Half-Baked Course Idea: Great Experiments

A couple years ago, we revised the General Education requirements at the college to require all students to take a "Sophomore Research Seminar" in their second year. These classes are supposed to be writing-intensive, and introduce students to the basics of academic research. The specified course components are pretty heavily slanted toward the humanities-- library searches, primary vs. secondary sources, and so on-- and don't really map that well onto research practices in the sciences.

A colleague in engineering managed to do a really interesting project-based class, though, and since hearing him talk about it, I've been idly thinking of possible ways to do something that fits better with science (usually at times when I really ought to be thinking of something else). This morning, while I was thinking about labs for my modern physcis class, it occurred to me that it might be possible to build a course around the idea of great experiments in the history of science.

In essence, this would be like "The Giant's Shoulders: The Class.

I thought of this because a lot of the experiments we do in the modern physics lab are basically updated versions of classic physics experiments. Today, we're going to do a Michelson Interferometer lab, for example, and in a couple of weeks we'll be doing the Photoelectric Effect, and the Compton effect, and so on. The labs that we do compress the great experiments into a three-hour lab period through the use of modern technology-- you can set up a Michelson interferometer in fifteen minutes using a laser, and use a computer to get fringe data really easily.

Back in the day, though, these experiments really took a lot of work. That's the inspiration behind the blog carnival, after all. And it seems like something that might be made to work as a SRS class.

The idea would be to have students pick one of the classic experiments in science from, say, before WWII, track down the original papers, and read them to work out how things were done (tracing back other references as needed). Then they would look into how the experiment could be updated using more modern technology, and what the pros and cons of the different versions are. Ideally, they would do some version of the experiment themselves, and write up the results as well.

This isn't a fully worked-out plan, but it might be able to do most of what an SRS is supposed to do. Arguments for and against:

Pro:

• It provides a way to usefully incorporate library research techniques in a class that is fundamentally about science, while also giving students some idea of how scientific research is done (which is not primarily in a library).
• It would get some technical information across, through the process of understanding what's involved in building and carrying out experiments, and could be fairly general-- I think about it in terms of physics experiments, but there are presumably similar things that could be done in chemistry, biology, and so on.
• These classes are not supposed to have prerequisites, which makes things a little dicey, but historical papers, modulo some changes in language, tend to be more comprehensible than modern ones, because science was less complicated back in the day.

Con:

• I'm not sure it would be possible to come up with enough classic experiments to allow an explanation through example, while also leaving enough experiments unexplained to make for student research projects. This could probably be done by offloading some of the teaching to students through in-class presentations and the like, but it would be tricky.
• I have very little idea what would count as classic experiments in chemistry or biology, or how you would go about replicating them.
• I'm not sure we have apparatus for enough different experiments to allow replication for as many as would be needed to cover a whole class (~20 students)

Anyway, it's something to think about at times when I ought to be thinking about something else...