Professor in Training has a cool post about how tardy reviewers are a pain in the butt of editors. She rightfully asks why someone would agree to referee a paper and then be several weeks late.
Being an author, I totally understand and agree. But, being a sometimes tardy reviewer myself, I feel I also need to speak in defense of my tardy brethren. We are very sorry to be late. We really mean it. We should certainly write an email to the editor, saying we need additional time. We actually say yes to a request to review because, upon cursory inspection of the paper, it appears interesting and within out field of expertise. We don't get much (if any) benefit from reviewing a massive number of papers every year, but we value journal review as an important service and are willing to put in the time. However, even with the best of intentions, life intervenes, and because paper review is a service activity with no recognition for the reviewer, it often gets demoted as people need to reprioritize. And boy, do we need to reprioritize!
Here's an example: I was recently on vacation and took a couple of papers to review with me. During the vacation, I got several more requests, of which I declined all but two. One was for a short paper which I reviewed promptly and one was for a longer paper, and I left it for after the vacation. When I came back from vacation, there was a major "fire" that needed putting out: I am a co-PI on a big, multi-PI grant that is up for renewal. There was a ton of work that had to be done over that very week on the grant proposal, because two big players were going to leave the country the following week, one of them for a whole year. So no paper review happened that week despite my intentions, but neither did some other important things, such as meeting with several of my students who got stuck with their projects while I was away.
Among comments to PiT's post, Dr. Girlfriend says it is just pain rude to say yes and then stress everyone else by not delivering in a timely manner. I agree, but...
Consider this instead — if people said yes to a review request only if they were 100% certain they could deliver by a given date, the editor would likely have to ask 20 people to get 2 to say yes and deliver; that alone would take a very large amount of time. Wouldn’t you rather have an occassional reviewer late? Remember that reviewing is an important service, but also a time-consuming activity that does not benefit the reviewer so yes it will get lower priority when more urgent stuff comes up. Ironically, when you are a reliable reviewer, editors tend to pile on you, so you end up being late. Personally, I would much rather my paper be reviewed by the editor's 1st or 2nd choice, even if the review is late, than the editor's 15th choice, because by that time we are likely moving far from the desired expertise of the reviewer. (See Odyssey's comment, who speaks from an editorial perspective.)
When I was a student I was always upset about late reviews of my papers; now that I review non-stop, I am no longer upset at people being late as I know what's going on. I am that reviewer who writes you a two-page report. Actually, I spend much more time on reports for unfavorably reviewed papers than the favorably reviewed ones, because it's important to give people something to work with to improve their papers. I try to complete my reviews in a timely fashion, but sometimes it ends up not being timely. Such is life. In fact, when I look at how long it takes me to get reviews back and how long I take to return them, I feel that the duration of the actual technical review correlates with the size of the paper (shorter papers review faster), time of year (high season for travel or not, grant proposal deadlines), and the quality of writing (if the paper is poorly written, my opinion is that it can delay the review, as the reviewer will likely try reading it several times, get pissed each time and just leave the manuscript for later. I do this, and so do several of my colleagues. So writing a nice readable paper will also increase your chances of getting the review back promptly.)
The review duration also correlates with the seniority of the reviewer (big shots get more review requests and are overall busier); to enhance the review speed, in some journals in my field, editors are encouraged to recruit more junior people (e.g. postdocs) as reviewers for precisely this reason — they are less busy and more likely to review in a timely fashion.
While I no longer wish all the worst to tardy reviewers, as I did as a student, I am still quite aware that publication speed makes a significant difference for the career prospects of students and postdocs. It is often favorable to choose a fast journal with prompt editorial attention over a sluggish one even at the expense of a couple of impact factor points. What does vary dramatically among journals is how long it takes the editor to make successful referrals and how prompt he/she is about making a decision once the reports are back; in my experience, editors who are practicing scientists (as opposed to career editors) as well as organized individuals have some of the best turnaround times I have seen. I am curious if this is universally true: dear readers, who do you think makes for an overall faster review process -- career editors or editors who are practicing scientists?