Does it happen? Can we see it?
Jan 19th 2006Wallabynerd alert
The basic problem in any experiment is choosing what to measure and how to measure it. Trying to determine the melting point of a solid, for instance, is difficult without a way to measure the temperature. Doing a pH titration using a visible indicator (phenolphthalein, etc.) is difficult if one is color blind. In effect it comes to “does it happen? can we see it if it does?”
This is not unrelated to the problem of “proving” the evolution of life from inert chemicals, or of “proving” intelligent design. Neither one is something that we can currently see happening, either because we lack the measurement tools to identify it, or because the timescale exceeds what any one researcher (or whole university, for that matter) can observe. At best we can look at the world now and try to identify signs that show either evolution or design. (I’ll avoid commenting here on the relative success or lack thereof; that’s not my point.) “Global warming” is perhaps another example — this time it’s because the measurable effects, so far, are relatively small and it’s not necessarily easy to separate different sources of variation.
The problem in my experiment is that the effect I’m trying to measure happens over the whole specimen, probably to an unequal extent at different spots. My measurement tool can only probe a very small spot on the surface of the specimen, and collects very noisy data at that. It’s very much like looking for a needle in a haystack — and limiting yourself to only staring at one straw at a time on the outside of the pile. It’s not quite that bad, actually; there could be multiple “needles”, and some of them might be close enough to the surface that they can be seen from outside. At least, that’s how I’ve been explaining the results I’ve gotten so far. 
The other possibility is that the measurement tool itself is not capable of detecting what I’m looking for. The happy news, as of yesterday, is that yes, it can detect the sort of thing I’m looking for. Testing polypropylene in the same way gives excellent correlation (r
Now to make it work in bone…