Throughout the Impact Challenge, we’ve touched on the importance of having persistent identifiers like DOIs for your research.
DOIs–digital object identifiers–make it easy for others to find your work by providing a permanent, unique identifier for each research output. That identifier will always redirect to where your work is stored, even if the URL changes, the journal you were published in disappears, and so on. All you have to do to make a DOI linkable is append “http://doi.org/” to the front of a DOI, like “http://doi.org/10.5061/dryad.585t4” for doi:10.5061/dryad.585t4.
DOIs also make it easy to track when and where your research is cited, discussed, shared, bookmarked, or otherwise used across the Internet. DOIs are widely used, understood by most researchers, and well-supported by platforms that track impacts across the Web.
Let’s dig into how you can get DOIs for articles, data, software, and other types of research outputs. It will set you up well for tomorrow’s Challenge, which will cover services you can use to track the impacts of your work using DOIs and other permanent identifiers.
DOIs for articles & preprints
Many journals issue DOIs for journal articles automatically. So, getting a DOI for your articles can be as easy as publishing with a journal that issues them.
If you’re planning to publish (or have already published) in a journal that doesn’t offer DOIs, that’s okay! You can archive a preprint or publisher-accepted postprint (peer-reviewed final draft of the article that’s not the formatted, published version) of your article on a platform that issues DOIs like Figshare, Zenodo, BioRxiv, or ResearchGate. Some institutional repositories also can mint DOIs. Here’s how.
Figshare, Zenodo, BioRxiv & some institutional repositories
All of these services work pretty much the same for issuing DOIs: you upload an article and a DOI is assigned automatically. We’ll briefly walk you through the process here using Figshare as an example.
- Login to Figshare and click the “Upload” link in the upper-right corner.
- Upload the article and click the “Add info” link.
- Add a description of the file (metadata). Be as thorough as possible when describing it; rich descriptions can make it easier to find your article using search engines.
- Some journals require that you add a statement to the archived preprint. It’s usually something along the lines of: “This is a pre-print version of the following article: [full citation pointing to publisher’s website]. It is posted here with the publisher’s permission.” You can usually find the statement on the “Author’s Rights” section of your journal’s website, and some relevant policies can be found on Sherpa/Romeo.
- Make the article “Public” (select the radio button for Public immediately to the left of the “Save changes” button.
- On the item record that’s now live on the Web, you’ll see your DOI:
The placement of the DOI will vary depending upon what platform you’ve uploaded it to, but the result will be the same: as soon as you’ve completed the upload process, a DOI will be automatically generated.
ResearchGate recently started allowing users to mint DOIs for articles that don’t yet have one, but it’s not done automatically:
- Login to ResearchGate
- On your profile page, click “Add your publications”.
- Select “All other research” in the pop-up box.
- Upload your article and add descriptive information, click “Save”.
- On your item record, click the “Generate a DOI” button at the top-right of the page.
- Confirm your publication details are correct and that the article doesn’t already have a DOI. Click “Generate a DOI” again.
- You’ll now see your DOI:
DOIs for data
You can also get DOIs for research data thanks to disciplinary data repositories like Dryad, KNB, and many others found on re3data.org. Plus, some national data repositories like ANDS will issue DOIs for data, too.
When should you mint a DOI for your data? Natasha Simons of ANDS says a DOI should be applied when:
- the data will be exposed and forms part of the scholarly record [this can be when you’re publishing supplementary data alongside a paper, “opening up” unpublished datasets, or otherwise making your data available to others];
- the data can be kept persistent [it won’t have to be removed from the repository];
- and the minimum DataCite metadata schema requirements can be met [you’ll need to provide information on the dataset’s Creator, Title, Publisher, and Publication Year; the Publisher information is communicated by your repository]
Getting a DOI for your data is usually as easy as just depositing your data. Nearly all data repositories that issue DOIs mint them automatically for new deposits.
Many repositories only issue a single DOI for a dataset, even if “versioning” (uploading of newer datasets, with the history of changes to the files preserved on the repository) is allowed.
But if you’ve got data that will be updated over time, you might need to use a repository that will issue a versioned DOI. Versioned DOIs can reflect what version of the data others are citing, making references to older versions of the dataset possible.
Dryad is just one repository that issues versioned DOIs. Here’s how it works:
- You upload your data to the repository and get your base DOI (e.g. doi:10.5061/dryad.585t4)
- When you upload a new version, Dryad will create a new suffix to the DOI that points to that particular version of the dataset (e.g. doi:10.5061/dryad.585t4.v2)
If you don’t have a national, disciplinary, or other specialized repository available to share your data, you can always deposit it to Figshare or Zenodo, where a DOI will be minted automatically.
DOIs for software
It’s easy to mint a DOI for your research software if you use GitHub to host your software and the connect it to your Figshare or Zenodo account. Here’s how it works:
- Choose the public GitHub repository* you want a DOI for
- Login to Figshare or Zenodo
- Connect your Figshare or Zenodo account to GitHub
- Use Figshare or Zenodo to select the GitHub repository you want a DOI for
- Check that your GitHub repository is set to communicate with Figshare or Zenodo
- Create a new “release” of your GitHub repository
- Head back over to Figshare or Zenodo and make sure the full description of your software package appears in the Uploads section, then submit your software
- An automatically-minted DOI will appear on the item record page. Here’s what that looks like on Zenodo:
In depth instructions to minting DOIs for software can be found on GitHub.
If you’re not on GitHub but want to mint DOIs for your software, you can upload your software and the accompanying documentation as binary files to Figshare and Zenodo, like you did with research data in the step above.
* Confusingly (for those of us more accustomed to data repositories and institutional repositories), GitHub calls specific software packages “repositories”.
DOIs for open peer reviews
An increasing number of journals and peer review platforms are issuing DOIs for open peer reviews.
If you’ve openly reviewed a journal article, there are two main ways you can get a DOI for your reviews:
- Review for a journal like PeerJ or peer review platform like Publons that issues DOIs automatically
- Archive your review in a repository that issues DOIs, like Figshare or Zenodo
DOIs will allow others to easily find your open peer reviews and also allow you to track discussions and reuse of your peer reviews across the Web, like you can with other scholarly outputs. That’s a major advantage over private, anonymous peer reviews, which are never seen beyond your editor and the article’s author and can rarely be claimed for credit towards the enormous amount of intellectual work they require.
DOIs for everything else
You can easily mint DOIs for your slide decks, posters, and even your blog posts if you upload them to Zenodo or Figshare, following the instructions outlined above.
Many of the limitations of DOIs are caused by human error. For example, though it’s ideal for your links to your work to use the DOI link (more on that below), you can’t control whether others will actually do it. That’s because research is often shared online using regular, easy-to-copy URLs instead of DOIs.
The best you can do is provide the DOI on the same page where the research output is shared. List it front and center, along with a preferred citation, so that anyone who shares your work will hopefully see it and follow your instructions.
It’s also bad form to create more than one DOI for a research output. So don’t mint a DOI for anything that’s already got one.
The final limitation is that we’re all counting on the publisher or service provider to keep the DOI record up to date with the DOI registration agency (most commonly Crossref or DataCite). And keeping records up-to-date is what ensures that DOIs point to the correct place on the Web (which you’ll remember is useful if URLs change or journals fold).
Most reputable publishers do this, but some publishers and repositories may not be as responsible (for example, as of this writing, as far as we know, ResearchGate doesn’t have a documented preservation policy). If you aren’t sure if the publisher’s archiving policy is up to snuff, ask them about it.
First, mint DOIs for your 5 most important research outputs that don’t already have them. Bonus points if some of those outputs are not articles.
Once you have your DOIs, use them:
- Put them onto your CV alongside your research products;
- Update your ArXiv preprint metadata to point to them;
- Put clearly-labeled preferred citations that include DOIs into your dataset or software documentation; and
- Encourage others to always link to your review using the DOI resolver link (these are created by putting “http://doi.org/” in front of your DOI; here’s an example of what one looks like: http://doi.org/10.7287/peerj.603v0.1/reviews/2).
Now that you’ve got DOIs for your most important research outputs, we’ll explore how you can use altmetrics and impact tracking services like Altmetric.com and Impactstory to discover how often they’re cited, saved, shared, discussed, and otherwise reused online. Stay tuned!