The ability of single-aliquot techniques to recover a known laboratory dose in coarse-grain potassium-rich feldspar separates is tested. It is shown that, for the samples under study, single-aliquot methods underestimate the known dose. It is deduced that this phenomenon is caused by an increase in electron trapping probability as a consequence of heating. The same phenomenon would lead to an underestimation in the equivalent dose in a natural sample. Multiple-aliquot techniques for equivalent dose determination may not suffer from this problem, but are not suitable for samples that might not have been completely bleached at the time of deposition. It is argued that testing for changes in electron trapping probability should become routine practice in the single-aliquot dating of feldspar.