Wednesday, November 06, 2013

blogging and academic honesty

Paul Weiss and the editorial board of ACS Nano seem to have ignited a bit of a controversy with their editorial published a couple of weeks ago.  The main message of the editorial seems to chastise bloggers and other social media users regarding accusations of scientific misconduct.  As with many positions taken on the internet, there is a defensible point of view (in this case, "It is better to try to go through an investigative procedure regarding misconduct, given that the consequences of a false public accusation can be very severe") that was expressed in an inartful way (coming across as scolding, and implying that people who criticize published work are likely not necessarily accomplishing science themselves).  As my mom used to say, sometimes it's not what you say, but how you say it that matters.   Clearly, as this extremely long thread on ChemBark and this response editorial by Nature will attest, there are differing views about this issue.  Here's my take (not that I have any privileged point of view - your mileage may vary):
  • Public accusations of misconduct or incompetence should never be made lightly.  If misconduct is really suspected, then the appropriate first course of action is to contact the relevant journal and ideally the research integrity officer (or equivalent) at the authors' home institution.  Public accusation should not be a first recourse.
  • Journals should deal with accusations in a timely way.  Years passing before resolutions of inquiries is unacceptable.  Authors stonewalling by refusing to respond to inquiries is unacceptable.  While authors should be given every opportunity to respond, it is also not appropriate to, e.g., refuse to publish a comment because the authors drag their heels on writing a response.  
  • No one has yet come up with a perfect feedback mechanism.  Things like pubpeer are better than nothing, but anonymous commenting, like anonymous peer review, is a double-edged sword.  Yes, anonymity protects the vulnerable from possible retaliation.  However, anonymity also leads to genuinely awful behavior sometimes. 
  • People are going to make public comments about the scientific literature - that's the nature of the internet, and in general that's a good thing.  I would hope that they will do this with due consideration.  The journals are free to encourage people to pursue concerns within the journal system, but it's not productive to imply that people who draw attention to suspect figures (for example) are somehow poor scientists or gleeful bullies. We're all on the same side (except the cheaters).

5 comments:

Anonymous said...

http://www.nature.com/news/neuroscience-my-life-with-parkinson-s-1.14084?WT.mc_id=FBK_NatureNews

Douglas Natelson said...

@Anon, that is an excellent piece of writing (though not particularly germane to this topic). I wish him well.

Anonymous said...

I would say there is one caveat in your approach: you directly advocate going to the journal/office of research integrity.
But why bypass the directly responsible person in the first step?
If you do that in a management chain issue, you are sure to p*ss off the bypassed person. (You don't directly go to the boss of your boss...)
And nowadays, if you feel not comfortable exposing yourself by contacting the authors, you can easily create an anonymous email address. As long as your questions are polite, open, and pertinent ("could you clarify this, because I do not understand"), you'll be okay.
The response of the authors will most likely be illuminating - either to the science or to their fraudulent behavior...

Anonymous said...

and then chembark posts another "suspicious data" post from JACS today. you can see he treadly more carefully, but still... i like your "it's how you say it assessment" - chembark is edging on being overly-aggressive in his posts after his essay to paul weiss

Douglas Natelson said...

@Anon8:07, you are right in principle, particularly for me at my stage of career now. Nothing to do w/ misconduct, but I have emailed authors directly in the past if I think that they've made a mistake or made an interpretation that I think is unjustified, for example. This has led to some good, useful discussions, though striking the right tone in email (respectful, not "you are wrong!", since I could always be the one in error) can be challenging.

That being said, if I was a grad student or postdoc, I can see being nervous about emailing an author and saying that I thought they had suspicious-looking figures. If I was really worried, I don't think I'd trust online anonymity. I do agree that the best approach would be to contact the most senior author, and strike exactly the tone you mention.

@Anon11:13, Paul@chembark clearly had his hackles raised by that editorial. He asks rhetorically at the end of his current post whether this (the new example) is how the system is supposed to work. The answer is "yes", I think, even if this is unsatisfying.

In the broader discussion, I really don't like how a number of the chembark commenters seem to think that confidentiality policies regarding misconduct investigations are intended as a way for universities, journals, and funding agencies to sweep things under a rug. That's just not true.