Let’s talk about how we write.
The primary output of astronomers are papers and proposals — we have to obtain data & money, publish results, and repeat. Here, I’m going to argue that we should write some of those words more collaboratively.
Our default method of writing was invented when colleagues communicated by postal mail with week-long reply time. Does that approach still makes sense in all cases?
Here’s the standard workflow for writing a telescope proposal. The PI consults with collaborators to work out technical details and the broad themes of the justification; tasks co-Is to generate analyses or to write subsections; circulates a draft 1-2 weeks before the proposal deadline; gets feedback from co-Is; and iterates until submission. Paper-writing is surprisingly similar. While data reduction and scientific analysis are often highly interactive, writing up the results usually follows the proposal workflow, with the lead author writing most of the words.
But we no longer must wait a week to receive letters from colleagues. Ours is an era of instant global communication, with free/cheap tools to collaborate.* So I pose a question: To what extent should we adopt more collaborative writing strategies? Or, as Kelle provocatively asked me, have we reached the end of the era where the first author writes almost all the words?
Before opening this up for discussion, let me justify what I view as the three key benefits of interactive, collaborative editing:
- It’s more efficient. Instead of one author who acts as a gatekeeper, there can be multiple simultaneous writers, without stepped-on toes. This multiplexing is during proposal season. I can write the introduction while my colleague inserts figures, and another colleague cuts & pastes the references. As a result, nobody’s standing around waiting for their colleague to finish writing a section and circulate an updated draft. Also (this is KEY), instead of sending send rounds of emails (“In Section 3, 3rd paragraph, 2nd sentence, do you mean the rms error or the error in the mean?”), you can simply write your question directly in the draft.
- Faster turnaround. As the lead authors proofread the draft, they can simply fix typos themselves, and mark bigger problems as inline queries, tagged with the name of the author they think could best solve them. On a recent paper, when we got comments from our later co-authors, the lead author posted the comments as a text file, so that the first four authors could work through them comment-by-comment. This distributed the tedious work of fixing commas and such across multiple authors, and shortened the time between a circulatable draft and a submittable paper.
- It’s more fun. Science can be isolating, especially on a project where your collaborators are external. Collaborative writing is one technique to draw your team together, especially if you’re geographically separate.
In an upcoming post, we’ll summarize results from experimenting with several different collaborative-writing tools, and seek your feedback about what tools y’all are using to get those proposals and papers written.
But for now, let’s ignore tools, and instead talk theory. What are your thoughts? To what extent do you find collaborative editing helpful? Are there hidden dangers? What strengths & weaknesses have I left out?
*(Getting your collaborators to respond is another story.)
“circulates a draft 1-2 weeks before the proposal deadline”
Oh what a wonderful world that would be! Lucky if we’re throwing around drafts 1 or 2 *days* before 😉
I’ve written most papers collaboratively – each person is assigned a section on their expertise, and it works pretty well.
The main stumbling block is getting everyone to understand the collaborative editing tools and to actually use them. In the last days/hours of any proposal, people usually go into a flight or flight mode and default back to their original method of iterative collaboration. Much good work can be done in the early stages, though, so it’s usually not a waste. But you do need an assigned ringleader (usually first authour) to keep the troops in line.
And yes, getting collaborators to give timely responses is, well, just difficult 🙂
The only thing that worries me about collaborative writing is that it can easily become leaderless writing, replacing coherent argument over pages with a collection of independently-authored, paragraphs that don;t stick together. Rather like replacing the webmaster with the wiki …
I’ve recently started collaboratively writing papers using a shared Dropbox folder. Inline notes are very convenient (and I’ve found it’s helpful to set up note commands in the tex file so that each author’s notes are a different color in the compiled ps/pdf file).
So far I’ve found the most important things are:
a) leaving perfectionism behind — I keep repeating to myself that my collaborators won’t judge me harshly for sharing draft-level writing (so far, this has been true); and
b) keeping the overall narrative in mind — I agree with Roy that true collaborative writing means it’s easier for the bigger picture to get lost, which is perhaps where Matt’s description of the first author as ringleader becomes important.
a) leaving perfectionism behind — I keep repeating to myself that my collaborators won’t judge me harshly for sharing draft-level writing (so far, this has been true);
— I think this should be put as its own post sometime soon. What are people’s opinions of early draft sharing. There is a lot of modes of thought on this and I for one would rather hear someone say you need to do this differently before I draw a lot of conclusions about it rather than after the paper reaches a final draft. I also believe that collaborators would rather sit and spend 20 minutes reading a section every couple weeks than a huge paper once or twice.
What about receiving “proper” credit for the work? If papers are largely written in collaboration, one can no longer assume that the lead author actually did the majority of the work and wrote the paper. Maybe I’m being naive, but I assume that when I see someone listed as first author, it implies that it is primarily (mostly?) their project that they took the lead on. How would one propose distinguishing authorship in a collaborative writing model? What have people’s experience been?
In one of my groups of collaborators, once the paper is in a good final draft form we all sit around the table with the document on the screen and we wordsmith and talk about how to say main points better etc. Its slow going sometimes but the improvement of the draft is far more than anything you will get from emails and comments. More people need to feel comfortable letting go of their words and get substantive feed back in a truly collaborative environment where we defend ideas and concepts until everyone has a clear picture of what is being said.
Roy brought up the worry that collaborative writing could become “leaderless” writing. What do people think about this? I’ve not experienced this problem, probably because, as Matt K. suggests, there’s always a lead author herding the cats, I mean co-authors.
I partly agree with Roy on this – I just wrote a grant proposal collaboratively using Google Docs, and while it was *great* to get our thoughts down on paper, and everything that we really needed in there was physically there, I ended up having to take it over 2 days before the deadline and rewrite the whole thing coherently. I guess we started out with true collaborative writing to get our ideas down, and then reverted to iterative writing to put the structure in place. Interestingly, I’m not the PI – I was just the person who got frustrated first!
I have to say, I am all for collaborative writing though – I think that seeing all of our thoughts in one place and getting people to read over it and put comments in-line was a brilliant way of doing it. I honestly believe that our proposal was better for it.
I agree with Roy and Carolyn – near the end of the process, almost always in my experience it takes one (or maybe two) dedicated author/editors to bring coherence, flow, and hopefully maybe even some eloquence to the document as a whole. But that’s not a knock against collaborative writing! The need to merge disparate contributions into a single unified voice is equally true whether those contributions arrive via email, red pen scrawled on printouts, or edits in the ether.
As for Anon’s query: I think it is erroneous to assume the first author did all the work. I’ve been on papers where that was true, but also on ones where the top three authors (of eight) shared the load pretty equally. And that’s to say nothing of large collaborations, or cases where person A took the data, person B developed the necessary analysis techniques, and person C realized a new topic to which those data and techniques could be applied… I am a huge fan of Nature-style explanatory statements at the end of a paper, that explicitly describe the contribution of each author: “Authors A and B led the data reduction and analysis and did the bulk of the writing, author C obtained the original grant funding, author D contributed one key insight in the analysis, authors E-G provided comments and useful discussion on the paper text, author H wrote a critical piece of software but was otherwise uninvolved, author G was on the original proposal but didn’t do anything since then”, etc. This sort of information is utterly opaque from merely an author list order. We should strive to do better.
Somehow the Document is the opposite workflow from the Software. The Document starts its life (so this thread indicates) as collaborative, everyone pitching in, then the team decides its direction, and the final apex product is articulated by a single voice. In contrast the Software, as I see it, begins as a shining solo idea showing the direction, and if that prototype is to succeed it integrates with a larger whole, a larger team — and thus the final product is collaborative. Opposite workflow.
You know, it all depends, as always. The collaborative writing workflow that I have in my head is similar enough to the traditional one now that I think it would not be too tough of a transition, but does indeedrequire a leader:
This is a relatively minor issue, but I assume I’m not alone in having co-authors who have various odd fixations with diction and grammar (e.g., changing an active verb to passive for no reason other than, I suppose, it sounds more “academic”; or objecting to “quasars at z>1” because “at” [in their minds] can’t refer to a range). When I get lists of changes and “corrections” from my co-authors, I usually use about half of their suggestions. And I get twitchy when other people change my writing without asking.
An extra potential problem is the political issues. Fortunately this is rare (at least I hope it is), but it does happen that a boss who is the nth author and who is only marginally involved with the actual work can end up having more input than the first author originally intended. It is one thing to ask for suggested changes and deciding if one applies them or not, and another thing to just having to take the edits that every co-author adds on a given section. I guess this can be dealt with if the rules are clear from the beginning, but there some things to be said for first author “privileges”
I am all for collaborative writing techniques, but maybe within some specific limits for difficult cases.
Hi, Eva (and everybody), I’m having exactly the kind of problem described in the comment above, and would like to know in your opinion what is the best way to deal with coauthors that stretch their privileges too far. I meant, I’m the main author (and rightfully so) of a paper, which first draft I alone write (after discussing with the others). The problem is that when I ask for comments, some co-authors thinks is their “duty” to rewrite anything they would have written differently as they please (that is, the whole paper). Being the less senior member of the group, I’m having a hard time rejecting their inputs, as if they were mere suggestions (I have been called from stubborn to arrogant). In that group in particular things have been that way for a long time (one of the heads is a classical micromanager, who really delegates nothing).
My question is: what should be the relations between main author and co-authors? From various sources, I have found that is common for the main author to write the first draft and then to have some rounds of co-author feedback (preferably as comments and not edits), followed by rewriting (by the main author), and so on, till the final version. That is the way I envisioned the process from the beginning, but for me it have been far from that.
Does anyone know if there is any published article, or any source of information, that clearly described that kind of collective writing process and authors interaction? I know it is not the only of such interactions process, but is the one that will suit our group better and best fulfill my need for independence and professional growth.
Thanks in advance.
M.