Authorship

Overview

Authorship is a contentious issue worldwide in the academic community. There are many definitions of what should constitute “authorship” of a scientific manuscript, and many of them are rife with elitism, cultural bias, and historical momentum. There are also a great number of field-specific conventions, so be aware that what is acceptable or normal in one scientific field or community (e.g. biology) may differ greatly from what is seen in another (e.g. literature or mathematics). I have strong opinions on this topic, but be aware that there are many opinions, and in some cases journals will also just clearly define what they consider criteria for authorship. This is becoming increasingly common, as is the addition of an “Author contribution” section within the manuscript itself which must be written in a defined format. The idea of this section is to basically make it clear what each author contributed, rather than relying on the set of informal conventions which have arisen over time, or defining a set of rules for authorship formally.

First and last authors

For our field (and I think biology in general) there is a strong convention that the first author writes the manuscript and the last author is the lead on the project (usually the principal investigator or PI). The middle authors are then ranked from second to second-last in order of contribution. For papers coming from collaborations between multiple groups there is a weak but increasing convention to have multiple first authors (co-first authorship) and also multiple corresponding or “last authorship” positions. The co-first convention is indicated with an asterisk which then indicates “these authors had equal contribution” or something similar. Unfortunately, it is not yet so widespread, and sometimes just ignored, particularly when citing the paper (this should actually change the in-text citation format to e.g. Li, Wang et al. instead of Li et al., but this is even more rarely seen). I have occasionally seen kind of split/declaration also for last author but this is to date pretty rare. It’s generally not as important for established PIs as it is for PhD students to be seen to have higher contributions for a particular paper. In other fields stranger and older conventions apply: authorship contribution purely in order from first to last is the most common, ignoring the relatively new “last author” convention, and some fields even still use alphabetical name order.

Corresponding authors

The corresponding author is the person who undertakes communication on behalf of the manuscript – traditionally of course this was who received manuscript correspondence by mail! However, “corresponding authorship” is sometimes now equated also with the last authorship position in terms of indicating the PI. This is not quite the same across all countries though: in China for instance the corresponding author (regardless of author position) is usually the PI, whereas corresponding authorship is secondary in e.g. Australia and sometimes bestowed even on PhD students if they are the ones communicating with the journal (submitting papers etc.). I have heard an argument against assigning corresponding authorship to PhD students and possibly postdocs, which is that the main point of this is to provide contact details for later communications from other researchers, and if this occurs many years on it is less likely that the contact details are still valid, since junior academics usually move around. I don’t find this a particularly convincing argument though, since realistically it’s not that hard to find contact details for anyone these days online (or shouldn’t be for academics at least).

Assigning authorship: research contributions

Individual research groups vary in how they assign contributions to manuscripts. If you look at the old literature, single author papers even in our field were very common (U 1935, Morinaga 1934, Stebbins 1950). Over time, we start to see more papers with multiple authors, until the current situation today where it is uncommon to see less than five authors on a manuscript, and ten or more is not uncommon (not to mention the occasionally hundreds or more on genome sequence papers etc.!). This is of course mostly due to just research groups getting larger: PIs tend to have more than one or two PhD students at a time now, and don’t spend a lot of their research time in the lab like they would have a century ago – competition is fierce and there are a lot of incentives driving the production of PhDs, which has been said better elsewhere by others. However, part of this change is relevant to the authorship debate: gradually, we are acknowledging the diverse contributions of multiple lab members to particular manuscripts. In many cases it is quite likely that some of these old “single author” manuscripts were supported at least by several technicians, and possibly also by junior lab members, but the only person who claimed authorship was the person who actually wrote the paper, i.e. the PI. Of course PIs have historically and up until the current date shown extreme overrepresentation of whatever the dominant demographic is (white men is classic in Europe and North America, but also e.g. Japanese of Japanese descent rather than of Korean descent in Japan – look up the story of “Nagaharu U”). So often people who would have had culturally no opportunity to undertake academic studies to do a PhD or become a professor would have become technicians instead; this was seen as a “lesser” job (prestige-wise at least; see e.g. the story of Rosalind Franklin). Hence, the exclusion of technical assistants from co-authorship has at the very least nasty historical and possibly current-day precedents we should be aware of (have a look around in Germany at the relative gender ratios of technicians to PhD and postdoctoral students and then to professors, for instance).

So what does and should constitute “authorship” of a scientific paper? Personally, I think the more we cling to this old idea based on writing books that “the author is the one who wrote the paper”, the more likely we are to perpetuate old and possibly damaging biases. Realistically, science nowadays in our field is done in teams. It’s becoming less and less possible for one person to do everything themselves: we rely on collaborations between teams of specialists to undertake research tasks, so that one person does not have to be an expert in everything. Over the long run this works better for everyone and is more efficient. However, it’s also at odds with many things: the historical and often current idea of the “PhD experience” for instance, as a training period where a PhD student should go off on their own and then (with some input from their advisor along the way) return in three or more years with a  complete PhD thesis, for one. But now many labs have a more hierarchical or complex structure, where different members fulfil different roles: the lab needs to apply for money through writing grant proposals, be managed scientifically and financially/administratively, and run day-to-day, including implementation of various protocols, data collection and supervision. These tasks are generally undertaken by not just the PI but by other more junior professors, research scientists, postdocs and/or technicians, as well as of course PhD, MSc and undergraduate researchers.

In my opinion, all contributions to a scientific paper should be acknowledged with co-authorship. Worrying about the contribution to the writing itself is old-fashioned and limited, and I have yet to see any particularly strong argument for excluding technicians, e.g. “just because they did the technical work” or “because they don’t need papers for their job evaluations”. Input and discussion of ideas, project implementation, technical assistance, supervision, data collection, statistical analysis, writing and critical revision of the manuscript, these are all potentially grounds for authorship. We don’t lose out by including more authors, particularly if all contributions are detailed to allow for a detailed dissection of who did what for each project/paper. “Gift” authorship is not okay – including people as co-authors who really did not contribute at all. But realistically, I feel this is less of a problem in our current system than under-acknowledgement of research contributions with co-authorship, particularly when under-acknowledgement disproportionately affects people from demographics which are currently underrepresented at higher academic levels.

Although not all journals require an “Author Contributions” section, I do see this as a way forward to appease academics with more conventional ideas of what “authorship” should entail, as well as giving fair credit to everyone involved in the project. Hence, for my research group, please just always include this section in the manuscript submission, whether it is asked for or not. I would hope we can also have frank and open discussions about contributions to each manuscript and hence author order. I may not always get this right! I may not directly oversee every stage of the project, and may just have a general idea of who was involved with each step, so please bring this up if you are even slightly concerned or feel it may be unfair.

Why do we care so much about authorship?

There is a high degree of importance attached to each type of authorship. First authorship is very important in our field: the number of first author papers you produce is the primary criteria for academic output at junior levels, such as PhD students and early stage postdocs. Experimental papers always count for more than review papers – reviews should really be considered as “bonus” papers, although they do help strengthen the CV overall, particularly if the number of first author experimental papers is already good. Reviews can also highlight that a junior researcher has good writing skills and a good understanding of the literature, which is definitely a plus. However, experimental papers are really the primary mode of currency. Co-authorship is also good in addition to first author papers, although a high number of co-authored papers but very few first author papers will be a red flag for postdocs (e.g. less than 3 first author papers when already a couple of years out from the PhD, unless they’re in very high impact journals, since these types of papers often basically swallow up the equivalent of multiple “standard” papers’ worth of experiments). On the other hand, for mid-to-late career postdocs (getting to 5 + years post-PhD) no or very few co-authored papers is a major red flag: it usually indicates that the researcher is not capable of working well with others and is never invited on to collaborative projects or involved in supervising students. For later stage postdocs and junior professor/group leaders, last author papers are really critical to show group leadership, and implementation of a project from start to finish including supervision of more junior researchers (usually PhD students). So with all this importance attached to papers, it makes sense that authorship is a very big issue…

Who should be first author?

First authorship is generally quite clear, although not always. I actually try to design projects so it is clear from the start who will be the first author, as in who is responsible for the majority of the work, mostly at least the data analysis and writing the manuscript etc. We don’t have to design projects this way (I would be delighted if several of my research group got together and planned out a joint experiment that utilised different skills of each member, for instance) but realistically this first authorship contribution is very important at the early stages of the academic career, so I plan projects like this so all PhDs and postdocs get the chance to get these important first author papers.

Sometimes in our group we have co-written review papers. I think this is quite a nice way to give everyone experience in writing and the publication process, so I doubt we will stop doing this. However, this does make it a bit more complicated: if everyone just wrote a section, how should we decide who should be the first author? To date I have just made a rather subjective decision based on who I think ended up writing the most or produced the biggest contribution, but I will have to think about this in future: probably the best option would be to discuss this before we actually start, then make the first author responsible also for some of the other manuscript tasks like referencing, checking for redundancies between sections etc.

Middle authorship

Co-authorship is pretty flexible, and here is also the question of “does it matter where you are in the authorship list”, if you aren’t first or last? I would actually have to say mostly no, it doesn’t really matter from a CV perspective. Second position is the “best” co-authorship position, but it’s still not so different to any other position on the list really. Realistically the best strategy to highlight co-authored papers in your CV is to describe in a dot point underneath the citation what your contribution to this paper was.

“Senior” author

Last authorship or “project lead” will normally be me. However, I will definitely give this up to anyone who designs and implements a project via supervision of another, more junior academic (postdoc supervising PhD or MSc/BSc, even PhD supervising MSc/BSc potentially, although this is probably not a good time commitment during the PhD since this is when first author papers count the most). Last authorship has to involve a fairly major contribution either to the project idea itself, or to the experimental design that will allow us to test experimental hypotheses. Writing the grant which employed the student definitely counts! Helping supervise and teaching experimental techniques to a student of a more junior level is grounds for co-authorship, but not for project lead status. Is this a bit arbitrary? Possibly, but I think it’s fair: this is generally what is assumed by last authorship position, at least in our field.