Statistical sampling is ideal in the following conditions:-
i) The population to be tested must be homogenous i.e. it must consist of items of the same kind, subject to the same level of audit risk.
ii) The population must be fairly large, otherwise the benefits of the technique will not be achieved. Statistical techniques need to be adjusted to work on small populations.
iii) Expectation of error must be low – i.e. control risk must have already been assessed as less than high (i.e. ICS is strong). High control risk requires impractically large sample sizes.
iv) The items in the population must be easily identifiable once selected so that checking of the items is easily carried out e.g. Items such as invoices must be pre-numbered and filed in sequence.