>is it possible to measure a system and control for the same factors that the system itself is trying to control for? It's hard to have a meaningful discussion

>is it possible to measure a system and control for the same factors that the system itself is trying to control for? It's hard to have a meaningful discussion of these issues in the abstract, but the short answer is: 1) it depends 2) bias correction is not a trivial matter Firstly, you have to have some sort of quantitative estimate of the bias. This is not always possible, and even in cases where it is, said estimate may be noisy or, itself, biased. Bias correction is useful and often legitimate, but its use should always be interpreted as weakening the epistemological rigor of the analysis. In lay terms: bias correction should be consumed with a grain of salt. General entry point into the subject matter: https://stats.stackexchange.com/q/31253/17054 https://stats.stackexchange.com/q/31253/17054