(i) Judgement and statistical sampling
Judgement sampling uses the auditor‘s judgement to select the number of items to be tested, which items to be tested, and to interpret the results. Statistical sampling uses probability theory to do the same. Some judgement is always used in statistical sampling in the assessment of materiality and in the determination of what constitutes tolerable error, for example.
(ii) Representative sample
A representative sample is one whose characteristics are the same as, or similar to, the characteristics of the population as a whole. All sample selection methods attempt to select samples that are representative.
For example, a sample of invoices that have not been properly authorised in 5% of cases will be representative of all invoices if the population as a whole also has around 5% of invoices not authorised.
(iii) Tolerable error
Tolerable error is the maximum error that the auditor is prepared to accept and still conclude that the audit objective has been achieved.
For example, in relation to receivables, the auditor may be prepared to form the conclusion that receivables are not materially misstated if sampling shows that the receivables population has a value that is within plus or minus, say, 5% of the figure in the financial statements.
(iv) Different methods of sample selection
Random selection requires the use of random number tables in order to select a representative sample.
Haphazard selection may be deemed to approximate to random selection provided that no bias is displayed.
Interval (or systematic) selection involves taking every nth item, starting at random. Monetary unit sampling is also a form of systematic selection.
Block selection methods (taking one full part of the population) will probably not result in a representative selection.
Block selection might involve obtaining confirmation of receivables from one region of the country only, for example.
NB: Only two examples are required.
(v) Extrapolation of errors
Errors found in a sample are extrapolated across the population as a whole, in order to enable the auditor to form a conclusion on whether the population is materially misstated. It is important to remember that there is not necessarily a direct, linear relationship between errors in samples and errors in the populations from which they are drawn.