A new survey reveals that economics departments only partially prorate credit for coauthorship. This creates an incentive for people to work in inefficiently-oversized collaborative groups, perhaps partly or even largely explaining the rise in mean number of authors per paper. Discussion from Andrew Gelman here. And coincidentally, Joan Strassman just posted on the same general issue, suggesting that it’s not a problem and that if anything people ought to make more effort to seek out collaborators with skills complementary to their own. Do you think there’s a problem with excessive coauthorship in ecology? If so, I suppose formal statements of author contributions are one way to address it, at least in part. Some journals now require these, and I’ve decided to routinely include them in the Acknowledgments of my papers whether they’re required or not.
Why you can’t “test” for multicollinearity. Briefly, because it’s a characteristic of your sample, not of the population. So it’s the sort of thing you can describe, but not the sort of thing you can estimate or make inferences about. (HT Economist’s View)
And finally, this cartoon has nothing to do with ecology, but as a (naturalized) Canadian I found the second panel really funny. “So, you’re saying that Canada is like a whole country’s worth of Oregon.” 🙂
I finally got a chance to read this article by Karen Lips where she talks about spending the past 15+ years researching the chytrid fungus that has decimated amphibians globally and trying to rally people to respond. It’s a powerful essay.
And, some last minute additions to my Friday links, both of which urge us to be more careful in the language we use. First, from Terry McGlynn, a call to stop referring to influential senior scientists as ‘silverbacks’. As Terry says, it’s sexist. Only males are silverbacks. Second — and very much related — is this explanation for the blog “Grandma Got STEM”. It points out that, when people say things like “explain it so that your grandmother can understand!”, that conveys the message that older women will have a difficult time understanding technical/scientific concepts. In other words, both of these posts are good reminders that we should be careful about how we speak.
Re: sexist language, it of course crops up everywhere. As a youth baseball coach, a common phrase I make sure to avoid is “you throw like a girl”. I’ve had girls on my teams who threw well. Plus, both girls and boys who throw badly don’t “throw like girls”. They mostly throw the way little kids do, before they’ve been taught to throw properly. Though I’m not suggesting the phrase “you throw like a little kid” should come into common usage…
Interesting — I used to coach basketball. I wonder if Brian used to coach, too? We could start a summer camp! Our slogan could be “Come learn to play from people who spend all day at their computers!” It would be sure to be a hit. 😉
Hey, I played baseball all through high school and spent a year on a college club team. I do actually know what I’m doing as a baseball coach! And you can ask plant ecologist Scott Meiners how seriously I took intramural softball in grad school. 🙂
Did you play basketball before you coached it? Brian’s tall enough that I can imagine he was at least pressured to play. I think all tall people get that pressure.
I think there’s some old post or comment thread where we talked about the most athletic ecologists, which is not a list I would be on. Frequent commenter Jim Bouldin was a much better baseball player than me–he played Div I baseball in the Big 10 for a while. My Calgary colleague Lee Jackson went to college on a golf scholarship. Theoretician Amy Hurford went to college on a basketball scholarship. Zac Long, whom I went to grad school with, went to college on a soccer scholarship, though as he himself would freely admit you’d never guess to look at him now. Dave Vasseur is a good enough skier to have spent many years doing ski patrol. My Calgary colleague Kyla Flanagan was a provincial champion in karate. And Henry Stevens, who was a postdoc in the Morin lab when I was there, was a professional ballet dancer before becoming an ecologist. So judging by that list, it’s actually not too rare for professional ecologists to have been NCAA Div. I-level athletes, which is really athletic. But I haven’t yet heard of anybody above that level–like national-level pro, or national- or Olympic-level amateur. Know of any?
most are probably too humble to talk much about a previous sports career. Running tends to be an academic friendly sport. I know of at least one olympic (US team) rower (physiol. ecol) and one world frisbee freestyle champion (physical anthropol). And of course there is a very famous world record holding ultrarunner (physiol. ecol).
More sporty ecologists: my undergraduate research advisor, Scott Creel, won the U.S. 50k running championship in 2003.
I’ve coached both baseball and basketball. I definitely don’t know what I’m doing so I only coach 5-9 year olds. My son has asked me to coach ice hockey but I had to draw the line there having never played 1 minute in my life.(although with 5 year olds its mostly about how to follow the rules enough that nobody gets hurt! I tried a pick and roll with 9 year olds but it was still a big stretch)
There are some interesting parallels between coaching and academic advising – a mix of teaching technical skills and helping to have the right attitude/mental approach. I enjoy both. .
Yes, I was just thinking about the parallels between coaching and advising. I suspect that one reason why I like baseball so much is that it’s a very technical sport. Just being naturally big/strong/fast/coordinated won’t get you nearly as far in baseball as it will in, say, soccer or American football. If you don’t have good technique in baseball, it starts holding you back pretty early on. Conversely, if you have good technique, you can go much further in baseball than in a sport that puts more of a premium on being big/strong/fast. So someone who isn’t super-athletic, but who knows stuff and knows how to convey that knowledge, is well-suited to coaching it. I’m looking forward to learning how to coach t-ballers and make things fun for them while also teaching them good technique from the get go (because bad habits become really hard to change later, as I can attest from trying to coach 13-14 year olds who’ve been practicing bad techniques for years)
Yes, I played lots and lots of basketball (and soccer, softball, and volleyball) growing up. I coached at basketball camps for three summers while in college. One of those summers I also coached softball to fill in some down time, but that was a bit more of a stretch! Now running is my main athletic pursuit — though, with a toddler and a newborn, not one that I do as much as I’d like!
maybe you can say “you throw like a chimpanzee”?
Or not. I think some commentator was fired from monday night football for saying some (african american) running back ran like a little monkey.
Who’s the ultrarunner? One of my best friends, Greg Crowther, is a molecular biologist and was an international-class ultrarunner before he hurt his Achilles.
Your examples also make me wonder if there’s some non-random tendency for really athletic scientists to study physiology, or maybe biomechanics. My friend Greg has previously done some exercise physiology work, if memory serves…
Bernd Heinrich is the ultrarunner. “Why We Run” is a reflexion on why we run that combines his biography and his knowledge of ecological physiology.
Many elite runners and xc skiers go into exercise physiology following their pro career. From there they either go into academia or coach or both. Unfortunately, few of them wander far enough to discover physiological ecology : (
The article on co-authorship was thought provoking. I didn’t have time to read the article in detail, but the argument for fully prorating credit for multi-author papers seems sound. I guess the tricky part is how should you actually prorate a multi-author paper? Just divide by the number of authors? In most cases that doesn’t seem fair for the first author, as generally they do at least half of the work. I guess you could give the first author 50% and divide the rest up among the other authors.
One other alternative would be to have the authors themselves divy up the credit, but I could see this being problematic, as bolder, extroverted co-authors may gobble up more than their due. The case where a couple alpha personalities work together on a manuscript may lead to a long drawn out battle over credit. I don’t think I would want to have to do this.
I also wonder how this might affect what Terry McGlynn refers to as “calling in the wolf”(http://smallpondscience.com/2013/02/18/calling-in-the-wolf/). If it would take you 20-40hrs of work to do something, but a co-author much less time, how do you value the contribution? If it is completely prorated, would the incentive for the wolf to help still be big enough?
How well do we prorate in ecology? I just finished my phd and have never been on any hiring or grant committee so I don’t know how ecologists tend to prorate. Through conversations I have the impression that you are judged primarily on your first-author publications. All your other publications are then combined into a sort of qualitative indicator on whether you collaborate too much or too little, and probably more importantly (for better or worse), who you collaborate with.
I think that the overemphasis on criteria for authorship, credit, and metrics isn’t healthy for anybody. Whenever anybody talks about whether or not it “counts” is principally in reference to tenure and promotion. We all know whether or not our contributions really count.
I have some coauthors that have contributed far less than myself, and I’ve been included on some papers on which my contribution was not huge. However, cooperation and collaboration is something that we should emphasize as long as it results in better science, and the more we talk about ‘credit’ the less we are doing science.
I would tend to agree with you on that. As I was reading the paper I had the feeling that there was too much emphasis on “credit”. But I guess the argument is in a competitive field like academia, “credit” strongly determines who gets to stay in academia. From an economist’s point of view if not correctly prorated, a scenario develops where the optimal strategy for furthering science and optimal strategy for your career diverge.
There is definitely something to this, but I don’t know if this is a big enough deal that the benefits of correctly prorating credit outweigh the extra time we spend thinking about it.
I think I agree, but instead of arguing about how to prorate credit, I’d prefer to dump the whole credit issue. The use of metrics to establish influence or prominence will always have some flaws that some individuals will be able to exploit, and others that will undervalue others. Also, the metric-based rewards system keeps people from working towards truly unconventional and risky work that often is the best. But that’s an old argument itself.
Re: “dumping the whole credit issue”
I don’t think we are going to dump the whole credit issue. When you are responsible for going through several hundred cvs for one open position you are going to use some criteria for abstracting what you see into some metric of quality. We can choose how explicit we are about the criteria of quality, but other than reading most of the papers written by every application, you are going to have to use some heuristics to come to a conclusion about the quality of the work a scientist has done.
Again I have never been on a search committee, but I imagine in ecology we are on the less explicit side of the spectrum. Each person who looks at a cv has his or her own fuzzy set of rules for evaluation quality (roughly how many pubs, where, on what, etc.). This less explicit method has its advantages. Because the metrics of how you will be evaluated are never stated it is hard to game the system by exploiting any one metric. But there is also a downside as the process is more stochastic and seemingly subjective.
I guess if it was my choice I would prefer this less defined way, but just because like you I do not want to see an overemphasis on any particular metric.
There’s an old comment thread on the blog (sorry, no time to go find it now) where Brian and I talked at length on how we approach our task when sitting on search committees. I should perhaps post on this sometime. It’s surprising to me how much misunderstanding there is out there about how search committees go about their business.
I’ll give you a bit of gentle pushback, Terry. I care about apportioning credit properly because it’s a fairness issue. It’s just unfair for people not to get credit–or blame–for that for which they are responsible.
And yes, cooperation and collaboration are great–but I don’t think that caring about fair apportioning of credit necessarily gets in the way of collaboration. If anything, I’d say it enhances it. Think back to the Wilson kerfuffle, where you lamented Wilson’s downplaying of his own collaborators’ contributions to his work.
One reason I like statements of author contributions is that they get around the issues that arise from trying to summarize people’s contributions and apportion credit via authorship order. A statement of author contributions is an objective, fair summary of what each author did. No need to guess what criteria the authors used to decide who got to be an author, or whether the first author’s contribution was bigger than anyone else’s, or whether the last author is only listed because he/she is the PI, or who actually came up with the idea for the study, or who actually did those fancy stats, or etc.
I would like to second Terry’s thoughts (and as I already posted on Joan’s original post). I think collaboration is really important. I think a good collaboration is greater than the sum of its parts. I think collaboration makes sense better and faster. All that said, collaboration is time consuming and hard work. So why would we be obsessed with making sure people who do this hard but valuable work don’t get “too” much credit for doing it.
There’s always going to be some senior person who just puts their name on papers. And medical researchers joke about “if you have coffee, your name goes on”. But really I think this comes out in the wash more than most people think. As academics we live in a world where there are only 20-30 people truly in our discipline enough to judge our work. This is a small enough group to smoke out true contributions. And although these disciplinary judgements feed in only indirectly to job searches (through papers accepted, letters of rec) they feed in rather directly to promotions, career awards, etc.
If somebody’s collaborating, let them have a few bonus points. It doesn’t really matter who gets credit, but it sure matters that large interdisciplinary problems are tackled, and that all problems are tackled efficiently by people who are highly efficient in each slice of the problem.
I collaborate a lot, particularly with two people (Carla Caceres and, especially, Spencer Hall). I definitely think it’s the case where we do better science thanks to collaborating — in my opinion, it’s definitely a mutualism! I realized recently that Spencer and I are coming up on 10 years of collaborating.
We sometimes include the contributions of each person, but it can be hard. For example, for a recent project we’ve been working on, Spencer and I both started out doing the modeling independently, using somewhat different approaches. (Reassuringly, despite use using somewhat different approaches, we both came up with the same approach for integrating the most novel part of the study.) Eventually, we decided his approach was better, and so went with his. So, I guess we would say something like “Hall developed the model in collaboration with Duffy” or something like that, which would sort of capture the collaboration but not really.
And as your example points out, despite all advice to the contrary, the best collaborations are the ones where you trust people enough that you don’t have to negotiate the credit up front
By working to refine the system, one validates it. It is about fairness. No matter how you slice it, using a quantitative metric to evaluate scholarship is unfair. A qualitative measure can be more fair.
Thanks for the uplifting women in science links, Meg!
Glad you liked them! I don’t intentionally look for women in science links every week; it just seems to work out that way!
Meg – I really appreciate the awareness you bring to social aspects of our career. The little bit that I spend in other work cultures, the more I’m impressed with academia, but other office environments may just be setting a low bar or maybe it differs at a large research university in say a math or physics or engineering (or even biology) department. I’m at a regional state university and because of the lack of intense pressure to bring in external research funds, we have a pretty family friendly environment. Or maybe I should generalize that to “quality of life” friendly environment, however one wishes to enhance their quality of life (family, outdoor recreation, travel, teaching and mentoring, community service, even more time in the lab). I often wonder about the constraints on how one strives for “quality of life” for faculty running large research labs.
Thanks for the comment! I’ve been considering a post on this topic. The idea of parental leave while running a lab is one I still am trying to wrap my head around. You’d think I’ve have it figured out, given that I’ve had two children, but it’s still not really clear to me what it means. There are definitely some constraints on work-life balance, I think, that come along with running a lab!
There is at least one reasonable method for divvying up credit on a multi-authored paper, outlined below. I saw this in (I think) a comment on a paper in Science or Nature but have never been able to track it down again. Whatever method you use should ensure that the sum of the author contributions equals one, to avoid the economics problem outlined above.
The equation is this: if author position is i, and there are n authors, then the author contribution is: (1/i) / (1/1 + 1/2 + 1/3 + 1/4 + … + 1/n). For two authors you get (1/1) / (1/1+1/2) and (1/2) / (1/1+1/2) respectively.
For example this leads to the following contributions (note each row sums to 1 as expected):
n author contributions
2 0.667 0.333
3 0.545 0.273 0.182
4 0.480 0.240 0.160 0.120
5 0.438 0.219 0.146 0.109 0.088
6 0.408 0.204 0.136 0.102 0.082 0.068
7 0.386 0.193 0.129 0.096 0.077 0.064 0.055
8 0.368 0.184 0.123 0.092 0.074 0.061 0.053 0.046
9 0.353 0.177 0.118 0.088 0.071 0.059 0.050 0.044 0.039
In some fields you can argue about whether the final author (senior author, planned, funded, conceived, etc.) should receive more credit, but this is difficult to automatically do since very often the final author did the least amount of work. One option would be to give the corresponding author and the first author equal credit if they are different.
Similarly, if there is a note “these authors contributed equally” then this could be taken into account.
Another complication is when there are lots of authors and clearly most are listed in alphabetical order. The rule I apply here is that if there are three or more authors in alphabetical order after the first author, then these authors are given equal weight. For example, imagine a paper with 6 authors, and the last five are in alphabetical order suggesting equal contributions. From the row above, we would get:
6 0.408 0.204 0.136 0.102 0.082 0.068
Now average the contributions of authors 2 through 6, so that you would get:
6 0.408 0.118 0.118 0.118 0.118 0.118
Again this adds to one, satisfying the principle of conservation of authorship contribution.
Andrew Gelman and the author of the econ study linked to in the post comment further on the coauthorship & credit issue: