Wikipedia:Improving referencing efforts

Why references are important? References are vital for the quality and usefulness of Wikipedia articles. If you don't believe me, don't take my word for it; At the Wikimania conference 2006, Jimmy Wales himself said he believed Wikipedia should focus more on the accuracy of our existing material instead of creating new material. Since then, we have broken the 5 million article mark and there is still much room for improvement.

We tell people to not rely on Wikipedia as their sole source of information but rather use it as a starting point for further research. For this to work, articles not only need external links to guide readers to further information, but also a list of sources we used to write the Wikipedia article so the reader can go and check our material against that of the sources to check our accuracy.

In other words, we need to make a joined effort to clean up the references in Wikipedia.

With this page I hope to spur the discussion towards a proposal we can use to actively counter this problem by choosing a middle ground between the group of people who advocate deletion of unreferenced entries and those who prefer keeping them without actively working on their improvement.

Current situation edit

Problem edit

There are lots of articles that have unsourced and questionable information. Some articles are totally unreferenced. Only a tiny percentage of articles have citations/sources for all important information.

Each month thousands of articles are marked with {{unreferenced}}: This article does not cite any sources. With a multiple-year backlog in this category only growing longer, there is little chance of editors catching up if something new is not done. (As of early January 2016, the total number of articles in this category was more than 217,000.)

The current WikiProjects that deal with this problem are Wikipedia:Unreferenced articles and Wikipedia:WikiProject Fact and Reference Check. Despite the combined efforts of these WikiProjects the backlog continues to grow.

Simply deleting all unreferenced articles would leave a huge hole in Wikipedia's coverage; that isn't going to happen. Deleting all unsourced statements also isn't going to happen: certainly the majority of such statements are true.

Solving the problem edit

Unlike stubs and deletion, efforts to improve referencing have yet to make full use of the fact we have a large group of active WikiProjects filled with editors who are interested in specific subjects or topics in a certain language area. For example, people who are interested in chemistry are more likely to know where to find sources for chemistry-related articles and judge how reliable chemistry-related articles are.

To improve the situation, fact-checking efforts should make use of WikiProjects, categorization, and other technical features to attract the enthusiasts and experts who can add references and remove incorrect statements.

Current viewpoints edit

  • The burden of proof for verifiability lies on the creator (initial author) of an article, but a lack of references doesn't mean an article is false, nor that it is non-notable. An article can only be deleted if no sources can be found after reasonable efforts have been made.
  • Trying to force editors to provide references may deter new editors from contributing good material - though a similar requirement is already in place on WP:AFC, when the notability of an article is challenged.
  • Referencing articles may require expert knowledge; whether or not it does, it is time consuming.
  • Information in articles always gets copied to many places on the Web. Leaving unreferenced information in articles for long periods of time can make it difficult to determine if existing information on the web came entirely from the unreferenced material, or preceded when the information was added to Wikipedia.
  • Having unreferenced material is bad for Wikipedia's credibility.
  • When dubious statements and articles are to be removed, a frequent response by other editors is often to ask those removing the statements and articles to help improve articles and referencing; those making this response don't volunteer to actually do any improving themselves.
Failed proposals
  • Previous proposals to create a special speedy deletion criterion to delete unreferenced articles after a grace period have found no consensus. (Wikipedia:Requests for verification)

Proposals edit

This collapse box contains proposals from 2008 and 2009, many of which have been acted upon. If you have new ideas, please share them discussion page

Barnstar incentive edit

A simple yet effective way to encourage people to embark on a quest to fact-check articles is to reward them for their efforts. I still get an addictive warm, fuzzy feeling when I receive a barnstar and editors have shown to go through great lengths to get noticed or appreciated. Just look at the amount of people who go through WP:FAC and T:TDYK to get their work featured on the main page. I could be mistaken, but on my perusal of Wikipedia:WikiProject Fact and Reference Check and Wikipedia:Barnstars, I was unable to find any barnstars that rewarded people for fact-checking or referencing articles that lack sources. A reward scheme would encourage people to do the job.

Educate editors edit

If new editors are educated about the need to cite their sources when creating an article and reward them for doing so, we will lower the amount of unreferenced articles coming in so we can deal with the existing ones. We can do this by adding a lesson to the Virtual classroom or, for example, update the tutorial with a new section.

Adding a citation template button to the toolbar edit

Suggested by MacGyverMagic

A lot of editors find the citation templates hard to use while using them helps people to provide the information needed to check a reference. The citation button can be added with the refTools feature in the Gadgets section of Special:Preferences, and it uses a fill-in-the-boxes method to easily insert a template with all the information needed. We should strive to have the button enabled by default.

Press release edit

Many hands make light work. If we were to make a press release that said "Wikipedia undertakes massive spring time cleanup" or something to that effect, we can attract new editors who are interested in working on the verification of articles. (See User:Naconkantari/cleanup for one proposal for a "spring cleanup".)

Bots and scripts tagging new articles edit

To ensure an article is checked by at least one set of eyes (preferably more) it's important to not let newly created articles slide by undetected. With the amount of articles that are created each day, it's important to use automated tools to our advantage. AlexNewArtBot (talk · contribs) goes through Special:Newpages and evaluates entries to see if they are suitable for inclusion in the main page did you know section. We can develop a bot that sorts articles based on their content and/or categories and leaves a list of articles to check on relevant WikiProject subpages.

Ais523 has created a script that allows editors to watch new entries in categories. If this script can be combined with the CatScan tool, it would be a breeze to find new articles that need citation in a certain category without the need for bots to post long lists to WikiProject pages.

New computer programs edit

It is a fact that Wikipedians are more likely to do a task if it's easier. Over the years, a variety of tools have been created to make for example vandal- and newpage patrolling. If a tool was developed that would making tracking down sources and adding them to articles easier, more people would engage in the task. An ideal referencing tool would combine the search capabilities of NPWatcher (NewPageWatcher) with citation programs like Magnus Manske's reference generator and Wikicite and the editing capabilities of AutoWikiBrowser.

Stable versions edit

The stable versions feature is not yet finished, but if implemented it will help a lot in lowering the influx of new articles that lack any sort of sources.

Changing norms and requirements for AfD deletion nominations edit

Suggested by Skomorokh

Similar to how Wikipedia:Arguments to avoid in deletion discussions is used to beat down especially poor arguments at Wikipedia:Articles for deletion, Wikipedia:Featured article candidates contributors are strongly encouraged to make only actionable opposes and Wikipedia:Administrator intervention against vandalism reports which do not conform to the criteria are removed, establishing stronger social norms with regards to referencing could help encourage and reward effort at improving reference quality. For example, an influential essay which discouraged editors from nominating articles at AfD without searching for sources first and discouraged commenters from arguing to keep based on the improvability of poorly referenced articles without making the effort to cite sources could go a long way towards a more collegiate and productive AfD process.

Suggested by DGG

Before nominating articles for deletion, it is already deletion policy recommended to try to see if they can be improved. But there is no actual requirement to do so, and it is routine for articles to be nominated as unreferenceable or notability not demonstrated, when a basic referencing effort would have shown adequate sources. If it were actually required to do such a basic search and present the results as part of the nomination, editors would immediately discover the possibilities for improvement and improve those articles that could be easily improved.

Suggested by Carlossuarez46

Before keeping an article which is either prod'ed or Afd'ed all material concerning its notabiliy must be supported by reliable sources; the burden -contrary to DGG's suggestion above- is on the creator who created the problem, not the person who discovers the problem.

Prioritising the article groups with the most need of referencing (i.e. BLPs) edit

Suggested by Casliber

I suspect that triaging editors in the direction of most need may be of value, namely Biographies of Living People. This category is possibly the easiest of the three broad contentious areas for editors in general to address, the other areas (national disputes, and traditional vs. alternative medicine) are likely to require a more vested involvement and background knowledge. Question is, how do we alert editors or keep this category 'on the radar'? Traffic gets reduced with each further 'click' of the mouse button, and I have found pages I never knew existed after being around here for over two years. My idea is to try and place BLP-related pages in more prominent positions to try and (hopefully) increase traffic to them. Ideas include:

  • Link to Wikipedia:Biographies of living persons in the Welcome template, plus short, friendly note on extreme importance of referencing in this scenario. Targets new users who are the ones least likely to use references.
  • Find somewhere to highlight link to page or category of unreferenced BLP articles (I can't find the page now, anyone is welcome to link here now) and place on Template:Editabuselinks or some other prominent noticeboard (or let run in signpost??).

If an article needs references ... add them edit

Suggested by WilyD
  • By the time a statement is identified as problematic - it can be referenced without much difficulty by almost anyone with experience. ~30-45 seconds in many cases. The problem is simply an attitude of unwillingness to accept responsibility. WilyD 19:52, 12 December 2008 (UTC)[reply]

Add links to sources edit

Suggested by Colonel Warden in 3 December 2008
  • I routinely add a {{find}} template in a {{notice}} box at the head of the talk page when I create or edit an article. I recently improved the {{rescue}} template to do the same. If editors did this more often then we would find it easier to locate sources. A bot might do this automatically.

Add instructions to cite sources to the message under the edit box edit

Suggested by Chaser in 15 December 2008
  • The edit box is currently a mere copyright warning. If this really is a priority, let's make it a visible message for everyone who clicks "edit this page". Something like adding after verifiable the phrase "and cited in reliable sources".--chaser (away) - talk 04:26, 15 December 2008 (UTC)[reply]
    • I'm glad someone else wants to do this. It was brought up here in June 2007 and never done, then I decided to be bold and make the change a few days ago and was promptly reverted. Rather than rehash the case I've made over there I'd like to ask anyone concerned about the absurd backlog of unsourced articles to have a read of the recent discussion and chip in. Even if it's mostly ignored a small improvement would be well worth a very short addition to the warning in my view. It's a very simple step with no changes to policy or software requried. — Trilobite 01:59, 30 January 2009 (UTC)[reply]

Actually enforce WP:V edit

Suggested by Rividian in 17 December 2008

WP:V says that challenged material can be removed and that the burden is on people who want to include that material to find references. This sounds like the solution to the whole problem. Instead of adding {{fact}} tags that are, in the vast majority of cases, likely to just sit there for years... why not actually remove challenged material? Why should challenged, unreferenced material have the right to stick in articles for ages? It shouldn't, core policy says that it shouldn't... yet it does.

Getting serious about removing challenged material forces people to find references if they want that content to stay. Apparently people are content to not look for references if their content just gets a tag added to it, but otherwise remains in the article.

How to do this? It's actually kind of simple. Make restoring any challenged material without finding a reference as egregious offense as restoring negative, unreferenced material about a living person. The kneejerk argument people make against this is basically "Well there's a lot of unreferenced content, we can't really expect every article to actually be based on cited sources." I say that if we expect to be taken seriously as a factual encyclopedia we can.

Research Honours - Research Apprentice / Research Master / Research Doctor edit

Suggested by SilkTork in 3 January 2009

Set up a scheme like DYK and GA and FA in which people are rewarded for levels of research on individual articles. 10 cites gets a person a Research Apprentice award, 25 gets a Research Master award, and 100 gets a Research Doctorate. People could put these awards on the top of their UserPages much like people do now with DYK, GA and FA awards.

Simplify the process of adding citations edit

Suggested by 66.57.190.166 in 5 January 2009

The process of researching and adding citations is lengthy and complex. Wikipedia makes it more complex with a combination of instruction creep and vagueness. The rules should be examined with an eye towards making them easier to follow. Some sort of technological fix should be implemented to make adding citations in the right format easier.

edit

To qualify for B-class an article needs to be verifiable from the references provided. GA has a few more whistles, mostly of the WP:MOS variety. Currently the B/GA articles have no visibility on the main page, but a single verifiable fact plus some bureaucratic qualifications are enough for a DYK. There's an advertisement gap on the main page between DYK and FA. Advertising articles that have been reviewed as factual should entice participants to write fully verifiable articles. Many articles don't have FA potential because of length issue; for many topics there aren't enough sources to write a FA-length article.

I propose a DYK-like box on the main page for GA-level articles: preview (look under featured pic).

Related proposal: unify peer-review/B-level review/GA review edit

Suggested by Xasodfuih in 6 January 2009

There are too many sub-FA-level review processes. All aim to produce fully verifiable content. A single review process should assign B or GA status depending on coverage and prose quality.

Use of WebCite edit

Suggested by OlEnglish in 18 January 2009

There should be more emphasis put on archiving citations with tools such as WebCite. Not just here in this WikiProject but throughout Wikipedia. I think it's very important for maintaining verifiability and reliability to preserve the countless web citations used on most articles. I'm guessing the vast majority of references cited are web sites, which because of link rot just aren't that reliable, and with link rot being so prevalent who knows how many featured articles' cites are now broken, and it'd take forever to check them all. That's why I think we should make a stronger effort to archive any web references we use as they are cited when writing new articles.

Plus it's very easy to do and only takes 30 seconds! Just enter the url to be archived at http://www.webcitation.org/archive.php and then use the new archived url with |archiveurl= and |archivedate=

Future edit

Once the problem is no longer the lack of sources in articles, we need to go a step further and start checking references on a regular basis to make sure they still point in the right direction.

How can I help now? edit

There are a number of different Wikipedia Backlog categories that help track the existing backlog of referencing activities. They fall into two major groups, Lacking References and Ensuring Accuracy through existing references:

Help now: Solve Citations Needed! edit

At the moment, there are over 517,344 articles with "Citation needed" statements. They are some of the easiest referencing issues to solve, because you are frequently looking for a targeted and singular reference. You can browse the whole list of these articles at Category:All articles with unsourced statements.

With 517,344 statements that need WP:Verification, sometimes it's hard to choose which article to work on. The tool Citation Hunt makes that easier by suggesting random articles, which you can sort by topical category membership.

I can help! Give me a random citation to find!
Q: Why not simply delete all unreferenced statements?
A: Unreferenced doesn't generally equal untrue, nor does it necessarily mean there are no references to be found. When faced with unreferenced information, an editor should make an effort to find sources before deleting material. The exception is when questioned material is libellous or otherwise harmful to people; in such cases, it is entirely proper to remove unsourced or improperly sourced statements, immediately.
Q: Why not increase the bar of notability so we get a lower, more manageable number of articles?
A: Changing the bar of notability would lead to the removal of thousands of articles, but the ones that would be removed aren't necessarily the ones that currently lack any sources.

See also edit