The tritium production rate (TPR) distribution in a fusion blanket assembly previously determined from measurements of tritium beta activity was remeasured using an independent electronic method. The results agreed within the experimental errors and confirmed the previously reported discrepancies with predictions based on three-dimensional Monte Carlo calculations and multigroup cross sections. The experimental agreement reduced the possibility that results based on the conventional chemical separation of the tritium produced could be subject to a common systematic error and confirmed the validity of the electronic method for TPR measurement. Detailed analysis showed that the discrepancies do not arise from possible neutron flux discrepancies but are most likely due to inadequate representation of anisotropic neutron scattering in the calculations.