Wikipedia:Citation underkill

Source: Wikipedia, the free encyclopedia.
A single reliable source is often enough to support a statement—however, it is best for statements to be sourced than to go unsupported.
A single good source is often enough to support a statement; however, it is best for statements to be sourced than to go unsupported.

The quality of Wikipedia improves when each statement is cited; our material is required to be verifiable with

inline citations
and plenty of them.

One cause for "citation underkill" is the thought that it does not matter when good content is unsourced: that

false beliefs, spreading errors in reasoning, and reinforcing widely held misconceptions
.

Without citations, it is difficult to know that material is not just made up. In order to maintain a neutral point of view, it is critically important for an article to be verifiable, especially when sources disagree. Changing single words can cause a statement that was sourced to become a statement which fails verification. When no citation is nearby, this error risks being missed. By not including a source when it can be, we make it hard to verify our articles and put their neutrality at risk, diminishing their encyclopedic value.

The integrity of content depends on where a citation is placed. Misplaced citations cause citation confusion, which makes it harder to verify claims. Placing citations where they clearly correspond to specific claims improves the verifiability in accordance with guidelines on footnotes. When no citation is placed to verify a claim or if the citation is made invisible (commented out), it decreases the verifiability of content, and readers may incorrectly hold that those statements are unsourced. Unsourced material on Wikipedia risks (rightly or wrongly) being considered as original research. It also may be deleted.

Controversial claims usually require only single citations, but additional citations may decrease the degree with which the claim is likely to be challenged. Citation overkill can occur when many (often weak) sources are used to support the same statement, which can give a false sense of authority. Using as many sources as you need to ensure verifiability is not overkill. In most cases, one citation to a reliable source for each statement is sufficient to satisfy verifiability.

Citations improve article content

The color of the sky varies.
The color of the sky varies.[1]

It is possible that an editor who is trying to promote an article to GA-class (good article status) might add citations to basic facts such as "...the sky is blue...".[3] This is a good thing, and the fact that the sky is not always blue does benefit from adding a citation. We can add citations for things that are well-known, and the source can contain additional information to benefit our readers. For content that failed verification, sky blue cases are not applicable because the content is not verifiable using the source presented. That means content that failed verification is a violation of Wikipedia:Verifiability policy. Rather than skimming a source and adding a point or two, it is beneficial to study the source, as it may have additional information. Further, the same source could be used for another page on a related topic. The How to mine a source help page is a how-to on maximizing the information obtained from a single source. If a claim is only verifiable via WP:PAYWALL then it is best to provide a citation. Where you may think citations are not needed they may be needed. For example, stand-alone lists such as List of electronic cigarette and e-cigarette liquid brands are required to be sourced in the same manner as other articles. Wikipedia has no firm rules, but by following the rules it is very possible to maintain a high quality of article content.[4] For example, see the Larry Sanger article. If a rule prevents you from improving a page, then the rule is wrong or you are wrong. Think twice before breaking a rule.

Citing common knowledge

Electric fans in South Korea commonly feature a timer, due to a widely held misconception that leaving them on while asleep can be fatal. If you really can't find a source to cite for your "obvious" statement, is it really true?
Electric fans in South Korea commonly feature a timer, due to a widely held misconception that leaving them on while asleep can be fatal.[2]
If you really can't find a source to cite for your "obvious" statement, is it really true?

One cause for "citation underkill" is the belief that something is such

false beliefs and spreading fallacies and widely held misconceptions. Without a citation, unsupported content may be deleted
because the content may be considered incorrect.

Wikipedia editors can make mistakes, and assuming what you think is common knowledge may not be accepted as common knowledge by others. Set an example by citing your content properly, whether you think it is common or uncommon.

Citations in the lead

Adding citations to the lead is done on a case-by-case basis. Providing citations in the lead can be very helpful, both for readers as well as editors. Without citations in the lead, our readers may think the content is not neutral or is original research, even if sourced in the body. We cannot expect our readers to always read the body to try to verify the content they read in the lead. Citations in the lead also help readers and editors find their way in the body of the article, when the citation supports the same or similar statements in different parts of the article. Contentious articles or articles on contentious topics benefit especially from citations in the lead.

Bundling citations

WP:CITEBUNDLE claims bundling citations has several advantages, without explaining in detail when bundling poses disadvantages. When multiple citations are bundled into a single footnote, especially when bundling all the citations at the end of the sentence or paragraph it may be difficult to verify the article content. Another argument against bundling is that it presents an extra step for anyone wanting to review the source for a claim. When different citations support different parts of the same sentence bundling can cause citation confusion. It is best to unbundle bundled citations and place them where they verify each statement when different sources verify different parts of a sentence or paragraph. Ease of verification helps readers and editors.

Wikipedia:Verifiability states "In Wikipedia, verifiability means that other people using the encyclopedia can check that the information comes from a reliable source." When different sources verify different parts of a sentence or paragraph, moving all the citations to the end of a sentence or paragraph makes it difficult to check whether each statement is verifiable. Verifiability policy also states "The cited source must clearly support the material as presented in the article. Cite the source clearly and precisely (specifying page, section, or such divisions as may be appropriate)." This means the content must be clearly supported by the cited source. An editor may think the content failed verification if the citations are misplaced.

In certain cases bundling citations may help readability, but if you're only talking about a handful of references—it likely doesn't. Bundling properly can be difficult and time-consuming, if you're not willing to put in the work—you're likely to leave a botched mess. It's better to leave citations visible unless you really know what you're doing.

Bundling correctly

When the sources verify the exact statement the citations may be bundled this way: The color of the sky changes at the beginning and at the end of the day.[1]

  1. ^ References:
    • Frank Staub (2005). The Kids' Book of Clouds and Sky. Sterling Publishing Company, Inc. pp. 10–. .
    • Lucia Ronchi (1 April 2014). The Semantics of Color Sharing the Laboratory with Color Vision. Vol. II. Fondazione Giorgio Ronchi. pp. 69–. .
Bundling incorrectly

Bundling all the citations at the end of the sentence would make it difficult to verify each specific piece of content when multiple pieces of content require verification from different sources or when 17 different pieces of content require verification from different sources such as the following sentence: Aluminum, barium, cadmium, chromium, copper, iron, lead, manganese, mercury, nickel, silicate, silver, strontium, tin, titanium, zinc, and zirconium have been found in the electronic cigarette aerosol.[1] For this particular case, it easier to verify each piece of content when each citation is placed where it verifies each claim.

If the sources verify different parts of the sentence or paragraph then bundling the citations will make it take longer to verify each statement. Therefore, putting all the citations at the end would make it difficult for a reader to know which piece of content comes from which citation. This is done on a case-by-case basis.

Hidden citations

If consecutive sentences are supported by the same citation, it is better for them to be all visibly shown. Hiding citations with the markup <!-- --> makes it difficult for our readers to verify claims. If citations are hidden it often is better to make them visible. References can occur and can become unhidden after each sentence, which is the preferred style for medical content. Hiding citations can cause confusion in the future.[5] For example, putting only one reference at the end of a section can require ongoing maintenance as other editors may mistakenly add {{cn}} tags or delete content that they believe is unreferenced.

Necessary repetition

To improve verifiability, material that is repeated multiple times throughout an article should have an inline citation for every mention. Doing so also increases the chance readers and editors will find the appropriate source for each statement. Without an inline citation next to each claim, it is difficult for readers to verify claims. Therefore, it is perfectly reasonable to provide inline citations repeatedly. For example: to state that the human hand has four fingers and one thumb[6] in multiple places in an article, you would do well to provide a citation after each mention. This can be accomplished by adding to the main citation a markup like this: ​<ref name=Latash2008>, while using a corresponding named-ref abbreviated citation like this: ​<ref name=Latash2008/>.

Citation underkill often occurs when:

  • Inline citation is provided only at the end of a paragraph
  • In certain circumstances, when all sources are placed at the end of a sentence

     An example of how to place sources in the middle of a sentence, in an example where it is appropriate:

Tomato products, such as ketchup, tomato juice, spaghetti sauce, and pizza sauce are high in lycopene,[1] which research indicates likely plays a role in protecting against cardiovascular disease and various cancers.[2]

Bundling all the citations together in one citation at the end of a sentence or paragraph often make it difficult to determine which citation verifies which claim. An extreme example where specific claims are verified individually is this:

Aluminum,[1] barium,[2] cadmium,[3] chromium,[4] copper,[5] iron,[6] lead,[7] manganese,[8] mercury,[9] nickel,[10] silicate,[11] silver,[12] strontium,[13] tin,[14] titanium,[15] zinc,[16] and zirconium have been found in the electronic cigarette aerosol.[17]

  • This example is extreme, and for certain situations can be rewritten, but to avoid readers or editors needing to scour through different sources—one citation is needed after each statement.

Wikipedia:Citation overkill suggests that repetitive use of the same inline citations is overkill and causes clutter, but the advantage of doing so helps an editor or reader quickly locate the citation and check to make sure the content is properly sourced. After an editor clicks to edit an article, it often states at the top, "Encyclopedic content must be verifiable." Removing a citation while citing WP:REPCITE, for example, may lead to difficulty in verifying a claim or it may even be perceived as a violation of Wikipedia's Verifiability policy. Placing a citation at the end of each paragraph instead of after each sentence within that paragraph may result in the content being tagged with a citation needed tag. It is better to place a citation at the end of each sentence to improve the ability to verify each statement. The content could be mistakenly deleted if someone thinks the content is unsourced.

To summarize, do not remove citations simply because they are being repeated. Citing each sentence improves an article's verifiability, which is preferred over paragraph citations. Editors should be cautioned against taking actions that make it harder to verify a specific claim.

Sentence merging

Merging a sentence without a citation with a sentence that does have a citation can render the newly formed sentence as partially failed verification. This can occur when a sentence was added into a paragraph without a citation. Merging sentences together without using the appropriate citations cause citation craziness. If you want to delete a sentence then be sure that the citation left behind is not misplaced. If you delete a sentence that verified only that claim then you should also delete the citation that verified that claim. If you leave behind the citation it may not verify the previous statement. If more than one source is being used to verify a claim and you are changing the wording that is verifiable to only one source, then be sure to remove the other sources that do not verify the new claim. When adding wholly new information, make sure it is cited to the correct citation. Citation hijacking occurs when adding new information before an existing citation where it does not verify the claim. When an editor rewrites a sentence that is properly sourced and adds one or more citations where there already is a citation at the end of a sentence it may cause a problem. The new citation may verify the new sentence but often the original citation used to verify the previous sentence does not verify the new sentence.

Citation placement

When multiple sources support different parts of a paragraph or passage it is important to place each citation where they verify each specific concept or idea. This also acts to increase the life-span of text on Wikipedia. If you write a paragraph, which later gets a new statement added in the middle of it—citing a different source – you will have citation confusion.

Simple claims can become confusing to verify. In certain circumstances, to avoid citation confusion it is easier to verify each specific claim by placing the citation where it verifies each claim, rather than place all the citations at the end of the sentence or paragraph.

A clear example of this can be found at Malaria, where one citation is broken up by another, and where citations at the end of a paragraph makes it more difficult to verify:

Most useful

The signs and symptoms of malaria typically begin 8–25 days following infection;

flu-like symptoms,[3] and can resemble other conditions such as sepsis, gastroenteritis, and viral diseases.[2] The presentation may include headache, fever, shivering, joint pain, vomiting, hemolytic anemia, jaundice, hemoglobin in the urine, retinal damage, and convulsions.[4]

Less useful

The signs and symptoms of malaria typically begin 8–25 days following infection;

flu-like symptoms, and can resemble other conditions such as sepsis, gastroenteritis, and viral diseases.[2][3] The presentation may include headache, fever, shivering, joint pain, vomiting, hemolytic anemia, jaundice, hemoglobin in the urine, retinal damage, and convulsions.[4]

Least useful (nearly pointless)

The signs and symptoms of malaria typically begin 8–25 days following infection; however, symptoms may occur later in those who have taken antimalarial medications as prevention. Initial manifestations of the disease—common to all malaria species—are similar to

flu-like symptoms, and can resemble other conditions such as sepsis, gastroenteritis, and viral diseases. The presentation may include headache, fever, shivering, joint pain, vomiting, hemolytic anemia, jaundice, hemoglobin in the urine, retinal damage, and convulsions.[1][2][3][4]

References

Benefits of proper citation are not limited to medicine, and a good example of proper use of citations can be found in the Featured article on Ukiyo-e (exhibited on the main page on 25 June 2017):

Determining at what prices prints sold is a challenge for experts, as records of hard figures are scanty and there was great variety in the production quality, size,[1] supply and demand,[2] and methods, which went through changes such as the introduction of full-colour printing.[3] How expensive prices can be considered is also difficult to determine as social and economic conditions were in flux throughout the period.[4] In the 19th century, records survive of prints selling from as low as 16 mon[5] to 100 mon for deluxe editions.[6] Jun'ichi Ōkubo suggests that prices in the 20s and 30s of mon were likely common for standard prints.[7] As a loose comparison, a bowl of soba noodles in the early 19th century typically sold for 16 mon.[8]

  1. ^ Kobayashi & Ōkubo 1994, p. 216.
  2. ^ Ōkubo 2013, p. 31.
  3. ^ Ōkubo 2013, p. 32.
  4. ^ Kobayashi & Ōkubo 1994, pp. 216–217.
  5. ^ Ōkubo 2008, pp. 151–153.
  6. ^ Kobayashi & Ōkubo 1994, p. 217.
  7. ^ Ōkubo 2013, p. 43.
  8. ^ Kobayashi & Ōkubo 1994, p. 217; Bell 2004, p. 174.

Citing different page numbers

Citing the page number or page numbers for the specific content used to source the statement or quotation in the article, makes it easier to verify the claim rather than a page range using the same repeated citation. If you are citing a book or PDF file then citing the specific page number or page numbers can be especially helpful for anyone reading the source.

One way to verify each specific page number without creating duplicate full citations is like this:

Enhancing the availability of drinking water can lead to clear benefits to health.

methaemoglobinaemia, in particular to bottle-fed babies.[2] It is recommended that water be absent of tastes and odors that would be unpleasant to most people.[3]

  1. ^ WHO 2016, p. 1.
  2. ^ WHO 2016, p. 6.
  3. ^ WHO 2016, p. 7.

A ​== Bibliography == section can be created for adding the main citation and a named-ref abbreviated citation is used for each page number. The named-ref abbreviated citation is placed specifically where it verifies the claim.

The main citation looks like this: ​{{cite web|url=http://www.who.int/water_sanitation_health/dwq/gdwq0506.pdf|title=Guidelines for Drinking-water Quality|publisher=World Health Organization|year=2006|ref={{harvid|WHO|2016}}}} and the named-ref abbreviated citation looks like this: ​{{sfn|WHO|2016|p=7}}.

Another way to provide a page number for each citation is by using the markup ​{{rp|}}. This is accomplished by placing it at end of the citation like this: [1]: 7 

  1. ^ "Guidelines for Drinking-water Quality" (PDF). World Health Organization. 2006.

Citation balancekill

Stating "Some weasels are white." may be engaging in unacceptable original research if the source does not explicitly state that "some" are white.
Stating "Some weasels are white." may be engaging in unacceptable original research if the source does not explicitly state that "some" are white.

Placing a citation after each idea or concept does not guarantee the content is verifiable. Consensus on Wikipedia does not magically generate accuracy. An editor may propose a change to consensus by discussion or editing. The verifiability of the content depends heavily on whether the content is in actuality verified to the source placed after each idea or concept. Furthermore, placing an inline citation where it verifies the content is important, but it is equally or even more important for the content to be neutrally written. By following Neutral point of view, Verifiability and No original research policies, citation balancekill (the sum of accurately sourced knowledge) is attainable. Altering the original meaning of the content may violate verifiable policy. The content is more accurate and neutral when including a modifier or weasel word supported by the source. Adding a modifier to a sentence not supported by the source alters the original meaning of the source. Weasel words or unsupported attributions are words and phrases that give an appearance that something explicate has been stated, when in actuality only a vague or ambiguous claim has been presented.[7] When a source only indicates a vague or ambiguous claim then the content added to an article should also indicate a vague or ambiguous claim.

Even changing single words or certain phrases can render content from being sourced, to content that has failed verification. Even when the content meets verifiability policy an editor may rewrite the content years later. For example, another editor comes along and adds a phrase or makes a modification to a sentence. Now, the new sentence basically says something that the source does not mention; effectively making the once verifiable sentence, a case of failed verification. This issue is often overlooked. If the source expresses a specified viewpoint such as using the word "some" then the content can also specify that viewpoint or similar viewpoint which avoids giving a misleading or vague impression. We can use the exact weasel word or a synonym to that word when the source has used that specific word. We cannot use the word "some" when the source uses another weasel word that does give a different viewpoint such as numerous or substantial. If the source does not use the word "some" or by combining different sources together to come to the conclusion that it is "some", then it probably is original research or a synthesis violation.

Do not combine material from different sources to reach or imply a conclusion not explicitly stated by any given source. This would be improper editorial synthesis to suggest a new conclusion. Combining sources to come to a new conclusion is original research performed by an editor. It is acceptable only when a

reliable source
has published the same argument in relation to the topic of the article.

Wikipedia:No original research (in particular, Synthesis of published material) clearly indicates we don't conduct our own reviews of the sources on Wikipedia.[8] For example, when two reviews verify the same content we can't state "Two reviews found...". This would in effect be combining material from different sources to reach a conclusion not explicitly stated in any individual source. The content must be able to be verifiable to a reliable source, not by counting of references that are present in a Wikipedia article.[9] It is not an allowable provision to include content that failed verification. Therefore, we can't state "Two reviews found..." unless an individual source stated it was "Two reviews...". Moreover, when there is no serious dispute between sources, the content should normally be asserted without in-text attribution.

For example, "There is some evidence that following this diet may lead to improvements in terms of body composition and metabolic effects compared with the typical

Western diet.[1]

Another example is the following: E-cigarettes are likely safer than tobacco.[1] This is incorrect. The part "likely" has failed verification. The following is correct: E-cigarettes are generally considered safer than tobacco.[1]

  1. ^
    PMID 25410845
    .

Editors mistakenly use unsupported weasel words even when they believe they are correct. For example, as of 18:05, 28 June 2017‎ Wikipedia's Electrical disruptions caused by squirrels states: "In the U.S., squirrels have been the cause of many power outages in Pennsylvania.[n 1]"[11] This is incorrect. The word "many" has failed verification. Adding up different sources together to come to a new conclusion is a novel synthesis. Another example is the following: "Some have described the Mannings as football's "royal family".[302][303][304][305][306]"[12] The word "some" has failed verification. Again, an individual source must make the claim. We don't tell readers what to think. Combining multiples sources to reach a new conclusion does not make it true. When multiple sources say the same thing it does not equate to "some". Even if you believe it to be true, it still must be verifiable.

Bear in mind that we are required to avoid copyright infringement and plagiarism. This requires us to use our own words to express the information we get from the reference. This means we paraphrase and thus words not found in the sources can be used even if they are not the exact same meaning. Changing the wording and rearranging ideas is also an important part of paraphrasing. The information should also be delicately summarized or rephrased without altering its meaning or implication. It is possible to construct an alternative wordings without violating the principle of verifiability.

Overciting content

Removing excessive citations is a reasonable option.
Removing excessive citations is a reasonable option.

Quality of citations, not their sheer quantity, improves article content. Unreliable sources can be removed, but it is better if possible to replace them with reliable ones in order to preserve article content. Multiple low quality citations following a statement does not make it more true. Over-citing content especially for non-controversial claims should be avoided.

Editors have cited Citation overkill as a reason for adding additional citations after each sentence for non-controversial claims. Citation overkill appears to support adding more than one citation after each sentence, regardless of circumstance. Editors have misinterpreted Citation overkill into thinking that it is okay to add more than one citation after each claim in any circumstance. This sort of "reference spamming" disrupts the flow of reading an article. Citation overkill states "If there is a good reason to keep multiple citations, for example, to avoid perennial edit warring or because the sources offer a range of beneficial information, clutter may be avoided by merging the citations into a single footnote." If the sources actually do contain beneficial information they could be used to cite other information instead of cluttering the article. If an editor insists on keeping the additional citations they can be commented out using the markup <!-- --> in order to avoid cluttering the article with needless citations. On the other hand, merging citations into a single footnote can also clutter the reference section.

Another type of reference spamming is called failed verification spamming. One way to spot failed verification spamming is when an editor who restores the additional citations refuses to provide verification on the talk page for the additional citations. Wikipedia requires that a citation presented for a claim verify the entire claim.[13] If the additional citations only partially verifies the claim then the additional citations still failed to verify the claim. It is not a valid argument to keep the additional citations when they do not verify the entire claim. However, if a sentence is making two separate claims it would be best to use a satisfactory source for each, placed following its respective claim. If one of the citations verifies the entire claim then you may only need to use one citation.

One citation after each sentence for non-controversial claims is usually sufficient. Adding more citations than needed can cause citation bloat. More than three citations for non-controversial claims or even controversial claims may be excessive. Reference spamming occurs when an editor adds the same citation multiple places in an article where it is unnecessary or does not verify the claim. Moreover, failed verification spamming occurs when more than one citation is used to try to convince others the claim is sourced when the sources presented does not verify the claim. For controversial claims one citation is usually enough for content that is likely to be challenged. If the claim is extraordinarily controversial then the content may require more than one citation. Adding more than one citation after each statement is done on a case-by-case basis. The purpose of a citation is for readers to be able to verify the content presented, not to persuade to other editors the validity of the content.

In certain circumstances, a very controversial claim likely to be challenged may retain as many as three references, cited in this form:

The benefits and the

health risks of e-cigarettes are uncertain.[1][2][3]

Feasibility and original reporting

You don't need to cite the entire library to back up an idea or a concept, but it is better to cite something than nothing. Your goal is to strive to cite the best possible source, and to cite in the best possible way.
You don't need to cite the entire library to back up an idea or a concept, but it is better to cite something than nothing. Your goal is to strive to cite the best possible source, and to cite in the best possible way.

Verification must be feasible. This means:

  • A statement must be able to be verifiable and not constitute original research.
    For example, not "You can verify this by walking around outside", or "I was listening to an interview". These by definition are original reporting.
  • The source must be identifiable.
    You can't source "that big book with the blue cover I have in the bookshelf in the downstairs bedroom"—while theoretically verifiable (by one editor), it is in reality not at all verifiable by readers.
  • Every statement must have its own verifiable reference following its claim.
  • It must be possible to figure out which source supports which statement.
    It is unreasonable to expect an editor or reader to go through dozens of sources in order to find out if the statement is supported by one of them.

This does not mean:

  • That the source must be available online;
  • That it must be free;
  • That it must be or remain in publication; or
  • That it must include a DOI or ISBN (though please do when possible).

Templates

Gallery

  • Instructional videos on using Wikipedia
  • Wikipedia video tutorial on reliability.
  • Wikipedia guidelines and policies.
  • Guidance for a new medical editor.

Userbox

Code Result
{{User:UserBox/NOTBLUE}}
This user believes Wikipedia
needs more inline citations
.
Usage

See also

References