Ranked set sampling: an auditing application

This study compares the statistical precision of simple random sampling with balanced ranked set sampling in an inventory valuation scenario. Computer simulation is used to calculate standard errors for the ranked set sampling mean, and those standard errors are then compared to the corresponding standard error achieved under simple random sampling. Results indicate that required sample sizes for a given precision are much smaller under ranked set sampling than under simple random sampling. This implies that simple random sampling is inferior to ranked set sampling in auditing scenarios involving the measurement of time consuming or difficult to gather data such as inventory observations, receivable confirmations, etc. Accordingly, auditors using ranked set sampling in lieu of simple random sampling can achieve the significant cost reductions associated with smaller sample sizes without sacrificing audit quality. This is a significant finding because current auditing practice is still using the simple random sampling methodology.