Computationally predicted N 1s core level energies are commonly used to interpret the experimental measurements obtained with X-ray photoelectron spectroscopy. This work compares the application of Koopmans' theorem to core electrons using the B3LYP functional with two commonly used basis sets, analyzes the factors relevant to the comparison of the computational with experimental data, and presents several correlations that allow an accurate prediction of the N 1s binding energy. The first correlation is obtained with a series of known nitrogen-containing functional groups on well-characterized organic monolayers. This approach can then be reliably extended to a number of nitrogen-containing chemical systems on silicon surfaces in which the nature of the chemical environment of nitrogen atoms had only been proposed based on a number of analytical techniques. In most of those cases, the XPS analysis is consistent with the proposed structures, but is not always sufficient for conclusive assignments. Third, it was attempted to also include N-containing systems on metals. Despite the admittedly oversimplified approach taken in this case (the metal surface is approximated by a single atom), the observed correlations are still experimentally useful, although in this case significant outliers are found. Finally, previously published correlations between experimental and theoretical C 1s data were reexamined, yielding a set of correlations that allow experimentalists to predict C 1s and N 1s XPS spectra with high accuracy.