One problem associated with the use of Computational Fluid Dynamics(CFD) in reactor modeling is the proper validation of the models. Proper validation in this context means that the physical fluid dynamic model, the mathematical implementation and the data used for validation must be consistent. The present paper addresses this issue and to provide appropriate relations between experimental method and modeling approach. A critical review of currently used measurement techniques for characterizing multiphase flow systems is presented. The interpretation of the data obtained from the various techniques is discussed as well as how these data can be used for validation of various CFD model formulations. Steady state models can be validated using time averaged data, making sure that the averaging time for the experimental data is long enough so that low frequency periodic oscillations also are evened out. If homogeneous systems are considered, then a volume average approach may be used for modeling. If the system cannot be considered homogeneous and steady, as is the most common case, then a dynamic ensemble averaging technique should be preferred. The validation of such models must be done with methods fast enough to resolve periodic fluctuating structures of interest. These methods are cumbersome and tedious to operate and the ergodic hypothesis may be invoked enabling the use of volume or time averaged data for the validation of ensemble averaged models.