The spectral responsivity, S, and the related spectrally resolved photon-to-electron external quantum efficiency, EQE, are standard device characteristics of organic solar cells and can be used to determine the short-circuit current density and power conversion efficiency under standardized test conditions by integrating over the spectral irradiance of the solar emission. However, in organic solar cells S and EQE can change profoundly with light intensity as a result of processes that vary non-linearly with light intensity such as bimolecular recombination of electrons and holes or space charge effects. To determine the S under representative solar light conditions, it is common to use modulated monochromatic light and lock-in detection in combination with simulated solar bias light to bring the cell close to 1 sun equivalent operating conditions. In this paper we demonstrate analytically and experimentally that the S obtained with this method is in fact the differential spectral responsivity, DS, and that the real S and the experimental DS can differ significantly when the solar cells exhibit loss processes that vary non-linearly with light intensity. In these cases the experimental DS will be less than the real S. We propose a new, simple, experimental method to more accurately determine S and EQE under bias illumination. With the new method it is possible to accurately estimate the power conversion efficiency of organic solar cells.