Overestimation in dose recovery results was observed at high doses (>1000 Gy) when a single-aliquot regenerative-dose (SAR) procedure is applied to the multiple-elevated-temperature post-IR IRSL (MET-pIRIR) signals from potassium-rich feldspar (K-feldspar) grains from the Nihewan Basin. Such overestimation results from the non-bleachable (residual) signals that were not appropriately corrected for. A signal-subtraction method was proposed to solve this problem, which involves building growth curves for the bleachable and non-bleachable signals. Here we modelled the effect of residual signal on equivalent dose (De) and found that even a small amount of residual signal may give rise to large overestimation at high doses. Our results highlight that residual signals may not only affect young samples but may lead to erroneous results for old samples. Our study also suggests that although residual signals may result in significant error in the dose recovery, they may have a very small effect on natural samples because their residual signals seem to be relatively small.