Placing credit where it is due

Apr 14 2011 Published by under Careerism, Conduct of Science

I've had a few interactions lately that have led to some pondering on the attribution of academic credit for papers. It all starts with the hilarious gyrations that promotions and tenure committees occasionally, maddeningly, go through.

As most of my readers are aware, in biomedical sciences we have a system in which there are four key elements to placing credit for a given academic paper.

1.) The first author. This person is generally a trainee and generally assumed to be the person who did "most" of the real work on the paper. This can be defined in many ways, from conducting the bulk of the experiments to the most important experiments to the drafting of the manuscript. Defined, that is, by the research team itself when determining who is deserving of the first authorship. Once the paper has been published, the assumption about the role of the first author is more nebulous but no less firm in attributing the academic credit.

2.) The senior author. Often the last author, often the primary mentor of the first author and often the PI of the grants identified as supporting the project.

3.) The "communicating author". Most often the senior author, less frequently the first author and very infrequently someone else*. If the senior author, this is just a reinforcing stamp on his/her seniority, particularly if there are several relatively senior people with their own laboratories contributing. If the communicating author is the first author, this can be an indication from the research team that the project is really all under the intellectual domain of the postdoc or graduate student in question. The senior author is saying "no, really, it was all my brilliant postdoc and she should get all the credit. ps, email her for reagents or mouse lines, not me".

4.) The grants identified as supporting the project and, by extension, the PIs of those grants.

P&T committees frequently find themselves parsing academic credit schemes not just from the biomedical perspective, but also from alternate academic traditions in which the number of other authors matters more than it does in biomedical disciplines, where the senior author is the first author, where single-authorship is important....or there is variance in other minutia. This can in itself be infuriating, after all, how hard is it to recognize that the tradition you trained in is only one of many equally arbitrary crediting schemes?

What kills me though is when they look at the senior BigCheez biomedical types who are up for promotion to Full or "University Professor of Swangage" or some such and say "Dude, there is no way you actually wrote the fifteen manuscripts you listed in the past year, we gotta divide this up somehow", while simultaneously looking at some poor schmuck Assistant Professor or Research Professor of -ology or whatnot and saying "Hrm, hrm, hrm. You have this senior author on all of your papers so we simply cannot, CANNOT I tell you, infer that you are operating as an independent investigator and you must simply be the tool of the BigCheez". A lovely little Catch-22.

Next, I've been pondering the attribution of credit from the grant-game side of the fence. Remember the post at the NIGMS blog which presented data on the number of publications per investigator, broken down by the total amount of NIH grant funding? Look at the second figure and the jump between one grant (~$200K in direct costs) and two grants (~$400K). The mean number of publications goes up by maybe two but look at the intraquartile range. The 75th percentile nearly doubles. I notice that the people claiming that productivity doesn't scale with the number of awards are talking mostly about the mean, not the range. And the trendline is ascending, not leveling off (as does the mean). Part of this may be, however, because some PIs, particularly those with reasonably basic research approaches, just go ahead and list every grant they have in hand on each and every paper that goes out.

The reason to do this, of course, is that when the NIH award is up for competitive renewal the PI lists all the resulting publications with the goal of looking "incredibly productive". And this works out exactly as planned, in my grant reviewing experience. The reviewer already has a Gestalt impression, in many cases, that the PI in question is "productive" because they notice the papers keep coming out. They review the Biosketch and the Progress Report listing the publications associated with the grant in question and say to themselves "wow! looks great". A perception of fabulous productivity covers a lot of ills in the actual grant proposal itself. Let me tell you though, the vast majority of reviewers are not calculating machines who attribute fractional credit based on how many grant awards are listed in the Acknowledgements section. Nor do they** look at the content of each publication to determine exactly which figures were attributable to the grant under current review. And of course, this reality screws the PI who has only the single award for whom each and every manuscript really was supported by the grant in question in its entirety.

I arrive at the ethical question from the PI perspective. What is fair? What's the minimum amount of time an award is active before it can be put on a paper? Or what sort of expenditure? Do you have to have paid some sort of research bill from it or is staff time enough? What about if it is only while you are revising a previously-submitted MS?

We all have to make our own calls, of course, and the point I made above about making the competing continuation look as good as possible is no joke. My standard is a loose one: I should have some plausible connection between the goals of the grant in question and the data in the manuscript. So yes, I have definitely submitted manuscripts that did not list all of my active awards. And I think that I've listed awards on papers only for which the connection was reasonable...but I'm sure opinions would vary on that one.

But how about start dates? This intrigues me. I have definitely seen competing continuation applications come through listing support for an article that as published 6-9 mo after the start of the award. In society level journals that take nearly this long from original submission to publication of the article. So I know dollars to doughnuts that no real research expenditures were being made on the immediately past interval of support. But the team may have been polishing up or revising the manuscript, right?

Personally, I don't think I have any minimum standard for the dollar contribution of a given award to the manuscript in question. My effort is primarily brain effort these days so, for example, if I'm working on writing or revising or fighting reviewers/editors, then I'm working on that manuscript. So technically, if the award is active and I'm having my effort paid from it, then that support is contributing. If I come up with some brilliant ideas that generalize across several of my research programs, that is a contribution. If a member of my lab is working up an assay or technique that we can apply to multiple projects, that contributes. I see no practical way to assess how much of a contribution has been made. So it seems like the brightest line we can draw is to determine if the grant is active before or after the manuscript has been accepted for publication.

The single-award laboratory is still considerably disadvantaged by this because of the perception factor. Really, if the grant score hinges on "productivity", we really should be able to divide the pubs by the number of awards or the amount of $$ or something. But I'm telling you, the reviewer mind can't really do this.

The perception of "amazing productivity" of the PI / lab is just too strong.

*a secondary "senior author", most typically. Another PI who appears as an author, perhaps in next-to-last position. Might be PI of one of the awards listed as supporting the project or supervisor of the second author.

**except the occasional jerk trying to make a point whom the rest of the study section promptly ignores as a nutter...

35 responses so far

  • Namnezia says:

    About assigning grants to papers published soon after the start date, I don't think this is necessarily a misrepresentation of the effort on the grant, because some of these papers may stem from preliminary data that you developed specifically for that grant and are experiments listed in the specific aims. Since it can be over a year from the time you first acquire preliminary data and submit your initial version of the proposal and the time it actually gets funded, its normal to expect that a given lab would continue working on this project and not sit around waiting for the proposal to be funded.

  • mat says:

    Huh, Namnezia? It still wasn't supported by the NIH award you are trying to get funded unless your University does some backwards accounting trick...

  • crystaldoc says:

    I am with Namnezia on this one. Some time back either DM or CPP (sorry forgot which) suggested a great model for writing grants where Aim 1 is ~75% done, Aim 2 is ~50% done, and Aim 3 (your "risky exploratory aim") is maybe 10% done when you submit the grant. This may well be what it takes to show "feasibility" and get the shit funded. And it can take way more than a year from the time you submit an initial grant to the time you actually get money-- the resubmission has to eventually get a fundable score, the freaking congress has to pass the freaking budget... but I digress. Chances are you have every piece of data for a pub, even a complete draft submitted to a journal, possibly serial submissions to several journals, before you get the grant bucks that will supposedly have paid for the science. If my (topically relevant) grant is active and paying for my time, for even one afternoon while I wrangle with the online submission site to upload the final resubmission of my manuscript and enter all author affiliations and e-mails for the umpteenth time, I have no qualms about listing that grant as supporting the research. Doing the experiments is only the first half of the research; getting the work written, accepted by the referees, and published can easily represent the other half.

  • becca says:

    Drugmonkey- interesting post. I'd never considered it much, but it seems like you've got a system where there is no good way to avoid 'double counting' publications.

    For example, let's say the world's foremost effort on bunny hopping writes an invited bunny hopping review, broad in scope and focusing on very recent developments in the bunny hopping field. Does it get credited to the grant on 'behavioral factors that affect variable height of bunny hopping in a population' or the grant on 'the effect of molecule X on frequency of hopping'? Why not both grants?

    I was just reading a post on how some pharma companies have taken to billing for using in-house expertise. I understand the desire to want to know what resources are most needed, and whether they are being used wisely. But it seemed like it was creating a really unpleasant culture. I can see logical attempts to fix the credit allocation for NIH grants running up against the same issues. I wish I saw a clear solution. Although ultimately, I'm beginning to wonder if the factors that determine the culture as a whole may depend more on the *amount of slack in the system* rather than *the specific metrics used to evaluate researchers*.

    mat- formally, I think there is a general problem with exactly what NIH award supports preliminary data generation for a new grant.

  • Reason number fuckjillion why you've gotta be a fucken numskull to stop writing R01 applications once you've "got your R01". The single-R01 lab is inefficient, wastes NIH resources, and is destined for the shitcan.

  • Sara says:

    As a managing editor at a bioscience publishing company, I appreciate this breakdown. I was vaguely aware of the positions as indicators of contribution, but not this specifically.

    Interestingly, I've had one paper recently in which the first author is a PhD candidate and the last author (there were only two) is technically the corresponding author, but I've only been in correspondence with the first author. (I.e., the corresponding author forwarded the first author the galleys for proofing and then the first author took care of payment as well.) I found it very odd. I wonder if I should just ask if they want to change the corresponding author...

  • Namnezia says:

    Because the PI doesn't want to deal with all this bureaucratic nonsense.

  • Beaker says:

    I see no way to prevent the practice of attributing multiple grants to papers because, whether or not the practice is justified, all parties involved benefit when it happens. The PI obviously benefits when the grants come up for renewal. The NIH programmatic people like to see that grants in their portfolios have produced lots of papers, and the study section likes to see that a grant they funded produced lots of papers (and should therefore be renewed).

    Having an unbiased person attempt to fairly match funding mechanisms to papers attributed to them is a big can of worms because the actual research that got done is usually not what was proposed. None of the parties mentioned in paragraph 1 want to go digging in the dirt to see see if the proposed research actually got carried out as described in a proposal 4-5 years prior.

    CPP acknowledges that the practice occurs, and he implies the best way to survive to get more grants so you can do it too! True that. But the fact that the practice is common provides no data either way to test whether a one-R01 lab is inefficient. I guess his point is that one-R01 labs are inefficient because when that inevitable break in funding occurs, people lose jobs, the lab progress slows or stops, and then it must be re-started if/when new funds arrive.

    The best we can do is figure out how to remove the bias against one R01 holders when we assess productivity/grant. For example we could allow only one R01 to be attributed to each paper when productivity is assessed. This would also produce a fake view of reality because multiple grants often DO support single papers. Moreover, this would require somebody to troll through all of the papers of the investigator and do the adjustment manually--probably not gonna happen.

  • Kaija says:

    When I was a grad student, my PI always wanted to be the corresponding author no matter what; he felt it was his place? job? as head of the lab. In truth, as Namnezia said, he really didn't want to deal with the back and forth required for receiving and responding to reviewer and editor comments, revisions, replotting of figures, approval of proof copies, etc, and it was a right pain to hound about "have you gotten anything from _____ journal in your email lately?" because if we didn't, important publication-related correspondence would get "lost" in his 200+ emails a day.

  • anon says:

    Re the corresponding author:
    I have previously been first author on a paper as above, writing the paper, doing revisions, and handling all publication correspondence etc etc while the 'corresponding author' on publication is the senior PI. In short, he only wanted corespondence from other scientists who might invite him to conferences or discuss the big picture science. I felt like a total lackey.

  • Pinko Punko says:

    I have heard that very occasionally the stuff all the grants with every paper come back to bite the PI, but perhaps the two grants were in the same section so it was very clear that all papers were listed on both and that the productivity did not match.

    Also, the tenure stuff is really maddening because some people would count papers with the post-doctoral advisor as zeroes. Just so dumb.

    As for putting papers on a grant that are published soon after the award- I don't see the problem in this. If the grant paid for the publication charges that is partial support of the work (and assuming the subject matter of the paper is appropriate to the grant).

  • Pinko Punko says:

    u r such a trollz

  • odyssey says:

    Also, the tenure stuff is really maddening because some people would count papers with the post-doctoral advisor as zeroes. Just so dumb.

    Such papers don't count, and shouldn't, because we want to see if someone is independent. Continuing to publish with a former advisor once you have your own lab* just gives the impression you're running a satellite lab of that advisor.

    * Except those manuscripts you didn't finish writing before taking your TT position. Don't spend too much time on those - they won't count towards tenure.

  • whimple says:

    Except those manuscripts you didn't finish writing before taking your TT position. Don't spend too much time on those - they won't count towards tenure.
    This is incredibly bad advice. What really counts for tenure is extramural funding and these easy-to-get-out-the-door papers give you productivity momentum on your first independent grant application (and something to do while all the new equipment is arriving). The ability to secure significant extramural funding is proof of your independence (and competence) so worrying about what the P&T committee might or might not want is a waste of time. Get all the papers you can, get funded as fast and as much as you can. That's all you have to know.

  • physician scientist says:

    At schools where you have to support 90-100% of your salary (+ fringe), most of us need to have 2+ grants supporting our salary. Given that one cannot realistically break down time into distinct periods when you only think about one project, I think its actually true that multiple grants support one manuscript.

  • anon says:

    I have a paper with my former post-doc supervisor in which about half the work was completed in my own start up lab. I was the first and corresponding author, the post-doc super was the senior author. I acknowledged that my own funding supported the work (which it did). But fuckin NIH study sections refuse to give me any credit for this. I initiated that project in the post-doc lab with the understanding that what stemmed from it was what I could take with me. I thought that's what post-doc training was for. I had plenty of other papers from that lab - some in glamour mags, but the supervisor made it clear that she wanted to hang on to that subject matter. That didn't phase me, since there were plenty of other things to pursue.

    In retrospect, should I have asked her to withdraw her name from that paper? It didn't feel right to me at the time, since her funding supported half of the work.

  • drugmonkey says:

    I object to the notion that publishing *with* a former advisor is 100% correlated with "running a satellite lab of that advisor". Also that this is categorical evidence for/against independence.

    One can have not an original thought in one's head, start up a new lab as a direct offshoot of the previous lab and look "independent" simply because the name isn't on the author line?

    Conversely, how many situations do we know of where the senior trainee is doing the heavy lifting, intellectually and otherwise, on a domain of research in the BigPI's lab....a domain the BigPI is only passingly familiar with? Yet because of grant issues and creation of the intellectual milieu it is entirely appropriate that the BigPI be on there as senior author. Nobody that is actually in the field is confused one bit about what the "independent" contributions are. So why do P&T committees have to punt on using their brains and downshift to bureaucratic formulae?

  • drugmonkey says:

    By "refuse" do you mean when you soberly and calmly make this point about ownership in the revised application they comment that they still don't buy it?

  • odyssey says:

    DM, I don't disagree. Obviously one can collaborate with a former advisor but still be independent. Or not. That's why I was careful to use the phrase "gives the impression." I was perhaps a little over the top by saying such papers shouldn't count.

    And P&T don't have to punt, but you know as well as I do that they often do. Consequently I would advise people on the TT to not continue to publish with former advisors.

  • odyssey says:

    Whimple, I didn't say don't finish those papers. I said don't spend too much time on them. There is a difference.

  • anon says:

    Where can I make this point? Is it common to write authorship contributions and/or support in the biosketch? thanks.

  • DrugMonkey says:

    well if you have gotten the criticism already, the intro to the revised application is definitely one place.

    the new personal statement in the biosketch provides a place for new-ish investigators to detail their independent contributions within the postdoc laboratory, I would think.

    judicious phrasing in the preliminary data section as well, particularly if you are referring to some of your work from the prior laboratory.

  • TeaHag says:

    I think that this issue could be addressed in the personal statement component of your biosketch. "I'm the perfect person to perform this research because this is an outgrowth of the post-doctoral work I performed in Dr. x's lab, which has evolved to the present, blah blah... fill in section here with how your insight and perspective is different from hers... and which resulted in my obtaining XXX fellowship (or whatever) leading to [current state of play] and publication. Whether that would work with study section... as unknowable as anything else!

    Be very wary of ever leaving off people who supported research at a financial level. They may not only have supplied the money, but also the authorizations for animal work, access to clinical samples or materials shared under an MTA.

  • drugmonkey says:

    a great model for writing grants where Aim 1 is ~75% done, Aim 2 is ~50% done, and Aim 3 (your "risky exploratory aim") is maybe 10% done when you submit the grant.

    That sure as hell wasn't me making this cockamamie suggestion...

  • bob says:

    That's why it's normally written "to whom correspondence should be addressed". It signifies who you should contact about reagents or with questions about the work, not who filled out a form for page charges.

    That's also why being "corresponding author" can confer some extra credit. It says that person is primarily responsible for the work, not that they put in some extra admin effort.

  • drugmonkey says:

    but for better or for worse, this tag has also evolved to represent something a bit beyond the person to email.

    back in the day, I think this tag was mostly about "who do I ask for a reprint", when this was much more common and there was no such thing as online pdf repositories for each journal.

    The lab head generally took on this duty because s/he could be presumed to still be at that address for years to come, whereas trainees would be off to newer pastures.

    Because of this, the "corresponding author" thing started to mean "the most important big cheeze amongst a list of authors that may include other big Cheezes". So then this became a matter of additional credit.. and therefore it was important for PI's to keep stacking up the "communicating author" pubs.

    It all makes my eyes bleed, even though I sort-of get how it all evolved.

  • Namnezia says:

    Plus I've known people who spend a huge amount of their research effort on finishing those old papers, at the expense of publishing independently, and have gotten themselves into trouble. So yes, publish as much as you can - but make sure you have several papers from your own lab, without your postdoc mentor.

  • Namnezia says:

    I would have asked her to please consider removing her name, because no matter what you say some people will never be swayed if they see your former mentor's name on there. I had a similar experience. If she really would like to keep her name associated (for example she invested quite a deal of time and money), then my advice is to publish this with as little extra work as possible from your own lab, maybe in a lesser journal, and then start the next phase completely independently in your lab.

  • crystaldoc says:

    In relation to the nature of the aims themselves, one important consideration is that the aims should be posed in such a way that each aim is at least partially independent of the others. This is because a typical reviewer knee-capper is "If aim #1 is not successful, then none of the other aims can proceed". My practice, when possible, is to have at least 75% of the experimental goals of aim #1 already complete and in the preliminary studies. Aim #2 can be not at all done, so long as the approaches are well-established and pretty-much guaranteed to work. Aim #3 is an opportunity to get a little more speculative and wide-ranging.

    OK, so it was CPP and I didn't get it exactly right, but not so far off and makes the point-- if aim 1 was 75% done when grant was submitted, resulting paper should not be far behind.

  • Pinko Punko says:

    One million huzzahs to DM here.

  • GMP says:

    About 6 months into my tenure track, I had a heart-to-heart with my PhD advisor (I didn't do a postdoc) and told him that I had been told by the folks at my new university that it would look bad for my tenure case if I continued to have papers with him for too long. We decided we'd wrap up one more paper with his name on it and that would be the end. No one was upset or disrespected. Well-meaning advisors realize you need to sever the umbilical cord (both in reality and on paper), and the person whose tenure is on the line needs to make sure (sooner rather than later) that everyone understands what must be done and why.

    I did a similar thing regarding severing co-authorship with collaborators: a student came to me from a friendly group, with an MS and a piece of code developed there. For the next three years we did a lot of new work that somewhat relied on the code he had brought and his previous advisors were on all the new publications. At one point their involvement became purely that of looking over the final manuscript and catching typos (worth an acknowledgement, not a co-authorship); that's when I exchanged a couple of emails with them saying that I thought perhaps it's time to part ways as we are moving too far from the topic where their original contribution lied. They were both OK with it, one even acknowledged that for a while he had been uncomfortable with essentially a courtesy co-authorship.

    A case in point of the PI being a corresponding author: the first author (student or postdoc) simply won't be at the current email address permanently, whereas the PI typically will. Sometimes students/postdocs feel robbed of ownership of the paper if they are not listed as corresponding. In those cases, whenever possible (often in my field) I request that both the student/postdoc and me, the PI, be listed as corresponding authors.

  • rknop says:

    Note that authorship sorting is partly a convention of the field. In large particle physics experiments, the first author is the author (out of potentially several hundred) whose name is first in the alphabet. The one who did most of the work on writing the paper itself is buried somewhere in the middle, and of all the authors, there is tremendous asymmetry as to how much they did or how central they were to the multi-year (or even decade) process that went into designing the gigantic detectors and processes for analyzing the data that led to this particular paper. That community depends on word of mouth (truthfully) to keep track of who really deserves credit for whatever scientific work went into a given paper.

  • ADA says:

    Odysey you should not give advice about situations you don't know about. Your university P&T policies apply to you not the entire internet!

    In my university, you write a paragraph about every paper explaining what you did on each paper. They are actually used in preparing your file. I had 6 papers in press or in prep with big cheezes in the middle in my last file,( me at the end, and students first) and it's been no problem to just say the cheezes just provided the experimental equipment and this is my line of research. I am going to continue publishing with my postdoc collaborators. I was already pretty independent when I went there and the collaborations did nothing but enrich my research.

    At another university, a friend had to wait until after tenure to finish a series of papers with her PhD advisor, which are now finally coming out after 6 years. They told her the papers wouldn't count, and even told her they would hurt. She's very successful, with awards up the wazoo, and amazing productive research program. I don't see what the university gained by delaying the publication of some good science.

    Bottom line is, there is no one answer to this. People have opinions about it, but there is a lot you can do by knowing what's expected and communicating your role in the work clearly come tenure time.

  • [...] Placing credit where it is due [...]

Leave a Reply