I have never been a big fan of comprehensive exams. In my opinion being able to perform on a test (of whatever form) has very little to do with what it takes to be successful as a scientist and most of the exam systems that I am familiar with have serious structural problems above and beyond this basic objection. I started a post on the problems with the comprehensive exam system at my university some time ago and will hopefully finish it one of these days, but for the time being I thought I’d point you over to Drug Monkey for his recent thoughts on the matter. The crux of his argument is:
Maturing through the career arc, I care less for this [the ability of the exam to protect the university’s reputation by preventing the graduation of incompetent hacks]. Mostly because I’ve come to realize nobody that is judging me now gives a rat’s patootie what University or Department of -ology appears on my doctorate. They care about the papers I have published. Period. Full freaking stop.
So if I were dictating a graduate program, I’d be looking to enhance the ability of the students to publish papers. This would pretty much rule out the examination approach.
While you’re over at his pad I’d recommend browsing around a bit. He and co-blogger Physio Prof aren’t ecologists but they are very sharp thinkers when it comes to life in academia. Plus, they introduced us to this, so what more can you ask for.