Previous measurements of minority carrier mobility by the Haynes‐Shockley technique have ignored one or more of the effects of the three‐dimensional nature of the ambipolar diffusion, and bulk and surface recombination. In some cases, the experimental conditions were tailored so that the effects of these factors were truly negligible, imposing restrictions on the respective experimental configurations. In general, each factor does influence the drift time from which mobility estimates are made. This paper shows how their effects may be conveniently and systematically included in the data analysis and gives examples of the processing of experimental data acquired fromp‐type silicon samples of different geometries. The suitability of the new technique for measurement of mobility in thin wafers of material is pointed out.