Where Do Scientists Learn to Write?

Janet asks "Where do scientists learn to write?" Well, actually, being a good academic, she asks many more questions than that:

  1. Do scientists need to write well? If so, in what contexts and for what audiences? If not, why not?
  2. Where do scientists really learn to write? What kinds of experiences shape their writing? Are these teaching scientists to write clearly and effectively? Are they entrenching bad habits?
  3. Where do you think scientists ought to learn to write? What are the most important things they need to learn about writing in a scientific context? What are the best ways to learn these things?
  4. Are scientists better off learning to write from scientists or non-scientists? Why?
  5. Are scientists better off learning to write in a classroom setting or in a more "realistic" setting? Why?

Update: Janet has more,
including factors that influence writing quality.

The short answer to the short question is: Grad school, at least for me. Everything I know about good scientific writing, I learned in graduate school.

Of course, I'm also an academic, so I can't leave it at that, not when there are all these other questions demanding longer answers...

1) Do scientists need to write well? Absolutely. It's hard to convince students of this, but good writing is essential. Lifting a paragraph from the guide to lab writing I use in the intro classes:

As a scientist, you can go into the lab and take data worthy of a Nobel Prize, but if you can't explain the results of your experiments clearly and concisely in written form, you may as well not have done them. The key to all of modern science is reproducibility-- for a result to be accepted as the correct result, other experimenters need to be able to reproduce the result. For that to be possible, you need to be able to explain to other researchers all around the world what your results were, how you got those results, and why those results are important. If you can't write clearly, you'll never succeed in communicating your results well enough to get the credit you deserve.

The other important context in which good writing is essential for scientists is grant writing. Modern science is expensive, and at some point, you're going to need to convince somebody to give you money to do your research. That will require writing a proposal of some sort, and your success will depend in part on how clear and persuasive your writing is.

2) Where do scientists really learn to write well? I've talked before about the "paper torture" process my research group in grad school used. Basically, when it came time to write up a paper, somebody would be designated to write a draft, and then there would be a meeting at which the co-authors would rip the draft to pieces. We would argue about every sentence, and sometimes about every word. These meetings could drag on for hours, but nothing went out the door without every author on the paper being satisfied that it was right.

It was a miserable process while I was in the middle of it, but it made me a better writer. I learned what to do, and more importantly what not to do by having my rough drafts torn up and stomped on in front of my face. To this day, I tend to write by doing a rough draft and then asking myself "Where would I have problems in paper torture?"

(It's not a fast process, though, and I've internalized it enough to keep me from ever being a really prolific blogger-- even in an informal context like blogging, I keep fiddling with the wording of posts as if my advisor were going to be taking a red pen to them later...)

3-5) Where should scientists learn to write? The final three questions are all asking basically the same thing, and it's a trick question. There's no one correct answer that will work for everyone-- different scientists will learn to write in different ways, and what works for one person may not work for another.

To give slightly more specific answers to some parts of this, I think that scientists should definitely learn scientific writing from other scientists. Ideally, you should learn from someone in the same field that you're going to be working in, but at the very least, you need to learn from scientists, not humanities types.

I would say something similar for any field-- law students should learn legal writing from lawyers, humanities students should learn writing from humanities faculty, etc.-- because each field of study has its own quirks and conventions, and the only way to learn them is to write within that field. Scientific writing is different than scholarly writing in the humanities-- for one thing, very few scientists write books-- and scientific writing in physics is different than scientific writing in chemistry or biology. (I know this because the chemists and biologists at work teach their students to do some things in lab reports that drive me up the wall...)

The best example probably has to do with writing length. Scholars in the humanities are expected to produce books, or failing that, journal articles that run to tens of pages, and their writing reflects that expectation. Most scientists write primarily for journals (other than the Ph.D. thesis, most scientists don't ever write full-length books), and papers running to tens of pages are fairly common. In physics, though, the top journal (Physical Review Letters) has a strict limit of four pages, including all figures, tables, and references. That limit forces a certain economy of style, in order to pack as much as possible into a short space without completely losing the reader, and causes physics papers to read very differently from papers in disciplines that don't regularly work under such tight constraints.

As for the appropriate context in which to learn writing, I would tend to lean toward a "realistic" setting rather than a classroom setting. That may just be me, though-- I never found formal writing instruction to be all that useful, and really only learned to write papers when I had to start doing "paper torture" in grad school. I think it's a little too easy to shrug off classroom assignments for the important lessons to really take hold, even if writing quality is graded heavily.

I'd love to find a way to simulate "paper torture" in a classroom context, but I think the mechanics of this would be really difficult. I'm not sure how to grade students on their participation in such a process, or how to get them to tak it seriously enough to really learn the important lessons. The goal would have to be to get them to torture their own papers, but I'm not sure that a one-term class for a single grade is really going to be enough to accomplish that. I'm open to suggestions, though...

(The other good "learn to write" strategy I would suggest is "marry a lawyer." These days, when I have something really important to write, I get Kate to go over a draft for me. She spots all sorts of problems that slide right past my eyes, and never fails to improve my writing.

(Of course, this is perhaps not the most practical strategy for students in general. Particularly because Kate's alreayd married to me, and you can't have her...)


More like this

During my office hours today, a student asked me whether, when I was a chemistry student, the people teaching me chemistry also took steps to teach me how to write. (The student's experience, in an undergraduate major in a scientific field I won't name here, was that the writing intesive course…
My Quantum Optics class this term is a junior/ senior level elective, one of a set of four or five such classes that we rotate through, offering one or two a year. We require physics majors to take one of these classes in order to graduate, and encourage grad-school-bound students to take as many…
I'm following up on yesterday's post on where scientists learn how to write (and please, keep those comments coming). First, Chad Orzel has a nice post about how he learned to write like a scientist. It involves torturing drafts on the rack, and you owe it to yourself to read it. Second, I'll be…
Via Bookslut, there's an article in the Chronicle of Higher Education about whether reading is really important: Is it always a good thing to read an entire book? When I was a graduate student, it dawned on me that I often had the most intelligent things to say about books I'd only half- or quarter…

I'm not a scientist, but in my capacity as an academic science librarian, I took a general chemistry course last fall to further familiarize myself with the terminology of the discipline.

In the course, there was an interesting set of writing assignments: three times during the semester, students were asked to use one of their labs to write a portion of a research report. They didn't have to write an entire report, but they did have to write a section or two (abstract, intro, methodology, results, discussion, and so forth) following ACS publication guidelines. In this context, I got to teach them a bit about doing background research for such a report.

I'll be honest with you, I'm not certain how well this translates to the graduate writing process. I did have to write research papers in library school, but most library science master's programs do not require a thesis. Also, by the time I began graduate school, I had a substantial freelance writing career behind me, so my major challenge was to shift my default tone to writing for a research audience. (To this day, a comment I ALWAYS get from journal editors is that my tone is a little too casual.)

But it might help.

By Genevieve Williams (not verified) on 07 Feb 2007 #permalink

My thesis advisor was (and still is) a stickler for good, clear, concise writing. It really does come from feedback, feedback, and more feedback. One paper I was writing with my advisor I ended up saving a 1 foot tall stack of various drafts (fortunately, we recycled). A few things I always keep in mind: all symbols and terms should be clearly defined, so an audience reasonably familiar with your field doesn't need to do lots of supplementary reading (i.e. I need to read ref. 9 of your paper to understand it, but in order to understand ref. 9, I need to read its ref. 2, etc.), and the paper should have a logical flow - after the introduction, every section and paragraph should in some sense answer a question the previous one has raised.

It doesn't hurt to stress these things to undergraduates, even on tests. I now make sure I tell the students that I'm treating their answers like research papers, in that they must be reasonably clear and concise, and not have any 'assumed knowledge.' Even if the answer is hidden there somewhere on the page, it doesn't matter if I can't find it or figure out where it came from after 30 seconds of searching. A lot of referees won't even search that long for an answer.

The only way to simulate 'paper torture' in a class room is to make them do 'paper torture'. If you do in class it should bet has be no harder to grade then classroom participation. As an undergraduate I have had to write several group papers and I felt I learned a lot from these assignments. It is the sort of assignment that ends up with most getting a good grade, but is that a problem if the students are learning.

When I took calculus our professor would give us group quizzes with very hard questions. I learned more taking the quizzes , in which we all got good grades, then in doing the homework.

I also don't post on blogs as much as I might because I am looking over my shoulder at what I am writing (changed this sentence four times already).

Good luck finding a good way to do paper torture in a classroom context. Even if it worked, your student evals would be terrible.

I do something that is sort of an approximation, where I force students do an outline, then a first draft (actually, there are three steps before these, but they are mostly proposal and research). After each of these, I grade the effort, then a week after they submit, we meet individually and I get to tear into it for about 20 minutes. Not long, but given that there are 24 students in the class, it takes me one long day of individual meetings to accomplish this. Then, before they hand in their final draft, they have to give a ten minute presentation in the class, during which anyone in the audience is encouraged to rip into the paper.

Final draft is due a week later.

By boojieboy (not verified) on 07 Feb 2007 #permalink

I am a recently graduate double major in physics and history. A response I got from an corporate HR representative on why I was turned down for an entry level technical job was due primarily to my history degree. Disregarding 3/4 of my CV deals with physics, the fact that I have a Humanities degree made me "uncommitted to a technical field." And people always bitch about science undergraduates lacking composition skills.

Stupid people do not know what they want from candidates. And where the hell do people who work in Human Resources come from? Argh.

Probably the main way I learned how to write science papers was by reading a lot of them. I remember that when I was writing for Nature, I gathered up a bunch of Nature papers from my discipline and analyzed their structure. Nature doesn't usually use headers to define the intro, background, data, results, so I had to pay attention to see how much space was given to each section. Then I wrote my paper using the same type of structure. I guess it worked, because it was accepted without any big glitches.

As a current undergrad, one of my main criticisms of my course is lack of teaching on how to write. Over my 3 years at Imperial, I have only had to write around 9 lab reports and a 2000 essay, a relatively tiny amount compared to soft degrees.

Each time I sit down to start writing, I always have to spend a a few hours 'finding my scientific voice'. It's a huge pain getting there, but when I do get into the flow of writing scientifically I do appreciate how useful a skill it is. When I do start to write my final year project report and PhD thesis, I'm sure I'm going to regret these years of cruising through with only maths in my pad.

I think everyone has missed something pretty important.

Good science writing is when you are able to get a specific point across in the most concise and understandable way possible. Great science writing is something else entirely. The best science writing is more similar to good literature. How do you learn to be a great science writer? Read and study literature!

As a science writer you have a story to tell! You might as well keep people interested and excited!

(and of course get your papers shredded, outlined, crossed out, ripped apart, crumpled and crapped on by as many people and as many times as possible)

This may only be relevant for grad students and postdocs, but in addition to paper torture (with a qualified torturer) a useful exercise is to review someone else's paper. Its very instructive to sit on the other side of the fence, and in science it often happens.
Perhaps such an exercise can be extended to an undergraduate classroom. But I've never tried.


See my post #3. I have reviewed both papers and also a talk we have to give as a senior to graduate for fellow classmates during my undergraduate career. It needs to start at the undergraduate level. Small classes help, but if you have to have large classes try breaking them in into groups.

Make them write papers in groups like the real world. How can one learn to peer review if one does not do it. Graduate school is too late, a good graduate student should already have many of these skills.

Good writing should start in high school with the basics (grammer, structure, etc.). The first time I taught microbiology to undergrads, I made them write a report about an organism of their choice. After two all nighters with a red pen, I just about gave up. Was it my job as a microbiology lab TA to teach the concept of verb agreement? I gave it my best shot, but eventually gave up.

Grad school should teach good science writing, but ultimately it comes down to "just do it". There is a misconception that writing should be easy. It is not. It takes time and skill and it is work. The more you do it though, and the more you get creamed by advisors and peers, the easier it gets.

Now I get a lot of papers to review. It always kills me to see excellent data get overwhelmed by poor writing. It is the writers job to convince me of an argument, not merely to present some data. I have no problem rejecting papers where I have to work to figure out a thesis.

Finally, I highly recommend every writer read this classic by Orwell. If science writers would follow these rules, the jobs of science readers would be consdiderably easier.