13

Wikipedia is a very useful resource, but:

  • sometimes it has unreliable information.
  • sometimes it changes.

Should Wikipedia be used as a reference to support claims in answers?

20

My point is -- Wikipedia changes

There is a simple fix: correctly citing Wikipedia means providing an oldid link. For example, to reference the “Great Moon Hoax” one would provide the following URL:

en.wikipedia.org/w/index.php?title=Great_Moon_Hoax&oldid=419939762

This is the proper way to cite a uniquely identifiable version of an article that will never change. The relevant link can be found for each article on the left-hand side at “Permanent link”.

Perhaps we should encourage that users use this URL format to cite Wikipedia.

Apart from that, Wikipedia is not an original source and therefore not always appropriate anyway. On the other hand, a lot of articles are well-researched, well sourced and provide a nice editorial on a difficult topic that is otherwise hard to find. Linking to such articles should not be discouraged: alternative sources may simply be impossible to come by.

  • 9
    On the other side: Wikipedia might get improved, with newer information, better information, corrections. – user unknown Mar 23 '11 at 9:54
  • 1
    I don't understand this passage: a lot of articles are well-researched, well sourced [...] Linking to such articles should not be discouraged: alternative sources may simply be impossible to come by." -- how can an article be well sourced, yet alternative sources impossible to come by? – Sklivvz Jun 30 '11 at 23:03
  • 6
    @Sklivvz You forget what encyclopedias do: collecting information and presenting them in an easily understood form. Most papers are horrible to read even for professional researchers, for a variety of reasons (but chiefly because researchers suck at writing, and because their job often isn’t to be understandable to lay people). I’ve already said that in my answer, too: “provide a nice editorial on a difficult topic that is otherwise hard to find”. – Konrad Rudolph Jul 1 '11 at 8:30
  • 4
    @Sklivvz furthermore, most sources could be behind a paywall, or worse - only exists in paper form and never digitized. It requires a library trip to verify such sources. (Note: not just any library - you might need a collegiate inter-library loan to have it retrieved from another city.) – user2547 Sep 5 '11 at 4:33
11

While I agree with Ustice that wikipedia is a good first source for material in answering questions, I also think we should encourage answerers to go deeper. Not so much because wikipedia might change (that can be mitigated by directly quoting the text as it was at the time of the answer), but more because primary sources are generally better. If all we are doing is linking wikipedia pages, we aren't adding much value.

I also think that wikipedia should not be the sole source in a good answer. It's a great starting place, but a good answer should reference primary sources whenever possible.

  • 3
    Small addendum: if your only source is Wikipedia, then it follows that either the question is too simple or that you didn't dig deep enough. – Borror0 Mar 22 '11 at 18:04
  • +1 because "adding [much] value" should be encouraged with the intent to protect this Q&A web site as an important contribution to the internet. (Copying from Wikipedia might even be comparable to duplicating another student's answers during an exam.) – Randolf Richardson Aug 22 '11 at 2:46
5

I think that it is good form to reference Wikipedia for answers, though it is better to go deeper. Usually in Wikipedia articles, you can find the sources for the info contained in the article, and it is helpful to list those and other sources as well. Wikipedia may not be a definitive source for info used in answers, but my guess is that almost everyone checks Wikipedia first before looking for other sources. It's an excellent place to start.

For any article that has been around a while, they have pretty stringent rules about citing sources. That isn't to say that those sources disappear, as I have found on several occasions where I wanted to use one of the sources on a Wikipedia page, and I got 404'd. Other than that problem, I think that Wikipedia is pretty acceptable.

  • 1
    You can often get a copy of the 404'd document by copying the URI into The Wayback Machine at: archive.org – Randolf Richardson Aug 22 '11 at 2:49
5

Wikipedia is not reliable enough to act as a primary resource for significant claims in an argument. There are some limited situations where it is appropriate, but when in doubt, do not cite it.

This is not intended as an attack against Wikipedia. Used correctly, it is a useful tool for starting research and finding other sources that you can reference. It is frequently correct. However, it sometimes relies on other resources that we would not accept as reliable. It is not the only source that we recognise as both useful and unreliable (see Other Sources) but has been raised as a question often enough to deserve special mention.

Situations where it is acceptable to reference Wikipedia:

  • When you are not making a claim, but using it to explain a technical or jargon term that some of the lay-readers may not be familiar with.

  • When the nature of the claim you are making is clearly uncontroversial, and is not the key aspect being investigated, e.g. the melting point of some chemical. However, if your only references are to Wikipedia, this should be a warning that your answer does not rely on empirical evidence.

  • When you want to demonstrate the notability of a controversial claim in a question. i.e. you are questioning something you see in Wikipedia. In such cases, changes to the page may invalidate the reference, so provide a permalink, i.e. one with an oldid field, like so:

en.wikipedia.org/w/index.php?title=Great_Moon_Hoax&oldid=419939762
  • 1
    Wikipedia is as accurate as the encyclopedia brittanica (nature.com/nature/journal/v438/n7070/full/438900a.html) – Avi Mar 14 '13 at 6:13
  • @Avi: Thank you. That doesn't redeem Wikipedia; it damns Brittanica. You may substitute Encyclopedia Brittanica for Wikipedia in the above answer, and get an equally valid version. – Oddthinking Mar 14 '13 at 6:20
  • Okay, what standard are you using for reliability then? – Avi Mar 14 '13 at 6:25
  • 2
    @Avi: It's a difficult question, but: primary sources beat secondary sources, peer-reviewed beats non-peer-reviewed, reproduced beats non-reproduced beats non-reproduceable, large samples beat small samples beat anecdotes, randomised and controlled beats uncontrolled, double-blinded (especially for medicine) beats single-blinded beats unblinded, meta-summaries beat individual experiments and, IMO, the Cochrane Collaboration pretty much own the top of the mountain! – Oddthinking Mar 14 '13 at 9:18
  • In what way are primary sources more reliable than secondary sources citing primary sources? Is the rate of error for a paper really lower than the rate of error for an encyclopedia article? – Avi Mar 14 '13 at 9:26
  • 1
    We're starting to get off-topic here and should probably take it to chat or another meta-question. We commonly see secondary sources getting it wrong (e.g. newspapers quoting from media releases quoting from articles). Unless it is doing further analysis (e.g. meta-summary, literature review), it's hard to see how a secondary source could have fewer errors than the primary source it is relying upon. We commonly see Wikipedia making uncited claims, and (less common, but not rare) providing references that don't support the claims. – Oddthinking Mar 14 '13 at 9:44
  • I understand the problem with secondary sources that try to summarize a primary source, but that's not what wikipedia is. News articles aren't really relevant to my question. If you think it's off topic that's fine. – Avi Mar 14 '13 at 9:46
  • I'm confused. Wikipedia frequently tries to summarise a primary source with a sentence or two. – Oddthinking Mar 14 '13 at 10:07
  • It combines many sources. It's against wikipedia's policy for an article to rely on a single source. You'll occasionally see notes at the top warning against reliance on a single source. As such, wikipedia's claims represent an analysis of a variety of primary sources, not an interpretation of one. – Avi Mar 14 '13 at 10:35
  • let us continue this discussion in chat – Oddthinking Mar 14 '13 at 11:49
  • Chat is bugging out on me. I'll try tomorrow when I'm not tired. – Avi Mar 14 '13 at 12:31
  • I find the argument against Wikipedia (and Britannica) quite lacking. I can link to primary sources from the 80s that say be careful about the amount of whole eggs consumed because it dietary cholesterol affects blood cholesterol. Of course the newest research is now saying that's not true. While a link to Wikipedia (and the contents of EB) will be updated overtime, linking to the primary source will point to now debunked science whereas linking to Wikipedia will not (even if it now contradicts the answer). – Andy Mar 11 '15 at 13:28
  • I think having an answer contradicted by an updated source is better than continuing to link to a now discredited source. It will call the answer into question and will force (via downvotes) it to be updated as well. Links to disreputed primary sources will continue to sit on skeptics and contribute to the spread of disinformation. – Andy Mar 11 '15 at 13:29
3

Wikipedia changes. That's an advantage of wikipedia.

An astrological book from 2000 would contain Pluto as a planet, for example, while Wikipedia has a correction. Konrad Rudolph explained, how to set a link to Wikipedia at the time, the link was given, which will help to keep the opinion, mentioned in the answer, in sync with the citation.

But if you link to Wikipedia, and - just for the argument - you say: Pluto is a planet, see: Wikipedia:planets. And there is written: No, no, it isn't! The reader will see a contradiction, should open a comment or edit the article. Of course, if nobody ever reads the answer or wikipedia, it will not matter at all.

But imho, it is better, people recognise the discrepancy, and can, with the date of the SE-answer, reconstruct the wikipedia-page of that day, instead of visiting an outdated wikipedia article.
Sometimes, the wikipedia-article-part might just have moved due to some restructuring going on. Maybe even then it might be better, to recognize it early, and correct the SE-text accordingly.

0

In many cases yes.

If the article in Wikipedia is of good quality with multiple references which would considered here as valid, there is no real reason not to accept it.

If the article in Wikipedia is controversial and with multiple [citation required] tags, then it shouldn't be used as a source.

  • 2
    This sounds like "I can't be bothered to dig to primary sources, so I will let other users verify them" – Sklivvz Jun 30 '11 at 18:30
  • 2
    I always consider it better to go to the primary source. I've seen it often enough that the summary in the secondary literature was misleading or even outright wrong. You should verify that the reference is actually supporting your claim instead of relying on someone else to accurately summarize that reference. – Mad Scientist Jun 30 '11 at 18:31
  • 4
    If it has multiple valid references, then use those references! You might as well argue for using a Google search as a reference when it turns up valid references; why throw indirection at the reader? – Shog9 Jun 30 '11 at 18:32
-1

If it has multiple apparently-valid references and looks well-researched, and you did not read them all, then copying the links to the references is a form of lying. Link to the wikipedia article, not to the links.

Yes, you could take the time to check them all and then link them. That's preferable.

But it is infinitely better to link the source you used, than some source that someone else claimed to have used but you did not.

Basically: tell the truth. If you just searched Wikipdia, then just link Wikipedia. It's as simple as that. If the reader doesn't like Wikipedia, that's fine - they can read some other answer, they can follow the link to wikipedia and go on a link-dive into each of the individial cites, etc ad infinitum. They'll do this if it's important to them.

But there's nothing here or elsewhere that says you must only answer questions that you are willing to take an hour of research to cite properly. And that would be extraordinarily counter to the whole idea of stackexchange.

  • This started out great! Quoting references you haven't read IS lying. But then you jump to suggesting that it is appropriate to answer questions without doing enough research to cite properly. No, not doing that is the whole point of this stackexchange. If you think it is counter to the whole idea of stackexchange, then so be it - this stackexchange is an outlier. – Oddthinking Oct 20 '17 at 0:52
  • In part, obviously, I agree with you: taking that hour is well worthwhile. And further, it's farcically pointless to half-ass an answer if there are multiple good answers that did spend the time. But given the choice between one answer that leads the asker in the right direction, and no answer, I'd go with giving the answer. – Dewi Morgan Oct 20 '17 at 14:55

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .