The difference between a sample estimate and the projected population characteristic at a specified sampling risk. This allowance is also the difference between the expected error rate and the tolerable deviation rate.