Opinion: Citation mutation, revisited

The discussion surrounding article references that acquire heritable mistakes continues

By | September 23, 2010

__This letter responds to a recent linkurl:post;http://blogs.plos.org/mfenner/2010/09/20/letter-to-the-scientist/ on the __PLoS__-hosted __Gobbledygook__ blog in which blogger and cancer researcher linkurl:Martin Fenner;http://www.mendeley.com/profiles/martin-fenner/ references a linkurl:story;http://www.the-scientist.com/news/display/57689/ on citation mutations that appeared on linkurl:www.the-scientist.com;http://www.the-scientist.com/ last week (16th September):__ Dear Martin Fenner,
I fully agree that this study is only an initial and somewhat anecdotal evaluation of citation mutants. My analysis has been limited to very few of the total number of citation variants of Laemmli's article. In fact, the exact number of variants could not be determined accurately using my approach. To obtain an estimate of the number of citation variants, I have searched the ISI database for the author's name, the journal and the year of publication (Laemmli, __Nature__, 1970). Any mutations located within these search terms have escaped detection. Such mutations can only be found by searching for a specific error. In this way I could, for instance, identify the variants Aemmli, __Science__ and 1907. As regards the stage at which mutations occur during the publication process, the distribution of the citation variants may offer some insights. For example, the page number mutation 680→777 has been limited to a few articles on closely related topics. Since these articles also cite one another, it is most likely that the authors have transmitted the mutation by copying the wrong citation from a previous article. In contrast, the mutation 680→681 appears more randomly across disciplinary boundaries, indicating that spontaneous typing errors are the primary cause. The incidence of this mutation has dropped from about 20 annual cases throughout the nineties to fewer than 10 in the last decade, possibly due to the wider use of reference management software. There are other examples in which the errors have clearly not been committed by the authors. For instance, the rate of the 680→685 mutation has risen sharply from fewer than 5 events per year during the early nineties to currently about 600 yearly events. In 2008 the mutations even exceeded the number of correct citations, according to the ISI database. Upon closer inspection, however, the majority of the listed articles do not carry the mutation, pointing to a systematic error at the database level. As to their phenotype, most citation variants are probably silent mutations due to the redundancy of the contained information. However, the fact that citation variants can be inherited may be an indication for a much bigger problem, which is that authors often do not read the publications cited in their work. I can only speculate how this would affect the accuracy of their interpretations. Kind regards, Christian Specht Christian G. Specht is a neurobiologist working on learning & memory and currently based at the ENS in Paris.
**__Related stories:__***linkurl:Opinion: Mutations of citations;http://www.the-scientist.com/news/display/57689/
[16th September 2010]*linkurl:Online access = more citations;http://www.the-scientist.com/blog/display/55437/
[19th February 2009]*linkurl:More articles, fewer citations;http://www.the-scientist.com/blog/display/54839/
[18th July 2008]


Avatar of: Katherine McCain

Katherine McCain

Posts: 1

September 23, 2010

The author commented that "Upon closer inspection, however, the majority of the listed articles do not carry the mutation, pointing to a systematic error at the database level." I have observed this in action with regard to citations to John Maynard Smith's book Evolution and the Theory of Games. My understanding is that TR has an automated citation checker that compares the OCRed citations in newly processed articles with "standard" citation strings that they have stored in a database -- and automatically "corrects" any mistakes caused by the OCR proces OR original author errors. This seems to work well most of the time, but in this case results in citations to JMS's book "corrected" from the appropriate--"MaynardSmith J, 1982,...." to "Smith JM, 1982..." even when the citing author had it correct. I notified ThomsonReuters of this some months ago. Not only is this a burr under my saddle (since it meant a great deal of manual data cleaning in my project) but, carried to an extreme, can affect authors' citation visibility (for those who care about "h indexes" and the like). \nRead more: Opinion: Citation mutation, revisited - The Scientist - Magazine of the Life Sciences http://www.the-scientist.com/news/display/57698/#ixzz10Nf4mvVO
Avatar of: Martin Fenner

Martin Fenner

Posts: 1

September 24, 2010

Thank your for the thoughtful response to my blog post. It puzzles me that so much depends on ciations (impact factors, grant decisions, etc.), but so little thought goes into why works were cited (or not cited). The mutation analysis is just the tip of the iceberg.
Avatar of: anonymous poster

anonymous poster

Posts: 4

September 24, 2010

Came across an identical review article in 2 diff journals, both from the same publisher (Elseiver) ...\n\nClin Lab Med. 2010 Dec;30(4):815-27. Genetics of psychiatric disorders methods: molecular approaches. Avramopoulos D. The practice of psychiatry has long suffered from the limited information available on the biological basis of mental disorders. This limitation is now coming to an end. Advances in DNA analysis technologies and in our understanding of the human genome, together with our new knowledge of the properties of the genome and significant efforts toward generating large patient and control sample collections, have paved the way for successful genome-wide association studies. As a result, reports now appear in the literature every week identifying new genes for complex disorders. Next-generation sequencing methods, combined with the results of association and perhaps linkage studies, will help us uncover missing heritability factors, achieve a better understanding of the genetic aspects of psychiatric disease, and devise the best strategies for incorporating genetics in the service of patients. PMID: 20832654\n\nPsychiatr Clin North Am. 2010 Mar;33(1):1-13. Genetics of psychiatric disorders methods: molecular approaches. Avramopoulos D. (same abstract as above) PMID: 20159337
Avatar of: James Southern

James Southern

Posts: 1

October 24, 2010

I've read many papers where, after you have read the cited articles used to support the conclusions, you find that the abstract appears to support the conclusion but the actual article does not and may not even be about the research topic.\n\nI think this is more like viral contagion than inheritability.\n\nAs the old saw goes:\n\nWhen you share references with one person, you are sharing references with everyone that they shared references with...\n\nAnd its an epidemic!

Popular Now

  1. Opinion: Why I Published in a Predatory Journal
    News & Opinion Opinion: Why I Published in a Predatory Journal

    My “colleagues” and I at the fictitious Arthur Vandelay Urological Research Institute were surprised to find our bogus “uromycitisis” case report swiftly accepted, with only minor revisions requested.

  2. Consilience, Episode 3: Cancer, Obscured
  3. A History of Screening for Natural Products to Fight Cancer
  4. March for Science: Dispatches from Washington, DC