Despite the growing use of atomic force microscopy (AFM) for the measurement of silicon wafer microroughness, no generally accepted method has been developed to deal with issues around accuracy and reproducibility. We review problems that affect these AFM studies and demonstrate the effect of probe tip size on AFM microroughness data. Without knowledge of AFM probe tip geometry, it is impossible to quantitatively compare Ra or RMS microroughness data between different measurements. An experimental solution is to characterize tip sizes during imaging and compare data taken with similar size tips. While this will significantly improve quantitation, it is restrictive in that data taken with different size tips cannot be easily compared. We propose a solution to this problem in the use of power spectral density (PSD) to evaluate microroughness with a “cutoff frequency” at the lateral wavelength where tip effects begin to affect the accuracy of the microroughness measurement. An example of this approach is described