Expected and Observed Error Rates in Performance Audits

For most performance audits, the auditor has no real basis for estimating the error rates (expected error rate) before performing the examination. Unlike financial audits, performance audits are not routinely repeated and results from several past audits are not available to predict the outcome of the current audit. As a result, performance auditors must choose another strategy for selecting a value for expected error because it is a required input variable for calculating a sample size (as part of a generalizable sampling approach).

It can seem counterintuitive, but the result of the examination, the observed error rate, does have an impact on the precision of the findings. This is because this precision is influenced by the amount of variance within a population, and the variance within a population reflects the rate of error. Variability within a population is at its highest when the error rate is about 50%. For any given sample size, the accuracy of estimates will be lowest, less precise, with a finding of about 50% error rate. As the error rate decreases, variability is reduced, and the estimate becomes more precise. More information on expected and observed error rates is in Appendix 1.

• PreviousConfidence Intervals
• NextPrecision Standards for Performance Audits