The determination of the light absorption coefficient by means of the integrating plate method is a very simple and reproducible procedure frequently used in various configurations. In this work it was calibrated with laboratory generated aerosol of different absorption and scattering properties generated from internal and external mixtures of carbon and a transparent substance. The light absorption coefficient was determined by an independent method which can be considered free of systematic errors. The light absorption coefficient determined by the integrating plate method was always higher than the value obtained by the reference method. It was strongly dependent on the ratio of light absorption coefficient to extinction coefficient of the aerosol. For weakly absorbing aerosols the integrating plate method gave excessively high values. For example, when absorption contributed 5% to the extinction, the absorption coefficient determined by the integrating plate method was too high by a factor of 2.8. Calibration factors for a variety of aerosols are given. The deviations can be explained by considering the transmission, scattering and absorption of light in the slab of deposited particles. Radiative transfer calculations show the same trend as the experimental values. It turns out that this deviation is inherent to all methods which determine the light absorption coefficient by transmission with simultaneous integration of the scattered light.