Dimerizations of delocalized neutral radicals may be endowed with quite significant activation barriers. The origin of these barriers is discussed in terms of a model that emphasizes the role of localization of the unpaired radical upon bond formation. Several examples are given in which the model is compared with the results of quantum chemical calculations including the coupling of allyl radicals and of benzyl radicals at various possible carbon sites. The dimerization behavior of radicals in the NADH family is also examined. The connection between the reasons that underlay the existence of the activation barrier and the principle of "nonperfect synchronization" is discussed. The dimerization of conjugated radicals indeed offers a precious example that can be used to decipher the reasons behind these behaviors, being devoid of the ambiguities arising from the simultaneous involvement of ionic and covalent states, significant solvent reorganization, and the contribution of extensive proton tunneling, in the mostly discussed case of proton transfer at carbon.