I agree that peer review is broken in the sense that it does not provide that expert and impartial evaluation that it is supposed to provide. With increasing specialization, it becomes almost impossible to find expert reviewers who do not have some sort of connection to the study in question. Complicating this further is the development of online journals who bypass peer review altogether.
What we may be seeing is a shift from peer review as a permanent absolute evaluation of the quality of a study to a provisional advisory to readers that may change as more people have an opportunity to weigh in. On a somewhat related front: Given the 'black drawer' phenomenon of pharmaceutical reviews (a strong publication bias towards positive results) I avoid when possible taking any medication that hasn't been on the market for at least ten years, allowing time for reports of side effects to accumulate. On Jan 16, 2014, at 11:27 AM, Mike Palij wrote: > On Thu, 16 Jan 2014 08:43:34 -0800, Paul Brandon wrote: > >Or maybe a really good situation, depending upon why the articles > >were retracted. > > Allow me to retort: > > (1) One thing this pattern implies is that peer-review is broken > in some significant way. It is not catching those articles that > should not be published (to echo Paul's point, this depends > upon the reason why it was retracted and I'm assuming that > some serious error exists with the research and/or fraud and/or > other unethical conduct has caused the retraction). > > (2) For fans of meta-analysis, if one "captures" a study early and > does not realize that it has been retracted, then their meta-analysis > will be biased. Just in the 200 articles I looked at, there were > several that came from the Cochrane database; for example: > > WITHDRAWN: Herbal medicines for advanced colorectal cancer. > Guo Z, Jia X, Liu JP, Liao J, Yang Y. > Cochrane Database Syst Rev. 2012 Sep 12;9:CD004653. doi: > 10.1002/14651858.CD004653.pub3. Review. No abstract available. > PMID:22972073 [PubMed - indexed for MEDLINE] > One wonders what happens to published meta-analysis if one or > more studies are retracted? Who changes the paper copies? > > (3) I am amazed at there are several articles that are published > "online/Epub" before coming out in paper version. Consider: > WITHDRAWN: Blm10-proteasomes antagonize mitochondrial > fission through degradation of Dnm1. > Tar K, Dange T, Yang C, Yao Y, Bulteau AL, Fernandez Salcedo E, > Braigen S, Bouillaud F, Finley D, Schmidt M. > J Biol Chem. 2014 Jan 8. [Epub ahead of print] > PMID:24285543 [PubMed - as supplied by publisher] Free Article > I chose this article because PubMed made it freely available as > an Epub (the "Free Article" identifies this; clicking on it now goes > to the publisher's website which indicates that the article has been > retracted but how many people got copies of the article before it > was retracted?). > > The "good" aspect is that these research articles were caught but it > does raise question of what percentage are "misses". > > I think there's trouble in River City. > > -Mike Palij > New York University > [email protected] > > > > On Jan 16, 2014, at 10:18 AM, Mike Palij wrote: > > > While looking through www.researchgate.net, I came across an > > article reference that began with the word "WITHDRAWN". > > I interpreted this to mean that the article had been retracted from > > the journal it had been published in for whatever reason (following > > through to the publishers website, there was a vaguely worded > > statement that the author had withdrawn the article consistent > > with some of the publisher's legalese points). I thought that this > > was a bit odd (e.g., that it was on researchgate which exists > > primarily as a self-promotional platform) and decided to go over > > to Medline/PubMed to see if there were any other article with > > this designation (the article was in a biomedical research area). > > I searched for "WITHDRAWN" in the title of the article and as > > of about 10 minutes ago, there were 1765 hits. PubMed does > > not provide stable URL for searches like this, so, the interested > > reader is encouraged to go to www.pubmed.gov and do the > > search themselves. Now, it is true that only those articles with > > WITHDRAWN at the beginning of the title appear to be > > retractions and there are some article titles that have withdrawn > > as a legitimate part of the title. Nonetheless, I set the number > > of articles per page to 200/page and it is clear that 80-90% of > > the articles are retractions (some have "article withdrawn" instead). > > > > The truly weird part is that in the first 200 articles listed (sorted > > by recently added) the oldest article is from August 2012 -- > > the rest of the articles are retractions from 2012-2013. > > > > WTF!?! > > > > Am I overreacting to this or is this a really, really bad situation? Paul Brandon Emeritus Professor of Psychology Minnesota State University, Mankato [email protected] --- You are currently subscribed to tips as: [email protected]. To unsubscribe click here: http://fsulist.frostburg.edu/u?id=13090.68da6e6e5325aa33287ff385b70df5d5&n=T&l=tips&o=32805 or send a blank email to leave-32805-13090.68da6e6e5325aa33287ff385b70df...@fsulist.frostburg.edu
