Controls

We went to a dinner tonight hosted by the company that makes the supplements for Beanz' store. I kind of like meeting natural health/food people, because they tend to be a nice blend of socially progressive political types, pragmatic small business types and intellectual types, without being too over the top in any of those areas. Although, they do tend to take themselves and what role they play in the world awfully seriously for people who sell organic pies to yuppies. But I'm just being cynical; we all sell organic pies to yuppies in one way or another. Grad students going into debt to spend your life studying post-Colonial theory? Bull shit! You teach luxury education to upper middle class white kids.
Anyway, we talked with this one guy who is the rep for the supplement manufacturer. He was interesting because he came across as a very typical glad-handing grown-up jock turned salesman, but he turned out to be a pretty interesting guy, with much more nuanced ideas about life and health than the initial impression I got. But he did say one thing that made absolutely, positively zero sense to me. He was talking about how powerful the placebo effect can be on people, and described how he used to talk with his ex-wife, a rep for a pharmaceutical company. He had said to her that the very fact that you rely on a placebo group as the control implies that you ultimately don't know why things work, and therefore all these hippy-dippy supplements might work to cure God-knows what!
To me, this was so assbackwards. The point of control groups is to prove that drugs actually do work beyond the placebo effect. Science isn't about trying things at random and relying on placebos to see what might happen, it's about trying to deduce what should happen and then using a control to prove whether or not the drug (or whatever) works as we expected it to. Controls don't show that science is afraid of not working, they show how vigorous the standard for success is.