Philip Guo (Phil Guo, Philip J. Guo, Philip Jia Guo, pgbovine)

Live Paper Editing

One of the best parts about becoming a professor is getting to run experiments ... on students. (Sorry, students!)

Here's an experiment I want to try on my own future grad students: live paper editing.

My senior colleagues concur that teaching students how to write technical papers is really hard and time-consuming. On his popular blog, Matt Might mentioned: “The hardest part of advising Ph.D. students is teaching them how to write.”

Beyond just teaching best practices for technical writing style, professors must also teach students how to write using the conventions and idioms of a given research community. This skill takes years to master, often by writing and reviewing lots of papers. And these conventions and idioms are often idiosyncratic, varying widely across communities.

An effective but time-consuming approach is for me to print out my students' paper drafts, meticulously mark up corrections and suggestions with my red pen of fury, and then talk through that feedback with them. They go off and make a round of revisions, come back to me, and repeat the process over and over again.

This sounds great when I have the luxury of time, but a lot of my paper writing interactions with students will be crunching for paper submission deadlines. We won't often have time for several rounds of detailed feedback and iteration, since there's always more programming to do, more experiments to run, more data to analyze, and more sections to refine ... right up until the deadline.

I'll probably have to directly edit my students' drafts while they do other tasks in parallel such as refining experiments. This is the most efficient course of action in the short term, but the problem is that students aren't learning how to improve their writing.

So here's a compromise: I'll have a student sit next to me for an hour or two while I edit their writing live in front of them. They can silently take notes and ask short clarifying questions, but they can't otherwise break my paper editing flow. At most, I'll think aloud to verbalize my editing decisions. I don't think I can edit for more than two hours in a row anyways, so it's not like they need to sit next to me for the entire day.

Here are some advantages of live paper editing:

  • students learn by watching me edit and justify my decisions
  • timeboxed to save time on both sides
  • possibly improved camaraderie that we are working together rather than me just giving distant, high-level feedback

And here are some possible disadvantages:

  • it's still slower than me editing on my own (maybe 30% to 50% slower)
  • student gives up a few hours of their time when they could be doing other tasks for our upcoming deadline
  • student doesn't get to practice editing on their own (not that they would have time to if we're in super-crunch-mode)

My hunch is that live paper editing will be almost as effective as marking up student drafts with a red pen of fury and having them revise repeatedly. And it will take substantially less time on both of our ends. But who knows; I haven't tried yet. Any volunteer victims?


Some interesting responses from the Internetz:

What about one-on-one meetings where you give a student live feedback on their writing?

That's also useful but different from live editing, since the student doesn't get to see me edit and think out loud.

What about pair writing, a la pair programming?

It might be nerve-wracking for a student to write while their advisor is sitting next to them, ready to pounce. Even I have a hard time writing when someone is staring at my screen. Programming somehow feels easier to do when paired.

I'm concerned that you're denying students the chance to practice editing on their own. That's the best way to learn.

Agreed. Nothing beats the student practicing by themselves, receiving targeted feedback, and repeating. However, that takes a tremendous amount of time. Also, I suspect that novice writers will be overwhelmed or demoralized by red pen marks of fury and not actually learn from their edits. They might type in the suggested edits as ordered and not reflect on why they are making those edits.

My live paper editing idea might allow the student to take advantage of the worked-example effect, seeing their advisor in action then mimicking later. A practical compromise is:

  • Do one hour of pair editing, which is like a student studying a worked example of me editing with think-aloud.
  • Then the student goes off and edits the rest of the section on their own, building off the momentum that I started.

This technique is sort of like, “Watch me do it, then you try.” It could work better than simply throwing them off the deep end from the outset.

Created: 2014-02-24
Last modified: 2014-02-28
Related pages tagged as assistant professor life: