How to be a more productive scientist
Or, why working like a nutter won’t help
Wouldn’t it be great if we could be more productive? In (modern) science this is the essence of existence. We are measured by our publishing of research papers and the getting of research grants. The more productive we are in writing papers and proposals for research money, the more successful we are measured to be. So I would wager that, every New Year, scientists from the most space-cadet graduate student to the loftiest, big-cheese Professor resolve to increase their productivity — by whatever name they call it.
Can it be done? Yes. And no. Or, to put it another way, nes. It depends on you. And whether you can change your “Productivity Factors” (oooo, I’ve always wanted to write one of those vapid self-help blogs with a million hits, and now’s my chance!)
Some scientists are staggeringly productive. Whenever scientists are measured by their output, we see the same pattern. There are a few nutters with output so extreme it makes the rest of us wonder how on earth they have time to eat.
How extreme? Let’s take a look. We are going to use the H-index (stop groaning at the back). The H-index is a useful way to measure productivity, as it captures not just rate of output, but the rate of useful output. After all, it’s no good if all of your copious output is turgid crap. We’re looking at you, Uwe Boll and Paul WS Anderson. (Event Horizon was at least half-decent). This H-index is a simple number: it is the number of papers with at least that many citations. So someone with an H-index of 4 means they have published at least 4 papers, each with at least 4 other papers which cited it. Of course, there can be cases where one of those has 1000, and one has 4, but nothing’s perfect. In short, the higher someone’s H-index, the more useful output they have produced.
(The H-index has all sorts of issues for comparing scientists: it will wildly differ according to research field, by age, and by a host of other factors. But is serves our purpose as a crude proxy for “useful output”.)
So, for the whole of academic neurosurgery in 2010, the distribution of H-index scores looked like this:
We can clearly see the hyper-productive academic neurosurgeons, way out there in the long tail.
Systems neuroscience also has these happy nutters. Take a look at these numbers (as of 30/12/2016). These are H-index scores just for citations in the last 5 years
Karl Deisseroth: 98
Gyorgy Buszaki: 92
Karel Svoboda: 75
Xiao-Jing Wang: 68 [theorist]
Larry Abbott: 56 [theorist]
Eve Marder: 48
Ed Boyden: 46
Ninety-eight papers each with a minimum of ninety-eight citations in the last five years? FIVE. There are many scientists who have read fewer papers than this in five years.
(This is not a new phenomenon. In trying to reinvigorate a moribund Royal Society, with the help of John Herschel, Charles Babbage did the most ungentlemanly thing imaginable to a Victorian scientist: he worked out how many papers each Fellow of the Royal Society had published. The outrage of it: measuring a fellow Fellow by his (yes, his, of course) output, like he was some kind of machine, rather than appreciating the quality of his mind. And his capacity to eat every course of a Royal Society banquet. Babbage worked out that a handful of Fellows published most of the Society’s papers — Davy, Faraday etc. And the majority had published precisely zero. Nada. Not a sausage. Disgusted, Babbage flounced off to help form the British Association for the Advancement of Science.)
How can this be? And what can we mere mortals do to get close to that level of output? (If, say, we wanted to, rather than taking the rather well-argued view that the pressure to publish is ruining science).
In 1957, Nobel laureate William Shockley noted the existence of the hyper-productive in science, and offered a way to understand it. Consider, he wrote, if we break down publishing a paper into its main steps:
F1. ability to think of a good problem
F2. ability to work on it
F3. ability to recognize a worthwhile result
F4. ability to make a decision as to when to stop and write up the results
F5. ability to write adequately
F6. ability to profit constructively from criticism
F7. determination to submit the paper to a journal
F8. persistence in making changes (if necessary as a result of journal action).
That gives us eight factors to produce one paper. Now, he says, imagine if our productivity was given, roughly, by the combination of those factors:
Productivity = F1 x F2 x F3 x F4 x F5 x F6 x F7 x F8.
Then the hyper-productive are not special. Say that Scientist A is equal to Scientist B in every factor, except that Scientist A is twice as good at F2 (working on the problem), F4 (knowing when to stop), and F7 (determination). Then their Productivity will be 8 times higher. For every one paper Scientist B produces, Scientist A will produce eight.
In this idea, if someone gets a bit better at every aspect of writing a paper, they become exponentially more productive. The hyper-productive are, then, just a bit better at every aspect than the “average” scientist, and their advantage accumulates.
Great! So this would mean all we need to do to increase our productivity is get just a little bit better at some of the “Productivity Factors”. Doesn’t seem so impossible now, does it?
And, actually, there are practical things we can do for many of Shockley’s factors.
Take the factors of the ability to think of a good problem and the ability to recognize a worthwhile result. These two factors need the same thing. New ideas and new insights all need time and space. And, first, fuel. Read, and widely, outside your immediate speciality; and for pleasure. Fill your subconscious with information, and then give it time to churn. Insights can then strike whenever the mind wanders freely. Baths, breakfasts, staring out of a window.
Here Shockley also offers his favourite idea to explain the hyper-productive. What if, he posits, we have some mental capacity to hold m ideas in our head at once, and we need some combination q of those m ideas to hit on a solution. Then the more ideas one can hold in mind at once, the exponentially more combinations there are to try: and thus the exponentially increased probability of hitting upon the solution.
What about the ability to work on the problem? That’s straightforward. PIs can hire talent; PhDs, post-docs, and PIs can train. Take an advanced course; challenge yourself.
Ability to write. Too much has been written about how to write. But the basic ideas are constant and simple. Writing takes practice. Write everywhere, write often; don’t be afraid to throw writing away. For style, read Strunk & White; for structure read Mensh & Kording.
Profit constructively from criticism. Like most researchers, as a PhD student and young postdoc I was deeply frustrated by reviewers’ criticisms of my papers. But part of scientific maturity is realising this reaction is unfair. Yes, some reviewers are just bizarre; some are out of their depth, and taking it out on the authors; some cannot see beyond their own, very narrow interests. But most criticism, I’d wager, is born out of desire to improve the paper. And indeed, the higher one aims, the better the reviews. Think of it as a collaborative exercise — especially when you are reviewing. Use the reviewers’ comments to improve your work. Use it as guidance. Give one day to rage, then move on.
Incidentally, there is nothing in Shockley’s list of Productivity Factors that is exclusive to writing papers, or science, or even academia. Stripped of the handful of specific nouns (paper, journal), his Factors work exactly the same for any creative process.
Like writing show tunes:
F1. ability to think of a good problem
“Hey, we need a song when Elsa runs away, to capture the rush of emotion arising from her escape from suppression and revelling in her freedom to express herself”
“What?”
“We need a catchy ditty so we know she’s not the bad guy”
F2. ability to work on it
“I’ll play the piano!”, “I’ll hum the words!”
“I’ve got my banjo!”
“Get out”
F3. ability to recognize a worthwhile result
“I got a melody line: hm hm *hmmm*.”
“OK, let’s try some words”
“Let them know, let them *know*’”
“Nah, try again”
“Let it snow, let it *snow*”
“You wanna get sued?”
“Let me go, let me *go*”
“Better, but we need something more vapid. Go 25% more vague”
“Let it go, let it *go*”
“YES! YES! YES! We’ll be rich! Rich! And parents everywhere will burn us in effigy!”
F4. ability to make a decision as to when to stop and write up the results
“I don’t think we need the thirteenth verse.”
“But I love that one, about the villagers’ rejection of unfair per-capita taxation plans and the subsequent litigation by the council.”
“It’s a kids film.”
“But we need something for the grown-ups, something involved”.
“Grown-ups are stupid. There’s a talking snowman. They’ll laugh like drains at that “out of shape” gag”.
F5. ability to write adequately
“Picture this: we bring in some violins in here…”
“Yes!”
“Cellos, an octave down…”
“Yes!!”
“..then the Ynigwie Malmsteen guitar sol..”
“No.”
“But he’s been working on it for months”
“No.”
“It’s a beautiful interpretation of the melody li…”
“It sounds like a cat got trapped in a guitar case that was on fire”
F6. ability to profit constructively from criticism
“Lose the Corey Taylor screaming verse”
“But it’s an expression of Elsa’s inner angst”
“uh-huh, and the bit where he screams “f you motherf-er” at the reindeer?”
“OK, point taken.”
F7. determination to finish the (song)
“One more time from the top, and this time we’ll nail that bloody note”
“Let the storm rage onnnnnnnnnnnnnnnnnnnnnnnuuurgghhhh…. I think I’ve torn something”
“Dag nam it!”
F8. persistence in making changes (if necessary as a result of editorial action).
“Kristen, Robert, I’m sorry, the studio has said that Future’s verse has got to go — and I quote “he just mumbles “Future” barely comprehensibly in an Atlantan drawl for two minutes”. I tried to explain that’s what the whole of Honest sounded like, but they weren’t having it.”
“No problem. We just put that in to wind them up. Anything else?”
“Yeah, they said the bit with a cat in a guitar case on fire has to go”
“What did I tell you?”
Any “productivity” model is likely a total fallacy. Reducing a complex process that takes years to a handful of factors, which independently multiply, is bunkum. But, like all competent models in science, this “Productivity Factors” model is useful: it breaks down a system into parts, suggests ways that the parts interact, and makes predictions. Better, bunkum or not, it doesn’t stop the advice that stems from thinking about the model from being useful. And it strongest prediction is the most hopeful one: a small, linear increase in a few things can make us exponentially better at the whole. Go try it.
If you liked this, please click the ♥ below so other people can read about it on Medium
