I tried to encourage my colleague that although this is disappointing it is just part of being a good scientist. I was reminded of an old post based on a paper, The Seven Sins in Academic Behavior in the Natural Sciences by Wilfred F. van Gunsteren.
One could even defend the proposition that a scientist with a sizeable publication record in science who has not published a single corrigendum is unlikely to be a good scientist. Either he or she has done such simple work that nothing could go wrong, or he or she has committed the fifth sin in science [neglect of errors found after publication.This is not an excuse for the sloppy work which is becoming more and more common in science due to the rush to publish the most surprising and spectacular results.
Results do need to be carefully checked and double checked.
But we have to face the fact that mistakes will happen... just like car crashes, burnt toast, stubbed toes, catching colds, ... Precautions should be taken, but inevitably they will fail sometimes...
The amount of checking should be in proportion to the importance and the surprise of the results.
I think my most significant errors were several in my first paper on hydrogen bonding. I pointed out the errors on this blog and in the second paper I wrote on hydrogen bonding. There was a fortuitous cancellation of errors that meant the conclusions of the first paper were still valid. Arguably, I should also publish an errata on the first paper.
I also think it is important that in public all co-authors take joint responsibility for errors. In particular, it is quite dubious for senior co-authors to shift the blame to junior co-authors. Many senior people are only too happy to take credit. They also have to accept liability.
What do you think about the errata criteria for a good scientist?
I also welcome your own stories about erratum.
In an experimental paper there are (usually) two parts: the data and the interpretation of the data. If you find that your data was wrong you should either correct or retract your paper (depending on the nature of the error). But if your data is 100% genuine and your interpretation turns out to be wrong, then this is just part of the scientific process. With hindsight someone (maybe you yourself) will publish another paper questioning the original interpretation and the community will slowly converge toward truth.
ReplyDeleteI appreciate that for theoretical and/or computation papers things are significantly different.
A strategy not mentioned in Ross' post is to "self-correct" minor errors in later publications. This can be appropriate if (a) there is a genuine need to publish later papers and (b) the item being corrected is relatively minor.
ReplyDeleteAt the same time, most journals make it very easy to publish corrections, so the "energy barrier" associated with doing so can be small.
The key issue, in my view, is how serious an error needs to be before a correction is warranted. It is likely that all papers contain errors of one kind or another, often very minor. If you think your own papers are immune, please let me know, since we are looking for papers to use in my department's PhD qualifying oral exams where they will be examined in minute detail by a dozen students.
Good article wrt your post.
ReplyDeletehttps://onlinelibrary.wiley.com/doi/epdf/10.1111/eci.12898
PERSPECTIVE
Improving the integrity of published science: An expanded taxonomy of retractions and corrections
Daniele Fanelli|
John P.A. Ioannidis|
Steven Goodman
Meta-Research Innovation Center at Stanford (METRICS), Stanford, CA, USA
Department of Methodology, London School of Economics and Political Science, London, UK
Departments of Medicine of Health Research and Policy, and of Biomedical Data Science, Stanford University School of Medicine, Stanford
https://onlinelibrary.wiley.com/doi/epdf/10.1111/eci.12898
Daniella F writes very well on these matters. His web site has very enlightening papers on this.
http://danielefanelli.com/publications.html