The Problem Of False Negative Results In The Use Of Digit Analysis

Abstract

Auditors are using the predictability of digit occurrence in recorded amounts as a tool to detect suspicious data (either erroneous or fraudulent). A key component in this type of analysis is the fact that digits naturally occur in predictable frequencies consistent with Benford’s Law. This study examines the effectiveness of common digit analysis techniques in detecting the presence of suspicious data. Using data created by graduate student subjects, I find that unless the percentage of suspicious data in the population is large (> 10%) digit analysis fails to detect the presence of suspicious data. Given this propensity to produce false negative results, auditors should not rely too heavily on digit analysis as a sole or primary fraud detection tool.