Perhaps we use my recent case of audit of revenue of a restaurant as example.
Turnover S$2.7m
Using Point of sales system.
Record the sales daily based on till roll total and reconciled to bank statements for cash, credit card, etc
The ledger record daily sales
All the till rolls are kept properly
Audit sampling: non-statistical sampling approach
I judgmentally determine 6 samples size for occurrence and haphazard selection from sales record. 6 samples size for completeness and haphazard selection from the date.
If I used statistical sampling approach, I would need to use the formula sheet to compute the sample size and random selection. (The outcome would be 80 samples for both occurrence and completeness and random selection.)
Since the SSA mentioned we can use either one approach based on auditor’s judgement, non-statistical is used on the basis that the characteristics of sampling unit is fixed at 365 days of a year. (This judgement was not based on sample size.) The objectivity of sampling is justified by the fixed sampling unit by date.
Otherwise, for a trading company. What sampling approach would we use? If I used non-statistical sampling and determine the sample size at say 25, used random selection. The challenge is how to justify the objectivity of this sampling approach in terms of the characteristics of sampling unit or other factors.