Aboukhalil (2014) examined ~24 million papers in the Pubmed database and found that during this time authorship has increase from a mean of 1 author per paper (in 1913) to ~5 authors per paper (in 2013).
Authorship trends in the Earth sciences have previously been investigated by Engelder (2007). However it seems that Engelder only looked at a sample of 100-150 papers per year from 1973-2007 in the journal GEOLOGY. Engelder also noted a trend of increasing authorship — the average number of authors rose from 2 to 4 between 1973 and 2007.
I was wondering if an increasing number of authors per paper is observable in other Earth science journals. To focus on a single subdiscipline, I used authorship data for all 1,425 articles from Journal of Geophysical Research – Earth Surface (2003-2016) that I retrieved from the Web of Science for a previous post on JGR-ES references section size. The distribution for the full dataset can be seen below:
(side note: two papers have 31 authors — see here and here)
Now let’s break down this authorship data by year to see if there is a trend toward more authors per paper:
Two observations to note here:
First, the mean changes from 3 to 4 authors in 2012. The slight trend for increasing authorship through time, and the number of authors is broadly consistent with Aboukhalil’s data (mean of 5 authors in 2013) and Engelder’s data (mean of 4 authors in 2007).
Second, 2010 seems to mark an increase in papers authored by 10 or more people in JGR-ES. To investigate this further, I extracted the number of papers with 10 or more authors for every year. These large author lists seem more infrequent prior to 2009 (only 4) compared to after 2009.
- I wonder if this trend is also seen in disciplinary journals with larger datasets (both in papers per year, or longer publication records) — GRL is my next target, with >20,000 papers since the 1974). Looking at the full dataset from GEOLOGY would also be good.
- Trends in rising authorship could reflect the sentiments of next generation faculty members, who prefer collaboration over competition see Trower (2012) for examples.
- Rising authorship could also indicate more complex or interdisciplinary work, though I think there are other ways to measure interdisciplinarity (I hope to explore this later).
(the last 2 plots were made using Tufte in R, a very cool reference)
Update (9/22/2016): a colleague pointed me toward a recent paper on authorship counts in Ecology journals (Logan, 2016)