A study of the dynamic inner collection efficiency method (DICE) by experiment and computer modelling of a-Si:H p-i-n structures is reported. The DICE method involves measurement of the spectral response of the device under dual-beam excitation and varying reverse voltage bias conditions, and allows, in principle, the calculation of a spatially resolved collection efficiency. It is found that introduction to an exact calculated spectral response of small random errors to mimic experimental measurement variation seriously reduces the spatial resolution of the analysis algorithm. More significantly, detailed computer modelling reveals the presence of photogating effects which can actually invalidate the DICE analysis, but which can also lead under particular excitation and reverse bias conditions to collection efficiencies far in excess of unity, a phenomenon which we have verified by experiment.