Retractions and corrections from Retraction Watch are now available in Crossref’s REST API. Back in September 2023, we announced the acquisition of the Retraction Watch database with an ongoing shared service. Since then, they have sent us regular updates, which are publicly available as a csv file. Our aim has always been to better integrate these retractions with our existing metadata, and today we’ve met that goal.
This is a companion discussion topic for the original entry at https://0-www-crossref-org.library.alliant.edu/blog/retraction-watch-retractions-now-in-the-crossref-api
4 Likes
If you want to explore the tutorial this is the specific resource. Make sure to visit the Learning Hub for other resources showing you different approaches to get metadata from the REST API
2 Likes
I’ve been trying to identify retractions by country using the API that go beyond the Retraction Watch database alone. But when ask for all retractions from Australian authors in the year 2022, for instance, it returns only three items, all of which are from the RW database. And I know there should be many more.
Here’s the query I’ve used:
works?query.affiliation=australia&filter=update-type:retraction,from-pub-date:2022-01-01,until-pub-date:2022-12-31
What am I doing wrong?
Thanks for the question. The update-type refers to the records you’re searching for, not how the records were updated. This means that there are only 3 retraction notices with an Australian author.
The best way to get what you want is to query for works with an update, then check the results filter them for retractions. We don’t have a filter that will directly return retracted works. You can make the query:
https://0-api-crossref-org.library.alliant.edu/v1/works?query.affiliation=australia&filter=from-pub-date:2022-01-01,until-pub-date:2022-12-31,has-update:1&rows=1000
There are 281 of these. I counted 5 with a retraction, so there numbers seem to be quite small in any case (there are about 54000 from Australian authors in total.
It’s also possible that there are more works from Australian authors that have been retracted but the affiliation metadata doesn’t include the country. These records would be missing from the results.
2 Likes
Thanks Martyn! That’s helpful and gives me something to try next.
From the RW database alone there are 58 retracted articles with Australian authors. If you search instead for Australian-authored papers with “RETRACTED:” in the title (as recommended in the recently released NISO RP-45-2024 standard) in other databases like Web Of Science it returns 64 articles, and in Dimensions, 61 articles. But although the numbers are similar there is only about 70% overlap in the results (by DOI). What I’m hoping is that I might be able to get close to the real number through CrossRef.
RW, Dimensions, and WoS all have some level of curation and are pulling in data from difference sources to supplement affiliation data. At Crossref, for better or worse, we only pass on metadata provided by our members. In the case of affiliations it can be incomplete, for example where there’s a mis-spelling or if only the affiliation of one author is included.
1 Like