What does ISA 530 say about Audit Sampling in para.4?
‘The objective of the auditor, when using audit sampling, is to provide a reasonable basis for the auditor to draw conclusion about the population from which the sample is selected.’
Define audit sampling.
Applying audit procedures to less than 100% of items within an account balance or class of transactions in such a way as to draw a conclusion on the account balance or class of transactions as a whole
Define statistical sampling
Approach to sampling that involves random selection of the sample items; and the use of probability theory to evaluate sample results including measurement of sampling risk (the risk of selecting a non-representative sample)
Define sampling units
Individual items constituting a ‘population’ (account balance or class of transaction) e.g., a single receivable balance within total receivables or an individual sale within total revenue
What is the key rule of audit sampling in regard to ‘sampling units’?
All ‘sampling units’ must have an equal change of being selected for testing.
What is sampling closely connected with?
Risk and materiality
Why does sampling affect detection risk?
Put simply, the more items within a population that are tested, the higher the chance of the auditor finding a misstatement
What will happen if the auditor does not test every single sampling unit in the population?
Is a risk that misstatements will not be detected
What is possible however carefully the sample is selected?
Possible the sample will not be representative of the population as a whole.
What does sampling always carry?
A level of detection risk.
What do auditor’s have to do to obtain the benefit of sampling and what is this benefit?
Determine appropriate level of risk
Benefit is that the auditor does not have to test everything
What will happen to detection risk, the larger the sample tested is?
The larger the sample tested, the lower the detection risk
What is non-statistical sampling?
The use of judgement to select a sample instead of a statistical technique
In some instances, rather than select a random sample, what may the auditor want to do?
Exercise judgement and test specific items
What is an example of why auditors will want to exercise judgement and test specific items rather than selecting a random sample?
Already observed that auditors should test all material items therefore, if the receivables’ ledger contains debts that are themselves material to the financial statements, they should be selected
Why are items selected based on materiality not statistical sampling?
Items have been selected with bias
Since items selected based on materiality are biased, what can the auditor not do and what should they do?
Not project the results of testing on these items to the rest of the population
What kind of approach is auditors selecting samples not randomly but using their professional judgement?
Non-statistical sampling - subjective approach
What is another example of non-statistical sampling?
Where auditor believes there is a greater risk of cut-off errors around year-end and therefore, focuses audit testing on the sales transactions just before, and just after, the year end
Define stratification
Dividing a population into smaller sub-populations, each of which is a group of sampling units, usually by value
When testing non-statistically, how must the remainder of the population be tested and what is this known as?
Should be sampled in a non-biased way - known as stratifying a population
Give an example of stratifying a population using receivables?
Instead of testing receivables as a whole, auditor might test 2 populations of receivable balances, individually material receivables balances and individually non- material receivable balances (total of which might well be material)
Using the receivables examples for stratification, how would sampling procedures work?
First material population - no sampling procedures - would test all material balances
Second non-material balances - apply sampling procedures and select balances for testing
In the receivables stratification example, what would then happen once tested only sampled balances from second population?
Results from sampled population could be projected onto rest of that sub-population to assess if total misstatements are likely to be material