Completely uncontroversial PI comment on paper writing

Sep 29 2016 Published by under Careerism, Conduct of Science

Go!

46 responses so far

  • Former Technician says:

    This describes my old boss completely. He didn't write anything anymore, just edit.

  • L Kiswa says:

    Don't first authors (postdoc/GS) do all the writing?

  • DJMH says:

    The only thing I find controversial here is "make sure they are right". By what, independent replication?

  • DrugMonkey says:

    Don't first authors (postdoc/GS) do all the writing?

    Nope. As with many of these little cultural things in science, practices vary tremendously. I imagine the full range is from PIs who never do anything other than edit to PIs who write the whole thing and barely let anyone else edit.

    Also as with many of these things, people express astonishment and occasionally outrage that any other laboratory does it in a way other than the way "it is supposed to be".

  • aspiring riffraff says:

    is it a question of point in career or skill in hiring?

  • UpstateNYProf says:

    Heh. Good luck "just revising" with new graduate students. For me, editing and training folks to write have become more of an undertaking than writing the manuscripts myself.

  • Postdocin it says:

    "is it a question of point in career or skill in hiring?"

    Curious what the more experienced have to say here. IME, boss wrote R grants, students/postdocs wrote papers, F grants, etc.

    My thought here is that PI sets tone. If PI wants trainees to do writing (and is willing to put in the time/effort/training) to help bring writing up to par, then it happens.

  • xykademiqz says:

    LOL "just revise them"

    Reading that first draft of the first paper of a newbie student is an unspoken circle of hell. That's not a paper, it's usually a steaming pile of $hit. It would be so much easier to scrap the whole goddamn thing and write it anew by myself. But we don't do that, because we want to train students how to write like professional scientists, which they do by the time they graduate. So other people writing the majority of the paper does not become a benefit at all until they are almost ready to leave the group; the rest of the time it's an exercise in hair-pulling and swearing over lack of subject-verb agreement, hyphens and commas gone AWOL (or sometimes multiplying like rabbits), adverbs letting loose... Just lots of unclear, crappy technical writing that needs lots of TLC to start looking presentable.

    Btw, I still sometimes write papers from beginning to end by myself, probably 1-2 per year. Usually it's writing up something based on the data collected by a student who left, a book chapter or a review article, or a conference proceedings paper. I love writing and I can do it fast. It's one of the ways in which I am well suited for being a PI, but that I admit I didn't know would be quite as important as it is when I was contemplating careers.

  • drugmonkey says:

    "Skill in hiring" presumes one has a broad choice of ~identically skilled prospective trainees that differ only in the ability to write papers. So all the PI has to do is skillfully suss that remaining talent out.

    I'm sure it is just like that in every other PI's experience.

  • A. Tasso says:

    Agree with @xykademiqz... Editing trainee papers is my least favorite thing to do because the whole time I keep thinking to myself "I could do this so much faster by myself".

    Now that I am taking on more mentees I understand how valuable I was to my mentor. Even when I was just starting out as a postdoc I could give him 98%-completed papers and he would barely have to do anything for that senior authorship. Unfortunately I haven't found any post docs who are as efficient and good as writers as I was at their stage.

    Perhaps ScienceGurl is just overestimating the ability of her hiring pool.

  • Agree with xykademiqz 100%. Sometimes the first draft of the first paper from a new student makes me want to gouge my own eyeballs out to stop the pain. The uncountable ways people can butcher language are all on display in first drafts from inexperienced writers. It is a good feeling, though, when that same person hands you readable first drafts of their PhD thesis chapters (even the ones that aren't manuscripts!).

  • Michael says:

    I echo xykademiqz and add that clear writing is one of the most undervalued skills in science. Our trainees here go through a fairly comprehensive scientific writing training, but it seems to make little difference. It's not just language issues per se (perhaps understandable if English is not the first language), it's not being able to logically arrange sentences and make coherent arguments. I ask most of my students and postdocs to start with a very bare bones outline of their paper (using just bullet points), assuming that this should not be a problem, since all experiments and their importance have been discussed ad nausea at that point and we usually don't start writing unless there is a complete set of figures. However, it's usually that stage where most of the problems are....

  • serialmentor says:

    This may sound like a tempting proposition, but I'd say I'm essentially at that point and it's not that much fun. I spend much of my days copy-editing and doing quality control, and it's mostly things like making sure the figures are all in the right order, all the supplementary files are available, things are consistently named throughout the manuscript, the manuscript uses consistent tense, the references are correct and complete, etc. There's such a long list of things that always needs to be checked, and apparently I'm the only one that really pays attention to all these details.

  • drugmonkey says:

    My thought here is that PI sets tone. If PI wants trainees to do writing (and is willing to put in the time/effort/training) to help bring writing up to par, then it happens.

    A significant factor here is not just putting in the time/effort/training but being able to absorb the delay. That in and of itself interacts with trainee motivation and gumption. Some PIs are in the position that a 6 mo plus delay waiting out the training process is of negligible impact on their lab's success. Some are not.

    We go through this now and again on the blog and I am reminded of a simple axiom.

    All trainees on the Internet are awesome, motivated ready-to-be-PI individuals being held back by their terrible PIs who won't train them with their time and effort.

  • poke says:

    "All trainees on the Internet are awesome, motivated ready-to-be-PI individuals being held back by their terrible PIs who won't train them with their time and effort."

    And all the PIs on the internet were outstanding writers when they were trainees, unlike the incoherent kids that pass for grad students these days, right?

  • L Kiswa says:

    Curious -- how does everybody provide feedback/iterate on manuscript drafts? I'll usually have the GS/PD prepare a full draft, but find that I end up essentially rewriting the intro/discussion in subsequent revisions. Certainly, it'd be more efficient for to me write those sections in the first place (I'd probably end up trashing my first draft on a subsequent iteration...:\ ). Looking back at my lab's first few papers, we've had an average of ~34 versions before initial submission. I'm OK with the process and its pace, but wonder if I need to rethink this.

    Particularly interested in how other newPIs handle manuscripts from their labs.

  • Ola says:

    I have found the only way paper or grant writing works with trainees is in-person. Sure, have them draft it, have them get everything (or just a few things) in place, maybe shuffle it past a few other people. But, when it comes to go out the door, every paper gets the "week sitting by the PI in the office" treatment. None of this "track changes" BS. No back and forth between Mac and PC. Just me and the first author, one working copy, one computer, one monitor, a pot of coffee and a lot of questions....

    "Do you really mean that?"
    "Go find a better reference for this"
    "What are you trying to say here?"
    "Where are the error bars for figure 2?"
    "What? You're telling me now that fig 3 data is only n=3, WTF?"
    "Have you got a copy of the instructions for this journal, so we can see if the abstract is too long?"

    On that last point, it still amazes me that trainees will often arrive to a manuscript writing party (!) without such information in-hand. As if I've got frickin' time to just go looking for that shittio on the internets in the middle of a writing session. I've also found in-person writing to be much better for getting the trainee's viewpoint across.... often some subtle point will come to light in direct one-on-one conversation, which is missed in a track-changes email volley.

    I don't know how to do it any other way. Call me a control freak, but it works.

  • Rheophile says:

    What fields tend to have PIs do 100% of the writing? Is this more common when you are closer to clinical work?

    I think Ola's right - I'd love to be able to do lots of careful work, reading through and editing together. Of course, I think 100% of the papers I wrote this year, I was in a different city than the other authors.

  • artnscience says:

    I agree with Ola: in person, painful re-writing is the only way that will help the person improve. But some people don't realize they need to improve - they don't want to take the time, and they whine and complain whey you try to help them.

    I have a trainee (postdoc) who likes to dump his tragic wreck of a first draft on my desk and then nag me every few days about whether I've finished fixing it yet. He also pouts when I point out the obvious errors and ask him to correct them before I work on his paper. He also complains that correcting the small details isn't really that important. And he likes to push back with stuff like, "well, when I worked in so-and-so's lab, they did it my way", or "I saw it somewhere in a published paper once so it must be OK" - usually about something ridiculously inappropriate, like publishing a figure with only n=2.

    This is not the way to impress your PI! Not surprisingly, I "can't find the money" to keep him in my lab much longer.....

  • Postdocin it says:

    "All trainees on the Internet are awesome, motivated ready-to-be-PI individuals being held back by their terrible PIs who won't train them with their time and effort."

    Thanks for the response, DM. While I know that there are many PIs who cannot afford to take the time delay, it never hurts to be reminded. And, while I do think most trainees (think they) want to be trained to write better, there are likely many who are not "trainable" in a reasonable timeframe for many. Which inevitably leads to the question of what is the purpose of grad school, postdoc, etc.

    Full disclosure: my writing stank when I started, and is potentially passable now.

  • I use track changes--I mark up the text, and then discuss my comments in person with the first author. I agree with Ola--there is no substitute for live discussions to make sure all is clear (both what the student meant to say and my comments on what is actually written). Early drafts are primarily about fixing the logic, and getting rid of "book report" writing (you know--"first we did this, then we did that, then we did this, then we did that" rather than letting the data tell a story). Later drafts are about fixing the language. For my students' first papers, it normally takes 15-20 drafts. Sometimes, I luck out and a student can already write well, but that is rare. Even with a good writer, the first paper takes 10 drafts. After the first author and I are happy with the manuscript, we send it to the other authors for comments.

    Ola, I am impressed that you can carve so much writing time during work hours for each manuscript. I often do editing at night when I can have uninterrupted time at home after the ProdigalKids are in bed, so a writing party is impractical.

    All these postdocs who can't write come from labs where the PI does the bulk of the writing. I think science writing is a core skill, so I take the time with my students.

  • JL says:

    "Even when I was just starting out as a postdoc I could give him 98%-completed papers and he would barely have to do anything for that senior authorship."

    Tasso, must be terrible to write papers with those senior authors who barely do anything. All they did was come up with the idea for the project, worked for many years to get it funded, hire the post doc and provide her/him all the resources and an environment conducive to the successful project. They shouldn't even be in the acknowledgments!

  • Dave says:

    Most students I work with cannot write a paper that makes much sense and that encompasses all the skills needed to do it, from data analysis, stats, figure preparation and referencing. For some, it's relatively poor English, but what I've been amazed about is how many students don't know how to draw figures or do T-tests. What are they getting taught in those first few years of grad school? I guess you have to train the good students piecemeal, and some students just aren't worth the effort...frankly....and you might as well just do it all yourself and send them on their way. I constantly tell students that if they do not develop their writing skills, it will be very difficult for them to make it beyond post-doc level. Most of them just don't believe me.

    One thing I also find difficult is explaining to some students how the submission/revision process will go down. For example, you most likely will have to do *that* experiment whether I ask you to do it, or a reviewer does!!!

  • Luminiferous Aether says:

    "I constantly tell students that if they do not develop their writing skills, it will be very difficult for them to make it beyond post-doc level. Most of them just don't believe me."
    ........... ............ ............
    "One thing I also find difficult is explaining to some students how the submission/revision process will go down."

    They will learn gradually, as they get burned along the way.

  • aspiring riffraff says:

    Grad student PI had me outline my first paper. Was helpful for structuring arguments, especially in the intro and discussion. Then I gave PI a complete draft we went through one section of the results together in PI's office. It was enough for me to figure out how to rewrite the rest of the paper. Later papers skipped the in person session (kept the outlining). I can hear this PI's voice in my head as I'm editing my work now, and it is very helpful.

    Postoc PI has you get a complete draft vetted by a labmate close to your project and one working on something unrelated. Then PI makes broad comments "you've stated the significance of this work very clearly in the abstract, but not as much in the discussion". Never does a line by line edit.

  • BugDoc says:

    @serialmentor: apparently I'm the only one that really pays attention to all these details

    me too. Even after pointing out certain errors repeatedly, typos, pasted methods from our old papers that didn't get updated, incorrect figure labeling, etc, etc, the next time I see a manuscript from the same person, I'll see all the same kinds of mistakes again (after asking them to double check before they give me the draft). Wondering if success at the faculty level selects for OCD....

  • David says:

    Good comments by all.

    One experience that I remind myself of all the time: as a TA in grad school I was grading lab reports. I was frustrated beyond belief at how terrible most of the reports were. I told my roommate that there was now way that I wrote that bad as a junior/senior in college. As luck would have it, I had retained copies of my labs for this exact class. Turns out I was worse. 3 years and good feedback can make a world of difference. [15 years later, lots more feedback, and tons of practice, and I think I'm finally a good writer.]

  • DJMH says:

    ^----- this. We all like to imagine our first grad school manuscript sprang fully formed from our foreheads but it was almost certainly unreadably horrendous. It's just that we've deleted that draft and forgotten we ever wrote it.

  • SS says:

    "We all like to imagine our first grad school manuscript sprang fully formed from our foreheads but it was almost certainly unreadably horrendous. It's just that we've deleted that draft and forgotten we ever wrote it."

    Thanks for this. Brought a big smile. Last September, I stumbled upon the first paper I had ever written. I had written it as an undergraduate who knew everything (nothing) and submitted it to a major journal. At that time, the cocky 20 year old in me had actually worried about underselling the paper because that journal was not one of the top 3-4 journals in the entire subject!

    I am glad the editor and reviewer did not laugh in my face, considering how ridiculous the paper was. Years later, when I got to meet the same editor in a professional setting, I only prayed he did not remember my name.

    Surprisingly however, although the writing was very immature, I thought there were some decent ideas. I updated those ideas into a more jargon heavy framework, fixed the writing and decided to submit it again. Got it published 2 months ago after receiving praise from two referees. Felt really good and nostalgic.

  • Luminiferous Aether says:

    "We all like to imagine our first grad school manuscript sprang fully formed from our foreheads...."

    I never had the luxury of that imagination, because although I am now faculty, I am still trying to forget what my PI said when he read my first grad school manuscript!

  • jmz4 says:

    @David
    Yeah, aging is a process of looking back and seeing how terrible you were. Having an archive of all your scientific writings is extremely humbling in that regard. There is always room for improvement, and your writing is never as good as you think it is/was.

    For all the PIs going on about typos. Don't you think your energy is much better spent actually critically thinking about the manuscript? If the typos distract you so much that you can't do that, then hire a proofreader.

    I find when teaching students to write scientifically, the old ways are the best. Have them fill out an outline for each section with the main points to be made in each paragraph. Then have them fill it out, and finally work with them to find awkward or confusing phrasing. *THEN* you worry about typos.

    Also, hands up, how many PIs here can actually make decent figures (vector graphics) on their own? Not a single one I know that has been a PI for more than ten years seems to be able to figure out how to work Illustrator or an equivalent software.

    @DM
    "All trainees on the Internet are awesome, motivated ready-to-be-PI individuals being held back by their terrible PIs who won't train them with their time and effort."
    -That's because most of our PIs don't read your blog. I suspect that your PI readership and your PD readership are at different types of institutions (after all, most PDs are at ILAFs).

    @JL
    "All they did was come up with the idea for the project, worked for many years to get it funded, hire the post doc and provide her/him all the resources and an environment conducive to the successful project."
    Eeeehh, your mileage may vary with that argument. The provenance of ideas in science is always hotly contested, many PDs have their own funding, and many PI-constructed environments are not the easiest to work within.

  • JL says:

    @jmz4
    Indeed, YMMV. But what's more common a PD who exaggerates their own contribution and feels taken advantage by the PI, or a PD who is self funded, found their own ideas and managed to get the work done and provide a 98% ready paper despite the difficult environment? I have seen many of the former (me included) and none of the latter.

  • vk says:

    @JL,

    Aren't the inferences you make based on anecdotes rather than any systematic study?
    I am no more inclined to assume that PI's are the ones who come up with most of the ideas, than to assume the opposite. In fact, if one goes by the common excuses given by PI's who are torced to retract papers, it is always the 'trainee's' fault or idea!

    The reality is that academia is set up for exploitation of younger researchers by more established ones, and human nature being what it is, I am sure a lot of exploitation does happen. Comments like the one put up for discussion by drugmonkey are pretty insensitive
    at best. Also, it is poor management practice to implicitly accuse people working for you
    of incompetence like the twitter commentator did.

  • Indeed, YMMV. But what's more common a PD who exaggerates their own contribution and feels taken advantage by the PI, or a PD who is self funded, found their own ideas and managed to get the work done and provide a 98% ready paper despite the difficult environment? I have seen many of the former (me included) and none of the latter.

    I WANNA PARTY LIKE IT'S 2008!!!!!

  • jmz4 says:

    @jl
    Sure, and we tend to prioritize the elements of the work that we actually provide. Like literally nothing could get done with out the funding but many PDs will think of that as a banality.

  • xykademiqz says:

    jmz4: For all the PIs going on about typos. Don't you think your energy is much better spent actually critically thinking about the manuscript? If the typos distract you so much that you can't do that, then hire a proofreader.

    Sure, and who exactly is going to pay for this proofreader? Should I use my personal funds, because it's definitely not getting charged on grants. The PIs do all aspects of editing because there is nobody else to do these things. And because the PI's brand is on the paper, the PI has to make sure that what comes out is in good shape.

    how many PIs here can actually make decent figures (vector graphics) on their own? Not a single one I know that has been a PI for more than ten years seems to be able to figure out how to work Illustrator or an equivalent software.

    I have been a PI for 12 years and can use Adobe Illustrator (and its open source counterpart Inkscape) as well as any of my students, and probably better than most. I often make figures for classes, my own talks, and group's papers; also, I routinely touch up the figures made by students. And I know several PIs who do the same.

  • drugmonkey says:

    I've been a PI for >10 y and I am always the best person in my lab with the vector graphics smithing of figures.

  • jojo says:

    "I never had the luxury of that imagination, because although I am now faculty, I am still trying to forget what my PI said when he read my first grad school manuscript!"

    THIS!!

    Although I think that as my PI had more students and realized we were basically all that bad or worse, their opinion of my writing has softened slightly. But not entirely by any means. :p

  • jmz4 says:

    "Should I use my personal funds, because it's definitely not getting charged on grants. "
    -Manuscript preparation fees can't be charged to grants? I've known faculty that have full time writing staff (grant writers mostly, but I'd imagine they help when it comes to papers), but maybe those people had to be paid off of a different source.

    "And because the PI's brand is on the paper, the PI has to make sure that what comes out is in good shape. "
    -Oh, I understand the importance of avoiding typos (I think DM had a whole post on the issue of typos in grants), but I just dislike when PIs use them as an excuse to delay critiquing a manuscript, or think that it is a sufficient contribution to the intellectual work.

    Also, can anyone explain CPP's comment?

    "I've been a PI for >10 y and I am always the best person in my lab with the vector graphics smithing of figures."
    -That's good to hear. My (55 ish) PI wanted me to go to a course 2 years ago to learn to use this "new program" called Adobe Illustrator that was apparently replacing Photoshop for "making pdfs". He also routinely asks me to rearrange power point slides to change fonts and sizes. Of course he got 2 R01s right out the gate and never looked back.

  • aspiring riffraff says:

    @jmz4 We trainees need to remember that our PIs are doing us a favor, not the other way around, regardless of our funding source or prowess with Illustrator (or Inkscape). Editing a manuscript to remove most of the typos so our arguments are at least comprehensible is the least we can do.

  • David says:

    "For all the PIs going on about typos. Don't you think your energy is much better spent actually critically thinking about the manuscript? "

    2 comments on this, 1) my grad PI made a big deal about typos saying they were giant red blinking lights that distracted him from doing his real job (reviewing the content). He required us to proofread our own work and get 1-2 lab mates to do the same prior to submitting a paper to him. May not work in all labs, but seems like a good thing to try (also gives the students extra opportunities to practice reviewing papers).

    2) A friend of mine has been working with a professional editor/instructor. The instructor's recommendation was to address only one type of comment per review. First review, look at big picture, 2nd review focus on specific wording, 3rd review focus on details (typos, etc.). Obviously this is hard to do in practice (for the reviewer) and takes a lot of time (since there are several iterations). But it can be really useful for the trainee, especially if it's an ESL student. No need to rewrite a whole paragraph if you are just going to cut it. The biggest obstacle (besides time) for my friend was making it clear that there would be multiple iterations.

  • drugmonkey says:

    jmz4- CPP is referring to the fact that this is a recurring discussion for this blog.

  • jmz4 says:

    ^Thanks, DM.

    "1) my grad PI made a big deal about typos saying they were giant red blinking lights that distracted him from doing his real job (reviewing the content)."
    -Interesting. I had a PI that was the same way about figures. I always assumed it was a stall tactic.

    That is definitely not my experience editing manuscripts. I send mine to a non-scientist for proofreading. They basically aren't reading for content, so the typos really leap out at them.

    I tend to go with style you outlined in 2, even when I'm writing my own papers.

    @aspiring riff-raff
    I can't tell if that is sarcasm...

  • drugmonkey says:

    well suppose it is a "stall tactic" on the part of a PI. Isn't the trainee's best move to fix typos and dress up the figures?

    Part of the frustration of this PI/trainee writing process was encapsulated by PP in a former go-round. It was about how pissed postdocs get sometimes when the PI tells them to can large parts that they've written. Despite the fact the PI and postdoc agreed on take/direction/strategy in advance. The Postdoc is pissed about their 'wasted time'. The PI knows that you write by WRITING. A LOT. And sometimes by being unafraid to ditch a lot of "work". And that you've only come to this realization by trying it out, creating that draft of the intended direction and reading it afterwards. Same thing with figure design. In some types of labs/work, with deciding what data goes into the paper.

    So postdocs have a tendency to avoid putting in the work. they want the back and forth with the PI to triage their effort. The PI is trying to train them in the often laborious process that they, the PI, has learned is required in academic writing.

    Anyway, I do a lot of writing for my job. I thought I was a decentish writer when I entered grad school, and I was*. But nothing got me honed into the academic writing process like those last few years of postdoc and the first several years of faculty appointment. It didn't happen through PI feedback. It happened through writing a lot of academic material.

    *and yes I am aware that my blogging does not testify to any such decency.

  • jmz4 says:

    " Isn't the trainee's best move to fix typos and dress up the figures?"
    -It depends, if the PI is going to make valid and important comments about how to order the figures, or how to normalize that graph, or which arguments to make in the intro versus discussion, then no. I'd rather not spend a long time making really nice, finished versions of things that are destined to get tossed.
    As I'm sure you're aware, honing your language takes multiple iterations, and it can be time consuming. But that is only worth the effort once the general flow and structure of the ideas are in place.
    So I'll let draft figures and manuscripts be a little sloppy, and then I refine them as the process moves on. Sometimes, yeah, you have to be willing to toss the whole thing, and that's precisely why it makes zero sense to be excessively fussy about them.

  • dougal says:

    I must be doing something wrong. I am an Endowed Professor with Tenure and I still write all the papers - from start to finish.

Leave a Reply