|
 |
|
EMBO reports 8, 9, 792–793 (2007)
doi:10.1038/sj.embor.7401053
Comment on the Correspondence by Cokol et al
Shi V. Liu
|
|
|
|
Shi V. Liu is Editor of Scientific Ethics
(http://im1.biz/).
e-mail: svl8se@im1.biz
|
|
|
|
The correspondence by Cokol et al (2007) confirms a previous finding that some journals with a high impact factor (IF) publish more retractions than those with a low IF (Liu, 2006f). Their explanation is that high-IF journals receive more scrutiny from the scientific community, and that therefore more publications are identified as flawed and subsequently retracted. Equally, they argue, low-IF journals probably receive less attention and are therefore likely to contain more articles that ought to be retracted, but instead go unnoticed. However, this assumption is wrong and the conclusions they draw for low-IF journals are not valid.
The model by Cokol and co-workers reflects a common view that a high IF means high quality. But the high IF for some journals is actually based—at least in part—on the high number of citations of their retracted papers (Liu, 2005a, 2007a). Rather than removing these 'negative contributions' from the IF calculation, these journals have continued to use their inflated IFs to promote their publications (Liu, 2006b).
I agree that more publications should be retracted, but I argue that most of these retractions would still come from high-IF journals. First, the high visibility of high-IF journals does not mean more (post-publication) scrutiny. Many scientists have lost their vigilance against flawed and fraudulent publications, in particular in high-IF journals. Their preference for citing articles published in high-IF journals not only deprives ground-breaking studies published in low-IF journals of the attention and respect that they deserve, but also artificially elevates the IF of some journals (Liu, 2005c, 2006h, 2007d,f).
Second, retractions in high-IF journals are often not a direct result of scientific scrutiny, but a side-effect of investigations into scientific misconduct (Liu, 2006e). The final retraction at the end of a misconduct investigation is often a reluctant action by these journals to save face, rather than the result of proactively identifying scientific mistakes in these publications (Liu, 2005b).
Third, 'top' journals often offer much stronger protection for their publications. They use the argument of 'limited space' as an excuse for refusing to publish criticisms of flawed publications (Liu, 2006a,c). Some journals do not even have a dedicated section or any other means of accepting and publishing criticism, or they allow the authors of criticized publications to decide whether corrections should be published (Liu, 2005e, 2007c,g,h).
Thus, even if we were to use standard criteria to identify and retract flawed publications, some high-IF journals would still have the largest number of retractions. The reason for the higher incidence of 'error-prone' papers published by these journals is not due to high post-publication scrutiny, but rather to lower standards of the pre-publication review (Liu, 2005d). Their publishing philosophy of following trends and seeking maximal 'eyeball effect' on 'hot' topics ultimately leads to more retractions (Liu, 2006g). Indeed, the events associated with retractions in some high-IF journals show that pre-publication scrutiny—at least for those publications retracted—was often sloppy (Liu, 2006d,e,i, 2007b). As a scientist, I have to give the same scrutiny to any publication, regardless of where it is published. It is therefore wrong to conclude that low-IF journals have not received adequate scrutiny and thus not retracted enough articles.
Cokol et al characterize the publishing process as 'stochastic'. Scientific research is aimed at finding the truth and there should be only one truth for a specific phenomenon or observation. Scientific publications should not reflect random guesses in submitted articles that are randomly selected for publication. Any observation or conclusion accepted for publication reflects new knowledge and insights, and thus cannot be an outcome of any stochastic process.
Similarly, retractions are also not stochastic events. It is true that scientific discovery is not a straightforward process and truth is often fleeting. However, many flawed publications can still be identified by evaluating the appropriateness of their methodology and the soundness of their reasoning (Liu, 2000, 2007e). Treating the publication of fabricated studies as 'stochastic' events in 'highly scrutinized' and 'space-limited' journals with high IFs might reduce the responsibility of these journals and degrades them to lottery machines.
Furthermore, citation rates are not uniformly distributed among scientific disciplines, which is a major reason for the different IF values of journals in different disciplines. Even for a general journal, the citation and retraction rates vary greatly between disciplines (Liu, 2007a). Thus, the argument for a positive correlation between the IF of a journal and its post-publication scrutiny level does not reflect the reality.
It is therefore not valid to assume that there should be a uniform retraction rate among disciplines and journals. It is also unfair to conclude that high-IF journals have over-retracted and low-IF journals have under-retracted. The outcome of a model depends on both its logical assumptions and its data input. The model by Cokol et al is wrong because it is based on misleading assumptions and uses flawed impact factors.
|
|
References
Cokol M, Iossifov I, Rodriguez-Esteban R, Rzhesky A (2007) How many scientific papers should be retracted? EMBO Rep 8: 422–423 | Article | PubMed | ISI | ChemPort |
Liu SV (2000) Logical fallacies and methodological mistakes in microbiology. Logical Biol 1: 25–31
Liu SV (2005a) Retracted papers: How to curtail their impact? Logical Biol 5: 128–130
Liu SV (2005b) Retractions should focus on the content rather than the conduct. Logical Biol 5: 272–273
Liu SV (2005c) Barking at the wrong tree. Logical Biol 5: 73–75
Liu SV (2005d) Chasing trends and pressing hot buttons: a typical case of top journals' low scientific standard. Logical Biol 5: 235–237
Liu SV (2005e) A public robbery of science in the public library of science. Logical Biol 5: 76–78
Liu SV (2006a) Cheating should perish from scientific literature. Logical Biol 6: 90–92
Liu SV (2006b) The impact factor scam in scientific publishing. Sci Ethics 1: 72–73
Liu SV (2006c) Limitation of space or limitation on views. Sci Ethics 1: 67–71
Liu SV (2006d) Schatten's misbehavior is more disgusting than Hwang's misconduct. Sci Ethics 1: 30–32
Liu SV (2006e) A slap on the face of "Science". Sci Ethics 1: 3–7
Liu SV (2006f) Top journals' top retraction rates. Sci Ethics 1: 91–93
Liu SV (2006g) Why are "top" journals often so unlucky in infecting top scandals? Sci Ethics 1: 49–52
Liu SV (2006h) Citation should credit pioneering researchers and original works. Sci Ethics 1: 94–96
Liu SV (2006i) Lessons not learnt or not revealed. Sci Ethics 1: 24–27
Liu SV (2007a) Hwang's retracted publication still contributes to Science's impact factor. Sci Ethics 2: 44–45
Liu SV (2007b) Nilsson and Science owe public more answers than a simple retraction. Sci Ethics 2: 41–43
Liu SV (2007c) PNAS: A permanent niche for aggressive sin. Sci Ethics 2: 7–20
Liu SV (2007d) Respecting published answers for important questions on epigenetics. Top Watch 2: 19–20
Liu SV (2007e) Logical review casts doubts on several publications in top journals. Top Watch 2: 23–24
Liu SV (2007f) Liu's message blocked by 'top' journals. Sci Ethics 2: 5–6
Liu SV (2007g) The misconduct of 'top' journals in exposing scientific misconduct. Sci Ethics 1: 17–19
Liu SV (2007h) Nature Genetics: a natural guard for flawed publication? Sci Ethics 2: 21–35
|
|
top  |
|
|
|
