Tuesday, July 8, 2014

Do Retractions Actually Hurt Journals?

I am not sure. In fact, I can’t help but wonder if STAP-like fiascos aren’t actually beneficial to a journal like Nature. Sure, the journal has been “soul-searching” and is now suddenly aware that fraud happens and will try harder to prevent it. But seriously, does this in any way damage the NPG brand? Are scientists going to be less likely to submit? Does it really hurt their reputation? Or is this much-coveted domination of the news cycle that is the kind of free advertising that corporations can only dream of?

I am not a whacko. I know that editors typically do everything within their power to avoid publishing bad science. In 99% of the cases, avoiding retractions is indeed the goal. I have written extensively about the problems of pre-publication peer review, and no journal is truly capable of ensuring that what is published is right. That’s a function of time. I also believe that the true problem with the glam journals is not so much the bad science they publish but the amazing science they reject.

Still, it bothers me to the core to see Nature play a victim here. Reading this NYT artice, it seems that Obokata is the convenient scapegoat that everyone is happy to destroy. I don’t know anything about Obokata. I have no inside information on this scandal. But I do know that many top scientists have a particular distaste for Nature. Over a decade, I have heard again and again the meme:

Nature sets up its authors to fail. They publish work they know they shouldn’t, then call for an investigation, and then publish the refutation.

Are Cell and Science any different? I doubt. When Science published the Arsenic paper, it sent it to reviewers that were likely to accept. And after enjoying the extravagant claims and press, Science published two papers refuting the original claim. Sweet. Here's what Bruce Alberts, editor-in-chief of Science, had to say about this:
We hope that the study and the subsequent exchange being published today will stimulate further experiments — whether they support or overturn this conclusion. In either case, the overall result will advance our knowledge about conditions that support life, an important outcome for science and education

Science, Nature, and Cell have the highest retraction rates of life sciences journals. And yet there is no indication that it hurts them the least bit. While publishing bad science doesn’t seem to hurt them, it does hurt the scientists. It hurts science. Crucifying Obokata is easy. But that does nothing to help the science enterprise. What would help is if we scientists stopped sending our work to the glam journals.


  1. Retractions do hurt a journal’s brand, but not much. That's the power of a strong brand: they help you buffer mistakes. Apple survived the Motorola Rockr. Microsoft survived the Zune. And so on.

    Given recent research showing that publishing in high impact factor journals significantly affects career prospects (at least in biomedicine; http://dx.doi.org/10.1016%2Fj.cub.2014.04.039), asking for people not to submit to those journals is just a wish. The incentives for people to submit are too strong.

  2. There is a rising of science publishing and share data websites where fellow scientists rate each others work and share - is this not more democratic and egalitarian than the current tiny circle of hi level science publications that charge an arm and a leg and are privy to high level, biased tenured faculty academic politics? Just Saying!