26 April 2014

An easy way to write less hurtful reviews: don't say "you"

I'll be honest: I've had my feelings hurt by scathing reviews more than a few times. In grad school I remember even crying over a review that I thought was particular pernicious. My skin has thickened a bit over time, though often in the not-so-helpful manner of dismissing reviews that I don't like as "they didn't get it," which defeats one of the two primary purposes of reviews in the first place (providing feedback; the other: making accept/reject decisions).

The thing that's hard to reconcile is that I really like most of the people in our community, and everyone I meet at least seems really friendly.

When doing mock reviews with grad students, I'll often tell them to keep in mind that there's a good chance that the author is, or later will be, a friend of theirs. It's possible to provide feedback to a friend in such a way that you don't hurt their feelings.

I've recently started doing something else (in addition to the above suggestion). I don't use the words "you" or "the authors" or even "I." The review of a scientific contribution is not about me and it's not about the authors. It's about the method, the experiments and the contribution. I see little reason why you need to mention anything related to the people involved. (One exception: "I" is often useful in hedging, like the previous sentence, which would be more forceful if I just said "There is little reason...") Perhaps we could even integrate this into START...

This is, of course, similar to the pop-psych advice of talking to loved ones about "actions" rather than "the person." For instance: "I hate you for spilling coffee and not cleaning it up" versus "I hate having coffee spilt on the floor." Or something. I'm sure others can come up with better examples.

My current approach is to write my review with this in mind, and then go back and search for all occurrences of my outlawed nouns, and rewrite these sentences. Often in the process of doing this, I become aware that in many of the cases what I've said really does sound like an attack, and with the very small edit this effect is removed or at least greatly reduced.

I realize I've now just given a pretty good signal for people reading reviews to see if they were written by me or not. Here's a solution: everyone should adopt this policy and then my reviews will no longer be so obvious.
But overall, I really think we should be nice to each other. Perhaps fewer people will depart from the field if they're not constantly battered down by harsh reviews, and then we'll all be better off.


Emily said...

Thanks for starting this discussion, Hal! I really do think there is room for improved reviewing in our community, and making it easier for the authors to take the critiques as constructive can only help.

I'm curious, though, whether a blanket prohibition on I/you is necessarily required. Do you think "I was not able to follow section 2" or "I think you mean X where you use the word Y" come across as attacks?

hal said...

@Emily: of course not all uses of I/you (especially I) are attacks. "I love your paper!" But I've found it's a consistently good way to find many of them. "I was not able to follow" is actually quite nice because it makes it my fault :). The second example, eh, I don't know. It's easy enough to rewrite, though. If "I" is okay, then "I think X should be replaced with Y" seems like a shorter and totally reasonable paraphrase.

winterkoninkje said...

Giving blanket advice is always tricky. But there is one big thing worth bringing up (even if I can't offer much help in resolving it). Namely, the "talk about actions, not people" paradigm can very easily slide into passive aggressive behavior (e.g., it's very easy to read/say "I hate having coffee spilt on the floor" in a way that makes clear that it is in fact an attack).
And it's passive aggressive behavior which really does the job of setting up a hostile work environment and scaring people away.

For this reason I think it's good to eschew the use of "I" at least as much as the use of "you" or "the authors". "I" can be very condescending and passive aggressive. It's fine in hedges ("I couldn't follow X", "I think Y is the case", etc) but it's definitely something to keep an eye on.

Arash Joorabchi said...

I never forget this review I got from a reviewer of IP&M journal at the early stages of my PhD:
"...The paper is written in very good English and is very legible, and this is the only good news. On all other counts, the paper is completely worthless, since it proposes a terribly naive algorithm, and it evaluates it on a tiny document collection and according a completely substandard experimental protocol. Above all, the paper displays the fact that the authors are utterly unaware of the basics of text classification. The paper is definitely unpublishable, in any scientific forum..."

Anonymous said...

I support and endorse your way.

Chris Brew said...

It particularly grieves me when work that is probably pretty good is undermined by deficient writing and confusing presentation. Since we don't typically have 'revise and resubmit' among our options for conference reviews, papers like that have to go into the reject pile. I always hope that another reviewer will fix my confusion, if it is just mine, but when they don't it is hard to know what to say. I am only guessing that the work is in fact better than the authors managed to convey, so there are limits on how enthusiastic the review can be. And, in claiming that the presentation is a problem, I am in fact making the stronger claim that it will be a problem, not just for me, but for the typical reader. Awkward!