Wikipedia:Attribution/FAQ

Wikipedia is not a publisher of original thought: all material published by Wikipedia must be attributable to a reliable published source. The threshold for inclusion in Wikipedia is whether material can be attributed, not whether it is true. For more details, see Attribution, which is proposed as policy.

This is a subsidiary page of Attribution, answering questions and offering examples that illustrate key aspects of the policy. If you don't find your question here, post it on the talk page.

The essence of Attribution
All material in Wikipedia must be attributable to a reliable, published source; that is, a reliable, published source must exist for it. If none does, the material is regarded as original research and should be removed. In reality, not all material must actually be attributed, but attribution is required for quotations and for any material that is challenged or likely to be challenged.

Doesn't Wikipedia care about truth?
Wikipedians do care about the truth, but we are mindful of our own limitations. We want to produce a high-quality encyclopedia, and by insisting on the use of reliable sources, we depend on the published research and opinions of informed commentators. Editors should ensure that all majority and significant-minority opinions are included in articles, in rough proportion to the representation of those views in reliable published sources.

A fundamental threshold for inclusion in Wikipedia is whether material can be reliably attributed—and this is independent of whether any individual editor holds it to be true. In particular, material that an editor believes to be true but that cannot be attributed to a reliable source should not be included in Wikipedia. Articles should simply present reliably attributed statements, views, and arguments, and then allow our readers to judge truth for themselves.

Wikipedians, as such, should not claim expertise; we cannot decide the truth in any field. But we can follow the consensus of experts. In some subjects, this consensus is likely to be the truth; in others, it is the best information available. Where experts disagree, all we can do is accurately report the debate, with the strengths and weaknesses of all sides. As human knowledge improves, and Wikipedia incorporates more of it, Wikipedia will become more accurate.

Does this mean we have to include every crank view that can get itself published?
It may sometimes be necessary to include views that one side of a debate regards as cranky. We have other policies; one of the most important of them is the Neutral Point of View. If there is in fact a widespread conflict between two points of view, we have to include both of them, so long as reliable sources have written about them.

This does not mean we have to include crankery in general. One of the chief requirements of neutrality is that Wikipedia may not give undue weight to any side, more weight than its support would warrant. In extreme cases, where the crank theory is rejected or ignored by all relevant, reliable sources, due weight may be no mention at all.

This is why the exception for dubious and self-published sources requires that they be used in Wikipedia articles largely written from sources independent of the subject. If there are no independent sources on a subject, Wikipedia should not have an article on it. It doesn't matter whether the crank has started a website, written a self-published book, or bought a publishing house: if no-one else has taken note of his theory, why should we?

For further information on this issue, including several examples of past community decisions, see WP:FRINGE and WP:SCIENCE.

What are primary, secondary, and tertiary sources?

 * Primary sources are documents or people very close to the situation you are writing about. An eyewitness account of a traffic accident published in a newspaper is a primary source. Parts of the Bible are primary sources. The White House's summary of a George Bush speech is a primary source. Publicly available databases, such as citation indexes and census surveys, are primary sources. Primary sources that have been published by a reliable source may be used for the purposes of attribution in Wikipedia, but only with care, because it's easy to misuse primary sources. For that reason, edits that rely on primary sources should only make descriptive claims that can be checked by anyone without specialist knowledge. Any interpretation of primary source material requires a secondary source.
 * Examples of primary sources include: archaeological artifacts; photographs; historical documents such as diaries, census results, video or transcripts of surveillance, public hearings, trials, or interviews; tabulated results of surveys or questionnaires; written or recorded notes of laboratory and field experiments or observations; and artistic and fictional works such as poems, scripts, screenplays, novels, motion pictures, videos, and television programs.


 * Secondary sources are documents or people that summarize other material, usually primary source material. They are academics, journalists, and other researchers, and the papers and books they produce. A theologian's account of what the Bible says is a secondary source. A sociologist thesis based on his research of primary sources is a secondary source. A journalist analysis of a traffic accident, is a secondary source.A New York Times analysis of a George Bush speech is a secondary source. Wikipedia articles should rely on reliable, published secondary sources wherever possible. This means that we publish the opinions of reliable authors, and not the opinions of Wikipedians who have read the primary source material for themselves.


 * Tertiary sources are publications, such as encyclopedias, that sum up secondary sources, and sometimes primary sources. Wikipedia is a tertiary source.

What if a self-published source disagrees with a third-party reliable source?
Where a self-published source in an article about itself disagrees with a majority view in reliable sources, the self-published source may be used to demonstrate the author's opinion, so long as doing so is consistent with the self-publication provision of WP:ATT, and the undue weight provision of WP:NPOV.

What kinds of sources are generally regarded as reliable?
Reliable sources are credible published materials with a reliable publication process; their authors are generally regarded as trustworthy, or are authoritative in relation to the subject at hand:


 * Books and journals published by universities and known publishing houses;
 * Mainstream newspapers and magazines published by notable media outlets;
 * Books written by widely published authors;
 * Mainstream websites published and maintained by notable media outlets;

Note that the reliability of a source depends on context; what is reliable in one topic may not be in another. A world-renowned mathematician may not be a reliable source on topics of biology.

Some criteria that can assist editors in evaluating non-scholarly sources. Note that these are not hard rules, and that sources needs to be always evaluated in the context of the article's subject:
 * Editorial oversight—A publication with a declared editorial policy will have greater reliability than one without, since the content is subject to verification. Self published sources such as personal web pages, personally published print runs and blogs have not been subject to any form of independent fact-checking and so have lower levels of reliability than published news media (e.g. The Economist) and other sources with editorial oversight, which is less reliable itself than professional or peer reviewed journal (e.g. Nature). Note that some of a publication's content, such as Op-ed pieces, commentary, announcements, advertising, etc., may have little or no editorial oversight, and could be treated as self-published material.


 * Declaration of sources—A source that is explicit about the data from which it derives its conclusions my be more reliable than one that does not. Ideally, a source should describe the collection process and analysis method.


 * Corroboration—If two or more independent originators agree, in a reliable manner, then the conclusions become more reliable. Care must be taken to establish that corroboration is indeed independent, to avoid an invalid conclusion based on uncredited origination.


 * Age of the source and rate of change of the subject—Historical or out-of-date sources may be used to demonstrate evolution of the subject but should be treated with caution where used to illustrate the subject. If no newer sources are available, it is reasonable to caveat use of sources with an indication of the age of the source.


 * Persistence—If a reader goes to the cited source to validate a statement, or to gain further understanding of the topic, the form cited should remain stable, continuing to contain the information used by the editor to support the words. In this sense a book or journal citation is superior to an online source where the link may become broken. Some web resources have editorial policies that lead to a lack of persistence; therefore, web citations should include the date in which the source was retrieved.

These issues are particularly pertinent to Wikipedia where various editors involved in an article may have their own expertise or position with respect to the topic. Not all sources on a topic are equally reliable, and some sources will have differing degrees of reliability in different contexts.

In general, a topic should use the most reliable sources available to its editors. Common sense is required to determine what sources to use; this guideline cannot be applied robotically. If you have questions about a source's reliability, discuss with other editors on the article's talk page, or if the source is already used in the article, you can draw attention to it with the unreliable template.

What kinds of sources are generally regarded as unreliable?
Some sources are generally unacceptable for use as references in Wikipedia:


 * An anonymous source is an unnamed person or a work created by an unnamed author. Anonymous sources are not acceptable in Wikipedia, because we can't attribute the viewpoint to its author. Anonymous sources whose material is published by reliable secondary sources, such as Deep Throat in The Washington Post, are acceptable, because Wikipedia's source in this case would be the newspaper, not the anonymous source.


 * An unpublished source is one that is not publicly available, or that has been distributed only through anonymous channels or forums, and for which a publisher cannot be identified. Unpublished sources may never be used as sources on Wikipedia.


 * An obsolete source is one that is out-of-date, or has been officially withdrawn or deprecated by its author(s) or publisher. Editors of articles on fast-moving subjects such as law, science, or current events should ensure they use the latest sources.


 * A confidential source, i.e. those sources that are considered confidential by the originating publisher may hold uncertain authority, as the original cannot be used to validate the reference.


 * A questionable source is one with no independent editorial oversight or fact-checking process, or with a poor reputation for fact-checking. This includes websites and publications that express political, religious, anti-religious, or racist views that are widely acknowledged as extremist. It also includes gossip columns, tabloids, and sources that are entirely promotional in nature. Questionable sources should usually not be used as sources except in articles about themselves; see the self-publication provision of the policy.


 * A self-published source is material, online or in print, that has been published by the author, or whose publisher is a vanity press, web hosting service, or other organization that provides little or no editorial oversight. The expression "self-published source" may also refer to the author of the material. Personal websites, blogs, Wikipedia, and messages on Usenet and Internet message boards are considered self-published. With self-published sources, no independent entity stands between the author and publication; the material may not have been subject to any form of fact-checking, legal scrutiny, or peer review. Anyone can create a website or pay to have a book published and then claim to be an expert in a certain field. For that reason, self-published material is usually not acceptable as a reliable source, with some exceptions.

Are Usenet postings reliable sources?
Posts on Usenet are almost never regarded as reliable sources, because they are easily forged or misrepresented, and many are anonymous or pseudonymous.

One exception is that some authorities on certain topics have written extensively on Usenet, and their writings there are vouched for by them or by other reliable sources. A canonical example is J. Michael Straczynski, the creator of the television series Babylon 5, who discussed the show at length on Usenet. His postings are archived and authenticated on his website, and may be an acceptable source on the topic of Babylon 5 under the self-publication provision of WP:ATT.

Are weblogs reliable sources?
In most cases, no. Most weblogs ("blogs"), especially those hosted by blog-hosting services such as Blogger, are self-published sources; many of them published pseudonymously. There is no fact-checking process and no guarantee of quality of reliability. Information from a blog may be usable in an article about that blog or blogger under the self-publication provision of WP:ATT.

Weblog material written by well-known professional researchers writing within their field, or well-known professional journalists, may be acceptable, especially if hosted by a university or employer (a typical example is Language Log, which is already cited in several articles, e.g. Snowclone, Drudge Report). Usually, subject experts will publish in sources with greater levels of editorial control such as research journals, which should be preferred over blog entries if such sources are available.

Blogs must never be used as secondary sources on living persons; see WP:BLP.

Are web forums and blog talkbacks reliable sources?
Web forums and the talkback section of weblogs are not regarded as reliable. While they are often controlled by a single party (as opposed to the distributed nature of Usenet), many still permit anonymous commentary and we have no way of verifying the identity of a poster.

Are wikis reliable sources?
Wikis, including Wikipedia and other wikis sponsored by the Wikimedia Foundation are not regarded as reliable sources. However, wikis are excellent places to locate primary and secondary sources. Many of them license content under the GFDL, which might be worth importing into Wikipedia, but once imported, the material is subject to Attribution and Neutral point of view.

Despite the above, some wikis sponsored by the Wikimedia Foundation are in fact intended to be reliable sources. Notably the Wikisource and Wikinews projects, which provide notable documents in the public domain and copyleft, and events in day-to-day news, respectively.

If circumstances require linking to a wiki page—for example, if the wiki itself is a notable project—it is best to use the permalink feature common on wiki software. Common wiki platforms, including the MediaWiki software that underlies Wikipedia, incorporate a feature allowing one to link directly to a version of a page as it existed some time in the past. To illustrate, this hyperlink points to revision 118386243 (dated 2007-03-28) of the article Encyclopedia, and will reference that individual revision indefinitely. When using the Cite Web template, specify both the date of the page revision you are citing and the date you retrieved that revision, as follows: This example would render as follows:

Are IRC, MySpace, and YouTube reliable sources?

 * IRC: Transcripts of chatroom sessions are not reliable sources because they are unpublished, and we have no way of knowing who the authors are. Transcripts are also easily forged or altered.
 * MySpace: MySpace is generally not acceptable even as a self-published source, because most of it is anonymous or pseudonymous. If the identity of the author can be confirmed in a reliable, published source, then it can be used with the caution appropriate to a self-published source.
 * YouTube: YouTube and other video-sharing sites are not reliable sources because anyone can create or manipulate a video clip and upload without editorial oversight, just as with a self-published website. In some cases, video clips published on YouTube may be acceptable as primary sources if their authenticity can be confirmed, or as a secondary source if they can be traced to a reliable publisher, but even then should be used with caution. They may also be used as a convenience link for material originally published elsewhere, such as Wesley Autrey's appearance on The Late Show with David Letterman. Be careful not to link to material that is a copyright violation.

How do I write citations?
Quotations and any material challenged or likely to be challenged should be attributed to a reliable source using an inline citation. This can be in the form of an embedded link, a Harvard reference, or a footnote. The level of citation to strive for is exemplified by the featured articles.

Don't let complicated citation structures prevent you from adding a source: if you don't know how to format the citation, provide as much information as you can, and others may fix it for you. Cite it!

What if a statement lacks attribution?
While any edit lacking attribution may be removed, the best practice is
 * 1) try to find a source for it;
 * 2) dispute the statement on the talk page, perhaps moving it there;
 * 3) add the fact template to request a citation, add the who template to request attribution; or
 * 4) remove it.

When should I tag unattributed material?
To summarize the use of in-line tags for unsourced or poorly sourced material:
 * 1) If it is doubtful but not harmful to the whole article, use the fact or who tags to ask for source verification, but remember to go back and remove the claim if no source is produced within a reasonable time.
 * 2) If it is doubtful and harmful, remove it from the article, and do not move it to the talk page, particularly if it concerns a living person.

All unsourced and poorly sourced contentious material about living persons should be removed from articles and talk pages immediately. It must not be tagged. See Biographies of living persons and Libel.

When should I use prose attributions?
In many case it is appropriate to include prose attributions for sources. A prose attribution is the explicit ascription of an assertion to a source in the article's text. For example, (taken from the article Milton Friedman):

"According to Harry Girvetz and Kenneth Minogue, Friedman is co-responsible with Friedrich von Hayek for providing the intellectual foundations for the revival of classical liberalism in the 20th century."

Prose attributions have the effect of distancing Wikipedia from a particular statement. Rather than conveying the message that Wikipedia endorses the statement, a prose attribution conveys the message that Wikipedia merely acknowledges that the named source has indeed made the statement. Prose attributions should be used to introduce direct quotations, and to credit a source. Per the neutral point of view policy, they should also be used whenever a statement is controversial, disputed, widely believed to be false, or concerns a matter of opinion.

In general, prose attributions are not precise enough to replace formal citations, although there are exceptions: Paradise Lost (V, 257–9) is more precise than a page, and independent of edition. Again, the following describes the sources precisely; whether it is brilliant prose, and how much of it should be put into footnotes, are editorial decisions:

"John Smith argues throughout Chapter IV of his book The Very Reliable Book (Reliable Press, 2005) that inline citations aren't always needed, while George Gordon completely disagrees in a lengthy essay ('Paging Smith') published in the June 2005 issue of Reliability Magazine. The 2007 edition of The Very Reliable Book repeats Smith's arguments, but adds, in a note on page 45, that 'Gordon has completely convinced me.'"

Redundant prose attributions may be excised for clarity; it is not necessary to prepend "according to..." to every place that a controversial source is cited. Where multiple sources state the same thing, a generic prose attribution, such as "According to numerous scientists...", may be used. Prose attributions that overstate the support for a position, which engage in value judgements, or which constitute ambiguous or unfalsifiable "weasel words" should be avoided (See Avoid weasel words).

What is a convenience link?
A convenience link is a link to a copy of a resource somewhere on the Internet, offered in addition to a formal citation to the same resource in its original format. For example, an editor providing a citation to Adam Smith's The Wealth of Nations might choose to include both a citation to a published copy of the work and a link to the work on the Internet, as follows:

"Smith, Adam. The Wealth of Nations, originally published 1776, this edition Methuen and Co, 1904, ed. Edwin Cannan."

In such a case, The Wealth of Nations is our source, not the website that is hosting a copy of it. It is important to be sure that the link offers a true copy of the original. Sites regarded as inherently unreliable, such as extremist websites or attack sites, should usually not be used for convenience links.

Will rewriting source material violate the NOR provision of the policy?
Original writing is desirable; it is original ideas and viewpoints that violate policy.

In order to avoid copyright violation and be licensed under the GFDL, articles must consist mostly of original prose. However, the ideas, facts, and arguments must have been published already by a reliable source. (Information may be copied verbatim from sources that are licensed under the GFDL, or are in the public domain.)

For example, an original plot summary is acceptable, because it merely summarizes the narrative and provides background for understanding the attributed critical and interpretive material that the article should contain. Condensing a complex theory into a few paragraphs is good writing, not original research, provided that the theory is described as accurately as a reasonable article length and the needs of an audience of non-specialists will permit.