Wikipedia:Verifiability

This page documents an English Wikipedia policy.

It describes a widely accepted standard that all editors should normally follow. Changes made to it should reflect consensus.

Shortcut

[[:|WP:V]]

Wikipedia should only publish material that is verifiable and is not original research.

The goal of Wikipedia is to become a complete and reliable encyclopedia. Verifiability is the key to becoming a reliable resource, so editors should cite credible sources so that their edits can be easily verified by readers and other editors.

One of the keys to writing good encyclopedia articles is to understand that they should refer only to facts, assertions, theories, ideas, claims, opinions, and arguments that have already been published by a reputable publisher.

It's important to note that "verifiability" in this context does not mean that editors are expected to verify whether, for example, the contents of a New York Times article are true. In fact, editors are strongly discouraged from conducting this kind of research, because original research may not be published in Wikipedia. Articles should contain only material that has been published by reputable or credible sources, regardless of whether individual editors regard that material to be true or false. As counter-intuitive as it may seem, the threshold for inclusion in Wikipedia is verifiability, not truth. For that reason, it is vital that editors rely on good sources.

Wikipedia:Verifiability is one of three of Wikipedia's content-guiding policy pages. The other two are Wikipedia:No original research and Wikipedia:Neutral point of view. Jointly, these three policies determine the type and quality of material that is acceptable in the main namespace. The three policies are complementary. They should therefore not be interpreted in isolation from one other, and editors should try to familiarize themselves with all three.

When adding information

Fact checking is time consuming. It is unreasonable to expect other editors to dig for sources to check your work, particularly when the initial content is questionable. The burden of evidence lies with the editor who has made the edit. Editors should therefore be specific, avoid weasel words, and provide references, linking to the source if it's online, and giving a brief citation in brackets after the sentence if it isn't. Giving citations in brackets after a sentence is called Harvard referencing. For example:

After the bombing, a spokesperson in Bali said you can't have total security.

This is difficult to verify. Many spokespersons may have commented on the incident, and it's unreasonable to expect someone to check all these statements looking for the one that matches. Consider instead:

Andi Mallarangeng, spokesperson for Susilo Bambang Yudhoyono, Indonesia's president, told reporters: "Suicide bombers in open places. You can't put metal detectors everywhere. You have to be realistic." [1]

This is easy to verify. A inline link to an article is provided, and the exact quote is given, so it can be fed into a search engine to find other sources for it. If there were no online sources for this quote, a brief citation could be given in brackets after the quote, like this: (The Guardian, October 3, 2005), with the page number if you know it. Then, whether the source is online or not, add a full citation in the References section at the end of the page:

==References==

"Suicide attacks leave 22 dead and 130 injured" by John Aglionby, The Guardian, October 3, 2005

Giving a full citation means that, if an online source is removed from the website at a later date, readers will still know how to track down the original article.

If you use documents in a foreign language, it is better, in the English Wikipedia, to reference both the original and its translation. The original provides the credibility, while the translation means others can check that the document has been used correctly as a source.

See Wikipedia:Cite sources for the related style guide regarding different ways of formatting citations (for example, by providing footnotes), although note that it is not policy: providing some information about your sources is more important than getting the format right.

Checking content

There are several reasons you might want to check the accuracy of an edit:

The author has a record of contributing inaccurate or misleading information.
The author has a conflict of interest.
There are other errors in the article, and the entire text needs to be checked.
The article is the subject of an accuracy dispute.
The article is about a contentious subject.
The subject area is one where errors are frequent.
The statement is implausible on its surface.
The statement is key to the entry as a whole.
The statement is overly vague.

Here's a suggested procedure for verifying content.

If you find a recent change and are not sure whether or not an edit is accurate, add the page and the diff to Wikipedia:RC patrol
If you feel the urge to remove a statement from an article, first check the bottom of the article for references.
If there are any, check the sources. If the sources are reputable or credible, and you can confirm the statement with reference to them, leave it in; otherwise, continue.
If there is a talk page, check that. The statement may already have been checked, so there's no need to repeat the procedure. However, if a reference or citation was only given on the talk page, move it to the article to help people who might want to check it in the future.
Use your common sense to work out what other resources would help, and check them. If you can find credible sources that support the statement using these resources, leave it in; otherwise, continue.
Move or copy the statement to the talk page, explaining that you have not been able to find a source for the statement, and stating what sources you have checked.
Optionally, check the article history for who added the statement in the first place, and leave a note on their talk page telling them that their statement is disputed, and directing them to the appropriate talk page.
Anyone may now feel free to try to find a source to support the statement and produce a citation or link on the talk page.
If you only copied the statement, wait for a period (probably at least one day), and if no-one has found a reference in that time, remove it from the article altogether.
If someone does find a reference, the statement should be put back into the article, with the newly found reference. To make it clear which statement used which reference, it might be worth numbering the references and then referring to them in the article like this[1] or like this¹. Alternatively, use Harvard referencing, which involves adding a brief citation in brackets after the sentence or paragraph — if it's a book by John Smith that was published in 2005, write (Smith 2005) — then adding a full citation for Smith 2005 in the References section at the end of the article (Smith, J. My story, Random House, 2005). If no-one finds a reference, the statement can remain on the talk page indefinitely.
Extraordinary claims require extraordinary evidence. Strong substantive evidence is required to support wilder claims. For example, a newspaper report may be sufficient evidence to support a sports result, but not to support a new detailed mathematical theory.

Dubious sources

For an encyclopedia, sources should be unimpeachable. An encyclopedia is not primary source material. Its authors do not conduct interviews or perform original research. Therefore, anything we include should have been published in the records, reportage, research, or studies of other reputable sources. Sources should be appropriate to the claims made: outlandish claims beg strong sources.

Sometimes a particular statement can only be found in a publication of dubious reliability, such as a tabloid newspaper. If the statement is relatively unimportant, then just remove it — don't waste words on statements of limited interest and dubious truth. However, if you must keep it, then attribute it to the source in question. For example:

According to the British tabloid newspaper The Sun, the average American has 3.8 cousins and 7.4 nephews and nieces.

Personal websites and blogs are not acceptable as sources, except on the rare occasion that a well-known person, or a known professional journalist or researcher in a relevant field, has set up such a website. Remember that it is easy for anybody to create a website and to claim to be an expert in a certain field, or to start an "expert group", "human rights group", church, or other type of association. Several million people have created their own blogs in the last few years. They are not regarded as acceptable sources for Wikipedia. See Wikipedia:Reliable sources for more information.

Sources in languages other than English

Because this is the English Wikipedia, English-language sources should be given whenever possible. However, this is not always possible or practical, for instance when the article discusses events or institutions pertaining to a country where English is not a native language. In such a case, focusing on English-language sources often implies focusing on outside sources, which may not be as well informed as the local sources, or may introduce systemic bias.

In the case where the original source material is not in the English language, there is a tension between accessibilty and verifiability:

On the one hand, this is the English language Wikipedia. Readers may not be able to read source materials in other languages, and thus require translations of those source materials into English so that they can read them.
On the other hand, translations, whether performed by a Wikipedia editor directly when writing content or performed by a separate translator who has published a translation of the source material, are inherently subject to error. A translation may be inaccurate or misleading. Readers have to be able to verify for themselves what the original source material actually said. Furthermore, readers require the ability to verify that the source material was published by a credible source and that it was peer reviewed, and thus require information about the original source.

Therefore, where the original source material is in a language other than English:

The original source in the original language should be cited, so that readers and editors can evaluate the reliability and credibility of the original source, can determine whether the original source was peer reviewed, and can verify that the article content is supported by the source material.
If any published translations of that material into English are used, they should be cited, so that readers and editors can evaluate the reliability and credibility of the translator, can verify the accuracy of the translation (by cross-checking with other published translations, by translating themselves, and by examining the reputation of the translator), and can verify that the article content is in fact supported by the translation used.
Where sources are directly quoted, published translations, where the reliability and credibility of the translator can be evaluated by readers, are preferred over editors performing their own translations directly. (If a published translation is erroneous, then the error can be pointed out. Editors should cite sources for such errata, too, of course.)
Where editors directly quote a non-English source, performing their own translations into English without reference to a published translation, then the original quote in the original language must be cited, or supplied on the page, so that readers can check the translation made by the editor.

This policy covers what should be cited and why. See Wikipedia:cite sources for style guidelines, and how to cite non-English sources and their respective translations.

Verifiability, not truth

Articles in Wikipedia should refer to facts, assertions, theories, ideas, claims, opinions, and arguments that have been published by a reputable or credible publisher. The threshold for inclusion in Wikipedia is verifiability, not truth.

A good way to look at the distinction between verifiability and truth is with the following example. Suppose you are writing a Wikipedia entry on a famous physicist's Theory X. Theory X has been published in peer-reviewed journals and is therefore an appropriate subject for a Wikipedia article. However, in the course of writing the article, you meet the physicist, and over a beer, he tells you: "Actually, I think Theory X is a load of rubbish." Even though you have this from the author himself, you cannot include the fact that he said it in your Wikipedia entry. Why not? The answer is that it is not verifiable in a way that would satisfy the Wikipedia readership or other editors. The readers don't know who you are. You can't include your telephone number so that every reader in the world can call you directly for confirmation. And even if they could do this, why should they believe you?

If you were firmly convinced that this new information should be published in Wikipedia, how would you go about getting it into Wikipedia? For the information to be acceptable to Wikipedia, you would have to contact a reputable news organization, and persuade them to publish your story, which would then go through a process similar to peer review before being published. It would be checked by a reporter, an editor, perhaps by a fact-checker, and if the story were problematic, it would be checked further by the lawyers and the editor-in-chief. These checks and balances exist to ensure that only accurate and fair stories appear in the newspaper. It is this fact-checking process that Wikipedia is not in a position to provide, which is why the policy of no original research is an important one. If the newspaper published the story, you could then include the information in your Wikipedia entry, citing the newspaper article as your source.

Obscure topics

Subjects which have never been written about in published sources, or which have only been written about in sources of doubtful credibility should not be included in Wikipedia. One of the reasons for this policy is the difficulty of verifying the information. As there are no reputable sources available, it would require original research, and Wikipedia is not a place to publish original research. Insistence on verifiability is often sufficient to exclude such articles.

Conclusion

However, just because some information is verifiable, doesn't mean that Wikipedia is the right place to publish it. See what Wikipedia is not.

See criteria for inclusion of biographies and auto-biography for some suggested criteria for inclusion of biographical articles.

"Доверяй, но проверяй (Doveriai no proveriai)" — Russian proverb (Trust but verify)