I'm hoping to make this kind of post a regular feature of my blog: "One Thing I Think I Learned Today." And I want it to be something substantial, rather than just some trivial observations. I'll kick it off by relating something I think I learned yesterday in the Seminar on Institutional Analysis and Development, where we were focusing on models and methods.
I learned that the "internal validity" of laboratory (and possibly field) experiments is largely determined by (or a function of) replicability of experimental conditions and outcomes, rather than the pure logic of the experimental structure. I find this interesting for several reasons, not least of which is its resonance with my recent work on cost-benefit analysis as a decision tool (see here). I find that the two "saving graces" of that otherwise malleable and manipulable tool are its twin formal requirements of transparency of assumptions and replicability. Needless to say, replicability is a prime component of all quantitative empirical and experimental work.