We present a study on the excited states of an ethylene dimer as to investigate the presence of and perturbation from low-lying charge-resonance states calculated by linear response density functional theory (DFT) using the B3LYP and CAM-B3LYP functionals. The calculations are compared to a reference CC3 calculation revealing a better description of the excited states by CAM-B3LYP than that of B3LYP. The K parameter introduced by Peach et al. [M.J.G. Peach, P. Benfield, T. Helgaker, D.J. Tozer, J. Chem. Phys. 128 (2008) 044118] does not always reveal the problematic charge-resonance states obtained with B3LYP. The generality of the problem is addressed by taking pentacene as an example. (C) 2009 Elsevier B. V. All rights reserved.