Measurement of the minority-carrier diffusion length (L) or lifetime (τ) in a nonuniform material, such as a photovoltaic (PV) silicon substrate, can pose major challenges. In general, a measured value is meaningful only if the size of the probe beam ≫ L, and the material is uniform in and around the region of measurement. When the measurement conditions differ from these, there is a net flow of generated carriers from the illuminated region into the unilluminated region, and/or an exchange of carriers within the neighboring regions. In these cases, the error can arise from two mechanisms: (i) recombination at the surface, and (ii) exchange of carriers within regions of different diffusion lengths. These errors can be minimized by the proper selection of the carrier generation conditions and sample preparation. By combining the experimental results with less rigorous theoretical analyses, the conditions for the measurement of local values of L (or τ) in a PV substrate can be determined. A network model can be used to integrate the influence of spatial distribution of L-values on a nonuniform substrate and predict the electrical characteristics of the large-area device fabricated on it.