Spatial statistics methods are used to determine the effect of the sampling scale and method on two measures of mixing: the coefficient of variance CoV and the maximum striation thickness. Three sampling methods: quadrats, probes and transects, were tested. Two CFD data sets were used as test cases: dispersion of floating particles in a turbulent stirred tank and laminar mixing of tracer particles in a micromixer. Over 100 probes are needed to track the evolution of the CoV, and the probe size should match the smallest mixing scale of interest. The final value of the CoV varies by up to a factor of 5 as the probe size increases. The most useful measurement is the one which changes the most in the later stages of mixing: intensity of segregation, or CoV, for the turbulent case; and scale of segregation, or maximum striation thickness on a transect, for the laminar case. (C) 2008 American Institute of Chemical Engineers AIChE J, 54: 3068-3083, 2008