Statistical variability

Statistical variability is the term we use to describe the fact that in science, we often need to use instruments which are not perfectly precise in our application. The consequence is that repeated measurements often result in data which are not exactly equal, but are similar. When this happens, we usually adopt the attitude that the quantity being measured is stable and that our variation is due to observational error.

This assumption may be false; the risk is that we ignore the possibility that the quantity is indeed dynamic and that the variation we observe is intrinsic to the phenomenon. Naturally, we prefer the simpler model of a stable quantity when it is tenable, but we must examine the phenomenon to see if it warrants such a simplification.

We may observe distinct members of a population which, ideally, we would like to consider interchangeable. Only in the arena of manufactured products do we approximate such a situation, and even there, the meticulous scientist finds idiosyncracy of sampled items. Hence, the differences between individual units becomes part of statistical variability.