Peer-review is often referred to as the "gold standard" of scholarly publishing: a rigorous vetting process that ensures the quality of the research that makes its way into print. And although the system isn't fail-proof--crap can make its way into top journals and every couple of years there's an actual scandal involving falsified data or obviously tendentious analysis--I do believe that, in the long run, the traditional double-blind review system mostly works.
But "in the long run" is a pretty big caveat, and I know no one who hasn't beaten her head against the wall of peer-review at least a few times. There are genuine horror stories of reviewers who make it their mission to block any work that challenges their own or that uses a theoretical model they disdain, but mostly there's just pettiness and obtuseness: reviewers who on some level can't "hear" arguments that don't match their own.
Indeed, although the publishing world has been good enough to me over the years, I've encountered as many obstructionist reviewers as generous ones, starting with my very first submission. As a grad student, I submitted one of my dissertation chapters to a top journal and got back a thirteen-page, single-spaced review. It was clear that the reviewer had taken a strong dislike to me, and hostility oozed from every sentence. He appealed continually to what "everyone knew" about the subject, sometimes going so far as to say, of actual, documented facts "this hardly seems likely." In one especially bizarre paragraph, he took it upon himself to lecture me on the Vietnam War and U.S. policy in Central America in the 1980s. (My essay was on Milton.)
Compared to later negative reviewers I've encountered, however, this one was easily sidelined: the journal editor told me to read the review carefully, make "whatever revisions you think necessary," and send it back. It went out to a second reviewer from a similar school of thought, but this time a professionally and intellectually generous one; I later met him at a conference, and as he shook my hand and introduced himself, he said, "I don't know if you could tell from my review, but I disagreed with pretty much every other sentence of your essay." He smiled, told me that it was a provocative and worthwhile argument, and added, "and boy, can you write."
You could say that this taught me to have faith in the system, and I do, but it's been tested routinely by both my own and others' subsequent experiences. The problem isn't so much with the bad behavior that anonymous review sometimes permits or with the way a single person can block good work for personal reasons. Those are problems, to be sure, but there are plenty of venues out there, and plenty of readers; good work will generally get published eventually, and the real test is its afterlife: how often it gets read and cited and grappled with once it's out there in the world.
No, the real problem lies with "eventually": scholarly time is always inefficient and unpredictable--it's hard to know whether the article you're writing will take six months or two years--but when your work is in your hands, at least you have some understanding of why it's taking so long and what comes next. This isn't true of peer review, which might as well be a black box: even the reports and the editorial decisions, once you get them, are not always self-interpreting. This is especially hard on junior scholars, who are always on a clock; they need that vita line ASAP because they're going on the job market, or they're approaching their third-year or their tenure review. Under those circumstances, even the usual delays--a reviewer takes six months instead of the promised four; a journal requires a second round of revisions--can be nerve-wracking, and the more capricious and unreasonable responses can damage careers.
There's not a solution here that I'm aware of; I don't believe that open-source peer review is a better answer--on the whole, I think it's likely to produce more conservative and crowd-pleasing rather than more innovative work--and I certainly don't advocate for the end of peer-review, but there are problems here that affect junior scholars disproportionately (although not exclusively: I've heard well-published full professors mention having written pieces that got savaged so badly they'd never had the heart to send them out again). I suppose one solution is, "submit early and often," but that too is hard on junior scholars, who tend to be focused on the One Big Thing that is their dissertation/first book and can't as easily work up little side articles.
Readers, what do you think? Are there ways to make the peer review process work more equitably and efficiently (beyond being an ethical and responsible reviewer oneself)--or do you have words of wisdom to give to grad students or recent PhDs stuck in peer-review hell?