Write a comment

This is a timely and novel study that discusses a topic of importance to the community. Although limited in scope to the life sciences, its conclusions are likely to be more broadly applicable and of interest to preprint servers, publishers, and scholarly infrastructure organisations. The main conclusion that more work is needed to communicate retractions across different versions of a manuscript is valid and well-supported by the data. 

The discovery that all retractions are in OA journals is fascinating and, to me, unexpected. As noted, the sample size is small, and it would be curious to know if it holds up across a broader range of subject areas and preprint servers. Is it because authors who deposit preprints have a tendency to publish in OA, that OA journals have more readers and therefore more scrutiny, or a weakness in OA review procedures?

Some specific comments: 

- The reason for more linked preprints from bioRxiv and medRxiv could also be stricter screening procedures for those preprint servers.

- ASAPbio focuses on life sciences and there isn't such a coordinated effort to promote preprints and set standards in other disciplines, although COPE has put out a discussion document (https://publicationethics.org/resources/discussion-documents/preprints). 

- The article would benefit from an expanded discussion of possible limitations and suggestions for follow-up studies.