CAREERS

POSTGRADUATES US university offers career- CITATIONS Publisher links citation rates to NATUREJOBS For the latest career
planning course for credit p.593 scientific peer recognition p.593 listings and advice www.naturejobs.com

perfect approach, but deciding on who gets an

GETTY

authorship credit, and how they are ranked, is a

crucial part of doing science responsibly.
Precise statistics on authorship disputes
are hard to come by, says Mario Biagioli, a
science historian at the University of Cali-
fornia, Davis, who has studied authorship.
Scientists may be reluctant to admit that they
have demanded undeserved authorship or
otherwise subverted the system, and the US
Office of Research Integrity does not track
such disagreements because they are not con-
sidered scientific misconduct, says Biagioli,
who co-edited the book Scientific Authorship:
Credit and Intellectual Property in Science
(Routledge, 2002). However, in a 2005 survey1
of researchers who had received a grant from
the US National Institutes of Health (NIH),
10% of respondents admitted to assigning
authorship “inappropriately”.

CREDIT CONFUSION
Questions of who deserves credit for a paper
are a fairly recent phenomenon, says Biagioli.
Once upon a time, a paper had one author,
maybe two. But with modern big science and
Bud Abbott and Lou Costello made disagreements about order in baseball the stuff of comedy legend. large collaborations, a study might have hun-
dreds or even thousands of authors — as in the
AUT HO RSHIP case of the ATLAS experiment2 at the Large

Who’s on first?
Hadron Collider at CERN, Europe’s particle-
physics laboratory near Geneva, Switzerland.
And what authorship means varies by
scientific discipline. For example, in par-
ticle physics, hundreds of researchers may
contribute to the development and mainte-
When scientists collaborate on an experiment and a paper, nance of a single piece of equipment, such
it can be hard to decide who gets the credit and how much. as an accelerator. At big physics labs such as
CERN, everyone who was working at the lab
when the discovery was made gets a slot on
BY AMBER DANCE important than what the other person did,” the author list — even if they haven’t seen the
says Kosslyn. paper, says Biagioli. The authors are usually

S
tephen Kosslyn first started to consider Such disputes are common. “As authorship listed alphabetically, regardless of how much
how author lists come together when he is our academic currency, it tends to be a hot- they contributed.
found himself mediating a dispute. A button topic,” says Karen Peterson, scientific In the biological sciences, by contrast, the
postdoc and a graduate student each wanted ombudsman at the Fred Hutchinson Cancer author list is often strictly ranked. The top
to be listed as the first author on a study. Research Center in Seattle, Washington. She spot is at the end of the list, where the principal
“They both had a case,” recalls Kosslyn. “It got says that one-fifth of the disputes she adjudi- investigator gets credit for running the lab. The
heated.” cates concern authorship. Similar conflicts are student or postdoc who actually did the work
Disagreements often happen when con- among the most common issues mediated by goes first. As for the authors in the middle, it is
tributors put in similar amounts of effort on the Committee on Publication Ethics (COPE), hard to tell whether they participated a lot or a
different aspects of a project, says Kosslyn, a says Virginia Barbour, the organization’s chair- little, says Biagioli.
psychologist at Stanford University in Cali- woman and chief editor of PLoS Medicine in The International Committee of Medical
fornia. For example, one person might have Cambridge, UK. Journal Editors (ICMJE), headquartered in
developed the idea for the project and the other Authorship disagreements can be miti- Philadelphia, Pennsylvania, has developed
performed most of the data analysis. “The gated with careful discussions, explicit lab authorship guidelines that are used by many
force of the dispute usually revolves around guidelines and a good understanding of journals and institutions. These rules state that
the feeling that whatever they did was more authorship practices in one’s field. There is no to be listed as an author, each researcher

2 7 S E P T E M B E R 2 0 1 2 | VO L 4 8 9 | NAT U R E | 5 9 1
© 2012 Macmillan Publishers Limited. All rights reserved
 CAREERS

must meet three key criteria: they must they know little about (see ‘Ghosts and guests’).

G. TODD/FHCRC
have been involved in designing the project, Gerald Schatten, a stem-cell researcher at
collecting data or analysing the results; they the University of Pittsburgh in Pennsylvania,
must have participated in drafting or revising learned that lesson when he lent his good name
the manuscript; and they must have approved to a high-profile but eventually discredited
the final, published paper. Many universities stem-cell paper by Woo Suk Hwang, then at
that have their own guidelines base them on Seoul National University. Schatten was inves-
the ICMJE’s wording, says Biagioli. tigated by his university, which cleared him of
Kosslyn has his own definition: the crucial misconduct, but chastised him for ‘research
element, he says, is creativity. For example, a misbehaviour’ because he failed to check the
researcher could work with study participants quality of the science3.
in the lab, but just be following a protocol. The decision to accept courtesy authorship
“Anybody could have run the subjects, so run- is a matter of preference, says Sneden. “Per-
ning the subjects is not enough,” says Kosslyn. sonally, if I haven’t actually contributed some-
To earn authorship, the researcher would be thing to the specific paper, I just won’t have my
intellectually engaged: they might point out name on it,” he says. In that case, he politely
a feature of the data that leads the team to Ombudsman Karen Peterson says that one-fifth of tells his colleagues that he shouldn’t be on the
reshape the experiment. The paper wouldn’t the disputes she handles are about authorship. list. “I make sure they understand that it’s not a
look the same without them. negative reflection on the paper,” he says.
to be senior author. They invited two trusted
THE AUTHOR IN QUESTION colleagues to mediate. TAKEN IN VAIN
COPE recommends that researchers decide The jury awarded the senior slot to Grou- Sometimes, the recipient of this courtesy may
who will be an author and what order they will dine, but he felt uneasy about it. He suggested not get the chance to bow out. A researcher
be listed in before they even conduct experi- that the other investigator be the correspond- who has been added to the author list with-
ments, and that the group revisits the author ing author, who communicates with the jour- out their permission might be surprised to see
list as a project evolves. A handshake isn’t nal and any scientists who enquire about the their name when the paper comes out, says
enough to seal the deal — researchers should work. “I consider corresponding author as Sneden, or even angry if they don’t agree with
keep author agreements in writing. equivalent, almost, to senior author,” says the conclusions. Those who find themselves
Whenever they occur, authorship discus- Groudine. Co-senior authorship is also an unexpectedly an author on a paper that they
sions need not be confrontational (see ‘Aggra- option, he adds. would prefer not to be associated with should
vation-free authorship’). Mark Groudine, But sharing credit too broadly can be risky. contact the editor of the journal, he recom-
deputy director of the Hutchinson Center, says Sometimes authors are listed more as a cour- mends. The editor will get in touch with the
that the parties in a dispute should sit down tesy than because they made a key contribution, study’s corresponding author, and decide
and try to talk the matter over. “People get says Chris Sneden, an astronomer at the Uni- whether a corrigendum is necessary to explain
so locked into their positions that they don’t versity of Texas at Austin, who will step down that the author in question was not involved
make the effort to understand the other per- from his post as editor of The Astrophysical with the work.
son’s point of view,” he says, “and therefore they Journal Letters at the end of this year. Accepting These kinds of conflicts shouldn’t occur.
don’t understand why it’s a dispute.” courtesy authorship is a “double-edged sword”, Corresponding authors are expected to have
If talking doesn’t work, Groudine suggests he says. If the paper becomes famous, “every the approval of their co-authors — but some
asking the opinion of an unbiased third party. author gets to claim credit”. But if it becomes don’t realize it. “People, do you read the publi-
For example, on one project he collaborated infamous, everyone gets a share of the blame. cation agreement that you sign?” Sneden asks
with another principal investigator. When it Researchers need to be aware of the potential his colleagues. (Often, the answer is no.)
came to writing up the paper, both wanted risks of adding their names to manuscripts that Increasingly, journals are attempting to keep
authors in line by asking for details on who
did what. In cases of fraud, those descriptions
C O N F LI C T R ES O LU T I ON should lay the blame at the right person’s door.
Biagioli agrees that delineating each per-
Aggravation-free authorship son’s contribution should help, but he says
that the descriptions are frequently too brief.
When many scientists work together, ●●If a contributor’s authorship is in question, As an example, he cites the study published
determining authorship isn’t always easy. it can help to consider what the paper would this month in Nature by the ENCODE Pro-
Here are some tips for settling the line-up. have looked like without their efforts, and ject Consortium4. It ascribes generic tasks
●●Make sure that you choose collaborators whether someone else could have made the such as “data analysis”, “writing” or “scien-
with whom you can work well. same contribution. tific management” to large sets of authors,
●●Discuss authorship early, and keep doing ●●Familiarize yourself with your institution’s making it impossible to tell, for example,
so often as a project evolves. Put it in writing. or journal’s authorship guidelines, or those who analysed which data. When scientists sit
●●When there are disputes, first try to talk of the International Committee of Medical down to plan a project — and ideally draft the
it out amicably and understand the other Journal Editors. Use them to back up your author list — they should also decide how to
person’s point of view. For example, try to case. describe everyone’s contributions, says Biagi-
work out how the idea first came about. ●●Be prepared to compromise or share oli. The relevant details will probably vary by
●●If you must approach your supervisor credit. discipline, he adds.
about an authorship decision that you ●●If you can’t agree among yourselves, In his own lab, Kosslyn has instituted a
don’t like, keep the tone inquisitive, not engage a supervisor, trusted colleagues or scheme to make authorship requirements
accusatory. Explain that you want to an ombudsman to investigate the matter explicit from the outset. As he listened to his
understand how authorship was decided. and make a recommendation. A.D. student and postdoc arguing their cases several
years ago, he started to develop what eventually

5 9 2 | NAT U R E | VO L 4 8 9 | 2 7 S E P T E M B E R 2 0 1 2
© 2012 Macmillan Publishers Limited. All rights reserved
 CAREERS

POSTGRADUATES
C O MME R C I A L PA P E R S Career-planning course
Ghosts and guests The University of Pittsburgh in
Pennsylvania has launched a course on
Authorship can be misused when there consultant to people who had taken career planning for graduate students,
is money to be made. Medical journals Vioxx and developed heart problems, or one of the first to offer degree credits.
contain a mixture of original scientific their families, in two court cases against Planning for Scientific Success aims to
findings and veiled advertisements for Merck, and he saw some of the company’s help students to identify and develop skills
drugs, and scientists and physicians internal documents6. “We were sort of based on their interests and values, and to
must read papers critically to understand shocked to find pretty rampant evidence create a lifelong career-development plan.
a medicine’s true merits, says Alastair that a lot of the trials were ghostwritten,” Steven Wendell, a molecular biologist
Matheson, a biomedical-research says Ross. “We would stumble across a and assistant director of the postdoctoral
consultant in Toronto, Canada. full draft of a manuscript that just said, office at the university, proposed the
Some pharmaceutical companies make ‘external author?’.” course. “The career problems I hear from
drugs and run clinical trials, then engage There are ways to identify traces of graduate students and postdocs are based
medical writers to draft manuscripts. guests and ghosts in a manuscript: “Check in their lack of a clear, authentic career
These contributors are often ghostwriters the small print,” says Matheson. That is vision,” he says. The course lasts for two
not listed as authors on the paper. Instead, where a medical writer or communications semesters and is required for oral-biology
the company’s marketing team finds company may be acknowledged. Funding graduate students at the University of
a big academic name to headline the from a drug-maker is another tell-tale sign. Pittsburgh dental school, but is open to all
project — even if this guest author makes “These are pointers to the likelihood that graduate students at the university. Each
no contribution to the paper apart from this is something originated and planned semester is worth one credit.
scanning the final version. Companies by industry prior to the involvement of
sometimes use the same technique to the headline authors,” says Matheson.
produce reviews promoting their latest Author disclosures are less helpful, he GRADUATES
medicines, says Joseph Ross, a physician
who studies health policy at Yale University
adds, because academic authors may list
several affiliations and it is difficult to tell Trouble with tracking
in New Haven, Connecticut. One survey5 which commercial relationship is relevant. Universities across Europe want to
found that guests and ghosts haunted With commerce and medicine intimately improve how they track graduates’ career
21% of papers published in six leading intertwined, it would be impractical for progression, says the European University
medical journals in 2008. academics to cut ties with companies, Association (EUA) in Brussels. In Tracking
“This vast production line of information says Matheson. But, he adds, when Learners’ and Graduates’ Progression Paths,
about drugs is passed off as the work academics are offered guest authorship, published on 13 September, the EUA finds
of academics rather than the work of “I would advise them, for the sake of their that if institutions follow career outcomes,
industry,” says Matheson. The companies reputation, to do two things”. First, he says, they can take steps to improve them,
get to advertise their products; the be more than a guest: make sure that your such as revising curricula or establishing
ghostwriters receive a pay cheque; and contribution is author-level. Second, insist strategies to improve communication
the academics get another line on their that company employees involved in the skills. But of 23 institutions surveyed,
CVs. But the patients and the integrity of study are also listed as authors. 77% did not systematically track PhD-
science all lose out, says Matheson. Matheson says it is the responsibility holders’ careers. Study co-author Michael
For example, Merck, a pharmaceutical of journals to make participation by drug- Gaebel, head of higher-education policy
company based in Whitehouse Station, makers more apparent. He would like to at the EUA, says institutions should
New Jersey, minimized reporting of the see papers marked right at the top with create a student database and team up to
risks observed for its painkiller Vioxx ‘commercial article’. He also suggests that standardize data collection.
(rofecoxib) until the drug was taken journals use labels to indicate who funded
off the market in 2004. Ross was a the study, and what drug it supports. A.D.
CITATIONS

became a 1,000-point system. The research- point system. “That,” says Kosslyn, “was the Nobel prizes predicted
ers who come up with the idea get 250 points, last heated dispute we had in the lab.” ■ SEE On 19 September, Thomson Reuters
split between them according to their contri- WORLD VIEW P.475 announced its annual ‘citation laureates’,
bution; writing the paper is worth the same. whom it deems likely to win a Nobel prize.
A further 500 points are available for design- Amber Dance is a freelance science writer in Since 2002, 26 of the predictions have
ing and running the experiment and analys- Los Angeles, California. come true. “We’re trying to demonstrate
ing the data. Researchers who score at least that there is a strong correlation between
1. Martinson, B. C., Anderson, M. S. & de Vries, R.
100 points make the author list, with each Nature 435, 737–738 (2005).
citation at high frequency and peer esteem
person’s point total determining their rank. 2. The ATLAS Collaboration J. Instrum. 3, S08003 in science,” says David Pendlebury, lead
Disagreements still occur; in those cases, (2008). analyst for citation-laureate selection
Kosslyn decides how the points are allocated. 3. Maris, E. & Check, E. Nature 439, 768–769 based in Eugene, Oregon. Each year,
(2006).
When the balance of contributions is unclear, 4. The ENCODE Project Consortium Nature 489, Reuters chooses up to nine candidates in
he does his best. However, it rarely comes to 57–74 (2012). each of the fields of chemistry, economics,
tallying points. “Usually it’s very obvious 5. Wislar, J. S., Flanagin, A., Fontanarosa, P. B. & physics and medicine. The 2012 laureates
DeAngelis, C. D. Br. Med. J. 343, D6128 (2011).
what the order’s going to be,” he says. 6. Ross, J. S., Hill, K. P., Egilman, D. S. &
include researchers in genetic regulation,
In recent years, no disputes have ever risen Krumholz, H. M. J. Am. Med. Assoc. 299, quantum teleportation and reducing the
to the level of the argument that led to the 1800–1812 (2008). speed of light.

2 7 S E P T E M B E R 2 0 1 2 | VO L 4 8 9 | NAT U R E | 5 9 3
© 2012 Macmillan Publishers Limited. All rights reserved