ABSTRACT
Underlying this table is an assumption that the poisson distribution, which is different from the normal distribution, is a reliable working approximation. Also, for most sampling situations the practical assumption is that not many errors are likely to exist in the population or be picked up in the sample. If either of these assumptions is invalid, perhaps because populations are small or error rates large (usually over 10 per cent), auditors are advised to seek expert advice, though it is generally considered that, if the auditor knows that error rates are high and large numbers are likely to be picked up from the sample, there is little point in sampling in the first place, as an unacceptable conclusion has already been reached judgmentally.
