations in a 2025 JMIR Mental Health study, 64% of the DOIs attached to fake references loaded real but unrelated papers. If you click through and land on something with a different title by different authors, the check has failed, not passed. Read the page you arrive at before you tick the box.

Confirm it on the journal or publisher's site

Find the article on the journal's own website, not just on a results page or a PDF mirror. Real articles have a landing page on the publisher's domain with an abstract, the full author list, and publication details. If the only trace of a paper is the citation itself, it probably isn't a paper.

While you're there, glance at the journal. A real article in a journal nobody has heard of, with a homepage promising review in 72 hours and a fee schedule on the front page, is a different problem from a fabricated citation but still a weak source to build an argument on. How to find credible and peer-reviewed sources covers how to judge a venue, and Unpaywall will find you a legal free copy of a paywalled article when the publisher's page is all you can reach.

Cross-check the details

Now compare what the citation claims against what you found: authors, year, journal, volume, issue, page numbers. A common failure mode isn't a wholly invented paper but a mangled one: real authors stapled to a title they never wrote, or a real article credited to the wrong year and journal. The rule of thumb librarians use: if two or more core fields don't match, treat the reference as unreliable, even if some of it checks out. To pull the authoritative version of a paper's details in one go, paste the title or DOI into the free Crossref metadata search, which returns the registered authors, year, and journal straight from the publisher's record.

Don't skip this step just because the paper turned up. A chatbot reference that exists is still usually wrong in the details: one analysis of 115 ChatGPT medical references found that of the half that pointed to real papers, almost all carried at least one error in the authors, year, volume, or page numbers, leaving just 7% both real and accurate. That study is one of several, and the rates run higher than most people expect across every model tested, as we set out in why AI makes up citations. "It exists" is the start of the check, not the end of it.

Confirm the authors actually exist

Search the lead author's name on Google Scholar, and look them up on ORCID, the registry researchers use to list their own work. A genuine researcher usually leaves a trail: other papers, an ORCID record, a university profile, a Google Scholar page in roughly the same field. An author who appears nowhere except in this one citation is a warning sign.

Check it hasn't been retracted

A real but retracted paper is its own problem: citing it as sound evidence undermines your argument. Search the title with the word "retracted," or look for a retraction notice on the article's page. The quickest way to be sure is the free Retraction Watch database: paste in the title or DOI and it will tell you whether the paper has been pulled. This step matters more than it used to. Nature reported that more than 10,000 papers were retracted in 2023, a record, so a citation that was sound a year ago may not be now.

A few signs should stop you cold, even before you finish the checks: no search result for the exact title, a DOI that won't resolve, a paper that exists only on the page that cited it, real authors attached to an unfamiliar title, or a suspiciously perfect fit where the "paper" says exactly what you needed it to say in a journal you can't quite find. Any one of those is reason enough to set the reference aside.

What kind of wrong is it?

A reference that fails the checks fails in one of four ways, and knowing which one tells you what to do next.

Fully invented

The simplest case. No one wrote the paper, no journal published it, and the DOI points nowhere. The author name is plausible, the title reads like real scholarship, the formatting is flawless, and the whole thing was assembled to look like a citation rather than to refer to one. This is the most common reason a search comes up empty.

A chimera

Real fragments stitched into a combination that was never published. The author is a genuine researcher in the field. The journal is a real journal. The year is reasonable. But that author never wrote that title, or the article appeared somewhere else, or the volume and pages belong to a different paper entirely. Each piece checks out on its own, which is what makes a chimera so convincing. You search the title and find nothing, because the paper as described doesn't exist even though its parts do. The way to confirm one is to search the lead author's name together with a keyword from the title: if the author is real and active but has never published anything close to it, you have a chimera.

Real paper, wrong link

The source genuinely exists and the DOI or URL is broken, so the link 404s or lands on something unrelated. The chatbot got the paper right and the identifier wrong. This one is recoverable: find the paper by its title and authors, take the real DOI from the publisher's page, and correct the entry.

Real paper, wrong claim

The trickiest case, because the citation survives every existence check. The paper is real, the DOI resolves, the authors are right. Then you open it and it has nothing to do with the claim it was attached to. It covers a related topic, or makes the opposite point, or the specific finding simply isn't in there. A real source pinned to a claim it doesn't make is still a misuse, and a marker who opens it will treat it that way. Only step four catches this one, which is why it's the step worth slowing down for.

When a source you can't find is actually real

Not every reference that resists a search is fake. Some genuine papers are just harder to surface, and these are worth ruling out before you delete anything:

• Paywalled articles. The paper exists but the full text sits behind a publisher login. The title should still appear in Google Scholar and the DOI should still resolve, even if you can't read the PDF for free.
• Preprints. A paper posted to a preprint server before formal publication may not show up where you'd expect a journal article, and it might not have a DOI yet.
• Book chapters. A chapter inside an edited volume can be cited in a way that looks like a journal article but indexes differently. Search the book title and editors, not just the chapter.
• Very recent work. A paper published in the last few weeks may not be indexed everywhere yet, so a thin search result doesn't always mean it's fake.
• Non-English sources. Work published in another language can be real and relevant while sitting outside the English-language indexes you usually search.

The test that separates real-but-hard from fabricated is simple: a genuine paper leaves a trail somewhere. The journal's own site, a publisher landing page, an author profile, a record on a preprint server. Find one independent trace and the source is probably real, so confirm the details and cite it. If the citation exists nowhere except in the text that handed it to you, that's your answer.

What to do when a citation turns out to be fake

Don't try to salvage it by reformatting or guessing the "right" version; you'll just bury the problem under tidier packaging. Delete it. Then go back to the claim the citation was supporting and find a source that genuinely makes it, using your library's databases or Google Scholar directly.

Work claim by claim rather than citation by citation. Swapping one unverified reference for another plausible-looking one keeps the exact risk you were trying to remove. For each factual sentence, ask what would actually support it, find a real paper that does, read enough of it to be sure, and attach that. Sometimes the claim survives with a solid source behind it. Sometimes you discover the point was overstated and needs softening. If you search honestly and find nothing at all, that is useful information: the claim itself may not hold up, which is far better to learn now than in feedback.

If one reference in a draft turned out fake, assume it isn't the only one and run the whole list. Fake references are one integrity risk; if you're also wondering whether an AI detector will flag your writing, that's a separate question worth understanding before you submit. For why these fakes appear in the first place, see why AI makes up citations, and for the settings and prompts that cut them down, how to get ChatGPT to cite real sources.

The free tools, and what each one is for

You don't need a paid checker for any of this. Each of these does one job in the routine above:

• Google Scholar confirms a paper exists and shows who has cited it. Your first stop, and the one that catches most fakes.
• doi.org resolves a DOI to the article's official page. A dead DOI is close to proof of fabrication.
• Crossref returns the publisher's own registered metadata for a title or DOI, which is the authoritative version of the authors, year and journal to check your entry against.
• ORCID is the registry where researchers list their own work, so it settles whether an author is real and whether this paper is theirs.
• The Retraction Watch database tells you whether a real paper has since been pulled.
• Unpaywall finds a legal free copy so you can actually read the paper and check step four rather than trusting an abstract.

Checking a whole reference list

The six steps are quick for a handful of sources. For a long bibliography, going one by one gets tedious, and tedium is where mistakes slip in. Two habits keep it manageable. Check as you write rather than at the end, because a reference you verify the moment you add it costs seconds while a list of forty at midnight costs an evening. And check in the order above, so a dead title search closes an entry before you've spent time on its authors.

Dedicated citation checkers can scan a list and flag the entries that don't resolve, which is a sensible safety net before you submit. Remember what they do and don't do: they confirm what's broken, they don't hand you a real source to use instead, and none of them can tell you whether a real paper actually supports your sentence. The most durable fix is to never let an unverified citation into your draft in the first place, which is easier when your writing tool only cites papers it actually retrieved. That is the idea behind an AI research writer that cites real sources, and we compare the options in the best AI tools for academic writing.