[R] Citing R in journal articles (or the failure to)

[Achim Zeileis]

Peter:

> I was reading a paper recently in which I was surprised to see an R 
> package of mine obviously used, without acknowledgement.  Indeed, R 
> itself was used without any acknowledgment.  So I contact the author 
> about these issues, who said (in part):
>
> Regarding the R packages, I used the "tweedie" and "statmod" for my
> analyses as you pointed out.
> The referee of this paper advised me that it is not needed to cite
> "R" because it is a kind of free software different from the
> commercial package such as SAS.
>
> What is the logic behind this?

No logic, rather obviously. The idea that one should only cite commercial 
packages (for which the user has paid with money) but not open-source 
packages (for which you have paid no money but coul "pay" credit) is quite 
awkward.

In the Journal of Statistical Software (http://www.jstatsoft.org/), we 
require that all software used (commercial or otherwise) is cited 
formally, i.e., by an item in the reference list (and not just a link or 
so). The main difference between R and many other packages is that it is 
rather explicit about the way it wants to be cited.

We observe that software is also cited increasingly in other journals and 
not surprisingly this is more commonly done in more computational 
journals. Acceptance of software references varies between journals, but 
my experience is that you will usually get away with "a few" references 
even in more theoretical journals. Personally, I typically add a 
"Computational details" section at the end of the paper and list 
packages/version I used.

> What can be done about it?

For your own work:
   - provide a CITATION file in your packages
   - insist on citing software in your own papers

For other people's work:
   - point users to citation("mypackage")
   - write to editors if you notice that your work has not been cited

> That makes me quite angry...

Sure, I would be as well (...and have been about papers that used but 
didn't cite my packages). I think there is not much you can do about the
article that has been published. But by pointing authors and editors to 
the fact that software is scientific work that should be published (just 
like other scientific work) will help to change attitudes about this. It's 
a slow process though...

Best wishes,
Z