Thursday, September 20, 2007


Peer reviews: Too much background information

Reviewing some conference papers recently has given me some thoughts that I'd like to put down. Here's one of them.

We all had experience in grade school with “padding” papers: adding unnecessary fluff to make the paper a bit longer. The stuff was easy to write, and we thought the teacher wouldn’t notice, and that we’d get a bit more credit for it than for turning in a shorter paper.

The teachers didn’t always tell us so, but they noticed.

Now that we’re submitting papers to peer-reviewed journals and conferences, many of us still do the same thing, and I think we still think that “they” — in this case, the reviewers — won’t notice.

We do, and I, at least, will tell you so if I’m reviewing your paper.

I can’t tell you how many papers I’ve reviewed that talk about spam solutions, and that spend at least an entire page — as much as 15-20% of the paper — giving the background, telling the reader what spam is, why it’s a problem, how serious a problem it is, how long it’s been a problem, etc. Let’s be clear here: anyone who will read your paper already knows all this. It’s just tedious, useless fluff that we have to get through, and as I read that stuff I’m saying to myself, “Yes, yes, get to the point already!”

Some brief lead-in is fine, and is even necessary for the flow of the paper, but the operative word here is “brief”. As I’ve said before, in other contexts, write to your audience. A technical paper that’s meant to be read by people who work in the field neither need nor want the well-known detailed background. If you’re explaining your work to the press, you’ll need to add some of it back in. It’s all a question of whom you’re writing for.

Of course, you do have to pay attention to the rules for the conference or journal to which you’re submitting your paper. Usually there’s a maximum size, but not a minimum — full-length papers for the Conference on Email and AntiSpam are limited to eight pages, for example, but the papers can certainly be shorter. I’m happy to review a paper that’s, say, five or six pages, or even four, if that’s what the details fit into. If there’s no minimum size (there is, for some journals), don’t pad.

Even if there is a minimum, don’t pad with fluff. Look for background information or details about related work that will actually be interesting to your readers. If you make us read junk, it will affect how we regard your paper. Wouldn’t you rather your readers find your paper fascinating, not dull and obvious?

This really can make the difference between getting your paper accepted and not... and between having your paper cited frequently as a reference and not. If you’re going to the trouble of writing it, take the time to make the most of it.


Maggie said...

I once read a paper in a math journal that went on for the entire paper saying "in this paper we will..." and never delivered on its promises. Basically the authors identified a bunch of problems that they said they would present solutions for, and they never did. I think some people expect their papers to be accepted based on their name and being friends with the reviewers, and in some cases, that's enough!!

Barry Leiba said...

Yes, some reviewers are star-struck, and not all are critical enough. On the other hand, some are over-critical, sometimes being gratuitously nasty. It takes all kinds, and on the whole, it's a reasonable system.