How one journal became a “major retraction engine”

If you think you’ve noticed more and more retractions at PLoS ONE recently, you’re not wrong.

The journal retracted 53 papers last year. That’s not a record — that belongs to a journal that retracted more than 400 papers at once — nor is it that many more than the Journal of Biological Chemistry, which retracted 39 last year. And it’s only about 3% of the year’s retractions. But it’s a dramatic increase, as this graphic shows:

Alison Abritis/RetractionDatabase.org

Based on the 22 retractions so far in 2019, PLoS ONE is on track to retract more than 60 papers this year. In the words of the pseudonymous Claire Francis, whose comments at PubPeer have been responsible for highlighting problems that led to some of the retractions, PLoS ONE is becoming a “major retraction engine.”

So what explains the jump? Best we can tell, there are two significant — and related — factors.

The first is Elisabeth “Eagle Eyes” Bik, a scientist formerly based at Stanford and now working as consultant in scientific integrity. Several years ago, she reviewed more than 20,000 papers for evidence of manipulated images. She and her colleagues found that about 4% of those papers appeared to contain inappropriate duplication.

Nearly 40% of the papers she reviewed were in PLoS ONE, Bik tells Retraction Watch, because it was an easy journal to screen. (It’s open access, as the title suggests.) So it makes sense that roughly the same percentage — 348 of 782, or 44.5% — of the tainted articles were in the journal:

I reported most of these cases to PLOS ONE between March 2014 and October 2015. So that means that their staff suddenly had to deal with almost 350 investigations, which is an enormous task for a small team that probably never had to deal with more than 1 or 2 cases at the same time.  They have been slow to follow up on these cases, but I have indeed recently gotten an increase in responses about papers that I had reported, and that are now being corrected or retracted.

Bik sent us the DOIs of the 348 papers, and 22 of them have been retracted, and one has been subjected to an expression of concern, according to a check of our database. And 55 have been corrected in some way.

Reviewing all of those papers requires resources at PLOS. Last year, as we reported, the publisher hired three new staff members to do just that, as well as other research integrity tasks. The move is one that a number of journals have made as they realize that the growing awareness of problems in the literature means more and more allegations will likely come.

In comments to Retraction Watch, PLoS ONE confirmed our hunch about what was responsible for the increase:

PLOS ONE has always been committed to high standards of research reporting and we take various approaches to identify and address publication ethics concerns. As explained in our recent blog post, we have devoted more resources to research integrity in recent years, and last year formed an Editorial team dedicated to Publication Ethics. The increase in the number of retractions we posted in 2018-2019 as compared to previous years reflects this shift in journal resources, which has enabled us to more efficiently address concerns and resolve a number of older cases. Notably, the recent increase in the number of retractions reflects the volume and timing of case resolutions in 2018-2019; it does not indicate a change in the percentage of articles published per year for which post-publication concerns have been raised. This percentage has followed a relatively constant trend, with some fluctuation based on reports we receive from readers.  

We anticipate that in the near future the number of retractions overall may continue to climb to reflect the Publication Ethics team’s work in resolving pending cases. We are also continuing our recent successful efforts around improving the reproducibility of research reporting through editorial guidelines and policies that aim to proactively address potential issues prior to publication.

We asked Francis for comment. He said:

The patient is doing well.

Like Retraction Watch? You can make a tax-deductible contribution to support our growth, follow us on Twitter, like us on Facebook, add us to your RSS reader, sign up for an email every time there’s a new post (look for the “follow” button at the lower right part of your screen), or subscribe to our daily digest. If you find a retraction that’s not in our database, you can let us know here. For comments or feedback, email us at team@retractionwatch.com.

7 thoughts on “How one journal became a “major retraction engine””

  1. oh,PLOS ONE should take care, be responsible for the true science as the retractionwatch has been doing all the time.
    I want to know,what does “The patient is doing well.” by Mr Francis, whether it means PLOS one want us has the patient to wait for its improvement?Thank you.

  2. It’s hard to tell what others really mean.

    From the graph of retractions it looks like the patient (journal) is improving.

    Perhaps the words were meant as encourgement for what the journal has been doing more recently. From this post, and an earlier post, we learn that PLOS has hired 3 people to review papers.

    Plos One does not mention Elisabeth Bik’s reporting of inappropriate image duplications, but imagine that retraction of 22 of the 348 publications reported by Bik is not by chance.

    1. I think journals would be wise to credit efforts by volunteers like Dr. Bik who are helping the journals clean up mistakes that slipped through their own quality checks. Most seem to ignore those efforts and some find post peer review threatening and push back against questioners.

  3. Good on PLOS for cleaning up their archives. It’s great that the efforts of these error detectives are having their work recognised and validated through editorial action.

    COI Disclosure: Have published a paper with PLOS. Would submit again.

  4. This really doesn’t surprise me too much. PLoS One is a general purpose journal and seems to farm out submissions to a wide variety of editors who may have little familiarity with the topic and who then may choose reviewers who have similarly little familiarity with the topic. Since they are tasked not with judging overall quality or interest but rather with finding actual errors, thereby “eliminating subjective assessments of significance or scope” as the website says, a lot of not-so-good papers get published.

  5. No surprise at all. The journal has more than 6500 academic editors + staff editors. I used to count manually but i just noticed that it is impossible – 181 pages of academic editor names (each page has about 30 names). Massive effort from PLoSOne to recruit these individuals…

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.