Automation decisions depend on measurements and models of real physical processes. Understanding measurement uncertainty, system boundaries and the limits of simplified models helps engineers interpret data and validate conclusions with greater care.