Views on authorship and author contribution statements: poll results, part 2

As promised, here are the results of our recent reader poll on author contribution statements. See part 1 of the results for respondent demographics and their views on authorship.

Part 1 revealed widespread disagreement about what contributions merit authorship. Authorship standards also are changing. And science is becoming increasingly collaborative. All of which would seem to be arguments against the traditional practice of trying to summarize credit and responsibility for a paper solely with an ordered list of authors, and in favor of author contribution statements. That’s been my thinking for a few years: I like author contribution statements and routinely include them on all my papers, even if the journal doesn’t require one.

But it turns out there are widespread mixed feelings about author contribution statements. Which I now share…

Most people (87%) only include an author contribution statement when the journal requires one. Didn’t realize that including them in all my papers put me in that small a minority (10%).

When I write an author contribution statement, I always write a draft and then circulate it to the other authors for their approval and any needed corrections. I had naively assumed this was how everyone did it. But a comment from Joan Strassmann (sorry can’t find it now) prompted me to ask how author contribution statements are written. I’m glad I asked, because there’s substantial variation in how authorship contribution statements are written. My way–with input and agreement from all authors–is most common but not the majority (40%). 20% of respondents said “it varies” in their experience. 18% said that in their experience the first author writes the statement without input or agreement from the other authors. The rest didn’t know or gave some other answer.

Most respondents (82%) think author contribution statements should be written with input and agreement from all authors. Which I was glad to see, since I struggle to see a principled argument for any other view. 12% said the first author should write the statement and that there’s rarely a need for input or approval from other authors, and 1% stumped for the senior author.

There’s disagreement on how accurate author contribution statements are. Only 41% of respondents said author contribution statements are always or usually accurate, in their experience. 24% said they’re sometimes accurate, 6% said they’re rarely or never accurate in their experience, the remainder weren’t sure. This surprised me. Until I read Joan Strassmann’s doubts about the accuracy of author contribution statements and saw the poll results, it had never occurred to me to worry about the accuracy of author contribution statements. Though I do wonder how much of the concern here is really with vagueness as opposed to accuracy (recognizing that the line between the two is fuzzy). Which brings us to the next question…

Most people have some criticisms of author contribution statements, but different people have different criticisms. I listed various real and hypothetical criticisms of author contribution statements, asking respondents to check all those they agreed with. Only 34% of respondents didn’t agree with any of the listed criticisms, but the remainder varied on which criticisms they agreed with. Many of the listed criticisms are hard to compress into labels on a graph, so I’ll list them:

  • 36% said author contribution statements are vague/imprecise
  • 28% said they encourage credit for a paper to be spread more evenly than it should be spread
  • 20% said they legitimize authorship for non-authorial contributions and encourage over-expansive author lists
  • 16% said they’re a tool for PIs to claim undeserved credit at the expense of trainees
  • 16% said they’re inaccurate (aside: that’s much less than the 30% of people who, in response to a previous question, said author contribution statements are sometimes, rarely, or never accurate. So a large chunk of that 30% presumably are concerned with vagueness rather than inaccuracy.)
  • 14% said they don’t add useful information to the author list
  • 14% said “they don’t tell me what I want to know”
  • 6% said their widespread adoption is inhibiting adoption of other, better ways of apportioning credit for scientific work

There’s disagreement on how useful and informative author contribution statements are. 36% of respondents find them useful and informative–a significant improvement over just looking at the author list. 40% find them useful only in certain circumstances, such as for papers with many authors or interdisciplinary collaborations. 18% find them rarely or never useful and informative. But that still leaves open the question of exactly what people find author contribution statements useful for. Which brings us to our next question…

People mostly don’t read author contribution statements for purposes of apportioning credit and responsibility. Rather, they mostly read them to satisfy personal curiosity–or not at all. When asked to check all the reasons for which they read author contribution statements, by far the most common choice was “when I’m curious to see who did what on a particular paper” (68%). Second was “when I want to ask a question about a paper and need to figure out who did what” (27%). No other reason for reading author contribution statements was chosen by more than a small minority of respondents. In particular, less than 10% read them when evaluating applicants for jobs, grants, scholarships, etc., not even just to help distinguish among the most competitive applicants. Even if you allow for the fact that it’s mostly only faculty who ever have to evaluate applicants, that’s a small percentage. And people who rarely or never read author contribution statements (21%) outnumber those who read them for every paper they read (12%).

Finally, there’s little support for the alternative of specifying each author’s overall percentage contribution to the paper, with the total summing to 100%. Which I was glad to see, because I don’t think it’s useful to try to be so precise about something that’s necessarily somewhat nebulous. 45% of respondents don’t want to see journals go this route. 22% would want it only if there was some standardized way of doing it (aside: good luck getting agreement on how to standardize that, given how much disagreement there is on who should even be listed as an author!). 16% would like it only as an option, not a requirement. 7% want to summarize author contributions in All The Ways and want to see percentage contributions listed in addition to author contribution statements. Only 2% want to see statements of overall percentage contribution replace author contribution statements.


This poll really made me stop and think, which is great. Previously, I’d been strongly in favor of author contribution statements but hadn’t really thought about why. I just had a vague sense that, because author ordership practices vary and are changing, it made sense to stop relying solely on author order to convey information about author contributions. But it turns out almost no one wants the extra information that author contribution statements convey, except for unimportant purposes. In light of that, I’m no longer in favor of author contribution statements. Because I’m sorry, but there’s no need to have formal statements of author contribution just to satisfy the personal curiosity of readers about who did what. And if you have a question about a paper, you can and should just contact the corresponding author, rather than reading the author contribution statement to try to figure out who to contact. That’s what corresponding authors are for. Conversely, the people who do have an important reason to care about apportioning credit and responsibility for scientific papers–those charged with evaluating applicants for jobs, grants, scholarships, and fellowships–mostly don’t read author contribution statements. Not, I would hasten to add, because they’re lazy or bad at their jobs–they’re not! But because they already have many other sources of information about the applicants they’re evaluating, to which author contribution statements add little or nothing.

It’s funny: until forced to stop and think by my own poll, I hadn’t realized that I’d never read any author contribution statement besides my own. Which tells you (and me!) all you need to know about how much I actually valued author contribution statements, as opposed to how much I said I valued them.

So I’ve changed my mind about author contribution statements. I’ll probably stop providing them unless they’re required, depending of course on what my co-authors want. Because there’s no point to them right now. Maybe there will be in future. Maybe some day everyone will use them instead of author lists and other sources of information to apportion credit and responsibility. And I suppose you could argue that since it doesn’t cost much time and effort to write author contribution statements, I ought to just keep doing them so as to be prepared for–and perhaps help bring about–the day when they serve some important purpose. To which, meh–I’m not an early adopter.* I’m not one to do something pointless in the hopes that one day it will cease to be pointless.

*I got my first smart phone last week, having previously used a cell phone more primitive than the ones in The Matrix. I also still purchase DVDs (not even Blu-Ray; DVDs). Insert your own Luddite joke here.**

**Brian’s a self-described Luddite too, making this blog 2/3 Luddite. Meg’s the only one here who tweets, uses Github, wears an Apple Watch, has a 3D 4K tv, and pre-ordered a self-driving electric Tesla car. Well, I assume she does all those things. As a Luddite, I see people like Meg*** as time travelers from a distant future. Every time I see her, I keep waiting for her to say this.🙂

***normal people

24 thoughts on “Views on authorship and author contribution statements: poll results, part 2

  1. Very interesting results! Following up on the bit about specifying each author’s percent contribution to a paper, with the total equaling 100%: it seems to me like that would be impossible to do, especially at the level of precision that having something on a scale from 0-100 implies. Let’s assume a paper has three authors. Authors 1&3 designed the study, but it was based on a study originally proposed in a grant proposal by Author 3 (meaning Author 3 is the one who secured funding for the work). Author 1 carried out all the empirical work. Author 2 developed and analyzed a model related to the empirical work. Author 1 wrote the first draft of the manuscript, but it was heavily edited by Author 3, with substantial input from Author 2 (especially on the theory portion). How would you possibly assign percentages in that case? The only possible unifying currency I can think of is person-hours, and that seems like a really bad one.

    • I agree that person-hours would be a terrible unifying currency. And that there’s no way to be precise enough about percent contributions to make them worth bothering with.

    • I really wish author contribution statements were more exact. For example, why don’t they just say exactly “Authors 1&3 designed the study, based on a grant proposal by Author 3. Author 1 carried out all the empirical work. Author 2 developed and analyzed the model. Author 1 wrote the first draft of the manuscript, which was heavily edited by Author 3, with substantial input from Author 2 (especially on the theory portion).”

      Part of my cynical side says the reason why we don’t get such statements is that the 10-20 author papers would get trimmed down substantially once the exact contributions were stated. If journals were concerned with space it could be a purely online only component.

      Part of me wonders if people don’t read author contribution statements because this info might be more accurately conveyed in letters of recommendation. If I were deciding a postdoc fellowship for someone who just finished a PhD I would really want to know how many of the thesis projects were initially the candidate’s idea vs the advisor’s idea. Maybe this info is likely in the letter of rec if the ideas were mostly the students’.

  2. If I am parsing this right, we have gone in my career of 20 years or so from a state where we had typically 1-2 authors and you know each made a large contribution to a state where we typically have 3-5 authors and their contributions are totally opaque. I don’t necessarily think being more inclusive on authors is bad innately, but the side effect of opacity of contributions is bad. I have to confess that these days on search committees when I see somebody from a big lab who has their name on 10 papers but is one of 10 authors on each it is nearly impossible for me to assess what they’ve done. I know some people just assume their contribution is zero, but that is not necessarily true and I don’t assume that. Others only look at first author papers, but I don’t like that either (especially as in some of those big labs first author is more about aggressiveness than relative contribution). But I don’t really know what to think anymore. This has moved things from just a straight show me what you’ve done CV to having to rely more on references (at least for early career folks) which introduces all kinds of subjectivity and potential biases.

    And I’m not at all surprised to find that author contributions haven’t solved the problem. I haven’t been on that many papers that include them, but the ones I have they often seem to me rather arbitrary and nobody really wants to have a candid discussion. I think probably the best of the lot are the PNAS style ones where there are five of six tasks (e.g. conceiving the paper, collecting the data, doing the analysis, writing the paper) and everybody who contributed at a stage is listed. But definitely still very imperfect. I totally agree a percent adding to 100% won’t work – its too high dimensional to put on a single scale as Meg said.

    • I agree with all of this. Especially the bit about references being more important than ever. And the bit about the issue being most acute for people from big labs. In a funny way I think the fact that I came from a modest sized lab and had no in-lab collaborations, and then did a “solo” postdoc, worked in my favor. Did I get my name on fewer papers than I otherwise might have? Yes. But nobody was ever in any doubt that I had my own research program.

      • I know more than one person who did their PhD or Postdoc in a large lab who felt like it worked against them. They got their name on more papers. But they felt like (and I think they could be right) that many CV readers basically gave them no credit for any of those papers and mostly assigned credit to the lab PI. But I suspect that varies with the reader – I think it would probably be viewed positively here at UMaine.

        And we have similar histories in that my PhD adviser did not put his name on student papers, so all my dissertation chapters were sole authored. My postdoc adviser also put his name only on genuine collaborations (of which we had several), but not things he had nothing to do with. Net result was I had quite a few sole author papers when I went on the job market, and I do think it helped. Nobody could doubt who wrote those.

      • As you say, what works for one person may not work for another. But I do wonder if there is a bit of frequency dependence here. If people like you and me stand out more these days (in a good way) precisely because we’re rarer than we used to be.
        Recalling that old post on ASN YIA award winners. They don’t publish any more than they used to, they just published with more coauthors. Which makes me wonder a little if people who do lots of within-lab collaboration are really publishing more as a result.

  3. Ana Marušić and her colleagues have done quite a bit of work on authorship and contributorship and how different contribution disclosure policies/procedures affect authorship validity. Including what happens when open-ended vs structured forms are used or when forms are filled out by the corresponding author vs other authors.
    Some of her papers – and others that may be of interest – can be found in the background reading list from the 2012 International Workshop on Contributorship and Scholarly Attribution
    Authorship problems are without doubt the most common things I get asked about, especially by ECRs, and range from everything from misunderstanding which contributions do/do not warrant authorship to authors being added for political reasons to very unpleasant situations where withholding authorship is used as a threat. Having open lab policies/procedures and early discussions in projects and collaborations (as noted in the IWCSA report) can help avoid some of these problems.

    • Thanks for the pointers. I agree that the way to avoid problems is for labs to discuss the issue often, and for collaborators to discuss it in an ongoing way, ideally starting before anyone does any work on the project. The “ongoing” bit is important to head off situations in which somebody says they’ll do X, where X would merit authorship, then doesn’t follow through and do it but still expects authorship.

  4. “almost no one wants the extra information that author contribution statements convey”

    Wait, wait. That might be true now, but perhaps that’s because we’re transitioning from one stable state to another. As Brian points out above, it used to be easy to see that 1 or 2 or 3 authors contributed a lot to paper X. But now there are more collaborations, more authors, and things are opaque. Who did what? Who did little? This is obviously a *problem* and the one that contribution statements are trying to help solve. However, as you blog here, contribution statements don’t seem to be much of a solution. Now.

    But that might not be because of the inherent lack of utility of contribution statements themselves. Consider that it might be instead because such statements are 1. not universally required and so can’t be used as a signal across all papers; and 2. not standardized, so can’t be used as a reliable signal. But undoubtedly things will continue evolving — either such that contribution statements overcome these hurdles, or else something better comes along — because the *problem* isn’t going away, and we as a community need a way to solve it.

    I write contribution statements not because I think my particular ones on my particular paper are important, but because 1. it doesn’t take much work, and 2. I want to signal to other people that this (or something like this) should become a cultural norm, because we need to takes steps to solve the problem.

    FWIW, I just started using the CRediT taxonomy (cheatsheet: ) It’s not perfect, but it’s a step towards standardization.

    • Good fair points. My brief remarks at the end of the post kind of address them a bit, but I’ll elaborate.

      I agree that there’s a problem here that won’t magically go away. But I guess I feel like author contribution statements have been around for a while now. They’ve been required for years at many leading journals. Yet I basically see zero sign in the poll results of movement towards a new social equilibrium in which people care about them for any serious purpose. Which I suppose might just mean that I’m too impatient, expecting practices to evolve unreasonably fast if they’re going to evolve at all.

      Re: voluntarily writing them so as to signal to others that they should become the new norm, I think that’s fine if you’re confident that author contribution statements are and should be the future. But I’m no longer confident that author contribution statements are or should be the future. So even though they don’t take much work, going forward I’ll probably just do them when required and let the evolution of cultural norms take care of itself.

    • Margaret,

      Thanks for the link: I will be suggesting to colleagues that we use it for our papers. I think something like this could be a good solution to the problem. I believe that the reasons you put forth are likely correct when it comes to why such a solution did not receive much support in the poll results.

      If, for whatever reason, contribution statements had gained a critical mass and become a cultural norm a few years ago, and if the leading journals had pushed through a standardized methodology that was taken up by most of the publishers, then I believe that the poll results would have indicated greater use of statements, if not satisfaction.

      • @Ken and Margaret,

        I’m not convinced that lack of standardization is the reason why author contribution statements aren’t used for purposes of assigning credit and responsibility. Maybe it is, but maybe it isn’t.

        I say that in part because lack of uptake of various other new science-publishing-related things. Advocates of those new things routinely attribute lack of uptake to lack of standardization. But in reality that’s usually not the limiting factor. Think for instance of online ms commenting systems. *Very* few people ever comment, and *very* few papers ever receive any substantive comments. Advocates of online ms commenting systems routinely attribute this lack of uptake to lack of a single well-designed commenting system. Trouble is, lack of uptake persisted even after PubMed added a commenting system:

        Seems to me it’s possible that many people just don’t care about the information author contribution statements provide, no matter whether it’s provided in a standardized way or not. If so, I suspect that’s for a couple of reasons. First, people often just want to know who did the bulk of the work, and/or what are seen as the most important bits of the work (whose idea was it?). Information that they can get from the author list, at least if there’s an agreed convention of authorship order that most everyone follows. Second, I think once the number of authors grows past some admittedly ill-defined point, the whole notion of individual credit and responsibility breaks down in a way that author contribution statements can’t fix. Think of the paper reporting the discovery of the Higgs boson, which had a thousand authors or something. Or papers from big genome sequencing consortia with hundreds of authors. Or maybe even some big collaborative efforts in ecology. There comes a point when assigning credit and responsibility to individuals ceases to be meaningful, no matter how you try to do it. It’s like trying to apportion credit and responsibility for Exxon’s annual profits among all of Exxon’s employees. Looking at Exxon’s org chart with everyone’s job description on it doesn’t help at all. And so for better or worse, much of the credit (or blame) tends to go to the CEO and a small number of other high-ranking employees.

        Hmm. That corporate analogy is perhaps worth developing further, might be some useful lessons there. And Brian may have some useful thoughts on the analogy between corporations and larger scientific collaborations, as he has a corporate background.

      • Okay, so, Jeremy, is there really even a problem then? If hiring/promotion/awards committees are getting information from other sources (letters, e.g.) about how people are making their scientific contributions, then do we even need that information in the publications themselves?

        Maybe the “problem” isn’t correctly identified. For me, I don’t like “author contribution statements” because I see the real problem as the problem of who to include as an “author” in the first place. I would much rather have a short author list (of the 1, 2, or 3 people who did the conceptualization, analysis, and writing) and a long and standardized “contribution statement” of *all* people who contributed, regardless of size or type of contribution. This may not seem to matter much to established research faculty. But it may matter to, say, undergraduate and graduate students trying to get small awards. And it definitely matters to natural historians, taxonomists, data managers, and the like who usually are not considered authorship-worthy, but whose work was vital to the work behind the paper. These people usually need to show *contribution* as part of their job, even if their primary job is not churning out research papers. And they’re getting screwed by the current system. A short-author-list-and-long-contribution-list system would provide fair and necessary credit to the worker bees who underlie the academic research system, while making it clearer who did the bulk of the work.

      • Hmm…interesting idea of an author list plus a contributor list. As you say, it might help stop the indefinite expansion of author lists and associated ambiguity of what “author” even means these days.

        Though what’s the difference between a contributor list and the Acknowledgments, where non-authorial contributors ordinarily are named? Just that it’s a list that (presumably) could and would be indexed, thereby making it easy for people to count up and list their contributions on their cv’s?

      • Yep, I want it standardized and indexed. Since folks typically put funding sources and other in-kind support in there, as well, it could get long/messy. So two separate statements makes sense to me. But I’m flexible, if there’s a nice standardized way to include everything in one lump of text.

      • I guess one other thought is that the word ‘acknowledgements’ comes from a time when primary researchers tended to do their own data management/natural history/taxonomic identifications/etc., so I think of the Acks section as a “thank you” section. Thank you for contributing in some *little* way to this paper.

        But these days, a huge amount of work — or vital piece of work — may be done for a paper by someone who doesn’t get authorship. I feel like this is a different sort of thing than just, “oh, hey, thanks.” It’s recognition that a substantial contribution was made by someone, even if that contribution doesn’t merit authorship.

      • “I guess one other thought is that the word ‘acknowledgements’ comes from a time when primary researchers tended to do their own data management/natural history/taxonomic identifications/etc., so I think of the Acks section as a “thank you” section. Thank you for contributing in some *little* way to this paper.”

        Hmm, I see what you mean but don’t know that I quite buy it. For as long as I’ve been writing papers (and I’m sure for a while before that), it’s been common for students and technicians to collect the data and be acknowledged for it. Which is a huge contribution even though also a non-authorial one on its own.

  5. I am curious how people on hiring committees view equal contributions i.e. when first two authors put an asterisk that they have contributed equally? I always thought it counts as a first author for both, but was surprised to hear recently from a tenured professor that he looks only at the first name on a paper, and people don’t really care about equal contribution notes. I am very curious what you guys think. Thanks!

    • I’m curious about that too. In retrospect, I wish I’d included a poll question asking people if they read footnotes in papers, and notes on cv’s, that state how authorship order was determined. In light of the poll results, I wouldn’t be surprised if such notes are widely ignored. Although even if they are widely ignored, I’m not sure how much it matters, at least for hiring purposes, because hiring committees have other sources of information to draw on. In particular, reference letters often will explain the person’s role in any collaborative research they’ve done.

    • I guess I would say I treat it as a first author (or more realistically as a “big” contribution – I’m not one to count first authors as a special category necessarily) if I notice – if its buried in asterisks and footnotes, I may or may not notice. So its more a notice it issue than a care about it issue. So if you’re going to do it, do it prominently.

      Notwithstanding my earlier comments about search committees I wonder if there is not a bit too much angst about how finely author credit is parsed by search committees. I literally don’t think I’ve ever seen a case where somebody being 2nd author or 1st author with equal credit on a paper would push somebody in or out of the next round. While publication history is a big factor, these tend to be in categories of “not enough”, “maybe”, “good” and “impressive”. Beyond that other factors (fit to job, references, teaching and research statements, teaching cred, etc) also weigh in. Its just not true exact counts are going to drive decisions. My earlier comments about being on search committees are how hard it is to know what to do with 10+ author papers. But even there, its not like being 3rd in the list 3 times and 7th in the list 5 times is that different than being 1st once, and 8th the rest. It is a qualitative assessment.

      • “Notwithstanding my earlier comments about search committees I wonder if there is not a bit too much angst about how finely author credit is parsed by search committees. I literally don’t think I’ve ever seen a case where somebody being 2nd author or 1st author with equal credit on a paper would push somebody in or out of the next round. ”

        THIS. I meant to say that earlier but got distracted and forgot to come back to it. Search committees make a holistic, qualitative assessment of all the lines of evidence available to them (even when they use scoring rubrics to convert those assessments into numbers). No one little thing like whether or not you were joint first author vs. second author on one paper is going to make even a tiny marginal difference to your odds of getting hired.

        Which I suspect is another reason people don’t care about author contribution statements, whether or not they’re standardized. Information so minor and granular (such as whether author X on paper Y contributed to the analysis *and* the writing rather than just to the writing) simply isn’t relevant to their deliberations.

      • I totally hear you, Brian. And as you pointed out before, I also have a friend who didn’t have a job with 48(!) papers on her CV because most of them had multiple coauthors.

Leave a Comment

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s