Wikipedia is an encyclopedia—that is, a comprehensive compendium of knowledge. The threshold for inclusion on Wikipedia is whether material is attributable to a reliable published source, not whether it is true. Wikipedia is not the place to publish your opinions, experiences, or arguments.

Although everything on Wikipedia must be attributable, in practice, not all material is attributed. Editors should provide attribution for quotations and for any material that is challenged or likely to be challenged, or it may be removed. The burden of evidence lies with the editor wishing to add or retain the material. If an article topic has no reliable sources, Wikipedia should not have an article on it.

Wikipedia's core content policies are Wikipedia:Neutral point of view, Wikipedia:No original research, and Wikipedia:Verifiability. These policies determine the type and quality of material that is acceptable in articles. Because the policies are complementary, they should not be interpreted in isolation from one another.

Reliable sources are credible published materials with a reliable publication process; their authors are generally regarded as trustworthy or authoritative in relation to the subject at hand. How reliable a source is depends on context. In general, the most reliable sources are books and journals published by universities; mainstream newspapers; and university level textbooks, magazines and journals that are published by known publishing houses. What these have in common is the process and approval between document creation and publication. As a rule of thumb, the more people engaged in checking facts, analyzing legal issues, and scrutinizing the writing, the more reliable the publication. Material that is self-published is generally not regarded as reliable, but see below for exceptions. Any unsourced material may be removed, and in biographies of living persons contentious material that is unsourced or poorly sourced should be removed immediately.

Original research refers to material that is not attributable to a reliable, published source. This includes unpublished facts, arguments, ideas, statements, and neologisms; and any unpublished analysis or synthesis of published material that appears to advance a position. Material added to articles must be directly and explicitly supported by the cited sources.

Note the difference between unsourced material and original research:

• Unsourced material is material not yet attributed to a reliable source. It is unattributed but may be attributable.
• Original research is material that cannot be attributed to a reliable source. It is unattributable.

The only way to demonstrate that material is not original research is to cite reliable sources that provide information directly related to the topic of the article, and to adhere to what those sources say.

Edits that rely on primary sources should only make descriptive claims that can be checked by anyone without specialist knowledge.

Primary sources are documents or people close to the situation you are writing about. An eyewitness account of a traffic accident and the White House's official text of a president's speech are primary sources. Primary source material that has been published by a reliable source may be used for the purposes of attribution on Wikipedia, but only with care, because it's easy to misuse primary sources. The Bible cannot be used as a source for the claim that Jesus advocated eye removal (Matthew 18:9, Mark 9:47) for his followers, because theologians differ as to how these passages should be interpreted. Edits that rely on primary sources should only make descriptive claims that can be checked by anyone without specialist knowledge.

Wikipedia articles should rely on reliable, published secondary sources wherever possible.

Secondary sources are documents or people that summarize, analyze and/or interpret other material, usually primary source material. These are academics, journalists, and other researchers, and the papers and books they produce. A journalist's description of a traffic accident he did not witness, or the analysis and commentary of a president's speech, are secondary sources. Wikipedia articles should rely on reliable, published secondary sources wherever possible. This means that we only publish the opinions of reliable authors, and not the opinions of Wikipedians who have read and interpreted primary source material for themselves.

Some sources pose special difficulties:

• A questionable source is one with no editorial oversight or fact-checking policy or with a poor reputation for fact-checking. Such sources include websites and publications that express views that are widely acknowledged as extremist, are promotional in nature, or rely heavily on rumors and personal opinions. Questionable sources may only be used in articles about themselves.
• A self-published source is material that has been published by the author, or whose publisher is a vanity press, a web-hosting service, or other organization that provides little or no editorial oversight. Personal websites and messages either on USENET or on Internet bulletin boards are considered self-published. With self-published sources, no one stands between the author and publication; the material may not be 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; visiting a stranger's personal website is often the online equivalent of reading an unattributed flyer on a lamp post. For that reason, self-published material is largely unacceptable.

Questionable and self-published sources should not normally be used. There are three exceptions:

1. Self-published and questionable sources in articles about themselves

Material from self-published or questionable sources may be used in articles about those sources, so long as:

• it is relevant to their notability;
• it is not contentious;
• it is not unduly self-serving;
• it does not involve claims about third parties;
• it does not involve claims about events not directly related to the subject;
• there is no reasonable doubt as to who wrote it;
• the article is not based primarily on such sources.

2. Professional self-published sources

When a well-known, professional researcher writing within their field of expertise has produced self-published material, these may be acceptable as sources, so long as their work has been previously published by reliable, third-party publications. Editors should exercise caution for two reasons: first, if the information on the professional researcher's blog (or self-published equivalent) is really worth reporting, a reliable source will probably have covered it; secondly, the information has been self-published, which means it has not been subject to independent fact-checking. Self-published sources, such as personal websites and blogs, must never be used as third-party sources about living persons, even if the author is a well-known professional researcher or writer; see WP:BLP. If a third-party source has published the same or substantially similar material, that source should be used in preference to the self-published one.

3. Carefully selected temporary links with regard to developing current events

"In the case of articles which chronicle a developing current event it is not a violation of Wikipedia policy to temporarily include links to blogs which contain contemporary opinion and observations about the event. A diverse mix is recommended, but the extent and selection of specific blogs is a matter of content to be determined by the editors of the article." See the ArbCom recommendation: Wikipedia:Requests for arbitration/Israel-Lebanon#Use of blogs.

Certain red flags should prompt editors to examine the sources for a given claim:

• surprising or apparently important claims that are not widely known;
• surprising or apparently important reports of historical events not covered by mainstream news media or historiography;
• reports of a statement by someone that seems out of character, embarrassing, controversial, or against an interest they had previously defended;
• claims not supported or claims that are contradicted by the prevailing view in the relevant academic community. Be particularly careful when proponents of such claims say there is a conspiracy to silence them.

Exceptional claims should be supported by the best sources, and preferably multiple reliable sources, especially regarding scientific or medical topics, historical events, politically charged issues, and biographies of living people.

You may cite your own publications just as you would cite anyone else's, but make sure your material is relevant and that you are regarded as a reliable source for the purposes of Wikipedia. Be cautious about excessive citation of your own work, which may be seen as promotional or a conflict of interest; when in doubt, check on the talk page.

Because this is the English Wikipedia, for the convenience of our readers, English-language sources should be used in preference to foreign-language sources, provided they are otherwise of equal suitability, so that readers can easily verify that the source material has been used correctly. Published translations are preferred to editors' translations; when editors use their own translations, the original-language material should be provided too, preferably in a footnote, so that readers can check the translation for themselves.

Material counts as original research if it:

• introduces a theory, method of solution, or any other original idea;
• defines or introduces new terms (neologisms), or provides new definitions of existing terms;
• introduces an argument without citing a reliable source who has made that argument in relation to the topic of the article; or
• introduces an analysis, synthesis, explanation or interpretation of published facts, opinions, or arguments that advances a point that cannot be attributed to a reliable source who has published the material in relation to the topic of the article.

Editors often make the mistake of thinking that if A is published by a reliable source, and B is published by a reliable source, then A and B can be joined together in an article in order to advance position C. However, that would be an example of an unpublished synthesis of published material serving to advance a position, and it constitutes original research.^[1] "A and B, therefore C" is acceptable only if a reliable source has published this argument in relation to the topic of the article.

Here is an example from a Wikipedia article, with the names changed. The article was about Jones:

Smith says that Jones committed plagiarism by copying references from another book. Jones denies this, and says it's acceptable scholarly practice to use other people's books to find new references.

Now comes the unpublished synthesis of published material:

If Jones's claim that he always consulted the original sources is false, this would be contrary to the practice recommended in the Chicago Manual of Style, which requires citation of the source actually consulted. The Chicago Manual of Style does not call violating this rule "plagiarism." Instead, plagiarism is defined as using a source's information, ideas, words, or structure without citing them.

The whole point of this paragraph is the conclusion that, given the Chicago Manual of Style's definition of plagiarism, Jones did not commit it. This is the editor's opinion; it is original research. If the paragraph attributed the opinion to a reliable source that specifically commented on the Smith and Jones dispute and made the same point about the Chicago Manual of Style and plagiarism, it would comply with this policy. In other words, that precise analysis must have been published by a reliable source in relation to the topic before it can be published on Wikipedia.

Editors may make straightforward mathematical calculations or logical deductions based on fully attributed data that neither change the significance of the data nor require additional assumptions beyond what is in the source. It should be possible for any reader without specialist knowledge to understand the deductions. For example, if a published source gives the numbers of votes cast for each candidate in an election, it is not original research to include percentages alongside the numbers, so long as it is a simple calculation and the vote counts all come from the same source. Deductions of this nature should not be made if they serve to advance a position, or if they are based on source material published about a topic other than the one at hand.

Citation exemptions have also been extended to plot summaries of novels, films, and related media. As Wikipedia's Manual of Style says, "The plot summary for a work, on a page about that work, does not need to be sourced with in-line citations, as it is generally assumed that the work itself is the primary source for the plot summary." Citations are, of course, still encouraged, and any interpretations, quotations, and secondary sources used must be cited in the article.

Pictures have enjoyed a broad exception from this policy, in that Wikipedia editors are encouraged to take photographs or draw pictures or diagrams and upload them, releasing them under the GFDL, or another free license, to illustrate articles. This is welcomed because images generally do not propose unpublished ideas or arguments. Also, because of copyright law in a number of countries and its relationship to the work of building a free encyclopedia, there are relatively few publicly available images we can take and use. Wikipedia editors' pictures fill a needed role.

A disadvantage of allowing original photographs to be uploaded is the possibility of editors using photo manipulation to distort the facts or position being illustrated by the photo. Manipulated images should be prominently noted as such. If they are noted as manipulated, they should be posted to Wikipedia:Images for deletion if the manipulation materially affects the encyclopedic value of the image.

Images that constitute original research in any other way are not allowed, such as a diagram of a hydrogen atom showing extra particles in the nucleus as theorized by the uploader.

Editors must take particular care when writing biographical material about living persons, for legal reasons and in order to be neutral. Remove contentious material that is unsourced or poorly sourced immediately if it's about a living person, and do not move it to the talk page.^[2] This applies to any material related to living persons on any page in any namespace, not just the article space.

Further information and examples: Wikipedia:Citing sources and Wikipedia:Citations quick reference

Any reader should be able to verify that material added to Wikipedia has already been published by a reliable source. Material that is challenged or likely to be challenged, and quotations, should be accompanied by a clear and precise citation, normally written as a footnote, a Harvard reference, or an embedded link; other methods, including a direct description of the source in the article text, are also acceptable.

Any edit lacking attribution may be removed, and the final burden of evidence lies with the editor wishing to add or retain the material. However, this policy should not be used to cause disruption by removing material for which reliable sources could easily or reasonably be found—except in the case of contentious material about living persons, which must be removed immediately. If you encounter a harmless statement that lacks attribution, you can tag it with the {{fact}} template, or move it to the article's talk page with a comment requesting attribution. If the whole article is unsourced, you can use the {{unreferenced}} template; for sections requiring sourcing, {{unreferenced section}} is available. Absurd unsourced claims and original research should be deleted rather than tagged or moved to a talk page.^[3]

• Wikipedia:Merge and delete, attribution within Wikipedia in relation to its licensing terms
• Wikipedia:Template messages/Disputes, maintenance tags that notify of a dispute

• Wales, Jimmy. Crackpot articles, July 12, 2003
• Wales, Jimmy. "Original research", December 3, 2004
• Wales, Jimmy. "Original research", December 6, 2004.
• Academic Publishing Wiki – a wiki welcoming original research
• Wikiresearch, a proposal for a wiki for original research.