Source: Global Science Scientific American
Over the past decade, the number of annual retractions in major journals around the world has increased tenfold. Among the reasons for retraction, academic fraud and academic misconduct accounted for about 60%. China has made a huge “contribution” to the world's largest retraction publishing unit, IEEE. Recently, Science magazine conducted an unprecedented and in-depth analysis of historical retraction events. Can this be the beginning of strict governance in the scientific research community?
Since 2000, the number of retractions in global reports has skyrocketed, and “retraction tides” have caused many intense comments from peers in the scientific community. More and more voices are calling for magazine publishers, magazine editors, and reviewers. Before the publication of the journal, people must carry out more severe checks and remove the articles that are false and unfavorable to scientific progress. This call also led to the birth of a website called “Retraction Watch”, which was founded by two journalists Ivan Oransky and Adam Marcus who have been working on health direction for a long time. Their purpose is very clear, in order to sort out the specific number of retractions and sources, and conduct a deep analysis of the causes.
The site has now built a list of retracted articles and was published in the search database last week (retractiondatabase.org). The retraction observation database is the most complete and in-depth retraction database available today, including more than 18,000 retracted articles and conference abstracts since the 1970s, and even a 1756 article.
The database can find the corresponding retraction file by inputting the corresponding content. (In the figure, search "ldquo; China" as an example). Image source: retractiondatabase.org
At the same time, the retraction data was supported by the international top journal Science, and an in-depth analysis of 10,500 retracted documents was carried out. The data shows that the number of retracted documents has increased year by year, but the retraction accounts for all published documents. The ratio is actually maintained at a low level; in addition to fraud, some authors have taken the initiative to apply for retraction due to their own errata; the report also reflects the phenomena and problems that are worth reflecting now.
Increased number of retractions and stable relative proportion
In the past decade or so, the number of annual retractions has been rising. In 2000, the average annual number was 100, and in 2014, it was an average of 1,000. However, the proportion of retraction is relatively low, and it is currently maintained at around 4/10,000. From 2003 to 2009, the retraction rate doubled and remained basically unchanged in 2012. Logificantly, the retraction rate should continue to rise in the increasingly rigorous literature review. Therefore, the retraction rate is still unchanged. It also reflects a phenomenon: the number of academic articles published from 2003 to 2016 may have increased. More than twice.
The trend of retraction rates in this century.
The database shows that about 60% of the articles involved in the retraction involve academic misconduct. Most of the 60% involve behaviors that the US government clearly defines as academic misconduct, including falsification of data, tampering with data, and plagiarism. What's interesting is that since 2009, there has been a new phenomenon of “peer review fraud” and it is showing an upward trend year by year. This phenomenon is mainly because the researcher provides the journal with a false peer review mailbox, allowing others to review their own articles. At the same time, a small part of it mainly involves academic ethics and reputation damage. Even if the government does not stipulate, the academic community still classifies it as 60% of academic misconduct.
The orange part of the picture above is academic misconduct, and blue is an academic error. The following picture shows the type of fraud, plagiarism (dark), unreal (middle), peer fraud (light). Image source: (GRAPHIC) J. YOU/SCIENCE; (DATA) RETRACTION WATCH AND NSF
Retraction of journals
Overall, academic journals covering retraction continued to grow, with only 44 academic journals in 1997. By 2016, the number has grown to ten times, reaching 488. Some of these journals have been retracted more than one article per year.
However, this is not entirely a bad thing. The increase in the number of retracted journals indicates that the publishing house is paying attention to the role of academic publishing. Nicholas Steneck, an academic ethics expert at the University of Michigan, said: “The increase in the number of retractions indicates that the editorial review capacity is improving, and the recollection of journal articles will also bring some shock to those who attempt to falsify. ”
In 2009, the International Publishing Ethics Committee provided a retraction guide to more than 12,000 journal editors and researchers around the world, which listed the criteria for how to decide whether to withdraw the draft. The following related actions can be used to perform the retraction: Unreliable findings (falsification of data), data distortion (calculation errors and experimental errors); data reuse (unreasonable use of data published in other journals); plagiarism; non-ethical research.
A publishing house retracted more than 7,000 articles
According to the retraction observation database, 40% of the retracted papers were from the same publishing journal, the Institute of Electrical and Electronics Engineers (IEEE), and all of them were counted as 7,300. It peaked between 2009 and 2011. The publisher mainly publishes abstracts of conference papers. For example, the 2011 International Conference on E-Commerce and E-Government in Shanghai withdrew more than 1,200 abstracts, most of which were from China, including physical sciences, business, technology and social sciences. For many articles that have been removed, the reason for the retraction given by the IEEE is always ambiguous. Usually after careful review, we found that the paper violated the publishing principle. & rdquo; for the reasons to smash the past.
IEEE chart of drafts for each year. Image source: Science
Biology research prevails picture fraud
In 2016, a statistical study focusing on image fraud analyzed a total of 20,621 academic papers from 1995 to 2014. Among them, 3.8% (770 articles) of data charts contain certain problems, and half of them involve image modification behavior. The study pointed out that in the past decade, the situation of being defined as academic misconduct due to picture problems has become more and more serious. For biological research, there are as many as seven sections in the research report for case analysis of different types of fake pictures. And the report pointed out that due to the limitations of our analysis of data types, the actual situation may be more serious. ”
The type of bio-image fraud, shown here is reuse. Image Source: The Prevalence of Inappropriate Image Duplication in Biomedical Research Publications
Individual retraction leaderboard
Despite the surge in fraud, 500 of the authors who retracted more than 30,000 databases contributed more than a quarter of the retractions. The number of retractions by 100 authors has reached more than 13 articles. The reason for their retraction is generally not an accidental error, but an academic misconduct.
Image source: Science
At the top of the retraction list, of course, the Japanese anesthesiologist Yoshitaka Fuji, which we reported before, has 169 articles. The following 2-9 were respectively (the names of the listed persons are all units before the retraction, and have now resigned):
German anesthesiologist Joachim · Joachim Boldt, 96;
Dutch psychology professor Diederik Stapel, 58 articles;
Peter Chen, a professor at Pingtung University of Education in Taiwan, China, 43 papers;
Yoshihiro Sato, Professor of Hirosaki University, Japan, 43 articles;
Hua Zhong, a lecturer at Jinggangshan University in China, 41 articles;
Shigeaki Kato, Professor of Biology, University of Tokyo, Japan, 39 articles;
James & Middot, Professor of Bentley University, USA, 36 articles;
Hyung-In Moon, Assistant Professor, East Asia University, Korea, 35 papers;
Hendrik Schön, a physicist at Bell Labs in Germany, has 32 papers.
In December 2009, the official website of the international academic journal "Acta Crystallographica. Section E, Structure Reports Online" published a report indicating that the lecturer and master's degree of the School of Chemistry and Chemical Engineering, Jinggangshan University Liu Tao, a lecturer and master of Huahe Institute of Technology, is the responsible person. The 70 papers published in the journal have fraudulent phenomena and make a one-time withdrawal decision. In the same year, Jinggangshan University made a decision on Zhonghua to hire and recover the prize money.
National retraction list
In absolute terms, the top two countries at the top of the retraction are undoubtedly the United States and China. However, analysts point out that it is unfair to directly rank in absolute terms, because the two countries invest heavily in research and will increase the number of retracted volumes in disguise. Therefore, it is more fair to invest in the corresponding number of retractions per dollar unit. After the factors were constrained, the rankings changed: Iran ranked first in the list, and 14 out of every 10,000 manuscripts were withdrawn; China ranked seventh, with a retraction rate of five ten thousandths.
Image source: Science
In these retractions, many studies may only lead to the end of a personal scientific career because of fraud, but there are still some, such as medical data fraud, which is likely to put patients who have already been in thin ice into a more dangerous situation. Once the data is published and adopted. The victim may rise to a group. Joachim · Bolt, a great scholar of anesthesia fraud, commented in the past: "The research he participated in may be suspected of data fraud and ethical issues." & rdquo; But only two of the 98 collaborative articles that he has participated as co-authors have been retracted.
These studies involve the use of hydroxyethyl starch (hetastarch) to control blood pressure in surgery and traumatic recovery, which has been used by hospitals since the 1960s. Although many scholars have said that it is not safe, it will cause kidney damage and death, but Bolt spares no effort to prove that it is safe by making false claims and publishing false documents. Therefore, this method has been used for a long time, and potential damage to patients has been impossible to measure. Christian Wiedermann, a nursing expert at the Australian Private University Health Institute, said, “Unfortunately, logically, patients will undoubtedly suffer more damage from Bolt. ”