Slashdot Mirror
Key Researcher Agrees To Retract Disputed Stem Cell Papers
sciencehabit (1205606) writes "After several months of fiercely defending her discovery of a new, simple way to create pluripotent stem cells, Haruko Obokata of the RIKEN Center for Developmental Biology in Kobe, Japan, has agreed to retract the two Nature papers that reported her work. Satoru Kagaya, head of public relations for RIKEN, headquartered in Wako near Tokyo, confirmed press reports today that Obokata had finally agreed to retract both papers. He said the institute would be notifying Nature and that the decision to formally retract the papers would be up to the journal."
Another article I read about this mentioned that she confessed to fabricating "at least some results". Now, there are various reasons why a researcher would fabricate results, from the pressure to publish to just literally being evil, but in this case how would she ever expect to get away with it? This is not like a paper in my sub-field, which if I'm lucky five people will ever read. EVERYONE wants pluripotent stem cells, so of course a simple method to create them is going to be tested and replicated over and over and over.
Do you know understand publication and peer review?
Nature peer review means that the data and methodology looked good and rigorous.
If they laid about the data, or some methodology it's very hard to know that unless if is really obvious.
This is why publication is only the beginning of peer review. After publication other experts can look at it and try to reproduce the results. This is also why the most interesting papers are the second papers.
The Kruger Dunning explains most post on
If the methodology looked good and the data looked reasonable, it'd pass the initial round of peer review. They don't recreate the experiment as part of the editorial peer review, they just look for things that were overlooked or that don't make sense. It's up to other labs to reproduce the results and subsequently publish their own papers.
Occasionally living proof of the Ballmer peak.
And this is Nature fer chrissakes; not the Journal of Homeopathic Chiropractic Aroma Therapy and Crystal Meditation.
Sigh, I'll point out what gets said it every one of these topics:
The point of peer review isn't to uncover fraud. That's the job for follow-up studies (like we saw in this case).
The point of peer review is to catch things like logical flaws. Do the conclusions follow from the data? Are there any obvious problems in the experimental setup? Did they mess up any math anywhere?
Catching fraud usually requires that the experiment be repeated. The initial peer review process of the initial article doesn't (and generally can't) do that. The whole point of the initial article is to put the experiment and the result into the literature and the community consciousness, THEN have other scientists attempt to reproduce the results. The fraud will presumably be caught then, if it exists. It worked in this case.
Now, I'll give you that science has a problem right now where not enough follow-up studies are being done. This means that there is way too much fraud, especially in the life sciences. But this is actually a case where the process seems to have worked pretty well.
Again, this is Nature we're talking about. Every time we get one of these situations, the apologists start up with "but peer-review wasn't meant to find that...", and yet the journals themselves are always chest-thumping about how everything they publish is infallible because it was peer-reviewed, except when it isn't, and then it's not their fault. Peer-review is just a crutch. It imparts a false sense on confidence where there shouldn't be any.
Peer review is not a crutch. It is a necessary, but not sufficient check and balance. That said, the peer reviewers MAY have performed less than admirably (hey, it happens). The part that really has turned up under closer review is her methodology is awful. Peer reviewers tend to work along the assumption that the researcher knows what they are doing. That assumption appears to be incorrect (recall the first three letters of the word). Looking a detailed Materials and Methods is amazingly boring and often not even possible because editors don't want to 'waste' space in their precious journal having somebody detail where they got a reagent from or exactly how they (supposedly) did things.
There is an increasing trend to require authors to put such details in the paper. Typically in a web based supplement (so it doesn't waste space in the precious journal). This trend has started for precisely these reasons.
Nature is going to eat some deserved crow on this one. Fortunately, that is the time tested recipe for improvement.
Faster! Faster! Faster would be better!
I'll have you know that the Journal of Homeopathic Chiropractic Aroma Therapy and Crystal Meditation practices peer review to the highest scientific standards.
To increase potency of the review, 15 scientists are tasked with reviewing each article. One of their reviews is then sent to 10 000 experts, who review the review. This step is repeated a few times. Finally, the resulting review is sent to our chinese editor (who at times is too busy editing other prestigious publications like the American Chinese Traditional Medicine and Voodoo annals and thus delegates this job to his team of highly-trained monkeys) who decides whether to publish or not.
Of course, this review cannot be questioned, otherwise it will never accurately review anything, as the trust between Journal and reviewers is broken.