Hello, and Happy New Year, Crossref community!
A little over a year ago, Crossref acquired the Retraction Watch database via an agreement with The Center for Scientific Integrity, the organization behind the Retraction Watch blog and database. Since then, we’ve made the Retraction Watch data openly available, first through our Labs API and more recently on our Gitlab repository.
The Retraction Watch database is the largest collection of retracted publications, and has a rich array of metadata, especially as pertains to the reason, type, and timeframe of each retraction. So, we’re very pleased to make it more readily available and excited to see how researchers end up using the data.
As the program lead for the Research Nexus, Martyn Rittman, noted here in the community forum that we have plans in the works to eventually incorporate the retraction data from Retraction Watch into our production metadata API. But, in the meanwhile, if you want to work with this data, you’ll likely be getting it as a csv file from the Gitlab repository.
We get occasional questions about the Retraction Watch database, and in reflecting upon them, I just wanted to note a few things to consider for new users:
- We have some very concise and useful descriptors of each of the column headers of the csv, in the READMe doc for our Gitlab Repository.
- For a more thorough deep-dive into the database, take a look at the Retraction Watch database User Guide, and its appendices for all the metadata fields; a detailed description of the reasons for retraction and the article types; and a comprehensive description of how the database was constructed.
- If you’re a regular user of our API, keep in mind that ‘retractions’ that are supplied in metadata records as part of the Crossmark metadata are not coming from Retraction Watch.
- Conversely, there are records that have a Crossmark update, with the update-type ‘retraction’ in our API, which you won’t find in the Retraction Watch database. They’re two separate sources of retraction information.
- We’ll be considering how best to distinguish Crossmark retractions from Retraction Watch retractions, in the process of incorporating the Retraction Watch data into our APIs and other metadata retrieval services.
- The Retraction Watch database csv is updated daily. There’s no way to retrieve a prior version of the data. If you think you’ll need to track when updates happen or items are added, consider cloning the git repository and pulling the recent changes day-by-day.
If you have used the Retraction Watch database, or are planning on using it in an upcoming project, we’d love to hear from you. Please feel free to share your experience and findings here in the community forum or reach us by email at support@crossref.org.