OMG: “756 additional authors not shown”

May 10, 2012 • 5:47 am

Alert reader coelsblog pointed out this astrophysics paper in a comment on one of yesterday’s posts:

It’s a paper with 815 authors, and they ran out of steam before they even finished listing the “B”s.  Is there really a need for so many authors? Did every one of the 815 make a genuine contribution? Perhaps some astrophysicists can explain to me the justification for so many authors.

If one assumes that each author is 5 feet 8 inches tall, then if you laid them all end to end the chain of gratuitousness would extend 4618 feet (nearly a mile), or 1.4 km.

125 thoughts on “OMG: “756 additional authors not shown”

    1. That’s a gratuitiously snarky comment and a slur on biologists as well. Don’t come into my home and urinate on the carpet.

    2. Hey! My wife is a biologist. They have standards. In her lab, if you didn’t contribute at least 10% of the work or make some other very signifcant contribution (like establishing a new line of mutations to compare to wild-type and the phenotype in question) you’re not on the paper.

  1. Jerry, read the section this was filed under again: “high energy astrophysical phenomena”. That doesn’t refer to the thing being studied in the paper; rather, it refers to what is required to get the publication done. Think about it – your metric is based on physical size. I think that’s wrongheaded.

    This is academic stuff, so the proper metric is intellectual wherewithal. If we consequently assume that each author has an IQ of 120, then we have a hypermind with an IQ of 97,800. I put it to you for your consideration that any study undertaken by the exertions of such a mind is going to be, at the very least, correct. You only think the presence of that many authors is gratuitous because you’re only using one mind which, I’m sorry to say, simply wouldn’t be capable of understanding the full gravity of the paper’s subject.

    Whatever else one might say, I think it is beyond argument that such a mind, and only such a mind, would fully justify the use of ‘swift’ in ‘swift follow-up’. QED.

    1. Don’t you need to reduce your hypermind by the overlap? Or is each rendition of 1+1 equal to an increase of intelligence?

      I’m going to gather 50 million idiots such as myself into a hypermind we’ll knock you 120+ IQ physicists right off the carpet.

      1. With me and the brain power of 1.5 million squirrels, there’s no limit to what we could accomplish. Why didn’t anyone think of this before?

      1. Excluding the references, abstract and authors the paper is 8,509 words long. Divide that by the 815 authors and you have each researcher contributing 10 words.

  2. I believe this paper is more about recent efforts to “crowdsource” parts of the scientific analysis. In exchange, they offered co-authorship of the resulting publication.

    I think this is pretty cool stuff, and is certainly an interesting effort to engage the public in science.

    1. What, nobody contributes to anything any more unless they get an authorship out of it? I’ve helped tons of people with their papers, including suggestions and rewrites, and not demanded that they slap my name on it. Sorry, but this seems ridiculous to me.

      1. Yes, but where do you draw the line between those who deserve an authorship and those who don’t for that kind of projects? It’s certainly easier to put everyone on the paper than to try to determine who should and who should not be on it, especially given that even then it would still be a large number of people.

        1. We draw the line all the time in biology. We TELL people if they’re going to be authors, usually in advance. When you ask somebody to look at a draft, or someone makes a helpful suggestion, we don’t automatically slap that person on the paper. Nor do I give authorship to the technicians in the physics department who built my fly cages. Is it not sufficient to get satisfaction with helping science move forward a tad without getting your name slapped on the paper? If you don’t draw a line, then everyone, including the person in the office who processes your grants, should be an author.

          Let’s face it, this is just a way to inflate people’s c.v.s.

          1. That’s true but it still does not answer the question how and where you draw the line for something like LHC, LIGO, 1000 Genomes, ENCODE, all the GWAS studies that regularly have 100 or more authors, etc.

          2. Technicians, secretaries, accounts etc all have different reward systems. For PhD-level scientists a major reward system is paper authorships.

            So what do you do if your project is sufficiently complex that you need PhD-level scientists to produce your data, does that get them onto the resulting papers?

          3. Let’s face it, this is just a way to inflate people’s c.v.s.

            In some areas of science where a very large projects aims at a specific goal that may produce only a few (but very important) papers, being 400th author out of 800 on relatively few papers may be the only thing they get on their CVs for multiple years of effort.

          4. To continue on the previous post – I don’t know if you have been involved in a consortium projects (although you seem to be on the 12 fly genomes paper, PMID: 17994087) but these projects usually involved dozens of labs each of which does different things. The PIs must be on these papers because, at least in the cases I have been involved in, they do the composition and writing (these things usually are quite political so this can not be left to the trainees. But they don’t do the actual work, that’s done by postdocs and grad students so those have to be included too. Now, let’s say you have 15 labs in a consortium, that’s 15 PIs, and at least 15 people, one from each lab; that’s 30 and by your criteria it is probably already too much, but usually it is a lot more people than that from each lab who spend most of their time working on different aspects of the project, so I am curious how exactly you would justify leaving them out?

            I understand and fully agree with the general argument that people who did very little should not be authors on papers. And it is true that a lot of people who end up on these kind of papers, did very little. But you can not leave out people who really contributed and there is no escaping the fact that their number is going to be large, due to the nature of these projects. Is it 3000? Probably not, but it is still going to be a lot of people.

            1. Now that’s a good point. My wife’s lab has never, in all the time been married to her, done anything like that. And they’re pretty strict, you have to do something meaningful.

              You just can’t clip fins or run PCRs or order pizza for lab meeting.

              1. Oops. Wasn’t done…

                Anyway, train of thought wrecked, something about her lab is more solo, or maybe two researchers from different schools/labs are cooperating on something. They’ve never done any of these really big projects with lots of people doing lots and lots of work on a single project.

          5. Nor do I give authorship to the technicians in the physics department who built my fly cages.

            I apologize if you think this comments breaks the “house rules” (this being your webblog, of course, you have the final say on that), but I think I must point out that if you think designing a collider component is at the same level as building a fly cage, then, sir, you are severely mistaken.

          6. Regularly draw the line in biology? That is vaguely amusing from my own narrow experience. I am listed on at least five >100 author genome related papers, like this one:

            http://www.ncbi.nlm.nih.gov/pubmed/10731132

            or this 300 author behemoth:

            http://www.ncbi.nlm.nih.gov/pubmed/15772651

            I’ve never seen or even communicated with most of these people. While I did make a pretty non-trivial contribution to these efforts, I do not think it would have been a crime to have omitted me from these papers.

            Nonetheless, isn’t clear to me who -should- be on these papers. They really were mob efforts. These are huge projects spanning multiple very large labs. The lab I worked in employed 200 people by itself. Should only the four or five lab directors be on these papers? Or people who contributed a figure? Or people who wrote more than X words of the text? There is not, in this kind of paper, a clear scientific idea, a hypothesis being tested, that you can trace to an individual or two. I really don’t know what to do with these cases. Apparently all the PI’s didn’t either, and so the author list swelled to huge proportions.

  3. I assume it’s more of a high-energy physics thing than an astrophysics thing. I think the largest papers I’ve seen (before this one) are usually data releases or similar, where everyone who helped with turning measurements into a catalog of galaxies (or whatever) is on the paper. (In my undergrad days, I worked on the variability of active galactic nuclei; we had a lot of collaborators, because having several observatories meant you could watch something nearly all the time; so most of the authors on the paper contributed to getting the data or doing the first analysis steps.

  4. Perhaps some astrophysicists can explain to me the justification for so many authors.

    No, I can’t really justify it. But it involves two very large collaborations (VIRGO and LIGO) both trying to detect gravitational waves (Nobel-prize stuff if they succeed). They have a policy of everyone in the team being on every resulting paper, and all being in alphabetical order. That is one way of resolving authorship issues!

    One can see some sense in this, if hundreds of people are constructing large machines costing hundreds of millions of dollars, all aiming for one outcome, then one can argue that all team members are contributing to the eventual science papers about possible detections of gravitational-wave events.

    The other solution would be to pick a handful of people who analysed particular bits of data, and not credit the hundreds of people who built the machine that produced the data.

    1. Or they can do what we do in biology: we acknowledge people who made small contributions in the damn ACKNOWLEDGEMENTS section. Not everybody has to be an author, and the many acknowledged authors could be listed somewhere, perhaps as a “consortium” or in the supplemental material.

      I continue to think that putting everyone as an author on the paper who makes the tiniest contribution is ridiculous.

      1. “I continue to think that putting everyone as an author on the paper who makes the tiniest contribution is ridiculous.”

        And, even worse, it seriously degrades the whole concept of authorship in the system. Are all these people putting these papers on their CVs? If so, we have some serious grade-inflation going on here and that can’t be good for the people in the discipline.

      2. I share your opinion to a large extent, but to put the counter-argument:

        In a field like gravitational-wave detection you have teams of many hundreds of people building the machines, yet the number of likely science-output papers will be small. Even if they detect gravitational waves they may only detect a handful of events, and thus only have a handful of papers (even though those few might get them a Nobel prize).

        So the result would be that many of the team would never be on any paper (in such a field it is not as though there is loads of subsidiary science, so that everyone can have their own programs as well as contributing to the overall goal), and yet the reward system in science is very much tuned to paper authorships.

        If they didn’t do something like this then they’d be asking PhD-level scientists to work for ten years or more and never have the opportunity to be on any resulting papers.

        1. If they do get the Nobel Prize it will be interesting to see who makes the cut from the several hundred researchers!

          1. Most likely the theorists that made the predictions. But for the LHC and the Higgs even this is a problem because if I recall correctly the original paper that predicted it had 6 authors. I don’t know what they’re going to do, perhaps they will have to wait for some of them to die it’s been a long time and they are quite old already

      3. But we don’t get much or any professional credit for acknowledgements. Maybe there’s something to be said for the astrophysics approach.

        1. Maybe there’s something to be said for the astrophysics approach.

          This isn’t widespread in astrophysics, it’s peculiar to that area. There is a big distinction here: If you’re involved in building some major telescope, say the HST, that is a general-purpose facility that can then be used for hundreds of different projects, then there is no need for this approach. Everyone can then pursue their own science program with it and write their paper as a sole-author or with a few collaborators.

          It’s when the large machine is aimed at one particular very major goal (say gravitational waves or the Higgs boson), and when there isn’t much scope for subsidiary science, and likely only relatively few papers will result, that you need to think about how to reward all team members.

          1. There’s still the problem of minor contributors getting essentially no credit in the “author with acknowledgements” approach. Is adding more authors really a horrible problem? Maybe there could be “major authors” (who think up the hypothesis or test) and “contributing authors” who provide some point of expertise (with perhaps just 2-3 hours of work) that makes the whole thing possible.

            1. Some journals ask for a succinct statement about the specific contribution of each author. That might double the length of that paper, though..

      4. I continue to think that putting everyone as an author on the paper who makes the tiniest contribution is ridiculous.

        That may be true, but it is not clear to me that the contributions were tiny at all, especially given that the LIGO team decided at the outside that everybody on the team is going to be an author on any results published.

        Also, as someone pointed out, biology has its share of large-author-sets too: one of Noor’s papers has him as the 151st author, and someone else pointed out that some genome sequencing experiments have 2900 or so authors. I don’t see how it is so clear that on such a large project, most of these authors just made a “tiny” contribution. Putiing them away in “damned ACKNOWLEDGEMENTS” section, I think, would then be very very unfair.

    2. If you look at the institutions involved, you see a representation of hundreds of institutions from all over the world. All of them gathered some data. As I understand this particular effort, LIGO & Virgo made one kind of detection, and the others verified or disqualified each detection based on their own measurements.

      See http://www.ligo.org/

  5. It’s the same in particle physics. Acquiring the data needs a huge machine which take a huge amount of work, tests, maintenance, code written etc.

    Sometimes the people making the paper are simply dumping some data into plots.

  6. If each author contributed, say, at least 3 words, pretty soon you’d have a real paper there. And when they win the Nobel prize, there will be like $1000 in it for each! Sweet!

  7. Let me suggest an, uh, “accommodationist” solution. 🙂 We all know those contributions weren’t equal, or even of the same order of magnitude (by labor, intellectual contribution, or any other measure). Yet I do think there’s some value in mentioning/thanking the contributions of people who may have helped out with your experiment. Having worked many late nights and put in 12-hour shifts on other people’s experiments, its a bit of an a**-hole move to ignore such folk completely.

    Now, books already have a section tailor made: the acknowledgements section. But journal articles don’t typically include this. ‘Acknowledgements’ for science articles typically refers to funding source identification, nothing more. Maybe we need to make a contributor acknowledgements section a formal part of artices: a place to name the folks who helped out, but may not necssarily qualify as authors.

    1. Well, in biology “acknowledgment” sections include people as well as funding, so that solution already exists!

      1. A lot of my colleagues are obseessed with “impact factors” – of their papers and of others. I’m wondering, Jerry, if a long established investigator such as yourself can get by NIH panel reviews when your students papers are included in your reports. Can an assistant professor expect the same kind of treatment these days?

        1. Well, I had the same NIH grant renewed repeatedly for 30 years, and often half the papers didn’t have my name on it. I’m told that the NIH and NSF don’t look at authorships of papers that were carried out under a given grant: all they care about is the quality and the number of the papers that came out of a given funded project. Now I can’t vouch for this personally, but it would be invidious if the granting agencies worried about whether the PI (principal investigator) was actually included in the authors). They care about whether a decent quantity of high-quality work was done.

          1. I have heard of grants submitted by two labs together being rejected because there were no published papers with both PIs on them, i.e. no evidence for successful collaboration in the past. And that happened because one of the PIs decided not to be on the papers from the other lab precisely for the reasons you don’t put your name on your students’ papers.

            1. Reasons for accepted and rejected grants are hard to pin down though I’ve heard of this happening too and for the same reason. Maybe it’s laburban legend!

  8. I feel this whole ‘too many authors!’ business is just coming from a bunch of PIs who are scared they are losing some mystical form of prestige now that people are actually getting credit for the work they did for the paper getting published.

    Although I agree it is shocking how many people it actually takes to ‘do science’.

  9. Or maybe the open access advocates would like to see all the tax payers names on the paper since they are paying for it.

  10. Immense author lists have been commonplace in high-energy physics papers coming out of big-project sites like CERN for a long time haven’t they? I’ve never kept track of actual numbers of authors, but I remember seeing huge numbers of authors in the contents of Phys. Rev. Lett. thirty years ago.

    Large numbers of scientists are needed to design, build, and maintain these facilities and it appears that the physics community hasn’t found a way of acknowledging these collective efforts other than by putting large numbers of these team members on papers.

    1. Good point! The really important point is the importance of the research behind the paper, not the name, nor the number of authors ..

      1. And when we say “numbers are not important” it works both ways: small numbers are not necessarily a good thing, large numbers are not necessarily a bad thing.

  11. I have a suggestion.

    We all know that there’s been some controversy over the accuracy of papers published in the scientific literature. We all know that there are papers where one scientist published fraudulent data that the other co-authors didn’t detect.

    This has lead to the reasonable suggestion that every author of a paper has to take responsibility for the content of the paper. It means that if your name is on the paper you have to attest to having read it and to agreeing with the content.

    Let’s just take this one step further and demand that journals only accept papers that are accompanied by individual signed statements from each of the authors. Every author has to swear that they have read the submitted draft, and the final revision, or else the paper won’t be published.

    I wonder how many of the 815 authors actually read the paper and are willing to stake their scientific reputation on the data and the conclusions?

    I’ve published papers with five other authors (three other postdocs and two PI’s) and I can assure you that it’s no trivial matter to get a consensus on what should be in the paper and how we should revise it for publication. I can’t imagine what it would be like to get such a consensus with 814 other authors!

    1. There is also the curious situation of people being on large consortium papers who then write their own papers that contradict the consortium results

    2. This does not solve the ‘big science’ problem others have noted above. If an experiment takes hundreds of people, years each of effort, and produces relatively few papers, are you really going to demand that each of them stake their professional reputations on the accuracy and validity of all the other team members’ work?

      IMO that would disincentivise working on such experiments so much that no young scientist would be willing to do them, and they’d never get done.

      We do not want to make the professional risk of contributing to a big experiment so much larger than the professional gain, that it actually stops scientists from doing the experiment.

      After reading more of the posts here, I am frankly wondering if any cure is better than the disease. Seems to me that the professional benefit of being one out of 1,000 authors on a paper is probably not going to be greater than what it should be.

      Maybe what we really need to do is find a better way to list the authors, i.e. via subsections like movie credits. Yes, you DO list the second junior grip along with the director. But you list them as ‘director’ and ‘second junior grip.’ There is no question about how much they contributed or what they did: it is right there, in the listing. Likewise, for something like LIGO you could list the various contributors as authors, but under headings such as “Hanford observatory design team,” “interferometer optical components team,” “data analysis team,” and so on (I’m just making those categories up and have no idea whether they’d be appropriate).

      1. I was thinking the same thing. With the single significant category of “Author” which generally means, “the dude that wrote the dang thing”, it’s confusing to laypeople and such a list of authors would make anyone just gloss over that section anyways, and makes such credit look like nothing more than a political credit grab. If they listed off each person and the title of the job they performed for the work, not only would credit go where it was due, but us laypeople would see exactly what goes into such research, including one-off construction of fly cages. I’ve provided technical review for software publications, and my name was presented as: technical review provided by… It would be a bit silly to put my name in there as author as I didn’t actually write the thing. What if Lawrence Krauss had to have Richard Dawkins listed as an author on his book, when all Dawkins did was write an afterword?

  12. I can only assume that this represents many teams of people that operate, maintain, archive, and process data, even if they had nothing to do with the paper at hand.

    I’ve seen something simiar, where the ‘author’ is listed as a Team name, like as “R Development Team”.

    I have to wonder, what /good/ does it do you to be author 546 on that list, and how do you put it into your CV? It’d have to be Evans et al, you couldn’t include every name in order to include yours, which you’d do on any paper that had a reasonable number of names. And if your contribution is so infinitesimally small that you can’t even include your name in the citation on your own CV, then what’s the point?

    1. Funny though. I can only speak about medicine, but usually the “big gun” is the either the third or the sixth author listed.

      Why? Because most of the time, when a paper is cited by another paper, that journal’s style truncates the author list at either 3 or 6 names, followed by “et al”.

      So, the bigwigs make sure they’re listed where their name will appear in citations.

      The first-named author is usually the person who did the heavy lifting. The third (or sixth) name is usually the senior researcher.

      1. Is it true that in some medical schools, the head of the department also gets an authorship, even if he/she did nothing?

        1. In most of the bio labs that I have worked in, the PI gets tacked on to the end of every paper’s author list. In a large research group such PIs get >>20 pubs per year. If you are not a “PhD-level” contributor you need a much larger contribution to be included (typically).

          BTW the term “PhD-level” is absurd and shameful. It is about the science, nothing else.

          1. “BTW the term “PhD-level” is absurd and shameful. It is about the science, nothing else.”

            That’s why it’s PhD-*level*, not PhD-*possessing*. Thus it means a contribution to the science at a level typical of material that could be presented in a PhD thesis (regardless of whether or not it is). Thus for example a student with no PhD can make a PhD-level contribution.

  13. It’s worth noting that the LHC produces roughly 20 million Gigabytes of data per year.. and rising. That suggests the scale of the experiments. As a small-group nuclear physicist, I have asked large-group colleagues if they miss the rewards and responsibilities of running their own experiments. The answer is an emphatic `no’.
    One said that the software that he and his student prepared was actually vital to the whole enterprise. He `owned’ a key part of the project… and there were many key parts. There will be probably be hundreds of PhD theses, justifiably, from the Atlas and CMS collaborations.

  14. In medicine, most of the large consortia have a writing team. Those are the people who compile and process the data and are responsible for the content.

    Contributors of data are listed in acknowledgements as members of the team. But they don’t get their names as AUTHORS of the paper.

    That’s the key word — authorship. If they didn’t write the paper or contribute substantially to the writing of the paper, then they’re not authors of the paper.

    Contributing data does not equal authorship.

    1. That is not the custom in my field. Writing the paper is only one contribution. Obtaining the data and analysing it are also contributions to the end-product paper, and deserve “authorship”, even if those people didn’t write any part of it. Why would one privilege the writing role over other contributions to the science?

        1. From m-w.com, meaning of “author”:

          1. One that originates or creates
          2. the writer of a literary work

          The word “author” is wider than the word “writer” (e.g. “the author of his own downfall”), and as used in science papers does not simply mean “writer”. Everyone significantly involved in the creation of a paper can rightly claim to be a “co-author” in the literal meaning of the term.

  15. Disclaimer: I am not a physicist. In the field I work in, more than 4 authors on a paper happen regularly.

    I think in large experiments such as this (this paper is from LIGO, if I skimmed it right), it is quite possible to have these many authors. People I know in experimental physics say that on a paper like this, it is common practice to include people who may not have been active in the actual experiment, but who helped designed the whole apparatus. Now designing physical experiments at the level of colliders, for example, is non-trivial contribution to science, and may require a lot of research on its own. And yet, there is no way, I think, that a Physics journal is going to publish this research. The reasoning goes that it is therefore natural to give physicists who pave the road for other to get credit when the results from the experiment start pouring in.

    In fact, I have heard the same reasoning from one of my friends in Bio-Engineering, where somebody’s PhDs might consist entirely of designing and creating a new device in their lab. Now, of course, there is no immediate “publication reward” for this, so it is considered fair (and ethical) to include their name as a contributor on future papers that crucially use the new device.

  16. I suppose this is a field predicated on the notion that very small things (quarks and such like) are nevertheless very important and deserve to be recorded. Perhaps they simply carry that approach across.

  17. If one assumes that each author is 5 feet 8 inches tall, then if you laid them all end to end the chain of gratuitousness would extend 4618 feet (nearly a mile), or 1.4 km.

    At least they’d be getting laid.

    {for a change}

  18. On a different note, I don’t think having a galactic author list is wrong in itself: it is just a bit different from the traditional method of working whether different (small) groups work in secrecy on problems to claim credit when they finally solve it. In mathematics, for example, there have been a few successful demonstrations of this approach. One was a community-generated proof of the so called Density Hales Jewitt (DHJ) theorem in additive combinatorics: several readers of Timothy Gowers’s blog collaborated (ranging from CS and maths professors to high school students and teachers) via blog comments to find a proof within a month or so (a very quick proof by usual mathematical standards). Of course, when they got around to writing it up (they being mathematicians, with their own ethics and standards on authorship), they simply omitted all the names from the author list, and instead put down the author as D. H. J. Polymath (Hint: Recall the name of the theorem being proved)

    1. I’m one of them!

      Did I make a significant contribution to this? Yes. Do I deserve authorship? Probably not.

      To some extent, though, I think the issue of giving too many people credit is a non-problem. What is the supposed harm? Diluting the credit of the people who really deserve it? Who really deserves credit for this giant mob effort? Or is the harm that it lets people who contributed little pad their CV’s and look more important than they are? This latter problem solves itself. When you have papers like this one, it does not really do much to pad one’s CV. I have a whole page of these kinds of credits, but no one cares. Or not much. It is only my first author, or at least short author list, papers that other people really credit to me. No potential employeer looks at my CV and says, “Human Genome!”, we have a star applicant here! They automatically see it for what it is, a massive effort on which I was probably a bit player. Only my mom is fooled.

      1. I have a whole page of these kinds of credits, but no one cares. Or not much. It is only my first author, or at least short author list, papers that other people really credit to me.

        I think it’s a matter of where you are in your career. Any level of authorship benefits you as a summer student or even technician. Grad students need at minimum a 1st author. Post-docs need to rack up their publication list, especially 1st and co-firsts for career advancement. What’s the general rule for assistant professors?

  19. I learned the following principle of granting authorship that I believe is reasonable and fair and which has served my colleagues and myself quite well (this at least applies to biomedical research done in either a single or a handful of labs; I’m not talking about large consortia or high energy physics). Completing a research publication (in experimental sciences) requires three steps: designing the research strategy, performing the experiments, analysis of the results/writing the paper. If a person took a significant part in two out of these three steps, they deserve an authorship. This principle eliminates the PIs who provided only advice and funding, and technicians who performed routine laboratory tasks, unless they were instrumental in designing and/or writing up the paper. It also eliminates collaborators who provided previously published materials without being engaged in the current research and people who simply helped with writing (they will be, of course, thanked in the Acknowledgments section).

  20. Journals can easily “solve” this “problem.” by adopting a simple rule:

    Any paper with more than 50 “authors” (pick your number) will be placed, without peer-review (with 50 authors you need and can find “peers?) into the public domain, the authors listed strictly alphabetically.

  21. Jerry did of course try to get his teddy bear as a signatory of one of his articles. IIRC the journal wouldn’t let him get away with it.

    1. Shine, C., N. Shine, R. Shine, and D. Slip. 1988. Use of subcaudal scale anomalies as an aid in recognising individual snakes. Herpetological Review 19:79.

      C. Shine and N. Shine are R. Shine’s dogs…

  22. I am astonished by all the colleagues who seem to think that it is even remotely possible that >100 people have contributed enough to a paper to be authors.

    Somebody should be an author because they have contributed something substantial, because they contributed something scientific, and because they participated in writing the paper. Helping to build a machine or being the lab leader should, on its own, not be enough. In some cases we are talking about papers here where each author would have been able to write two words on average, tops.

    That is just completely ridiculous. I know that in many more competitive areas the boss is always on the paper even if they don’t have the foggiest idea what it is about. I know that there are vast differences between areas of research, with some thinking that everybody who as even one co-author must be too incompetent a scholar to write a publication on their own and others seemingly unable to publish anything with less than at least twelve authors. But arguing “that is the way it is in particle physics” does not cut it. If something is indefensible it does not suddenly become defensible just because we play the “it is their culture” card, if you know what I mean.

    1. Helping to build a machine or being the lab leader should, on its own, not be enough.

      Wait, so working out a collider’s design is not more considered “scientific” enough? Would you kindly define what then is this stuff real “scientists” do that a mere physicist trying to optimize a collider design to get the best possible experiment is not doing?

      1. Easy: (1) Have the engineer(s) publish a paper on the technical solution you found for whatever design problem there was; (2) do not have the engineers on the papers using the machine to test hypotheses that some of them potentially don’t even understand. (Not saying that what I describe is the actual situation for an actual consortium in real life; just as an example.)

        I mean, I could not have done my work in the lab without the invention of the thermocycler or the sequencer, but the people who invented them or the people who built them don’t get to be on my papers either.

        1. Now suppose that you had an idea that couldn’t be achieved with any of the off-the-shelf machines available, but a highly skilled colleague designs and builds a bespoke machine specifically to achieve your aim. It works, and you write the resulting paper. Would you really not want your colleague as a co-author?

          1. I would assume that the designers and builders of all those bespoke machines should be authors on any papers describing the specific problem, engineering solution, and any tests carried out to show whether the machine works.

            …or whatever seems appropriate to the people directly involved in writing the paper, of course. If they all sign an appropriately worded statement that they contributed, and do not object to inclusion of other named authors who may have done more or less.

            Putting dear old Cooper Shine on as first author is more egregious (shame on you Rick!) and probably wouldn’t get past many journals these days.

    2. The thing you’re missing is a case that this is “indefensible”. In experimental physics, being an ‘author’ of a paper means you contributed, as a scientist, as part of a team on the building of, running of, and analyzing data from a machine. For large modern experiments that involves hundreds or thousands of people. This is well understood by people in the field. It’s not like a prospective employer looks at a publication list and says “Ah, I see you were an author on the 2011 results at the LHC. How long did it take you to build it?” Of course, it’s often just cited as, e.g., “the CMS collaboration”, but the full list of what that means has to exist somewhere.

  23. Further to Larry Moran, by the way, how about another suggestion to perhaps reduce the number of authors to sensible figures?

    When it comes to evaluating people, one should simply divide the paper’s worth by the number of co-authors. So if you are a taxonomist publishing a single author paper in the Journal of the Piddlecreek Natural History Museum with an Impact Factor of 0.3, you get 0.3 brownie points. And when you are one of a hundred authors of a Science paper, Impact Factor fluctuating around 30, you also get ca 0.3 brownie points. Same if actual citations of the papers in question are the metric.

    Sounds fair to me, and should change the incentives for gaming the system quite quickly.

    1. Sounds fair to me, and should change the incentives for gaming the system quite quickly.

      I am far more astonished at colleagues so very eager to assume bad faith. What makes you think the decision of the LIGO group to include all participants in a large experiment (in which no experimenter might have had a contribution substantially larger than another experimenter) was made in order to “game the system”? Why is that your null hypothesis?

      1. This was more referring to situations where the boss gets their name on every paper even though they made no contribution, or where people supply the slightest bit of help and then expect co-authorship even though they are too lazy to even proof-read the manuscript.

        But I stand by what I wrote above: if you have a manuscript of 4,000 words with 200 authors, there is simply no way all of them could have made a sufficient contribution to the manuscript to merit being listed as authors. That is what acknowledgement sections are for, as others have repeatedly pointed out (and, in the case of technical staff without scientific knowledge, salaries, one could add).

        1. if you have a manuscript of 4,000 words with 200 authors, there is simply no way all of them could have made a sufficient contribution to the manuscript to merit being listed as authors.

          A bald assertion doth not an argument make.

          How does the word count come in at all? This is not a novel we are talking about. A large experiment at a place like CERN or even an experiment like LIGO is an affair, where at the end of the day, it might be hard to pin point whether contributor X_1 did more work than contributor X_2. In this case, the only possible solution to me seems to be to just say the author is, say, the “LIGO consortium” and list the people who form that consortium.

          Anybody who is foolish enough to count a contribution to a large experiment like this to the same extent as a single author breakthrough is probably the one who has a problem with the system (or perhaps has the wrong “bean-counting” system altogether).

      2. Also, it is definitely a way to game the system. If a lab A of ten people has a policy of putting everybody on every paper even if they only stood around in the background and smiled at the principal investigator, and another lab B has the PIs publish the papers alone, and their evaluation is conducted on a “look, she has a publication in PNAS!” basis, then the scientists from lab A will be evaluated ten times better than those of equally productive lab B.

        And scientists aren’t dumb, we generally realize what we have to do to get the most out of what we do under our area’s evaluation and hiring policies.

        1. For what it is worth, places like CERN and LIFO are not 10-people garden variety experiments. They are experiments requiring expertise of thousands of people working on several different things. Are you trying to say that the only science worth doing is where every contributor understands every aspect? If yes, then sorry, that age is long past.

          Also this:

          and their evaluation is conducted on a “look, she has a publication in PNAS!” basis, then the scientists from lab A will be evaluated ten times better than those of equally productive lab B.

          … is a problem with the “bean-counter” system of publications. A more thorough (and less lazy) method of evaluation will factor in the fact that the experiment behind the paper was conducted in collaboration with a 1000 others. Why should the blame of the evaluator’s lousy work-ethic fall on the experimenter? If you are going to say that that is the way the world work now (and I am afraid there is some degree of truth to that)
          then a more pressing ethical problem is to combat this lazy and silly method of evaluation.

      3. Unthreading a bit…

        A bald assertion doth not an argument make. How does the word count come in at all?

        Because we are not talking about contributors here, which logically would also include accounting staff managing the grant. We are talking about “authors” of scientific papers. And word count enters because, again, in the example above every author would have written 20 words on average. Either that would results in an incoherent, unreadable mess, or >95% of the people on the list cannot possibly really have been authors.

        For what it is worth, places like CERN and LIFO are not 10-people garden variety experiments. They are experiments requiring expertise of thousands of people working on several different things.

        The same could be said about every area of scholarship. The historian writing a book on the Roman Empire could not have done their work without library services, and would perhaps not have come to work without the bus driver. That does not mean that the librarian and the bus driver have contributed substantial scholarship and writing. This is what we call “division of labour” and it has been around since the Neolithic.

        Out of curiosity, in your world, what are acknowledgement sections for? The funding agency and then nothing?

        Are you trying to say that the only science worth doing is where every contributor understands every aspect?

        Of course not. I am trying to say that the only people who deserve authorship are the ones who understand the topic the paper is really about and have participated in writing the manuscript. This is not a difficult concept to grasp.

        It is even highly problematic if you do not understand the paper, and are just on it because you fiddled around with a tiny sub-problem or gathered 0.3% of the data: What if it turns out that the principal investigator, the one who coordinated everything, manipulated the analyses and committed fraud? Being an author, you have essentially put your good name behind the conclusions because you were unable to notice the fraud, and that may fall back on you.

        Anybody who is foolish enough to count a contribution to a large experiment like this to the same extent as a single author breakthrough is probably the one who has a problem with the system (or perhaps has the wrong “bean-counting” system altogether).

        Excuse me, but I think you will find that that was precisely the thrust of my comment above that you originally objected to. So complete agreement there.

        The problem is that at the moment, people make these skewed comparisons. The taxonomist in my example is looked down on because he never publishes in anything above an impact factor of one point something, but the colleague in the next building gets tenure because his name is on a couple of high-impact papers, no matter if it was as one of fifteen authors. People see 10/15 as more prestigious than a 1/1, for some weird reason.

        1. “We are talking about “authors” of scientific papers. And word count enters because, again, in the example above every author would have written 20 words on average.”

          The word “author” is not a synonym of “writer” — and that is where much of the comment on this thread is going wrong. I’ve posted a dictionary definition of “author” up-thread.

          You can be an “author” of a paper (meaning involved in its creation) even if you don’t contribute any words or any writing.

          As I ask up-thread, why privilege writing over the many other tasks in a scientific project that lead up to a paper?

          1. I would turn it around and say that if somebody has contributed something substantial and scientific to your study, you are honour-bound to offer them a co-authorship – under the condition that they participate in writing. At a barest minimum, proof-reading the manuscript and saying, “yup, this makes sense to me, I think it is okay, and I will sign it”.

            Writing the manuscript is crucially important. Understanding what you put your name on is crucially important, for the reasons explained above.

        2. We are talking about “authors” of scientific papers. And word count enters because, again, in the example above every author would have written 20 words on average.

          You seem to be contradicting yourself here. So now, scientific contributions are less important for being credited as an “author” of an experiment than the act of actually writing the results down?

          Because we are not talking about contributors here, which logically would also include accounting staff managing the grant.

          That’s completely beside the point, and irrelevant. I made it very clear what kind of “contributions” I was talking about: actually scientific expertise going into the design/analysis of experiments.

          That does not mean that the librarian and the bus driver have contributed substantial scholarship and writing. This is what we call “division of labour” and it has been around since the Neolithic.

          Again, the same non-sequitur as above. If you think the librarian contributed as much as the guy who figured out how to analyse 1 TB of data, then there is something wrong with that way of assigning credit.

          Out of curiosity, in your world, what are acknowledgement sections for? The funding agency and then nothing?

          In my field (which is not phsyics), authorship issues were settled long long ago, for most intent and purposes. We follow the Hady-Littlewood rule, and ours being a mostly theoretical discipline, it is seldom the case that more than 4 people can claim to have contributed to a given result. We do sometimes have HUGE groups in related fields (see D. H. J. Polymath for example), in which case they are credited as such.

          The problem is that at the moment, people make these skewed comparisons.

          Yes, I agree, that is the main problem. But then why not address that problem, rather than this particular non-problem (I am referring to large projects at CERN and LIGO etc here, not to PIs putting their names on papers for no reason except funding the proposal).

    2. I think this prorating is *exactly* what happens already, qualitatively, if not quantitatively. As I said earlier, only my mom is fooled into thinking that my being one of 300 authors on a genome paper in Science or Nature is a big deal. No one else is fooled even a little. As you say, scientists are not dumb. No one looks at my CV and says, “Look, five papers in Nature and Science! This guy is a hot shot!” Anyone actually in the field will mentally adjust those line items on my CV down to almost marginal status. You don’t get a post-doc, faculty position, and certainly not tenure, based on papers like this. For those purposes, five consortium genome papers in Nature do not, I think, typically come close to even one first author paper in an IF 3.0 journal. So I think that part of the supposed problem already solves itself.

      1. My personal experience does, unfortunately, include people who behave exactly the way I describe and systems of evaluation that work the way I describe.

        (Although admittedly, the university at which I got my PhD had a very funny system in its internal evaluation: first author, assume to be the one who did the actual work, got 33%, last author, assumed to be the prof, got 33%, all other authors got the remaining 33% divided by their number.)

Leave a Reply