A series of molecular dynamics simulations of prepolymerization mixtures for phenylalanine anilide imprinted co-(ethylene glycol dimethacrylate-methacrylic acid) molecularly imprinted polymers have been employed to investigate the mechanistic basis for template selective recognition in these systems. This has provided new insights on the mechanisms underlying template recognition, in particular the significant role played by the crosslinking agent. Importantly, the study supports the occurrence of template self-association events that allows us to resolve debate between the two previously proposed models used to explain this system's underlying recognition mechanisms. Moreover, the complexity of the molecular level events underlying template complexation is highlighted by this study, a factor that should be considered in rational molecularly imprinted polymer design, especially with respect to recognition site heterogeneity.