Let me present a short true story, not the joke. Few years ago the heads of my institute asked authors to declare their (subjective) contribution (in %) to every multi-author paper they wrote. What happened? In one case of 4 author paper each involved author declared his part as 90-95% - clear absurd. On the other hand it became clear that, at least in this very case, their parts were nearly equal. Nevertheless the idea was abandoned and never exercised again.

The "Nature" requires authors to state their individual contributions. However, those declarations are treated very loosely. Sometimes we see a detailed list: who was the idea originator, who made samples, who made measurements, who analyzed them, who produced figures and who finally wrote the whole paper. Most of the time, however, this declaration is much shorter: all the authors contributed equally.

It would be interesting to know whether the makers of a >100 author publication do indeed know each other? Did they ever meet? If not in person, then - maybe - at least through correspondence channels? It would be ridiculous to call a 300 person group a "team", when they do not know each other.

I personally never saw a paper written by more than about 10 authors. Could you send a link to a paper written by 500 authors ?

V K! Make sure you choose up to 5 TOPICS above. You have selected only 2 thus far. The more exposure, the more the participation and the wider the spectrum of participation. I'll be reviewing this post routinely.

In present day research, some problems are so hard that they cannot be solved unless a dedicated team works together on it. My view is that it is not meaningful to ask how much each author contributed, because without the collaboration, there will simply be no paper!

For multicentric, multi-institutional study one need to have more collaborations hence all authors get due credit for that!!!

Reuven Ianconescu- may be in the answers to follow several such papers will be listed. But I will not because I do not want to individualize the problem. I would only exemplify by saying that there are several international collaboration in which some people perceive a concept and then several countries join the people and the authors become several hundreds.Because the principal perceiveers are authority and projects are big requiring enormous funding, nothing wrong in collaboration. The problem arises when comparisons majority of the authors is made with those who do not work in such groups, rather perceive their own problems and perform their own experiment or theory or drafting. My question concerns how to distribute credits.

"Do not judge, and you will not be judged" (Luke 6:37).

@Sergei, we'll be judged by faculty/university, not by Luke, I admit (or hope).

@Sergei, good that we pass no judgement. Sometimes we are on the otherside and devise a transparent and objective criterion. Then we need to open this closed question.

This is not a (purely) scientific question because it inevitably involves personal relations between co-authors. In any collective paper there is always a leader (formal or informal) who is ultimately responsible for this publication. Formally, this person is in a position to distribute credits for each collaborator of the group and may convey this info to any interested person OUTSIDE this group. However, this criterion can be objective only IF this person is not biased (for some reason) towards a particular co-author. Therefore, in my opinion, it's virtually impossible to devise any transparent and objective criterion which would completely ignore a "human factor".

In the U.S. outside of academia, the first author is generally the lead contributor with decreasing contributions for each subsequent author. In academia the professor is generally last author but the first author grad student probably did most of the work. It is not true in elementary particle physics, where I think they are just listed in alphabetical order. If you contact any of the authors, they can tell you who the lead was. I have seen many many cases where everyone's name was put on any paper published by the group, whether they had anything to do with it or not. This is because people count publications for hiring, tenure etc.

Dear Reuven Ianconescu,

>

The following newsletter deals with the so-called "hyperauthorship".. You will find a mention of several papers each having over 3000 authors. .

http://archive.sciencewatch.com/newsletter/2012/201207/multiauthor_papers/

@Reuven, ScienceWatch ran an article in July 2007 on papers with very large numbers of authors. Here's the link:

http://archive.sciencewatch.com/newsletter/2012/201207/multiauthor_papers/

As you can see, papers with 500 authors are not at all uncommon in fields such as high energy physics where an entire laboratory is involved in each discovery.

Some journals, such as Nature, require statements at the end of each paper explaining each author's contribution. Some fields use ordering of the author list to indicate each person's role in a paper. In soft condensed matter physics, for example, the senior author in charge of the project typically is listed last. The junior author who did most of the work typically is listed first. Such conventions differ between fields, however. In cases where it's important to determine a particular author's contributions, obtaining letters of recommendation from senior co-authors and other authorities in the field often is the only way to go.

OK, I just wondered: one page has about 3000 characters, and each name to list takes in average 12 characters, so one needs one page to list 250 authors

Now the ball is set rolling. As you notice rightly said especially in high energy physics this is the problem of several hundred authors listed alphabetically. In an outside world, each author claims full credit of that high impact. There is no problem when comparison is made amongst their group, but if that norm is applied amongst all groups having varied specialties these several hundred authors claim (all) full credit which is highly inflated in terms of total number of publications also. I agree that generally we follow a norm that a graduate student who does the overall work under directions is the first author, whereas the senior professor assumes the last and corresponding author. But no such thing happens in the big groups I am talking about

@Sergei, it is not a scientific question but ethical one in science. We will have to address this. Some of us may not be able to appreciate this problem at this point of time. In the present scenario of quantitative assessment of h index etc., scientists are invariably judged by the number of papers and their citations. The groups I am talking of (that is consisting of these several hundreds of authors) publish may be more than multi-dozens of papers in a year - even if out of these 20 active authors publish 2 papers a year they add up to 40 per year, but now each of the several hundred authors gets a claim of same magnitude. These big groups are of course cited innumerably. You can't imagine if you feed in the names of any author the h index value turns out stating too high to be computed. The important question of ethics arises who you think is to be given credit? I had the impression that in research the originator of the idea of a problem and how to solve it is most significant. Then comes active participants and then who drafts and writes the paper. If you are building a huge ship, each one participates because some 10 collaborator brought some ideas together. But then when each of them further distributes the assignment downstream to further 100s, the total mass becomes big. Each one being qualified as a scientist gets an authorship but naturally other questions come up. One tempts to ask if all the authors are even aware of the title, the colleagues names and son on, forget the knowledge of contents. This is a special problem and is non existent for groups limited to numbers of the order of 10. Therefore, this issue of dividing the credits needs to be addressed to.

This is a fascinating discussion. I had no idea that there were papers with 100's of collaborators.

"Do not judge, and you will not be judged" - I like this adage very much but object to its gratuitous and condescending use. If we do not judge by the cumulative experiences of our human experience then there would be no meaning to morality. INDIVIDUAL JUDGEMENT OF GOOD & BAD! Me, I would refrain from quoting archaic verses from the Tora, Quran, Bible, Veda etc. just to keep this discussion mundane and down to earth. Invoking the almighty or the infinite intelligence is a loosing battle for many of the rest of us. Thanx!

If someone is truly a follower of "Do not judge, and you will not be judged" principle, they would have no business even being involved in such contentious discussions.

BTW we are not passing judgments but expressing our opinions. I do see Sergie's point in, "Therefore, in my opinion, it's virtually impossible to devise any transparent and objective criterion which would completely ignore a "human factor"".

Keep on blogging!

@VK, I understand that you're also a condensed matter physicist (like me). And, to be honest, I'm also not happy with "train-like" author's list in high energy and nuclear physics. Yes, high h-index, hyper number of citations. That is all true and there probably exist "free loaders" among those numerous co-authors who did not deserve any credit at all. What you are actually suggesting is to concoct a recipe to expose these "bad guys". Well, perhaps for some really important situations (when a person is considered for a lucrative research funding or for a tenure), all his/her publications with "99 co-authors" should be simply discarded unless he/she is the first author.

BTW this is not as serious as some of the controversial honorary degrees, Nobel prize awards and Doctorates bestowed upon so many by so many others. I see a list of authors as a "promoters list". These folks must at least know the condensed content of the publications to promote it.

A positive spin! BS & redundancy reduction? Instead to 200 publications we see 200 authors on one publication!

We are moving rapidly towards unhampered Global collaboration. So we better be open minded and versatile enough to accommodate for rapid changes due to technology & communication.

I've just seen an ad for 'support scientist' and I wonder what that might be, might well be one of the 500+ authors of the HE physics papers. Are they simple but valuable lab assistants, PG students, or another group working for the research project? So, I'll ask the RG community in a separate thread about that. I don't want to interfere this discussion.

"Reuven Ianconescu" If your co-authors were from southern India, you'd need more than the page. Just my last name alone has 16 letters. "Ananth" is only the first 6 letters. Besides, I have two other 10 plus letter names (First & Middle). ID theft is not that trivial.

I want to thank V K Jindal for posting this issue, as you noted, I was not aware of the existence of multi author papers. I gave some thought to this issue and my conclusion is as follows:

Almost any journal limits the number of pages of articles, either by rejecting too long articles, or by making the author pay for extra pages. The logic behind this policy is to leave place in each issue for some nominal number of papers, in other words allowing too long articles diminishes the number of articles per issue.

Same logic should apply to the number of authors. Each journal should decide the maximum number of authors per article that it accepts, either by rejecting or by requiring fees for extra authors. The space taken by the authors list should by a small fraction of the paper, and if it takes the same space as the paper itself it is like reducing information rate to 1/2 - isn't it?

The question is if such policy could be unfair for some authors - I think not, but if anyone thinks differently, please comment.

Hi Reuven, Authorship on a paper is determined by the authors' roles in writing the paper. The usual test is: if the paper could not have been written without a researcher's direct contribution, then that person should be an author on the paper. Not including a key contributor in the list of authors would be unethical and would create a misleading historical record of the discovery. Some would say that this happened to Rosalind Franklin in the paper announcing the discovery of the structure of DNA. That's why journals should not set limits on the length of author lists.

This leaves open the question of how to present long lists of authors. Different journals have different solutions. Physical Review journals, for example, include all authors, using superscripts for their institutional affiliations. There was a time when Physical Review Letters required papers to adhere to a strict length limit, including the author list. The resulting space constraint meant that some very important papers could not fit in PRL. The policy now is to accommodate long author lists with an extra page, if necessary.

The question of fairness is a separate issue. The editors of journals want to publish important results. If a result is sufficiently important to merit publication in a particular journal, the editors will happily make room for the author list.

Hi David, I fully agree with your claims, however I still see some open fundamental questions:

1. Anything has a limit (although it may be very big): the number of files in one's computer, the size of the universe (probably), etc. So what is the maximum number of authors per publication ? I don't really now the answer, but my feeling is that it should be less than 25. How to count them, in the next point.

2. When considering someone's indispensable contribution: would one also include the designers of a special apparatus ordered from a third party ? Or the people which helped with the lab setup ? Many contributing people are addressed in a special "thanks" section and only researchers who directly contributed (in mathematical development, simulation, experiment, etc) are authors.

To summarize: I can hardly imagine an article to which 1000 researchers contributed in the sense mentioned above.

We need to understand that there are (at least) two kinds of communities in science- the one who work in smaller groups and collaborations who know what each collaborator is doing and may be share information and drafts of papers, and the other hyper number groups running into half a thousand kind. In the latter case it is virtually impossible to exchange draft papers and only professional authors write the papers. At least we could devise a means to assess contribution based upon different parameters rather than existing h index etc. or leave this category separate while comparing with others. A ship designed and perceive by some would need help from several hands from seats to painting, for its completion. But is everybody an equal owner of this ship that is the question.

@Maxim Kotsemir , you have hit nail on the head.

So how do you propose a real situation where there are actually multitudes of qualified contributors be handled?

Some of you seem to be indicating that only the individual or individuals that actually put the verbiage together rather than the "technician" be attributed. Where did you come up with that rationale?

After reading some of your comments I'm realizing the significance of this issue. I think, that is a challenge for those that are using these statistics to "pass judgments" to resolve. I'd think it is they that have to figure out how to assign credit in such case to the individual that they are evaluating based on the rest of the subject's credentials. Though to be objective!

This discussion runs the risk of obscuring the long-term goal of listing authors on scientific papers, which is to establish the historical record of researchers who participated in the research. How to interpret this record in the short term for professional advancement is a separate issue.

In establishing the historical record, the authors themselves almost certainly are the best judges of who should be counted as an author and who should receive other forms of acknowledgment. Massive collaborations come together to solve specific problems, and they work toward their goal collaboratively. The specialists who design the instrument, who build it, who run it, who analyze the data and who interpret the results all are essential to the outcome. Their contributions should be acknowledged. That is why such collaborations often list large numbers of authors.

Regarding professional advancement, everyone understands that some authors will have contributed more to any multiauthor paper than others. In cases where this makes a difference (hiring and promotion, for example), it is natural to ask a candidate to explain their role in their publications. This is necessary even for dual-author papers!

Ultimately, the "credit" for an important discovery will go to the small number of people around whom the project coalesced and ran to fruition. The primary record in the peer-reviewed literature preserves the names and the institutional affiliations of all of the researchers who participated in the work.

I should mention that I am an experimenter working in a field comprised of small research groups. Most papers in my area have fewer than five authors. I chose this field in part because I enjoy being involved in all aspects of a project. That does not reduce my appreciation of big science, or my appreciation for the people who work in those large collaborations.

I think the situation is much worse (when it comes to judge individual contributions) than in movie-makers business. There we also have pretty long "pay lists", including hair-makers, but they are often ending with phrase "and others". Here, however, the problem is solved in a more elegant way: first comes the name of a film director and the screenplay author, sometime the movie producer. Only later the actors' names are shown, perhaps with the most important one (or two) distinguished with longer exposition time and/or bigger letters. But even the list of actors often ends with "and others".

It may be therefore suspected, that in case of big experiments (i.e. lasting few years), even the crew working three shifts to keep the instrument running 24 hours a day is listed as "authors". Among them post-docs, spending 2-3 years at this very experiment lasting 6 years overall. There is no way to say "who is who" when the authors' list is ordered alphabetically and is longer than, say, 25 names.

The broadly understood "pay list" may be the real reason behind such phenomena.

Yes, the cooperation is a must, but there is someone, maybe a small group, say 3-5 persons, who know what they're actually doing and why - the real authors. The other crew/staff members, say those responsible for uninterrupted supplies and alike, should be paid for their efforts (as they already are), but not credited with authorship.

1 / 0 · 33 Answers · 167 Views

I concede each contributor here is genuine. So please do us all a favor and pick a "Green" or a "Red" up top. These statistics help circulation and hence more informed participants in this discussion.

"If you're happy inform your face!" Rodney Dangerfield

BTW this discussion has no risk of affecting anything but some open-minds of readers at the most. Surely not the way "business is done". Individuals and groups will do whatever is in their best interest. The "judgers" though are the ones in a quandary.

I want to return to the original question of credit, and to suggest a method to calculate it. Say we give a GRADE to each article, and only for our example, say the grade is between 0 and 1000.

We may suppose that an article of GRADE=100, written by 2 authors required less work from each one than another article of GRADE=100, written by 1 author.

So we consider the individual contributions as additive, and without the knowledge of how much each contributed, we assign to an author for one publication the GRADE divided by NUMBER_AUTHORS.

So if a scientist published in 1 year 80 publications (!), but each one had 1000 authors, and say each of those publications received a grade of 500, this author will receive the credit of 80*(500/1000)=40.

Now another scientist published in 1 year 2 papers in which he/she was the sole author, each paper got the grade of 100, so this author will receive the credit of 2*100=200.

It turns out that we have to be open to allow each community to make its own rules for giving authorship to any number they deem fit. The question raised had not the intention to disturb this freedom. It posed a serious problem of how to distribute credits of such publications where all authors are stated equivalent - they are arranged alphabetically. Yes the community is aware of the individual weight but when it comes to inter-community comparison which most often is done on the numbers like h index, citations, and impact factors of journals then it becomes an issue. Especially when the administrators just are semi-literate on this issue.

May be authors are type 1 and type 2 and categorized in publications in case they want to maintain alphabetical order. They agree to break as a paragraph after the principal authorship. But I know this wo'nt be agreed to as the authorship is country based having financial overtones.The collaboration continues because of agreements.. Once the financial bosses decide to fund and decide the collaborating groups, academicians do not bother about such details as long as their work continues.

My simple solution: don't do it. I think it is impossible to separate contributions of authors. Their contributions are all in a quantum superposition so to speak :-)

A better question is how can we fight those bean counters that make us do fatuous and pointless exercises.

If authors have mutually decided to coauthor a paper, then the best metric is to always reduce it to an equal contribution. If we impose that rule, then in the future the number of authors per paper will reach some sort of Nash equilibrium and we will stop seeing papers with 500 people on them.

Let me present a short true story, not the joke. Few years ago the heads of my institute asked authors to declare their (subjective) contribution (in %) to every multi-author paper they wrote. What happened? In one case of 4 author paper each involved author declared his part as 90-95% - clear absurd. On the other hand it became clear that, at least in this very case, their parts were nearly equal. Nevertheless the idea was abandoned and never exercised again.

The "Nature" requires authors to state their individual contributions. However, those declarations are treated very loosely. Sometimes we see a detailed list: who was the idea originator, who made samples, who made measurements, who analyzed them, who produced figures and who finally wrote the whole paper. Most of the time, however, this declaration is much shorter: all the authors contributed equally.

It would be interesting to know whether the makers of a >100 author publication do indeed know each other? Did they ever meet? If not in person, then - maybe - at least through correspondence channels? It would be ridiculous to call a 300 person group a "team", when they do not know each other.

@Derek, yes so was also planned at our end but did not succeed. simply fractionalize the score by the number of authors! If that happens to be done by the people at Web of Science, Google Scholar and Scopus this will have a significance. Otherwise it will not be accepted. Let them fractionalize and recalculate h index by considering above say 5 authors papers as a fraction. They may devise a means differently for deliberate first author (not by way of alphabatical) and deliberate last author who deserve better part of the fraction,

@Marek, in the coming times, a declaration from the principal author or corresponding author should be incorporated which should state what fraction of 1 each author has.

@Marek. Not only does my institution make us allocate a % contribution, but:

a) When adding the contributions of all authors it always adds to >100%

b) We have been doing this for years

c) We still do it.

Conclusion: fire the administrators who make us do these ridiculous exercises.

That's not going to work! As the very idea of assigning what one thinks of others' contribution in itself is subjective. This is a principal reason why "each involved author declared his part as 90-95%", as Marek points out.

In my opinion, it is a matter of magnanimity in many cases. There is too much emphasis on the "objective" in this case rather than the "subjective" as is in this case of "credit for research". Seems like it depends on the field, matter, collaborations involved. With all the open access to information as of to-date, it would be tough to hide the real facts. So, those in the "subjective judging business" need to be aware and cognoscente of many of the underlying facts, as astutely pointed out by all of you in this discussion.

The proposal presented by Reuven Ianconescu seems quite fair: give each author a credit equal to 1/n, if there are n authors (his GRADE=1 here, for any published paper, regardless of the Impact Factor of publishing journal). Maybe increase this share to 2/n for the first, last, corresponding or otherwise distinguished author. I don't see a big problem, if my "productivity" calculated this way is fractional, not integer. So what? Just call it "the effective number of papers authored", or "contributivity" in place of "productivity". Are you feeling the subtle difference?

Surely, the Hirsch index would have to be evaluated differently - but that's another story.

All of us probably know smart authors with publication list length exceeding 200. At the same time we are quite sure that those authors are no contenders to Gauss or Euler (they gave us less than 40 papers each)), isn't it? From this point of view, the proposed measure would be certainly more adequate (just?) than many others. Anyway, Cauchy will be a winner - he authored more than 1400 single-author papers.

But "smart" authors (including good organizers, heads of labs, etc.) would not be hurt too much.

The bad thing is that such kind of evaluation (of a single researcher) is very hazardous to cooperation, it may nearly kill it. This is counterproductive, as the well cooperating team is certainly more effective than a single researcher.

Rightly said, no doubt easy means by fractionalizing with say 50% weight reserved for principal authors- like chosen first author and last corresponding author and may be a few more, the rest 50% divided equally amongst remaining authors does offer a solution which needs acceptance. After all if there are 1000 authors they also publish 1000 papers in a given time of months. The authors who believe, apart from this collaboration, they are doing good independent work and publishing in smaller local groups also, they get full credit from these anyway. The additional advantage of big collaborations if needed to be merged with vast majority of non-heavy collaboration groups such solutions will need to be considered. It will discourage giving unnecessary authorships as it hurts their own workers. The worry that it may not discourage big collaborations is perhaps unfounded as there are additional advantages of such collaborations which outweigh the fictitious overrated numbers and indexes. They keep participating in their research meetings and are exposed to newest things, perks, socialization and so on and this promotes new ideas for independent research.

Prof. Jindal, a very good question, but... I think it takes the matters too far. The problem with multi-authored papers in high-energy physics is that is is impossible to pinpoint who found what, so the collective is credited with the finding, in addition to the head of the group. The scientists at CERN for example are all looking for a variiety of particles, and these can pop up in anybody's shift, and there many people in each shift, so it makes sense to list all participants. If you watch an old Hollywood movie you see the credits at the end...and they say who played what and who did the costmes, who wrote the music, etc. So all contributions are listed, but with the exception of the director, the producer and the main actors the other names are not remembered except by imdb... It is important, as we know, to attribute credit, but really in such large groups there is not much point in asking people what was the percent of their contribution. At the movies you could list who did what, in authorship it's not that easy. The merit is in participating in this effort. The need for quantification is something that kills the fun out of publishing, and it's something relatively recent. What should matter is the fact that there was an acknowledged contribution to a project that eventually resulted in a particular publication. These things were never intended to be measured like this. In my view such quantification is pointless except that administrators think that 'there must be order and quantification.' Perhaps following the discussion here many of the authors in multi-authored papers should appear in acknowledgements rather than as authors, but that would not count at all, as this is not measured.

Btw, I am not sure how these papers are eventually written and published. Do all the authors read the paper? That's another question of course, but I really think that the attempt to quantify everything is not the way to go. Btw, besides particle physics I am not sure in what fields hundreds of contributors are the norm.

It will be slightly out off-topic but well in line with some comments by Rafael.

I'm a part of the ever-changing team searching for gravitational waves (Einstein@home project). Ale we do is "renting" our computer cores (and other resources, like time, disk space, etc., of course) for astrophysical data analysis. Each data chunk is sent to at least two computers for analysis. Everything is managed automatically by computers and there are literally thousands of participants involved in this huge task. Majority of us has in fact no idea what is being calculated and how. No gravitational waves had been detected until now (after few years!), but there were some true discoveries of interesting celestial objects. Guess, who were the authors of some papers on that? Usually 5-7 persons, more specifically the ones, who know the subject very well, those who "keep an eye" on the entire enterprise and 2 or 3 random individuals who happened to analyze the "proper" data chunk, often not the scientists at all (say system administrators, able to offer ~1400 cores of underused computers, in contrast to my contribution of just 2 cores). Isn't this situation similar to the one in high energy physics? Shouldn't such the big teams, with members never acquainted, follow the similar publishing policy?

Yes Marek, I think they should, and it probably depends on who is running the actual program. I have never participated in such big groups, the maximum number of authors in my papers is seven...:-)

I think however, that all you need to say in your cv is that you participate in the program. In my view, one cannot really quantify the contribution in such projects unless you are among those who decypher the results (which is what happens in the CERN projects I mentioned). In addition to the researchers they have an enourmous tech support system and those guys are rarely mentioned except in acknowledgements...

@Rafael and Marek

We now appreciate that there is a genuine problem here of credit distribution. The modern days of quantification of progress indices like h index, number of citations, number of publications etc. are applied to scale scientists. But perhaps when they devised such index, they forgot this category of mega multi authorships. A ten year experienced researcher averages out 2000 publications in premier journals as a results of this 1000 author publications, infect all of them. Without judging what they do, we must appreciate that new parameters need to be defined. It is left for people like Hirshch if he invented h index and so on..

Vijal, don't blame Jorge Hirsch for poor performance (inapplicability?) of his h-index to papers with more than, say, two-digit number of authors (and, consequently, to citations of such papers building positions of their authors). Note, however, that if you happen to appear (once!) on the authors' list of such a publication, and later it is cited 1000 times then your h-index will not increase at all or only by 1. Quite a different situation is when you "author" 50 papers of this kind. I have never tried to see what is the common part (of authors, of course) of few such publications but it seems likely that only the unquestioned leaders appear there frequently. The rest is "plankton" as we call in Poland the inactive parliament members. Their names remain mostly unknown to the rest of citizens. The problem is real if my suspicion is incorrect - but is there anybody able to say something objective about hundred's authors' lists intersections?

Marek you are unaware of the magnitude of problems and number of such publications shared. These mega author publications are in thousands as mentioned in my earlier note. The result of such magnitudes is when you try to find h index of these authors the answer you get is too high to compute!