Hearing about retractions, involving both the regular news media and scholarly journals, is becoming more and more commonplace. With the general proliferation of misinformation across social media sources, this may even be expected by many readers. However, the increase in retractions from our historically “most trusted” information sources may have more to do with our increased technological ability to detect certain types of retraction-worthy offenses (for example, the re-use of already-published text without proper attribution).   

Why should retractions matter to MSK authors?

Retractions should be a concern of all MSK authors for two reasons: 1) it is not good for an author’s reputation if they have a retraction associated with any of their works, no matter how “small” the offense; and 2) it is equally detrimental if an author inadvertently cites retracted works (for example, after incorporating data from retracted works into their analyses/meta-syntheses, etc.) as it could put their own research results into question and at risk of retraction.

What can MSK authors do to protect themselves?

There are a number of things that MSK authors can do to protect themselves:

1. Educate themselves about what constitutes a retractable offense
   
   When most people think of retractions, their minds generally go to the most notorious of these in recent history that was proven to be the result of research misconduct– the 1998 Wakefield Lancet paper that reported on fraudulent research. What most folks don’t realize, however, is that a paper can be retracted for a much less serious offense. For example, if an author self-plagiarizes themselves (i.e., uses their own previously-written words to describe something that they have published on before and not properly citing their own past publication).
   
   To get a better appreciation for the variety of reasons for retraction, see:
   
   – Retraction Watch Database User Guide Appendix B: Reasons
   
   – Wager E, Barbour V, Yentis S, Kleinert S. Retractions: guidance from the Committee on Publication Ethics (COPE). Croat Med J. 2009 Dec;50(6):532-5. doi: 10.3325/cmj.2009.50.532. PMID: 20017220; PMCID: PMC2802086.
   
   
2. Take advantage of available writing support tools, like plagiarism detection software
   
   When most people think about plagiarism detection software, they immediately think back to their college days and – if those days were in the more recent past – the Turnitin software that their academic institution may have subscribed to in an attempt to curb student plagiarism and the use of “paper mills”. Many research institutions, including MSK (to request an account, send email to resources@mskcc.org), as well as academic publishers, have started to make tools like iThenticate available to their researchers and authors as part of their Author/Writing Support resources and services. These tools give authors the opportunity to check their manuscript before they officially submit it for publication to a journal, in order to double-check if any sections need additional paraphrasing or could benefit from the insertion of additional citations to other works.
   
   
3. Find out if an article you are considering citing has been retracted by:
   
   a) Searching PubMed
   
   To limit a set of search results from a PubMed search to those citations that have been flagged as retracted articles, authors can simply add:
   
   AND (retractedpublication[Filter])
   
   …to their PubMed search strategy or use the filter menu options. To use the filter menu on the left, choose “Additional Filters” to select the “Article Type” option Retracted Publication, and then select this option (once it is visible in the menu) to actually apply it as a Limit to the search results.
   
   
   
   
   
   
   
   
   
   
   
   
   
   
   
   
   
   
   
   
   
   b) Searching the Retraction Watch database
   
   The Retraction Watch database is also available for all to search for free on the Internet. Be sure to have a look at the User Guide for more details about its search capabilities and contents.
   
   
   
   
   
   
   
   
   
   c) Choosing a citation manager that flags retracted citations in your library
   
   The problem with retractions is that they can occur at any time after the paper has been published, often years later. As such, performing a final check for retractions – just as you are inserting citations into your manuscript’s bibliography – is recommended but admittedly can be time-consuming. Fortunately, several citation management tools have begun to incorporated this functionality into their products (by partnering with the Retraction Watch folks), including: Zotero, Papers, and the EndNote 20.2 software update (that will incorporate retraction alerts that “will now notify you of any retracted publications in your library”).
   
   Note: At the time of this writing on 11/23/2021, the EndNote 20.2 update has not yet been universally pushed out to all users of MSK’s institutionally-licensed EndNote software.

To lean more, be sure to Ask Us at the MSK Library!

International Open Access Week is a good time to revisit preprints and their growing role in the biomedical scholarly communication landscape. Although embraced by researchers in fields like physics for decades, it has only been in the last few years that researchers – and funding agencies – in the biomedical sciences have begun to become more serious about using preprint servers.

How are preprint servers the same or different than open access (OA) journals?

The most important difference is that manuscripts posted to a preprint server have not been confirmed by peer-review, whereas OA journal articles published in reputable scholarly journals will have gone through a rigorous peer-review process before being published. As such, taking extra precautions before citing research that appears only in a preprint version may be merited – for example, checking that it has not been listed in the Retraction Watch database.

That said, most publishers allow manuscripts that have been previously posted as preprints to also be submitted to their journals for eventual publication as a peer-reviewed article. There is even a Preprint/Journal Manuscript matcher tool now available that can help authors who have posted to either bioRxiv or medRxiv preprint servers to use a text-matching automation tool to identify good journal contenders for their preprint server-posted manuscript.

Although both OA journal articles and preprints are freely-available to readers to view and download, posting to a preprint server is free for the authors, whereas most OA journals charge an Article Processing Charge (ACP) or publication fee. Researchers can search the Directory of Open Access Journals (DOAJ) for OA journal ACP information.

Both document formats are now also accepted as works that can be cited in NIH awards reporting. The NIH has in fact expanded their working definition of publication to better accommodate “Interim Research Products” like preprints. As per NIH guidance:

Publication: A “Publication” includes (a) published research results in any manuscript that is peer-reviewed and accepted by a journal¹ or (b) a complete and public draft of a scientific document (commonly referred to as preprint).²

It is also important to note that most preprint servers will assign a DOI (digital object identifier) to the preprint manuscript that will be different than the DOI that may eventually be assigned to the final published article. As such, the two versions can and should be treated as separate “citable” items that can both be included in a researcher’s author profile(s) and CV.

Select preprints have also begun to be indexed in PMC and PubMed, initially as part of a pilot project for COVID-19 research, but “NLM will expand the pilot to include preprints resulting from the broader spectrum of NIH-supported research as curation and ingest workflows are refined, automated, and made scalable”. 

Last – ORCiD has also added features and functionality to accommodate preprint citation information in their author profiles. A preprint work type category has been added, as well as the ability for preprint servers that are ORCiD members to transfer citation information into author profiles. Furthermore, linkages within ORCiD can be created once the published article citation information related to that preprint becomes available.

To learn more about preprints, be sure to check out NLM’s new self-paced tutorial on Preprints or Ask Us at the MSK Library!