As part of my teaching assignment last year, I supervised a lab class that was known to be in need of serious improvement. Last year, I ran it as is to get a feel for the issues, and this year (which means this summer, since I also teach a lecture course in the Fall) I am redesigning the course now that I've seen the problems with it. Now that I am teaching a full load, I've decided to front-load this one, and spend a lot of time on it now, pulling back on it after the course is set up with (hopefully) only minor fixes in the future (like redoing sections of the lab manual or grading rubric, which is easy vs selecting and testing new labs, which is hard).
Rather than wade through education journals, my strategy is to ask some friends what their departments do, and also use Google to see examples of labs run at other schools. I found it hard to decide if students could actually do the labs discussed in the education journals, and went with a more battle-tested approach. The labs are still a bit cookbookish, I think, but this is a low level class where the students need to learn lab safety and technique as much as seeing concepts from their lecture classes demonstrated in a hands on way. I am also incorporating more writing into the course (see writing rant here), since I think technical writing is an important skill too. I've also been thinking about why we run these labs in the first place, so I am focusing on labs that highlight concepts that I've seen students struggle to master in lecture courses.
I have to say that this is turning out way better than I thought. I am only picking new labs that sound fun to me--if I took this class now, I would LOVE it. We are doing labs with a higher fail rate where the trade-off is a more interesting experiment. Just doing the supervising of a lab course is most of the bad things about teaching (the whining, the problems, the paperwork) without the good things. At least now I get to have some fun!