Reducing the complexity of system models by averaging fast subsystems has a long history in applied mathematics in general and for stochastic models in particular. The fast components of the model determine an occupation measure, and the averaging argument seeks to replace this occupation measure by a simpler measure. Classically, the simpler measure has been identified as a limit of the occupation measure as some parameter in the model goes to infinity. This averaging argument will be discussed along with a recent approach by Cotter and collaborators that identifies an averaging measure that appears to give a more accurate approximation than the classical limiting argument.