John Ionnadis and colleagues have a remarkable finding about scientific authors:
we estimated that there are 15,153,100 publishing scientists (distinct author identifiers) in the period 1996–2011. However, only 150,608 (<1%) of them have published something in each and every year in this 16-year period (uninterrupted, continuous presence [UCP] in the literature). This small core of scientists with UCP are far more cited than others, and they account for 41.7% of all papers in the same period and 87.1% of all papers with >1000 citations in the same period.
This is a disturbing finding, for two reasons. On the one hand, from the viewpoint of an individual scientist it’s frightening that only a tiny fraction of scientists can sustain a publishing career. On the other hand, from a social point of view it’s possible that we are wasting an enormous pool of human talent.
The details of Ionnadis’ study have been questioned, but overall the finding rings true. To my knowledge, every distribution describing scientific literary productivity is highly skewed, following something like a power law or log-normal distribution. For example, most articles get very few citations and very few papers attract many citations.
So from the point of view of an individual scientist, how does one adapt to such a harshly selective environment?
Brian McGill points out that William Shockley was one of the first people to notice the extreme skewness of scientific productivity. Shockley won the Nobel Prize for Physics as the co-inventor of the transistor (he was also a notorious racist). Shockley noticed that most scientists have few career publications while a tiny fraction have a large number. He speculated that publishing success would depend on at least 8 different factors.
Some of these are obvious, such as the ability to think up a good scientific question. Others are less obvious. For example, it’s really important to be able to respond constructively to critical reviews from referees. But Shockley’s model of the steps in scientific writing is not the main point. The key thing is that is that scientific projects involve collections of heterogeneous and often difficult tasks and that you or your team have to succeed at all of them to get a publication.
It’s instructive to think about what it means if the chance that you will succeed at solving one of these problems is independent of whether you will solve the others. Then the probability that you will succeed at the whole task is the product of the probabilities that you will succeed at each individual task. This has many implications:
- Your overall probability of successfully publishing a paper will be less than your chance of succeeding at your worst task (your rate limiting step). Perhaps you have trouble starting a first draft. Attend to this. Any single significant weakness kills your productivity. And in a 8-step process, you have an 85% of being in the bottom fifth of the distribution of the skills that are relevant to at least one step.
- Having mediocre skills at everything is no good either. Suppose you had a 75% chance of success at each step of the process. Then your chance of seeing a scientific project through from initial idea to publication is only about 10%.
- Conversely, there are real gains to be had if you can get better at every step. A 10% improvement in your performance at each step of the process cumulates to a better than 200% improvement in your total productivity.
Shockley argued that if only a few scientists excel at each step of a complex process, then the mass of scientists will be relatively unproductive and a few will have enormous productivity. So if you want a scientific career, you must either develop every critical skill or establish close collaborative partnerships with colleagues who can compensate for your weaknesses.
I don’t think, however, that this is the only or even the best way to think about optimizing your scientific career. It’s cynical to focus only on the quantity of publications. It’s also bad career advice. There are important career returns if you can solve a difficult problem and get that work into a rigorously reviewed journal. For interesting thoughts on how to focus your life on excelling at difficult things, see Cal Newport.
We also need to look at the problem of the scientific 1% from a social point of view. What happens to the 99%? Do they drop out of science completely, or do they contribute in ways that do not lead to publications? Do our scientific institutions make optimal use of the 99%’s motivation, intelligence, and energy to improve human health? This takes us to consideration about what makes a learning health system, and beyond the scope of this post.
