Testowanie całej populacji AI w quizie audytowym

Odpowiedzi i wyjaśnienia pojawiają się w miarę rozwiązywania quizu.

1. When an auditor runs a three-way match on every purchase invoice instead of a sample, what is eliminated for that attribute?
   • The need to evaluate exceptions
   • The risk that the data extract is incomplete
   • Sampling risk
   • The need for IT general controls
2. Why does testing 100 percent of recorded purchases not prove that liabilities are complete?
   • Unrecorded liabilities never appear in the recorded data being analyzed
   • Analytics cannot read invoice amounts
   • Purchases are never matched to receiving records
   • Sampling standards forbid full-population tests
3. Before running full-population tests, what step does the guide say proves the population?
   • Asking management to confirm the file is complete
   • Running a Benford test on the amounts
   • Selecting a random sample of 25 items
   • Reconciling the extract to the general ledger or trial balance using record counts and control totals
4. A cutoff test flags 4,000 items, most of which look like timing differences. Which response fits the guide's approach?
   • Ignore the flags because most are probably harmless
   • Group the flags by shared cause, investigate each group, refine the rule and rerun it
   • Report all 4,000 items as misstatements
   • Switch back to testing a sample and discard the analysis
5. How should an auditor treat an item flagged by a full-population test?
   • As an item that needs investigation, not automatically a misstatement
   • As a confirmed misstatement to be booked
   • As evidence that controls failed completely
   • As a sampling error
6. An auditor examines 25 items from a group of 2,000 flagged items and draws conclusions about all 2,000. What applies to that step?
   • Nothing, because full-population testing was already done
   • Only the rules on IT general controls
   • The audit sampling requirements
   • Benford's law thresholds
7. What has been a growing focus for auditing standard-setters and regulators regarding full-population testing?
   • Banning audit sampling entirely
   • How auditors should respond to large numbers of items flagged by technology-assisted analysis
   • Requiring Benford's law in every audit
   • Eliminating the need for performance materiality
8. Why can an invoice that passes a perfect three-way match still be a problem?
   • Three-way matches cannot be run on complete populations
   • The test was not run by a licensed CPA
   • Receiving records are always unreliable
   • The test checks only the rule written, and a fake vendor can still produce matching records