Samples that are thick with respect to the measuring apparatus can pose problems in the measurement of directional-hemispherical transmittance. Due to the lateral displacement of light that these samples induce, traditional small-beam techniques are often no longer applicable. Broad-area irradiation offers a solution to these problems, but, in practice, imperfections in the extent and the uniformity of the irradiation can create errors. This paper reports measurement of the ＇＇detector-integrated inscattering function＇＇, which describes the lateral displacement of light, for a number of thick samples over a range of incidence angles. This function is used to determine the minimum area that must be irradiated for an accurate broad-area irradiation transmittance measurement. The errors caused by transverse non-uniformities in the irradiance distributions produced by a state-of-the-art broad-area irradiation spectrophotometer are also reported. Errors caused by transverse irradiation non-uniformities are estimated, and transmittance measurements made with small-area and broad-area irradiation are compared. It is shown that in most cases small-area irradiation transmittance measurements underestimate the transmittance of thick samples due to outscattering, and that broad-area irradiation measurements using state-of-the-art irradiation optics can contain large errors of up to 20%, due to irradiance non-uniformities, particularly at large angles of incidence.