AUDIT-SPECIFIC CALIBRATION OF A SAMPLE EVALUATION METHOD USING COMPUTER MODELING

Auditors need to assess sampling risk accurately when they are examini ng only a portion of the transactions making up an account balance. Re search over the past thirty years has provided auditors with a large n umber of statistical methods for assessing sampling risk. The accuracy of these methods is unknown for many audit populations. This study pr esents the simulated reference distribution (SRD) method that allows a uditors to assess the accuracy of a statistical method with the actual population being audited and to adjust the calculated sampling risk a ccordingly. In applying the SRD method, auditors first select a sample and evaluate the nominal sampling risk using the statistical method o f their choice. Then they construct a hypothetical audit population wi th errors that sum to an amount close to materiality. This hypothetica l audit population is based on auditors' knowledge of the various sour ces of error that could affect the account balance and lead to a mater ial misstatement. Computer simulation is used to construct a reference distribution of nominal sampling risks. The previously calculated nom inal sampling risk is calibrated using this reference distribution. Th is paper illustrates the use of the SRD method with a case study. The results of this investigation indicate that the SRD method may be a us eful addition to the auditor's tool box when: (1) the objective of the audit sample is to reduce audit risk, (2) the error rate in the popul ation is low, (3) the marginal cost of auditing per sampling unit is h igh, and (4) the accuracy of the sample evaluation method is unknown f or the population being audited.