[godelski]

  > A lot of physics experiments are very indirect, and you're not actually measuring the 'thing' you're looking for. 

This is actually true for any experiment. Just at a macro scale the proxy is often less complicated. A simple example of this might be measuring how big a piece of paper is. You get a ruler and measure it, right? Your measurement is an approximation of the ruler's measurement, which is an approximation based on the standard meter. I have a bunch of rulers and I can tell you that none of them measure exactly the same. Usually it doesn't matter though because the difference is much smaller than the level of uncertainty and we're usually measuring with enough accuracy that we don't care.

What physicists do is a generalization of this same thing. Usually much more accurately than your paper experiment. But yes, in high energy physics (HEP) you're usually measuring very indirectly and based on theory. This is a big part of the Von Neumann's Elephant thing. Fitting data is easy, explaining it isn't. The casual nature is the critical aspect