Page 73 - Becoming Metric Wise

63
Publishing in Scientific Journals

lead to one of the authors or an accomplice. When these researchers are
then approached to perform the review they of course accept and write a
glowing review (in reality the accomplice does). In 2017, Springer had to
retract 107 articles with faked reviews published in Tumor Biology, a jour-
nal included in the WoS and the Journal Citation Reports (Stigbrand,
2017), but these articles are not the only ones that were retracted for this
reason.

3.3.4 Retractions

According to the COPE, retraction should be reserved for publications
that are so seriously flawed (for whatever reason) that their findings or
conclusions should not be relied upon (Wager et al., 2010). Retraction is
then the measure taken to safeguard the integrity of the scientific litera-
ture. Retractions can result from honest mistakes, sloppy work, or incom-
petence. Yet they can also result from plagiarism and fraud.
According to Amos (2014) plagiarism accounts for about 10 17% of
all retractions, while duplicate publication accounts for another 14 17%.
Other articles in her study were retracted for reasons of human error, or
were not mentioned. She further observes that unethical publishing
practices cut across all nations.
Although retracted articles should play no role anymore in scientific
research and consequent publications, Neale et al. (2010) could provide a
categorization of the use of articles affected by scientific misconduct not
just retracted ones. They considered 102 articles from the biomedical lit-
erature officially identified as resulting (at least partially) from scientific
misconduct. Of these 102 articles, 86 were cited at least once after retrac-
tion. Next they performed a content analysis of a stratified random sam-
ple of size 603 among the 5393 articles citing these tainted articles. Only
17 (or 2.8%) of the 603 included a reference to the corrigendum, i.e.,
retraction, erratum, or comment. Luckily, it turned out that none of these
articles retracted as the result of misconduct had affected clinical practice.
Conrad (2015) worries about false discoveries (in physics, but his
thoughts apply to many other fields: how many possible steps that might
lead to cures for cancer have been announced?). Pressure to be first let
scientists announce “results” they, or another team, have not confirmed
yet. Conrad also note that making data public increases the risk that these
data are used by scientists who do not know how they were obtained or
are not familiar with restrictions of the instruments used to collect these