Audit Sampling & Analytics

Sampling is where audit rigour is won or lost. When an auditor cannot test every transaction, the sample must be large enough, drawn properly, and projected back to the population so the conclusion is statistically defensible. ISA 530 and ISSAI 1530 set out exactly how — sample size derived from materiality, expected error and confidence, with selection and projection documented. A guessed sample of 'twenty invoices' supports nothing.

In public-sector audits this matters because conclusions about irregular expenditure, SCM compliance or revenue completeness rest on the sample. If the sample size is arbitrary or the selection is not traceable to source, AGSA cannot rely on the testing and the finding collapses — or worse, a real misstatement goes undetected because the population was never properly covered.

Defensible sampling turns testing into evidence. Monetary-unit and statistical methods derive the size from materiality and risk, stratification focuses effort where the money is, and every selected item drills back to the source transaction. Projected misstatement is computed automatically, so the conclusion about the whole population is mathematically supportable.