The size of representative microstructural samples obtained from atomic force microscopy is addressed in this paper. The case of an archetypal one-dimensional nanolayered polymer blend is considered. Image analysis is performed on micrographs obtained through atomic force microscopy, yielding statistical data concerning morphological properties of the material. The variability in terms of microstructural morphology is due to the thermomechanical processing route. The statistical data is used in order to estimate sample size representativity, based on an asymptotic relationship relating the inherent point variance of the indicator function of one material phase to the statistical, size-dependent, ensemble variance of the same function. From the study of nanolayered material systems, the statistical approach was found to be an effective mean for discriminating and characterizing multiple scales of heterogeneity.