In this study, we investigate the effect of substituents in determining the modes of one-electron reductive cleavage of X-NRR' (X = Cl and Br) molecules. We achieve this through comparison of the calculated gas-phase electron affinities (EM) and aqueous-phase one-electron reduction potentials (E degrees's) for a range of nitrogen-centered radicals ((NRR)-N-center dot') with the corresponding EA and E degrees values for Cl-center dot and Br-center dot. The gas-phase EM have been obtained using the benchmark-quality W1w thermochemical protocol, whereas E degrees values have been obtained by additionally applying free energy of solvation corrections, obtained using the conductor-like polarizable continuum (CPCM) model. We find that the N-halogenated derivatives of amines and amides should generally cleave in such a way as to afford (NRR)-N-center dot' and X-. For the N-halogenated imides, on the other hand, the N-brominated derivatives are predicted to produce Br-center dot in solution, whereas the N-chlorinated derivatives again would give Cl-. Importantly, we predict that N-bromouracil is likely to afford Br-center dot. This may have important implications in terms of inflammatory-related diseases, because Br-center dot may damage biomolecules such as proteins and DNA. To assist in the determination of the gas-phase EAs of larger (NRR)-N-center dot' radicals, not amenable to investigation using W1w, we have evaluated the performance of a wide range of lower-cost theoretical methods. Of the standard density functional theory (DFT) procedures, M06-2X, tau-HCTHh, and B3-LYP show good performance, with mean absolute deviations (MADs) from W1w of 4.8-6.8 kJ mol(-1), whereas ROB2-PLYP and B2-PLYP emerge as the best of the double-hybrid DFTs (DHDFTs), with MADs of 2.5 and 3.0 kJ mol(-1), respectively. Of the Gn-type procedures, G3X and G4 show very good performance (MADs = 2.4 and 2.6 kJ mol(-1), respectively). The G4(MP2)-6X+procedure performs comparably, with an MAD of 2.7 kJ mol(-1), with the added advantage of significantly reduced computational expense.