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?
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:
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.
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.
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.
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:
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.
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.
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.