09 July 2018

Giving an article a DOI does not make it citable

With preprint servers and data repositories like Figshare gaining in popularity, a related myth is also gaining in popularity: “Giving this article / dataset a DOI makes it a citable academic product!”

Exposition for bystanders: DOI is short for “digital object identifier.” I usually describe it as “a serial number for a journal article.” You often seen them tucked at the top or bottom of academic papers,  a lengthy number beginning with 10. But as the name indicates, it’s mean to be usable for any kind of “digital object,” not just papers.

The beauty of a DOI is that if you have it, you can type in “https://doi.org/” followed by that number beginning with 10, and it will take you straight to the paper.No having to go to Google to find the journal, then drill down to the volume, then the issue, and so on.

A Twitter search shows people have been wrongly making the connection between “having a DOI” and “something that can be cited” for at least seven years. Here’s a tweet from 2011 (emphasis added):

#DataCite offers DOIs for making data sets citable and getting credits for its reuse.

A few example: here, here, and here. And also here, here, and here. And people have been pointing out this is wrong for the same amount of time.

What is “citable” is an editorial policy set by journals individually.

Most journals have long traditions of allowing you to cite things that are not part of the formal scientific literature. After all, journals existed well before links ever existed, never mind the DOI standard.

But not every journal will let you cite whatever. If a journal says, “We will only allow citations to peer-reviewed journal articles,” saying, “But this has a DOI!” is not going to make any editor change her or his mind.

I don’t quite understand why people think this. I suspect this myth arises because it plays into scientists’ obsession with formalism and simple “If this, then that” rules. Maybe people are confusing the DOI number itself with, “Anyone who goes to the bother of giving DOIs to things is probably an organization that is fairly large, stable, and has its act together.” But that exploitative journals regularly give their articles working DOIs shows that it can’t take that much to assign those number and get the links working.

It’s another example of the information vacuum around academic publishing: publishers make up new stuff and assume academics will figure out how it works.

While I’m here, Lyndon White pointed out another DOI myth: that the link “never breaks.” They can, although in practice I’ve found them to be far less susceptible to link rot than publisher links, which get rearranged every few years, it seems.

Related posts

Innovation must be accompanied by education
Why can’t I cite Mythbusters?

1 comment:

Mike Taylor said...

My own most recent paper (Taylor 2018) cites two of my own SV-POW! blog-posts. None of the four reviewers, nor the handling editor, raised an eyebrow. Truly, we're living in the future.