Compilation of the inelastic mean free path (IMFP) values obtained from elastic peak electron spectroscopy (EPES) indicates a large scatter of these parameters, Furthermore, the measured IMFPs deviate from the values resulting from the predictive formulas, i.e., the equation TPP-2M and the equation G1. The differences in the measured IMFPs originate from the statistical error of the measurement and the systematic error of the EPES method. The contributions to the systematic error arise due to instrumental factors, reliability of the Monte Carlo model, and accuracy of the input parameters for the Monte Carlo algorithm. In this work, the authors investigate the possible effects of the sample surface roughness, the texture, the grain size, and the standard selected on the IMFPs determined by the EPES method for four Ni and two Si specimens. The Cu and Ag standards are used in the EPES method. The scatter between the measured IMFPs obtained in the present work and predictive formulas IMFP values is determined. The sources of errors are extensively discussed.