Is there an inflation in the number of authors per paper?

Question

In most (or at least many) fields of academia, peer-reviewed publications are essential. For a compilation thesis, no papers means no PhD. For tenure, you need papers. To get grants, you need papers. Universities may distribute internal funds based on the number of papers per group. In short: publish or perish.

On the other hand, it is quite cheap to offer someone co-authorship. Send a nearly finished manuscript to a colleague/friend at another university for review... colleague reads it, offers some advice, perhaps just minor. First author offers co-authorship in return, and colleague has another co-authored paper for possibly less than a day of work. One can discuss if it is the right thing to do, but that is not my question here. It happens. (NB: I am not suggesting such has happened in the examples listed below!)

Criteria for co-authorship differ per field, but some papers have a lot of co-authors. Perhaps due to having an instrument that was used in an inter-comparison/validation study. Some examples of papers with lots of co-authors, not particularly extreme:

^{_{Kasai, Y., Sagawa, H., Kreyling, D., Dupuy, E., Baron, P., Mendrok, J., Suzuki, K., Sato, T. O., Nishibori, T., Mizobuchi, S., Kikuchi, K., Manabe, T., Ozeki, H., Sugita, T., Fujiwara, M., Irimajiri, Y., Walker, K. A., Bernath, P. F., Boone, C., Stiller, G., von Clarmann, T., Orphal, J., Urban, J., Murtagh, D., Llewellyn, E. J., Degenstein, D., Bourassa, A. E., Lloyd, N. D., Froidevaux, L., Birk, M., Wagner, G., Schreier, F., Xu, J., Vogt, P., Trautmann, T., and Yasui, M.: Validation of stratospheric and mesospheric ozone observed by SMILES from International Space Station, Atmos. Meas. Tech., 6, 2311-2338, doi:10.5194/amt-6-2311-2013, 2013.}}

^{_{Milz, M., Clarmann, T. v., Bernath, P., Boone, C., Buehler, S. A., Chauhan, S., Deuber, B., Feist, D. G., Funke, B., Glatthor, N., Grabowski, U., Griesfeller, A., Haefele, A., Höpfner, M., Kämpfer, N., Kellmann, S., Linden, A., Müller, S., Nakajima, H., Oelhaf, H., Remsberg, E., Rohs, S., Russell III, J. M., Schiller, C., Stiller, G. P., Sugita, T., Tanaka, T., Vömel, H., Walker, K., Wetzel, G., Yokota, T., Yushkov, V., and Zhang, G.: Validation of water vapour profiles (version 13) retrieved by the IMK/IAA scientific retrieval processor based on full resolution spectra measured by MIPAS on board Envisat, Atmos. Meas. Tech., 2, 379-399, doi:10.5194/amt-2-379-2009, 2009.}}

(Again, I would like to stress than I am absolutely not implying that there is anything inappropriate about these two examples!)

On the other hand, I rarely see papers written by sole authors, and I have the impression that such papers were more common in the past — but I have no evidence thereof.

Is there an inflation in the number of authors per paper? In other words, is the number of authors per paper increasing and if so, does this reduce the value of a co-authored publication?

Related: What is the average number of articles written per author in a year and has it increased recently?

Answer 1

I take inflation to mean that the number of co-authors grows faster than what the content/effort of the research merits: that would imply that if there is a non-negligible amount of such an inflation, it does affect the value of the authorship.

• Number of co-authors per paper is increasing
• Co-authorship inflation is perceived as a problem
• Perceived amount of contribution depends greatly on the position in the author list: first, last and corresponding author are perceived as contributing much, middle authors are perceived as contributing only a bit.

---

Long version:

First of all, the number of coauthors per paper is clearly increasing, e.g. pubmed provides statics about such questions:

Let me mainly take the optimistic position and list sensible valid reasons for increasing numbers of coauthors.

Some are well known and widely discussed

• large combined scientific efforts like big instrumentation, the practical implementation of such long author lists widely and somewhat controversely discussed.

• Increasingly interdisciplinary research

• But I think that also the density of researchers has increased, which greatly facilitates collaboration.
  E.g. I'm in a 100 000 inhabitant university town. The university has about 20 000 students and 7 000 employees (incl. professors - not sure whether this count includes technical personnel or only research staff). That alone is more than the whole town had inhabitants in 1900. PLus we also have a university of applied sciences and a number of non-university research institutes. I'm in of those reseach institutes, with about 300 employees. So there are several thousand researchers with whom I can collaborate even person to person by bicycle/foot.
  This high concentration of researchers facilitates intra- as well as interdisciplinary collaboration. These papers then naturally have more co-authors. Say, an "instrumentation" group develops a customized sensor for a group tackeling some application and yet other people develop the data analysis for the paper.

• In addition, email, skype and cheap travel (plus I'm in the luxirous position that there are basically no legal travel restrictions as I'm German and EU citizen) makes it much easier than, say, 30 years ago to know, meet and collaborate with colleagues from all over the world.

• Specialization, particularly now that I'm at such a big institute.
  E.g. where I'm now I usually receive readily prepared samples for measurements. Actually, being specialized on data analysis I often receive just the measured data (and I'm very lucky if people bother to have a chat beforehand on the design of experiments with me). Someone else prepares the samples and someone who mainly works on instrument development does the measurements. On contrast, where I was before everyone did all of that for their own topic and samples (of course also having emphasis on some part of this work flow). Of course all these people here contribute significantly to the paper.
  But it also means that there is a continuous distribution size of contributions. I've somewhere seen a notion that weights the papers by 1/total no. of authors.

Of course, also abuse of co-authorship, such as honorary authorship, does happen, and maybe the specialization can become a salami-slicing of contributions.
I very much like the possibility of including a "contributions" section and decided to do that whenever possible. I think it can help checking against the abuse. At the moment (in my field), I think the existence of such a paragraph alone is a quite strong sign of no abuse of co-authorship.

But I think there are also valid reasons that mean that nowadays more authors are on a paper without the amount of work of the different persons involved having changed:

• Nowadays, sometimes technicians who did a lot of the work (and often also contribute to the development of the lab methodology) are mentioned.

• Also I believe that students who do research nowadays have a far better chance to end up on the author list.

• Maybe a gray zone, which also depends on customs/tradition is how to deal with the higher-up levels of supervision:

  • Vancouver says: providing funding alone is not sufficient (and of course the DFG goes along the same lines)
  • German tradition says: head of the institute is responsible for all that is going on in his/her institute, and thus is always included.
    To be clear here: this does not mean (and AFAIK has never meant) only an organizational responsibility, but a scientific responibility, i.e. supervision of the project. The gray zone IMHO comes from the fact that the proper contribution can superficially look similar to improper (i.e. no proper contribution) -- it is difficult to judge from the outside: A very good supervisor may guide in a way that is hardly perceived. If this good supervisor is looking after a good student, after putting his intellectual facilities to the project may find that the good student does well, and not many changes are needed. This is a proper contribution. Yet it superficially can look very similar to a bad supervisor who does not contribute his intellect to the project or paper and just waves everything through - regardless of whether the input is good or not.

---

I think this paper is interesting:

Wren et al.: The write position. A survey of perceived contributions to papers based on byline position and number of authors, EMBO Rep. 2007, 8(11), 988–991.
DOI: 10.1038/sj.embor.7401095 PMCID: PMC2247376

In addition, we also asked respondents for their perception of general trends and attitudes towards authorship of scientific publications. Forty per cent of the respondents (35/87), for example, agreed that granting authorship to someone who does not meet journal authorship criteria was a common occurrence. Half of the respondents also agreed that author inflation makes it significantly harder to judge whether or not a candidate merits promotion.

While this does not answer the question whether there objectively is an inflation in co-authorship, it means that this is at the very least widely suspected and perceived as a problem.

Also, the outcome of that paper IMHO boils down to: perceived as authors are the first, last and corresponding authors, the middle authors are generally perceived far less.

Personally, I share the suspicion that a significant amount of co-authorship abuse happens. However, my field is small and I think I have a reasonably good overview of what is going on. This includes a (subjective) idea of where I'd suspect honorary authorship or small contributions and on the other hand also some idea of who likely contributed what (specializations) to the paper. In addition, of course the listing of the instituions makes a lot of that clear (e.g. if someone from a statistics department, someone from a clinic and someone from a spectroscopy lab is listed that gives me a very good guess who did what).

Answer 2

Is there an inflation in the number of authors per paper? In other words, is the number of authors per paper increasing and if so, does this reduce the value of a co-authored publication?

Bit late but somewhat inspired by this question I did some research on author inflation within PLOS journals (due to their nice API).

The full write-up is here, should you be interested, but the TL;DR is that author inflation does indeed appear to be happening, at least in this selection of journals mostly from the life sciences in recent years.

Here are linear regressions per journal of yearly mean number of authors per paper:

In one of the comments I was linked to a much longer-term study which revealed the same trend in a prominent chemistry journal.

The second part of your question is harder to answer but cbeleites has given some good insight and references. +1 for a very interesting question.

Answer 3

"On the other hand, I rarely see papers written by sole authors"

A lot depends on the field you are researching. In the sciences you'll often find papers written by 6+ authors - probably because they're a collaborative effort between a team which might be spread across several institutions. In the arts, however, it's not uncommon to find single author articles - especially in fields such as classics and ancient history. If you look at the publications by staff in that department at the University of Manchester you'll find many single author papers:

http://www.alc.manchester.ac.uk/subjects/classicsancienthistory/people/

Looking through the 100+ publications I cited in my ancient history articles, I can find only one which had more than one author, whereas in my computer science thesis there were only 12 with one author - most of which were unreviewed technical reports.

Answer 4

There are two types of dangers when it comes to publishing as a basis for evaluations. One is certainly as you mention more authors included although they have not fulfilled the basic criteria as for example outlined by the Vancouver protocol. A second effect is so-called salami-slicing where the results are sliced to produce as many publications as possible. There are tendencies such as these and journals have started to act against the by requesting disclosure of contributions by the authors. Salami-slicing should be corrected through the review process and may be more difficult to identify since reviewers and editors do not know the full extent of any particular project.

Against all this is the fact that science has over roughly the past century (different in different disciplines) steadily moved towards larger groups and consortia performing research. This results in many co-authors, particularly on papers synthesizing results from the larger projects. The number of authors have therefore increased but due o several and opposing reasons.

The value of co-authorship has therefore also changed over time. I believe the view of co-authorship varies between disciplines, maybe even a lot. In the disciplines with which I am familiar authorship alone is not sufficient to value a paper. For better or worse, we also look at the impact factor to try to assess the value of co-authorship. This means it may be possible to value a co-authorship of one key paper as more valuable than first authorship of another more run-of-the-mill paper. What this implies is that valuations are not necessarily simple arithmetic although that is certainly how it is often treated. In terms of a thesis, there was a time, not too long ago (when I finished my PhD), when single authorship was looked as the only acceptable form but now, it is a rarity. we do however, require all papers to be listed with a detailed author contribution.

Clearly the main problem is different kinds of free authorships. as this becomes common so will actions to reduce the problem. Top journals have started this and I am sure many others will follow. At the same time the reasons for "cheating" must also be reduced which puts responsibility on persons evaluating applications where publications constitute a basis for decisions.

Answer 5

As one of the authors of one of the two shown examples I can definitely say that the long authors list is the result of a large cooperation: 19 of the over 30 authors listed come from more than 10 internationally distributed institutions. Validations are always large undertakings involving many data sources from other groups, and each group has to do some work for such a publication. Especially in the space/satellite segment, data acquisition and result retrievals cannot be done by a few persons, these are decade long processes with many people involved. So the trend to many authors is just a reflection of the fact that research is getting much more complex in effort, money, and material involved, the times when sole researchers can produce scientific results of large impact in their ivory towers are long gone. No surprise at all, and surely, not a sign of some "science fraud"...

Answer 6

After reading the following article, your count does not bring much of a surprise.

The paper, published in the journal G3: Genes Genomes Genetics, names 1,014 authors, with more than 900 undergraduate students among them.

The corresponding author was questioned as to whether everyone did make sufficient contribution.

The paper’s senior author, geneticist Sarah Elgin at Washington University in St. Louis, Missouri, says that large collaborations with correspondingly large author lists have become a fact of life in genomics research. “Putting together the efforts of many people allows you to do good projects,” she says.

If you really want a look at the paper, it is available here:

Leung, W., Shaffer, C. D., Reed, L. K., Smith, S. T., Barshop, W., Dirkes, W., ... & Yuan, H. (2015). Drosophila Muller F elements maintain a distinct set of genomic properties over 40 million years of evolution. G3: Genes| Genomes| Genetics, 5(5), 719-740.

Well, if 1014 isn't enough, then how about 5000+ authors:

Only the first nine pages in the 33-page article, published on 14 May in Physical Review Letters1, describe the research itself — including references. The other 24 pages list the authors and their institutions. (Ref.)

The paper has exactly 5,154 authors and is the paper to have the largest number of authors ever known. You can find that paper here:

Aad, G., Abbott, B., Abdallah, J., Abdinov, O., Aben, R., Abolins, M., ... & Abulaiti, Y. (2015). Combined Measurement of the Higgs Boson Mass in p p Collisions at √s= 7 and 8 TeV with the ATLAS and CMS Experiments. Physical Review Letters, 114(19), 191803.

Back to the question, does increase in the number of authors decrease the value of the co-authored publication?

The honest answer would be, it depends. It depends on the field of publication as well the impact of the research produced along with so many other factors.

Scientists are trying to popularise the word 'hyperauthorship' as an umbrella term to cover such papers.