Date: Mon, 12 Jul 2004 10:34:33 +0100 From: "David H. Foster" To: cvnet@mail.ewind.com Subject: CVNet - Order of Authors: Summary Many thanks to all who took the trouble to reply, some at length, to the following enquiry posted 8 June 2004 on CVNet. > Conventions governing the order in which authors are listed on a > journal paper vary with discipline. Some common rules are these: > > 1. Alphabetical order. > 2. Descending order of contribution. > 3. Descending order of contribution but "senior" (most experienced) > author last. > 4. Descending order of need (for employment, promotion, etc). > 5. Random order. > > I am interested in seeing what practice researchers in vision follow. > If there are enough replies (to me please at d.h.foster@umist.ac.uk, > not to the list), then I shall post a summary. Not all respondents chose a single method. Where methods were used equally often, I have given them equal fractional scores (e.g. 0.5 for method 2 and 0.5 for method 4) and where methods were used unequally often, I have interpreted "usually" as 0.7, and "sometimes" as 0.3, the fractional scores summing to unity. The total scores for each method were as follows (based on 108 responses). Alphabetic 1 Contribution 45 Contribution + senior 55 Employment 7 Random 0 Although Random received zero score, some respondents reported using it after ordering the first few authors by another method. I have attached some (slightly edited) comments where more general beliefs or issues have been addressed, and, in particular, some explicit proposals for changes in practice. I have not included comments dealing with personal experiences when this might cause embarrassment or when the same comment has been made by someone else. The ordering is alphabetical, by forename. David Foster email: d.h.foster@umist.ac.uk web: http://www.op.umist.ac.uk/dhf.html --------------------------------------------------------------------------- Descending order of contribution is the only defensible system. But it only works if there is a convention. There seems to be a recent trend borrowed from big biology labs of "senior authors" being last and thus giving themselves a different status from the rest. I am very much opposed to this. It means that there is no way of knowing how much actual contribution the great man made. I think all authors should have equal potential status with actual order depending on contribution. One of the advantages of American/British/Australian science as opposed to European is that young scholars can be independent researchers and are not regarded as necessarily subordinate. Barbara Gillam b.gillam@unsw.edu.au --------------------------------------------------------------------------- I vote for #2, with one exception: Oral presentations should always have the presenter's name first. In big labs where the "senior" is mainly involved in grant acquisition, this will automatically include #3. There are cases in which the ranking of contributions is not quite clear. In these cases order may vary across publications. However, a sensitive issue may be to work out and agree on the ranking of relative contributions (see below). #1 would favor Z. A and K. Aaaadi and be unfair to W. Zziwambazza (Medline examples). I do not like this order, which was the style of some journals some time ago (still is?). (Therefore, I am also not particularly happy with the new style of Vision Research to list multiple citations in alphabetical order.) #4 would be unfair to co-authors and to the whole scientific community. Almost everyone needs publications for something (employment, promotion, grants, etc.) and it would be unfair if those making a minor contribution were given an incorrect advantage in overall rating estimates. (To include people with no contribution at all would be against the ethical rules in science.) A critical issue would be how to rate the individual contributions to a paper, and a related question (not listed above) is about which kind of contributions are considered to make it onto the list of authors. Many technicians, for example, contribute a lot but rarely become co-authors of a paper. You cannot argue with responsibility; many technicians are responsible for their work and probably much more responsible for their part of a paper than people who just followed instructions of their supervisor or just provided a drug or antigen. In my opinion, a major contribution would be to write the paper, i.e. to give the outline of research and put it into scientific context. I would feel very strange about a first author who might have done a lot of the measurements but had not taken care of writing them down. If you are uncertain about the ranking of contributions and the later order of authors, talk about this early. Another point: If you write a paper about your view of things, giving your interpretation of data and your philosophy behind, it should be your paper, not the paper of a group. If you refer to experimental data of the group you should cite them properly making clear what is their contribution and what is your interpretation. This does not hold for original publications about experimental data, which should always be published by the group. Christoph Nothdurft hnothdu@gwdg.de --------------------------------------------------------------------------- In my view, there is only one useful order, which is # 3. The first author writes the main draft of the paper. If there is an order of need, the authors should make sure that their contributions reflect that order (or are all equal). I'd like to take this occasion to make a plea for a rational order of text citations in journal articles. More and more journals seem to be switching to alphabetical order for text citations. This option is both invidious and irrelevant. Invidious because it means that Anstis always gets cited before Zeki (to pick familiar examples). Alphabetizing may be convenient, but it is perpetually prejudicial. On the other hand, the classic approach of citing in date order not only avoids prejudice, it is beneficial on conveying the order of priority. It just seems absurd to me cite, say, Banks (2004), Ogle(1952), Tyler (1973), Wheatstone (1838) for disparity pattern effects. So is there any agreement that we should pressure the journals to change this prejudicial policy? Christopher Tyler cwt@ski.org --------------------------------------------------------------------------- I use a hybrid of your 2 and 3. One major problem in assigning the order of authorship is with the definition of "contribution"; is it time spent on the project, or is it the original idea or is it the solving of technical or intellectual bottlenecks? The practice of placing the 'most experienced" author last probably represents the situation where this author provided the original idea for the work as well as the funding for the study but the person placed first on the list of authors was the one who did the lion's share of the actual work including solving technical problems and taking a lead in the writing of the paper. This is the practice that I normally follow. On the other hand, when the person who did most of the physical work does not solve most of the problems that develop during the study or did not contribute much to the writing of the paper, then this person may lose senior authorship. Donald Mitchell@Dal.ca --------------------------------------------------------------------------- Almost everywhere there is descending order of contribution. In many biological disciplines, the last author is the head of the laboratory, i.e., who gets the funding, but that may still not contradict the "descending order of contribution" rule. Of course you may remember the old J. Physiol. policy of alphabetic ordering of authors. It was fully understood by most consumers of J. Physiol. articles at the time, but not by the wider scientific community in America. Gerald gwest@socrates.Berkeley.EDU --------------------------------------------------------------------------- I think this is an interesting question, and actually quite an important one because its answer implies a view on the nature of scientific effort and achievement. Rather than writing an essay on it, can I suggest a policy, as follows:- "The names of the authors of a paper can be listed in any order, but a brief note explaining its basis (such as one of those suggested below) must be included." My suggested list, based largely on yours, would be:- 1. Names are listed in descending order of the amount of actual work contributed to this particular paper. 2. Each author has indicated with superscript letter(s) the area(s) he or she contributed most to, eg: b for boss; e for experimenter; tia for theory, interpretation and analysis; pw for presenter of the results and writing the paper. 3. Descending order of contribution but "senior" (most experienced) author last. 4. All authors take full and equal responsibility for the correctness of the results and we are listed alphabetically. Horace hbb10@cam.ac.uk --------------------------------------------------------------------------- Recently in the case of research students it is becoming the rule here to have the supervisor last as communicating author. I think this is a good idea. The practice of having the lab head on the paper regardless of contribution is indefensible in my view. The best practice seems to be to agree up front on who does what and who goes where, but memories get blurred,(self-)assessments of contributions tend to be upwardly variable, and things rarely go as envisaged. I have had at least one bad experience with this in the past. There is an article of agreement among biomedical academic publishers about the authorship of papers you might like to refer to, as this offers good guidance on these matters. http://www.icmje.org/ No doubt you are aware of this. An international standard or at least a strong body of evidence of best practice under the auspices of the Royal Society or such an organisation would be very helpful to everyone, as would be a series of high profile scholarly articles on the topic. A paper I found at Stanford http://www.stanford.edu/dept/DoR/rph/rph_pdf/2-8.pdf addresses the issue thoughtfully and shows that practice varies and that there are no simple answers. It is still their current policy even though it is dated 1985! Ian I.E.Holliday@aston.ac.uk --------------------------------------------------------------------------- These are often complex issues, particularly today with growing author lists and various pressures. For years I have tended to put my students first to give them opportunities for recognition. In this, I followed one of my senior mentors, Walter Stanley Stiles, FRS. However, when I was active in grants management in my lab, I was told by a NIH grant site visit team that they were sorry I was no longer active in research. They had noticed I had placed my name last on a series of papers! On a new research venture I have originated, on occasion, I will put my name first on the first paper in a series. That is, I identified this as coming from this lab. Then I followed the usual practice of putting my name last on subsequent papers. I agree that in an oral presentation, the speaker goes first. In some instances, if there is a long string of names, I have placed my name second if I had made a meaningful contribution. In short, I have not found a comfortable solution to the problem(s) you raise. I do object to lab heads placing their names on papers where their contribution has been minimal or non-existent. This is a conundrum where there is really no right nor wrong. Jay Enoch jmenoch@socrates.berkeley.edu --------------------------------------------------------------------------- I am a member of the university Research Council and we are having a difficult time with all these different systems in different fields of research (and we do not always believe what applicants tell us). My own policy has changed a bit over the years: 1. It used to be system 2 rather strictly, except perhaps that the supervisor sometimes let the graduate student take the first position, even if the latter's contribution was not the largest (because our PhD requirements stipulate that only 1st author papers can be included in the thesis). 2. Under influence of the more biomedical traditions, I am currently using a mixture of system 2 (when submitting to psych-journals) and system 3 (when submitting to neuro-journals). As you may know, our colleagues in the biomedical labs are fighting almost more heavily to obtain the last position than the first! 3. Under influence of increased publication pressure (e.g., for chances of postdoc positions), elements of system 4 have crept in over the years, not as much in terms of order of authorship but just authorship in the first place: When other members of the lab have contributed by software assistance, data acquisition (e.g., being subjects in long and boring psychophysical experiments), or discussion, they could sometimes be listed as 2nd or 3rd author, even if their role would have been acknowledged in a footnote in the past. Other labs that have used this policy have blossomed because of this policy, simply because their postdoc candidates have twice or three times as many publications (even though the overall output from their labs has been the same). As far as I can tell no-one in our field uses system 1 or 5. Johan Johan.Wagemans@psy.kuleuven.ac.be --------------------------------------------------------------------------- I believe that criteria for authorship is a more important issue (and more often abused) than order of authors. I like and use the guidelines at www.icmje.org In our dossiers and annual reports at OSU, we have to give a narrative description of our intellectual contribution, e.g. 1. I designed the experiment (which was carried out by the graduate student co-authors), and wrote the article 2. I identified the patients for the study, administered the drug regimen, reported results to the consortium and reviewed the draft manuscript. 3. I completed and wrote the literature review for the paper, shared equally with the co-author in the analysis and interpretation of the data, and reviewed the complete draft manuscript. Some journals require a similar explanation. Mark MBullimore@optometry.osu.edu --------------------------------------------------------------------------- We tend to use criterion 3: descending order of contribution but the senior author last (unless he/she wrote the paper, then first). One important question concerns who has to be included as an author (as opposed to be acknowledged). I think the rule is that whoever made a significant intellectual contribution should be offered an authorship. But then, what is a significant intellectual contribution? Another issue is: what about former contributors who left the project but whose data are significant enough to be included in the paper? Ruxandra Sireteanu sireteanu@mpih-frankfurt.mpg.de --------------------------------------------------------------------------- While I prefer descending order of contribution, the standard here seems to be first author is the major worker, the last author is the "senior author" (which in Biology seems to mean the person whose lab is most involved), and the others are then filled in, mainly by level of contribution. This also seems to be a widely assumed ordering, since on NIH study section individuals often look to see if a person is either first, last or "in the middle", with "in the middle" being less desirable. Steve sburns@vision.eri.harvard.edu --------------------------------------------------------------------------- We follow #2 rule making it better to be second last than last. Terri L. Lewis lewistl@mcmaster.ca --------------------------------------------------------------------------- I believe the custom is quite different in different disciplines. For example in medical and cell biology it is my impression that the senior author, or the author in whose lab the research was conducted, becomes the last author almost routinely. In medicine there is the very bad habit, seen far too frequently, of putting the chair of the department on as the last author even when it is clear that he/she had little or nothing to do with the manuscript and the research specifically! Tony Adams aadams@socrates.berkeley.edu --------------------------------------------------------------------------- My rule (and as far as I know the rule that is generally accepted as normal and fair this side of the Atlantic) is this: A. If the paper comes from my lab only: the people who did the actual bench work in descending order of contribution, then me. B. If it is a collaboration between two or more labs: the actual "doers" in descending order of contribution, then the bosses in ascending order of contribution, so that the person who was the driving force of the project is the last author. Seva Vsevolod.Gurevich@vanderbilt.edu ---------------------------------------------------------------------------