Hundreds of scientists publish a paper each 5 days
Authorship is the coin of scholarship — and a few researchers are minting quite a bit. We searched Scopus for authors who had printed greater than 72 papers (the equal of 1 paper each 5 days) in anybody calendar 12 months between 2000 and 2016, a determine that many would contemplate implausibly prolific1. We discovered greater than 9,000 people, and made each effort to rely solely ‘full papers’ — articles, convention papers, substantive feedback and critiques — not editorials, letters to the editor and the like. We hoped that this may very well be a helpful train in understanding what scientific authorship means.
We have to be clear: we’ve no proof that these authors are doing something inappropriate. Some scientists who’re members of huge consortia may meet the factors for authorship on a really excessive quantity of papers. Our findings counsel that some fields or analysis groups have operationalized their very own definitions of what authorship means.
The overwhelming majority of hyperprolific authors (7,888 writer data, 86%) printed in physics. In high-energy and particle physics, initiatives are completed by massive worldwide groups that may have upwards of 1,000 members. All members are listed as authors as a mark of membership of the staff, not for writing or revising the papers. We subsequently excluded authors in physics.
Who gets credit? Survey digs into the thorny question of authorship
Of what remained, 909 writer data have been Chinese language or Korean names. As a result of Scopus disambiguates Chinese language and Korean names imperfectly, these might have wrongly mixed distinct people. For 2016 (when disambiguation had improved for Chinese language and Korean names), a minimum of 12, and probably greater than 20, authors based mostly in China have been hyperprolific, the most important quantity from any nation that 12 months. We imagine that this may very well be related to Chinese language insurance policies that reward publication with money or to attainable corruption2,3.
Due to the disambiguation points, we excluded these names from additional evaluation, in addition to group names and instances through which we discovered errors (reminiscent of journalistic information objects misclassified as full articles), duplicate entries, or convention papers misassigned to an organizer.
This left 265 authors (see Supplementary Info). The variety of hyperprolific authors (after our exclusions) grew about 20-fold between 2001 and 2014, after which levelled off (see ‘Hyperprolific authors proliferate’). Over the identical interval, the overall variety of authors elevated by 2.5-fold.
We e-mailed all 265 authors asking for his or her insights about how they reached this extraordinarily productive class. The 81 replies are supplied within the Supplementary Info. Widespread themes have been: onerous work; love of analysis; mentorship of very many younger researchers; management of a analysis staff, and even of many groups; intensive collaboration; engaged on a number of analysis areas or in core providers; availability of appropriate intensive sources and information; end result of a giant venture; private values reminiscent of generosity and sharing; experiences rising up; and sleeping just a few hours per day.
Supply: J. P. A. Ioannidis, R. Klavans & Okay. W. Boyack
About half of the hyperprolific authors have been in medical and life sciences (drugs n = 101, well being sciences n = 11, mind n = 17, biology n = 6, infectious ailments n = 3). Once we excluded convention papers, virtually two-thirds belonged to medical and life sciences (86/131). Among the many 265, 154 authors produced greater than the equal of 1 paper each 5 days for two or extra calendar years; 69 did so for 4 or extra calendar years. Papers with 10–100 authors are widespread in these CVs, particularly in medical and life sciences, however papers with the tons of of authors seen in particle physics are unusual.
Health tips for research groups
Supplies scientist Akihisa Inoue, former president of Tohoku College in Japan and a member of a number of prestigious academies, holds the file. He met our definition of being hyperprolific for 12 calendar years between 2000 and 2016. Since 1976, his identify seems on 2,566 full papers listed in Scopus. He has additionally retracted seven papers discovered to be self-duplications of beforehand printed work4. We looked for information articles in Google detailing retractions for the subsequent 20 most hyperprolific authors and located just one different writer (Jeroen Bax) to have one retracted paper.
The 265 hyperprolific authors labored in 37 international locations, with the very best quantity in america (n = 50), adopted by Germany (n = 28) and Japan (n = 27). The proportion from america (19%) is roughly much like its share of printed science. Germany and Japan are over-represented. There have been disproportionally extra hyperprolific authors in Malaysia (n = 13) and Saudi Arabia (n = 7), international locations each identified to incentivize publication with money rewards5.
Hyperprolific authors additionally tended to cluster particularly establishments, typically as a part of a standard research. For instance, Erasmus College Rotterdam within the Netherlands had 9 hyperprolific authors, greater than some other establishment. Seven of them co-authored principally papers associated to the Rotterdam research, a virtually 30-year-old epidemiological venture, or its successor Era R research, which have adopted a number of well being parameters in hundreds of older adults and yielded hundreds of publications. 5 hyperprolific investigators from Harvard College in Cambridge, Massachusetts, additionally typically co-authored papers associated to cohort research. Eleven hyperprolific authors throughout totally different establishments have been on one massive cohort research, the European Potential Investigation on Most cancers and Diet; different massive epidemiological research have been additionally represented. Hyperprolific authors have been additionally concentrated in cardiology and crystallography.
These organic and medical disciplines with many hyperprolific authors exhibit totally different patterns from these present in particle and high-energy physics. Papers with tons of to hundreds of authors are the norm throughout a group of many hundreds of scientists working in initiatives based mostly at CERN, Europe’s particle-physics laboratory close to Geneva, Switzerland. In crystallography, papers are likely to have few co-authors. In epidemiology and cardiology, lengthy lists of authors seem solely in relationship to particular analysis groups that appear to have a convention of intensive authorship lists.
Kid co-authors in South Korea spur government probe
This raises the query of what authorship entails. The US Nationwide Institutes of Well being, for instance, has tips on the actions that qualify: actively supervising, designing and doing experiments, and information acquisition and evaluation exterior “very fundamental” work plus drafting the manuscript. Amassing funds or distant mentorship don’t qualify. A lot of the 6,000 authors in a current survey throughout many geographical areas and disciplines felt that drafting a paper, decoding outcomes and analysing information ought to qualify for authorship, however attitudes diverse by area and area6.
Authorship standards
Maybe probably the most extensively established necessities for authorship are the Vancouver criteria established by the Worldwide Committee of Medical Journal Editors in 1988. These specify that authors should do all of 4 issues to qualify: play a component in designing or conducting experiments or processing outcomes; assist to jot down or revise the manuscript; approve the printed model; and take duty for the article’s contents.
The Worldwide Committee of Medical Journal Editors doesn’t rely supervision, mentoring or acquiring funding as enough for authorship. We did observe that some authors appeared to turn into hyperprolific on turning into full professors, division chairs or each. It is not uncommon and maybe anticipated for scientists who assume management roles in massive centres to speed up their productiveness. For instance, scientific cardiologists publish extra papers after they assume director roles (regardless of heavy scientific and administrative duties). Sometimes, the acceleration is beautiful: on the peak of their productiveness, some cardiologists publish 10 to 80 occasions extra papers in a single 12 months in contrast with their common annual productiveness after they have been 35–42 years outdated. There was additionally typically a pointy lower after passing the chair to a successor. One other research famous comparable patterns 20 years in the past7.
One surprising consequence was that some hyperprolific authors positioned many publications in a single journal. Outstanding on this regard have been Acta Crystallographica Part E: Construction Stories On-line (relaunched in 2014 as Part E: Crystallographic Communications, with transient structural information stories now printed in IuCrData) and Zeitschrift für Kristallographie New Crystal Buildings. Three authors have every printed greater than 600 articles within the former (Hoong-Kun Enjoyable, Seik Weng Ng and Edward Tiekink); three authors have every printed greater than 400 papers within the latter (Karl Peters, Eva Maria Peters and Edward Tiekink). Three different authors (Anne Marie Api, Charlene Letizia, Sneha Bhatia) printed many papers in single complement problems with Meals and Chemistry Toxicology centered on critiques of perfume supplies.
Journals listed in Scopus are usually thought of to be high quality journals. The quotation influence of hyperprolific authors was normally excessive, however there was massive variability: with a median of 19,805 citations per writer (vary: 380 to 200,439). The median variety of full papers per hyperprolific writer in 2000–2016 was 677; throughout all hyperprolific authors, final writer positions accounted for 42.5%, first writer positions for 7.1%, and single authorships for 1.4%. Throughout the years, the median proportion of papers with center writer positions (that’s, not a single, first or final writer) was 51%, however diverse from 2.1% to 98.5% for particular person authors.
Our work to determine hyperprolific authors is admittedly crude. It’s primarily supposed to boost the bigger query of what authorship entails. Whether or not and the way authorship is justified unavoidably varies for every writer and every paper, and norms differ by area. It’s possible that generally authorship will be gamed, secured via coercion or supplied as a favour. We couldn’t assess these patterns in our information. We didn’t study contributorship statements8, which aren’t archived in Scopus. Nonetheless, even contributorship statements will be gamed and may not be correct.
Additional work is required to discover find out how to finest normalize these information and what’s the optimum stage of normalization: for instance, adjusting for vast self-discipline, comparatively slender area and/or extremely particular analysis staff.
What authors say
To higher perceive authorship norms, we e-mailed a survey to the 81 hyperprolific authors of 2016 (see Supplementary Info). We requested whether or not they fulfilled all 4 Vancouver standards. Of the 27 who accomplished the survey, most stated they didn’t (see ‘Survey’). Nearly all of the responders have been from US and European establishments. The one two responders from elsewhere said that they failed Vancouver standards in most of their papers. It’s possible that the survey underestimates the proportion not assembly Vancouver standards.
Not all authors had accepted the ultimate variations of their very own papers, however all thought of approval of the ultimate model obligatory for authorship. Fifty-nine per cent (16 of 27) stated that they’d contributed greater than some other listed writer for 25 or extra of the papers they authored in 2016.
Peer review: Troubled from the start
Responses to the query “What, in your individual phrases, do you assume ought to be required for authorship?” usually mirrored a requirement for “important contributions”, but in addition dissatisfaction with how authorship was assessed. One scientist stated, “I personally don’t rely them as ‘my papers’ and don’t have them on my CV as such, as there’s a distinction between being a ‘named writer’ versus a ‘consortium member’ authorship.” One other noticed that authorship was typically awarded for seniority, and one other that higher distinctions have been important. “I feel there ought to be ranges of authorship — and never these implied by order!” It will likely be fascinating to watch how improvements in assigning credit score, reminiscent of data citation or formal writer contribution taxonomies, may alter authorship conventions. Authorship norms can differ inside every area and even inside every staff. For instance, some groups in epidemiology and cardiology apparently provide authorship extra generously; others follow stricter (and possibly extra acceptable) authorship standards. For the same process and contribution, one cohort research would possibly credit score 20 authors, one other would possibly give credit score solely to three individuals or none. For instance, genome-wide research usually embrace many dozens of authors. As a dramatic counter-example, one current publication of a genome-wide research had just one writer9, and apparently that researcher did the identical quantity of labor for which maybe dozens would get authorship credit score in comparable papers spearheaded by totally different groups. Some proof means that the rise within the common variety of authors per paper doesn’t mirror a lot the real wants of staff science because the strain to ‘publish or perish’10.
Extensively used quotation and influence metrics ought to be adjusted accordingly. As an example, if including extra authors diminished the credit score every writer acquired, unwarranted multi-authorship would possibly go down. We discovered that the 30 hyperprolific authors who appeared to profit probably the most from co-authorship numbered 6 cardiologists and 24 epidemiologists (together with these engaged on inhabitants genetics research). (For these scientists, the ratio of their Hirsch H index to their co-authorship-adjusted Schreiber Hm index was greater; see Supplementary Info.)
General, hyperprolific authors would possibly embrace a number of the most energetic and wonderful scientists. Nonetheless, such modes of publishing may also mirror idiosyncratic area norms, to say the least. Free definitions of authorship, and an unlucky tendency to cut back assessments to counting papers, muddy how credit score is assigned. One nonetheless must see the overall publishing output of every scientist, benchmarked in opposition to norms for his or her area. And naturally, there isn’t any substitute for studying the papers and making an attempt to know what the authors have completed.