Field data on accounting error rates and audit sampling

Changes in the accounting and auditing environments since the passage of Sarbanes-Oxley (SOX) and the creation of the Public Accounting Oversight Board (PCAOB) have likely impacted accounting population error rates and audit sampling procedures. We investigate accounting population error rates and auditor sampling performance for a unique proprietary post-SOX dataset of 160 audit sampling applications for a large auditing firm. Compared to previous research, we find that in the post-SOX period, accounting data error rates are significantly lower in magnitude and frequency, sample sizes are significantly larger, and auditor performance has improved in evaluating sample errors. The low error rates suggest criticisms in prior research regarding audit sampling techniques may not be a serious concern for most audit sampling applications. Furthermore, we provide evidence that auditors' application of audit sampling is more consistent with the requirements in professional standards than previously reported.