An equation used to describe the calibration of an instrument is either of an arbitrary nature or based upon physics. The latter, which we call behavioral modeling, often provides a much better fit to the calibration than traditional, statistical, curve-fitting techniques. In applying behavioral modeling to a pressure transducer and to a radar receiver, we found that the behavioral models gave smaller residuals than even fairly high-order polynomials. Using fewer terms, they provided insight and information about the physical behavior of the devices in question, which otherwise could not have been obtained from arbitrary models. While behavioral modeling cannot be used for all instruments, the advantages in using them may be beneficial.