Statistical sampling uses random selection to select samples and then assesses the result using probability theory. It involves generating a random number and size of interval. Samples has to be large enough to be true representative of the universe. If errors are discovered, the auditor can increase sample size.

Non statistical sampling is referred to as judgemental sampling. The difference between this two types of sampling is that, sampling risk of a statistical plan can be measured and controlled while in non statistical, plan cannot provide measurement of sampling.

The basic similarity of these two is that, both sampling approaches require exercise of auditor's judgement during planning, implementation and evaluation of sampling plan.

Simple random sampling

Systematic sampling

Stratified sampling

Simple random sampling

Auditors uses this sampling method where he can pick a subject without replacement. Auditors can use statistical tables or computer program to generate random numbers.

Systematic sampling

Here, every "Kth" case (I.e. every number picked after a certain interval) in population frame is selected for inclusion in the sample. Sampling frame is a list of all those within a population who can be sampled.

Stratified sampling

It requires the population to be divided into small groups called strata where random samples are taken from each stratum. It involves the following steps;

(a) Identify the population

(b) define criteria for stratification

(c) list population according to the defined strata/ subgroup

(d) determine the required sample size and appropriate representation in every stratum

(e) select using random numbers and appropriate number of subject in each stratum.

The advantage of stratified random sampling is that it ensures inclusion of all subgroups.