Changing retraction criteria for Science family
Extensive corrections could lead to a retraction for lack of confidence
In this week’s editorial, Jake Yeston, Valda Vinson, and I address a topic that comes up a lot about papers that get extensive corrections.
With the advent of websites like PubPeer where image sleuths find figures in scientific papers that might have been inappropriately altered, we regularly hear about issues that need to be addressed. Most errors that crop up are due to inadvertent changes or aesthetic changes from the past that are no longer appropriate under today’s standards. These can usually be handled by simply substituting a correct image for one that may be problematic and the issuance of an Erratum or Correction. However, we are obligated under COPE guidelines to determine whether the author response to the concerns was “satisfactory” before deciding on a course of action. Whether the response is satisfactory is a matter of opinion that doesn’t always involve agreement with the authors, but it is ultimately our decision to determine that.
Our earlier Editorial Policies gave two criteria for when these errors could lead to retraction of a paper. The first was if there was demonstrated misconduct, and the second was if the main conclusions of the paper were invalidated by the error. We have added a third criterion in our updated Editorial Policies which states that if there is an accumulation of errors, the editors could lose confidence in a paper and retract it. Specifically, we state “An accumulation of errors identified in a paper may cause the editors to lose confidence in the integrity of the data presentation, and the paper may then be retracted.”
While we have always had the latitude to retract a paper for this reason under the ‘satisfactory’ rubric, we are specifically calling this out now to alleviate any confusion.
Occasionally, when we contact authors about possible errors, we get outstanding cooperation and a quick correction or retraction. That’s great when it works.
But unfortunately, a lot of the time, we get stonewalling from the authors and their institutions. That is extremely unfortunate. Science is done by human beings, and errors by authors and journals are inevitable. Only by forthrightly correcting them can we demonstrate that science is a self-correcting process that leads to the right answer in the long run.