Many questions and answers will involve citing journal articles and other scientific works.
Clearly, it's desirable if there is a link to a free online version of the article. However, such links (e.g., from an academic's website) often disappear over time. So it is also desirable if at the very least, enough information is given to enable someone to find the article even if the article disappears.
The following represents the preferred style for references and citations on this site. Broadly speaking, the site follows APA style. Following the style is optional, but all users are encouraged to edit posts that are missing one or more of the elements discussed below.
References should be in APA inline format. This allows the reader to know where a claim is coming from without forcing them to read a footnote1
As mentioned in Smith and Jones (2008)
or
As seen in several studies (e.g., Smith & Jones, 2008)
A full APA Style Reference List is preferred at the end of the question or answer. The paper title should be a link if possible and the DOI should be provided (and also a link) if possible. By providing the full citation a user can find the article even if the posts links break (which is common) and the DOI is not known. These references can be separated from the body text of the question/answer using the Block Quote.
Mercier, H. and Sperber, D. (2011). Why do humans reason? arguments for an argumentative theory. Behavioral and Brain Sciences, 34(02):57-74. doi: 10.1017/S0140525X10000968
Smith, J. and Jones, D. (2012). An article about stuff. Journal of Stuff, 28(3):71-74.
Google Scholar makes it easy to quickly obtain an APA formatted reference based on a quick search (see this guide).
If possible a Digital Object Identifier (DOI) should be provided. Other sources such as PDFs may be removed and result in broken links; DOIs are eternal.
Link the DOI with https://doi.org/ and then the DOI number goes on the end of that.
For example, the DOI number for Mercier & Sperber (2011) above is 10.1017/S0140525X10000968 and that number is best linked to https://doi.org/10.1017/S0140525X10000968 making the linked reference
Mercier, H. and Sperber, D. (2011). Why do humans reason? arguments for an argumentative theory. Behavioral and Brain Sciences, 34(02):57-74. doi: 10.1017/S0140525X10000968
If you have found a freely available copy of the article such as a PDF or full text link, please use it in addition to the DOI, or in place of the DOI if no DOI is available.
1: Footnotes are a pain in the butt and harm the flow of reading hence they are not used in standard APA
<sub> method? it works just fine with multiple lines for references.
I think we should provide DOI links (i.e. http://dx.doi.org/...) for peer-reviewed articles as they are standard and stable. Of course it is good to point to an open copy but (as it may disappear) as an additional source.
See e.g. a related thread on meta.tp.se (there is an encouragement to use arXiv links (e.g. http://arxiv.org/abs/...) as it is both open-access and points DOI, but it works only for physics, mathematics and cs papers).
Is there a preferred way to cite articles in a question or answer?
One think to mention here, mostly for any users finding this answer who are not familiar with Stack Exchange, is that it is not acceptable to simply link to an article as your answer.
That is, good answers are more than just links, they contain content as well. This is doubly true if the link goes to some paid source. I for one have no access to paid sources of cognitive science articles. But even for those who do, as you said, the links may go dead.
Always include relevant links and provide proper citation / attribution, but always quote the relevant parts of what you link to.
As an example, I'll quote what user Brian R. Bondy posted in his answer to a question about how to answer questions with a link from Meta Stack Overflow:
Summarize but give credit to the source via a name and link.
Links alone can get broken over time. A summary can be improved upon by others who edit your post in the future.
Posting links alone helps the person asking the question, but for future people it's nicer to explain in the answer.
I think we should use inline APA style whenever possible.
This is of the (Watson, J, 1960) format. We can make the citation a link as well, though SEO wise the title of the article should be the link. This allows us to provide a proper academic reference without harming the flow of the question or answer.
In my opinion posts should be edited to include APA style references to journals to maintain an academic appearance and ease of reading throughout the site.
I have a few initial guiding principles that I've been using in the first few days of private beta:
You can see some examples of this approach in this answer, of which I quote one example below:
See for example this paper on "the role of deliberate practice in chess expertise" PDF by Charness, Tuffiash, Krampe, Reingold, & Vasyukova, (2005).
I also quite like the idea of APA style where you include both inline citation and a full reference at the end, but I also realise that adding formatted complete references is a bit more work. I also think that there is an issue of the optimal placement of the link to the PDF. Do you put it inline or at the end of text reference?
In response to Jeromy Anglim's answer:
provide enough of a citation somewhere in the text to enable someone to find the article even if the url breaks
That's a very good guideline which results in a true benefit.
I actively try to give the reader a sense of what they are about to click on. E.g., if you are going to click on a link that is a wikipedia article, then I'll let people know that, instead of just converting a word to a link.
Agreed to some extent. By using APA style as in Ben's answer it's clear you're referencing a scientific paper. For other sources, wiki/webpages like you say, it's best to mention them prior to quoting them as e.g. I did in my amateuristic answer.
On the other hand, I do not feel terminology links should be announced, as I did e.g. for online disinhibition effect. They provide a quick link for people who aren't familiar with the subject, and the source of the information is less important. If the link shows up dead they can still google it.
As to the other points Jeromy mentions, I am not entirely convinced. I would definitely also link to paid links, even if it is just to be able to read the abstract, simplify searching the paper and as proof you're not just answering from the top of your head.
I wouldn't worry about dead links in general, as long as enough context is given to understand the answers without having to follow the links. This is a general SE guideline which we definitely should follow as well. When links do turn up dead after a while, it wouldn't be too difficult to find working links again, provided that all the other guidelines are followed.
