Three tests have been performed to investigate whether cross sections determined experimentally from reflection electron energy loss spectroscopy (REELS) are more accurate than the Universal cross section for background correction of electron spectra. In the first test, the shape of background corrected spectra from a thin layer and from an infinitely thick layer of the same element were compared for the Ag3d, Au4d, and Yb4d peaks. It was found that the Universal cross section is more accurate than the experimentally determined REELS cross sections to determine consistently the peak shape of background corrected spectra. In the second test, the intensity ratio of the two components in the Au4d, Cu2p, Zn2p, and Yb4d doublets were compared to theoretical photoionization cross sections. For both the Universal and the REELS inelastic cross sections, the obtained intensity ratios agree well with theory. In the third test, we compare with theory the peak intensity ratios to Au4d of major peaks from Si, Cu, Ag, Yb, and Au obtained by using the Universal and the REELS cross sections for background correction. For the metals, the two cross sections give peak intensity ratios that are equally close to theory to within the expected uncertainties in the theoretical peak intensity ratios. However, for Ag with application of the Universal cross section the deviation from theory is slightly larger. For Si the REELS cross section is clearly most accurate.