Experimental design

Ken asked if an experimental-design class was offered where I teach.

The short answer: not for the undergraduates. For the grad students (of which we have few and we're only a Master's institution), there's a course called "Research Methods" which includes some elements of experimental design. It's sort of a redheaded-stepchild course that is taught by different people from different departments - I have a colleague who teaches it once in a while, but it's most often taught by folks from the Psychology department (apparently they have a harder time making "load" than we do, which surprises me).

Some of our students complain that it's very psychology-centric when it's taught by one of the psychologists.

I DO try to put in a lot of experimental and research design stuff in my ecology class, and also in the stats-for-biologists class I teach. (Rule #1, I tell them: before you start collecting data, have some idea of how you plan to analyze it. I used to be kind of the "consult person" among my fellow grad students for stats stuff - because I knew more stats and had more stats background than any of the other students, and because I was considerably less scary than the official biostats prof. I once had someone come to me with data that she had collected with NO plan for analysis. NONE. The data were a big mess. There was almost nothing she could do - I think I told her she could calculate something called Brouillin's Index and that was it - because she hadn't even ensured that her plots were randomly sited.)

I never had a class in experimental design - well, an official-type class, I've gotten elements of it in many, many classes I've taken, and I've READ a lot. I think that's actually one of the weaknesses of the large-university science program in the U.S. Or at least it is in biology departments. Perhaps that's because in larger universities, there's kind of the assumption that "The good students will go pre-med, and the not-so-good students, we don't really need to worry about them." The problem is, there are some good students (like me) who can't stomach the sight of blood and really have no interest in being a G.P. or something, who actually WANT to go into research.

Grad programs SOMETIMES have them - mine didn't but luckily I worked closely with the stats guy (who wasn't all that scary if you weren't a complete idiot; you just had to tell yourself, "He's brusque, he's rude, it's his schtick" before going into his office. He was kind of a proto-Greg House now that I think of it.) And my advisor was pretty good so I learned the stuff I needed sort of by doing it. But I do tend to think a class where you're doing "little" research - that's not the whole next 3 years of your life, so if you screw it up, it's not so horrible - is a good thing for undergrads. If anything, it might open a few folks' eyes up that there's more to life than either being an M.D. or a gel-jockey for Abbott Labs. (That was kind of the assumed dichotomy when I was an undergrad at a large Research I institution: the vast majority of students became M.D.s (if they were good) or learned some technical skill (running electrophoresis gels) if they weren't, and so the "oddballs" who actually liked working with plants and such didn't count as much. Looking back on it, I probably should have applied to the School of Natural Resources instead; the SNR classes I took were more in line with my interests - and that was where I got my really good stats background.)

Maybe it's different in engineering or other science programs; I hope it is.

(If I were 16 again, and were starting over from scratch and wanted to do something different? I'd pay more attention in Pre-Calc and Calc and then go for Materials Science Engineering or something like that. The MSE folks always seemed to have interesting labs, trying to make stuff fail and playing with different kinds of ceramics.)