Friday, February 15, 2008

A-Z-ist

When I was recently serving on an interdisciplinary committee that had to read dozens of CV's, the inevitable topic of authorship order conventions arose. In order to fairly evaluate these CV's from across the entire spectrum of Academe, we had to know what the convention was in each field.

Even when told that the convention in, say, economics is for alphabetical ordering, it is difficult for faculty who are not in alphabetically-ordered fields to overcome a bias against someone whose name is always in the middle of a pack of authors. I have read with great interest about studies that show that even within fields in which alphabetical ordering of authors in the norm, there may be a subconscious bias against authors with surnames late in the alphabet, and therefore negative consequences for hiring, promotion, tenure, and awards.

On my committee, as we evaluated one CV from a professor in a field with alphabetical ordering of authors, a committee member was very frustrated because the article citations listed the CV person's name first, followed by 'with A. Person and B. Person'. This committee member wanted to see the citations written out in the correct order. I said "But we know the correct order because the order is alphabetical." Even so, my colleague insisted that seeing the authorship written out in the correct order was essential. He felt that the CV person was somehow being slightly dishonest by making his name the most prominent in each citation. Well, given that the CV person's last name began with R, I didn't blame him. If he wrote out the citations completely, he would have been last or near-to-last in them all, and this was not a field that valued last authors. It seemed like a lose-lose situation to me.

Regarding fields in which being last author is customary and prestigious, indicating the research group leader: I also saw on my committee that people in fields in which last author implies the least contribution to a paper have trouble overcoming a bias about the (in)significance of last authorship.

In one of the 57 ethics training courses we have to endure at my university, I was told that even if I provide the funds for a particular research project that is conducted in my lab, unless I contribute something else to the study (e.g. ideas, data, writing), I cannot (= should not) be a co-author, not even as a caboose author. Conflicted feelings and ethical issues related to last authorship led to discussion in my committee of whether someone whose CV had his name listed last on every single publication had really done any of the work. I think it was obvious that he had, based on other information we were evaluating, but again, people in fields in which last authorship = the least contribution had trouble overcoming a negative reaction to last authors.

I have participated in both alphabetically ordered and non-ordered authorship cultures and think that there are problems with both schemes. Alphabetical ordering seems simple enough, but if anyone with a surname after D is going to be disadvantaged in their career, then that system is not as objective as it might seem.

Ordering by relative level of contribution has its own complications because in some cases you have to decide the order of authors who contributed at similar levels to the paper. And who should be first author if one author supplied the ideas and one supplied the data?

In some cases, the addition of a third author to a paper means a major decrease in exposure (and therefore name recognition and therefore possibly career advancement) for the second author, who is relegated to being part of the 'et al.' rather than having their name listed more prominently in journals that have such citation formats within the text.

Maybe we should publish more single-author scholarly articles? No, that's not good either because that makes us look like we don't work well with others, and are not multi-, inter-, or even transdisciplinary. Single and few-author papers were, however, important to me earlier in my career because I wanted there to be no doubt that I had played a major role in these publications. I think this is particularly important for early-career women scientists.

These days, I don't care where my name falls in a multi-author paper for which I was a mid-level contributor, though sometimes having my name buried in the et al. of a paper can lead to strange situations. Several times in the past year, I have had people talk to me about a particular paper on which I was a co-author, and I initially assumed that my role was known because my name is on the paper, albeit somewhere in the middle of the author list (which is not long). The repeated use of "they", as in "they said this" and "they concluded that" is a clue that this assumption is flawed. I like to say "Yes, and their paper is so well written, and I particularly love Figure 4" etc. I don't actually mind it if someone doesn't instantly recognize my authorness -- I certainly don't have the author lists memorized for all the papers I read -- but it would not be so entertaining if being recognized for my contribution mattered to my career more than it does at this stage.

32 comments:

Dr. Jekyll & Mrs. Hyde said...

The biological sciences, at any rate, follow the "first name did the benchwork, last name gave the money and Big Thoughts, and anyone in between, your guess is as good as ours" format.

It always reminds me of the description of democracy as the worst form of government ever devised, except for all the others.

Suresh said...

Hmm. I do my CV like the person you mentioned, and my last name starts with a V !! I wonder if that will affect me in some way.

David Moles said...

Maybe some day when everything's electronic we'll be able to randomize the order every time you look at the list of authors.

Anonymous said...

1) It is important to have few-author publications at the beginning of their careers for both men and women.

2) If every field adopts alphabetical order then there would be no problem, at least after a few years... at the beginning it might be hard to remove the bias from the people in fields used to order by contribution.

3) If academics are supposed to be really smart, how come a simple name-ordering problem causes bias?

Anonymous said...

I'm in an alphabetical field. When I got engaged, a more senior person demanded to know my future-husband's last name, and, hearing that it was 15 letters ahead of mine, told me of the career enhancing potential of changing my name. I was startled, to say the least.

--Dr. H., formerly Miss W.

YAMP said...

Math is an almost always alphabetical field. It means there is no arguing about who thinks who did what. If you want to know the value of my contributions you can look at my record or ask around. I am all for conflict reduction.

I have a P. colleague who once published out of order and had to explain at tenure time why that had happened as if there were something untoward going on -- which there almost was. I am a W. and have declined re-ordering on many occasions. I worry that f I am author first on one paper and last on almost all others it will call in to question my contribution on most papers.

Chinese co-authors are great for X. and Z. names and I did even manage to publish a WXYZ paper!

Schlupp said...

Does the alphabetic order reduce conflicts? In my field, there is often discussion if Dr Xyz, who basically just passed by the lab, should be an author at all. As it is, one can just put the name in and it doesn't really hurt a lot. But if Dr Xyz's contribution thereby became indistinguishable from Dr Abc's who did all the work, we might see more fights instead of fewer.

How do the alphabetical fields judge an author's contribution to his/her papers? Is this information conveyed in recommendation letters?

Anonymous said...

I disagree with a lot of these comments. I'm in a "1st author contributed most and last author is group leader/big idea person" field and I happen to like this model quite a lot. I don't think it would be a good idea at all for every field to move toward the alphabetical ordering system or randomization.

I like for people to know when I contributed the majority to the paper and when I just played the supporting role (both roles are important for successful researchers). Of course some ambiguity is still present in this system, but it is the only system I know of that allows two forms of recognition (1st author did most of the work and last author is the group leader/big ideas person). For scientists' careers, name recognition really matters, and this is the best way to do it. I refuse to believe that anyone is above being recognized for his/her work.

When you guys out there in the alphabetical world say that other researchers must ask around to figure out an author's level of contribution to a paper, you are implicitly recognizing that knowing 'who did what' really does matter. Why not take out the rumor element of recognition and put it out there in the open. It seems to me that the alphabetical system will start more conflicts than it resolves.

Buffalo Sally said...

While difficult, I think the order of names should be in the same order as contributions to the work. If not, I would be lost in the shuffle. -Buffalo Sally W.

Anonymous said...

I am in an interdisciplinary field on the boundaries of alphabetical and amount of work. Some of my papers are alphabetical and others are effort based. There is a modified alphabetical convention that has arisen in this subfield: first author is the one who did most of the work and actually wrote the paper, all others are alphabetical.

My job prospects have all been in fields that are purely effort ordered. I have a name in the bottom dregs of the alphabet. The alphabetical papers have been an issue in at least one job search that I know about.

For my tenure review, I was told to write about the ordering in my research statement and to put footnotes on the papers that are alphabetical.

Rachel said...

I'm not a scientist but I'm fascinated by biases.

I think the "First author contributed most and last author is group leader/big idea person be renamed the "movie credits" approach.

It's how stuff is done in Hollywood:

Starring.....George Clooney and Featuring Martin Sheen.

Although in a scientific context perhaps "Starring Danica McKellar and featuring Stephen Hawking" might be more apt on an Erdos-Bacon level.

Anonymous said...

Has anyone seen single author papers by women in science? I occasionally see single author papers by male theoretical physicists but not women. It would be cool if more women were writing single author papers.

Anonymous said...

To Anon above, I myself wrote a single author paper and am a woman. It is one of the most cited papers in my publication list!

Anonymous said...

To the Anon above-
You rock! I think 2008 should be the year of single author papers in science by women!!
Let's go get 'em Ladies!

EcoGeoFemme said...

I am unlikely to have more than a handful of single author papers during my career (if I stay in this field), which is a little sad to me. Interestingly, single author papers in my field seem to be increasingly rare, even over the last 10-20 years.

My boyfriend will have mostly single author papers, but is looking for ones to write with his advisors so he can firmly attach his name to theirs. I think given the choice, I'd choose my situation over his.

So in response to anon 2:38, I think it depends on the field. My BF's advisor is a women who has many papers on which she is the sole author.

Ms.PhD said...

Fantastic post.

To the anon who asked about scientists being smart but biased, those two things have nothing to do with each other. Intelligence is not the same as self-awareness.

Several people have been referring to the Implicit Association Tests lately, you should go take a few and learn about bias.

Also saw an interesting blurb in Nature about how adding additional authors decreases the amount of credit awarded the first author, but has no effect on the others.

Am really tempted to publish my latest papers as a single author. They would be the last papers I ever publish, if I do. The only single-female-author-papers I'm aware of in my field are reviews.

lymie said...

The medical field is generally an effort first/seniority last author list. Some in the field are exceptionally stingy about sharing collaborative authorship. Our research group always tries to include representatives from all contributing research centers. Many journals in the medical field now require a detailed "contribution" list of who did what. That might work well in other fields. It was implemented in medicine because of the nefarious influence of pharmaceutical $$.

Saxifraga said...

Interesting post. I have a question for those of you in favour of "first author did the work and last author provided the funding and big ideas" ordering. Is associating the "big ideas" with the last author really fair?

I am a postdoc on project money and I see the point in recognising the efforts of my PI in getting the money in the first place. But my work is not a small part of an assembly line. I do work that ties in directly with the big picture of the research project (for which the PI is a natural coauthor), but I also do work that is rooted in my other sub-field, on ideas that are mine rather than the Pi's and has nothing to do with the PI's expertise or big ideas. I would be very frustrated if crediting the person who came up with the money also automatically gave all intellectual credit to that person.

I'm in a field where we use a combination of who contributed what and alphabetical ordering. The first author is always the person who did the majority of the work. Following authours can be added outside the alphabetical ordering according to level of contribution (second author did more, third author did less). Alphabetical ordering is used for the rest who did more or less equal amounts of work and/or who all did significantly less than the individually ordered authors. This system is flexible and allows for various degrees of contribution to be acknowledged.

Anonymous said...

I am in an pure maths, which is strongly alphabetical. You usually assume equal or comparable contribution from all authors. Those who just did a small bit would get an appendix in their name; if the bit is too small for an appendix, they get their name on the relevant lemma, but it doesn't count as publication.

If you just put the money and/or the idea, you get thanked for. No authorship. I am used to this and find it not too bad.


In my particular branch of mathematics single name papers are not rare, and single female name papers not too rare.

Full disclosure: I am actually the first author of all my joint name papers except one :-).

franglais said...

Excellent discussion on this post.
I am used to and like my field approach to ordering authors, based on relative contributions. If contribution is the same, then go to alphabetical order. However, there is no way to distinguish the two, so the ABCs are inherently favored and the XYZs are not helped over time. I think the alphabetical order should reverse every year (ex: even year normal, and odd year reversed) to make the alphabetical ordering less permanent without going to the randomization scheme.

But wait, now the middle alphabet people may complain. I give up..

Female Science Professor said...

It's interesting that people in fields with alphabetical author conventions assume that because everyone knows this is the convention, that author order doesn't matter. I used to think that, until I read recent studies showing that order does matter, even in fields with entrenched alphabetical tradition. There are also studies showing that authors in the middle get substantially less credit; for example:
http://www.physorg.com/news113482761.html

Anonymous said...

When you guys out there in the alphabetical world say that other researchers must ask around to figure out an author's level of contribution to a paper, you are implicitly recognizing that knowing 'who did what' really does matter. Why not take out the rumor element of recognition and put it out there in the open. It seems to me that the alphabetical system will start more conflicts than it resolves.

What drives me nuts is that no ordering system gives any substantial information whatsoever. Having a contributions section that explains what each author actually did would solve this problem. This is simply way too much information to encode into the name ordering. Even with six authors, the possible orderings can't even convey ten bits worth of information. It's ridiculous.

One problem is that each field's approach is based on assumptions that don't universally hold. For example, in pure mathematics it is assumed that all authors contributed in exactly the same way (i.e., they didn't have different, specialized roles in the collaboration) and both worked as hard as they could. Often, this is true, but of course not always.

In biology, it is assumed that one author came up with the idea and secured the funding, one did all the hard work, and the others made valuable but less demanding contributions.

In both cases, the standard ordering for the field works perfectly for the default type of collaboration, but fails miserably for anything that doesn't fit the pattern.

Does the alphabetic order reduce conflicts? In my field, there is often discussion if Dr Xyz, who basically just passed by the lab, should be an author at all. As it is, one can just put the name in and it doesn't really hurt a lot. But if Dr Xyz's contribution thereby became indistinguishable from Dr Abc's who did all the work, we might see more fights instead of fewer.

Mathematics uses what we often call the Hardy-Littlewood rule. Anybody who in good faith joins a collaboration will be listed equally as an author, regardless of the relative contributions they end up making (although if you really didn't contribute much, the respectable thing to do is to ask not to be an author). Anybody who doesn't pull their own weight will not be invited to collaborate in the future. If you see several joint papers by the same group, then the assumption is that everybody was satisfied with the contributions.

Mathematics does not generally have well-defined research groups within a department that publish together by default. Basically, everybody is a free agent, as opposed to lab sciences where a postdoc may join someone else's lab to work on his/her projects for several years.

Mathematicians also rarely use "et al.", since it is grossly unfair to the omitted authors. I was once forced to use it by a certain journal editor, and it didn't seem worth withdrawing the paper over this issue, but I sort of regret not having made a principled stand. That was a pretty eccentric editor by mathematics standards.

How do the alphabetical fields judge an author's contribution to his/her papers? Is this information conveyed in recommendation letters?

Yes, but surely that is the same as in every field. Nobody hires job candidates or awards prizes based solely on the author ordering. If you are going to require letters of recommendation anyway before making any decisions, why not make slightly more use of them?

Besides, there are lots of clues. If two people collaborate repeatedly, we assume they are both contributing. (This rule could be taken advantage of if both authors are corrupt, but so can all forms of author ordering.) If someone collaborates with many different people, we assume they are probably good - why else would everyone want to collaborate with them? These guidelines aren't perfect, but I think they work about as well for judging people as author ordering does. In other words, they give a first, crude opinion.

It's interesting that people in fields with alphabetical author conventions assume that because everyone knows this is the convention, that author order doesn't matter. I used to think that, until I read recent studies showing that order does matter, even in fields with entrenched alphabetical tradition.

I've been a little skeptical of some of these studies, at least the ones I've seen, because of issues of interdisciplinary work. For one study I saw, which I can't locate right now, it compared two adjacent fields with different ordering conventions. Perhaps economics and some other social science? I wonder whether people in each field see enough papers from the other and have enough contact with researchers from the other field that reputation and ordering effects are transmitted.

I'd be really curious to know what happens with pure mathematics. I doubt the alphabetical ordering has much of a biased effect on reputations in math, but this is nothing but a gut feeling.

Vodalus said...

I'd be really curious to know what happens with pure mathematics. I doubt the alphabetical ordering has much of a biased effect on reputations in math, but this is nothing but a gut feeling.

This could well be true if the vast majority of papers are single or double authors. Negative effects of being late in the alphabet would only arise from people being consistently listed at the end of a long list. In other words: it's much easier to remember (and pay attention to) 2 names than it is for 6.

JSinger said...

In one of the 57 ethics training courses we have to endure at my university, I was told that even if I provide the funds for a particular research project that is conducted in my lab, unless I contribute something else to the study (e.g. ideas, data, writing), I cannot (= should not) be a co-author, not even as a caboose author.

I find it hard to imagine what could be more corrosive to ethics than these "ethics training courses" that teach guidelines that everyone in the room knows bear no resemblance to actual practice.

I don't actually mind it if someone doesn't instantly recognize my authorness -- I certainly don't have the author lists memorized for all the papers I read...

Agreed, which is why I'm having trouble understanding why you're marveling that this happens. *I* don't remember some of the papers on which I'm buried in the middle of the author list and certainly don't expect anyone else to. And even if they did know that I contributed a reagent and some edits in return for my 6th-of-11 spot, it still doesn't seem outrageous to refer to "their" work when the paper *is* fundamentally their work.

Female Science Professor said...

In fact, I contributed substantially to the paper in question -- to the ideas, data, and writing. But who's to know? My point was that it doesn't matter to me at this point in my career, but it would if I were an early career scientist.

Anonymous said...

This could well be true if the vast majority of papers are single or double authors. Negative effects of being late in the alphabet would only arise from people being consistently listed at the end of a long list. In other words: it's much easier to remember (and pay attention to) 2 names than it is for 6.


That's a really good point.


I find it hard to imagine what could be more corrosive to ethics than these "ethics training courses" that teach guidelines that everyone in the room knows bear no resemblance to actual practice.


Depending on how broad the audience is, it may be an important message for some people. For example, in mathematics, providing funding is never enough reason to be an author. In some areas of computer science, it can be in practice. In theoretical computer science, on the border between the fields, the community norms are strongly against it but it occasionally happens. I agree that nothing learned in an ethics training course will shift behavior in a field in which everyone does this routinely, but it may help with compliance in a field in which people think they really shouldn't but may feel tempted.

Teutsch said...

Appearing always as first author on papers proves only one thing: the author is good at arguing to become first author. Fighting over who gets to be first only detracts from what really matters, namely the research. Contributions from collaborations should be encouraged under any circumstances; a joint work is not a comptition. Typically one's reputation is already known anyway, or can be found out by asking co-authors

yellowfish said...

This is interesting... in my field its always first does the work, last supervises with the rest in rank order of work done in between.

When applying to grad school my boss at the time told me his pubmed-advisor-selection criteria... he said you should see initially some middle author publications, then a long string of first authors, then last authors. But, the important thing to check is that they should still have SOME first author pubs or else you know they don't do much active work and might not be very involved, although if they are still ALWAYS first author you know they don't let their students write up any of the papers and you'll have trouble I kind of liked it, it was like lit review espionage, and it totally worked (of course, the alphabetical thing would totally throw a monkey wrench in this system.)

PhysioProf said...

"It is important to have few-author publications at the beginning of their careers for both men and women."

This is untrue for the biological and biomedical sciences, where it is important for trainees (grad students and post-docs) to have first authorships and beginning PIs have last authorships. No weight at all is given to the number of middle authors.

Anonymous said...

"I wanted there to be no doubt that I had played a major role in these publications. I think this is particularly important for early-career women scientists."
More important for early-career women as opposed to early-career men? Why?

Female Science Professor said...

I can't tell you how many times I have been on a search or awards committee in which committee members questioned whether an early-career female author had played a major role in the research (even if authorship order indicated she had), but this same question does not arise for male authors in the same situation.

Anonymous said...

Like Dr. Jekyll, I'm also in one of the fields with the "first name did the benchwork, last name gave the money and Big Thoughts, and anyone in between, your guess is as good as ours" format.

I don't have a magic formula for a good system.

On the one hand, we have several joint publications with other groups where 2-3 people would all qualify for first authorship (different aspects / techniques pulled together) - and, even if we publish in a journal that allows "equal contribution", the first name will still get the credit later.

On the other hand, we have several names on our papers either for purely political reasons or with very minor contribution - in an alphabetical system it would be unfair to mix "Dr. Supplied Reagent A" with ones who actually did the bulk of the work.
(Including them cannot be avoided.)

I think we found a pretty decent balance by having more or less the same collaboration partners for several papers, and rotating first- and senior- authorship between the groups for each publication.
Also, often for internal papers where the main concept was provided by PostDoc, the Big Boss relinquishes senior authorship to him.

I tend to lean towards the non-alphabetic system as more fair in fields with more authors who contribute to papers in *very* different ways and to very different extent.

As long as the search committee has a good idea which field uses which system, things should work out (or so I hope).