Statistical Support of Forensic Auditing

Document Type



Published by INFORMS: Institute for Operations Research in Interfaces, volume 26 issue 6, 1996. Bryant users may access this article here.


INFORMS: Institute for Operations Research

Publication Source



This article investigates the Lloyd's of London case of stolen gold using statistical support of forensic auditing in Rhode Island, taking a sample of Sammartino's House of Diamonds' sales receipts and randomly choosing one of the jewelry items listed on each receipt. The auditors presented the results to James Campise, the attorney representing Lloyd's in early 1990. Campise requested to make a report visually compelling by using charts and graphs that jury members could easily understand. The authors constructed two hypotheses to test the results of the forensic audit. The first hypothesis was designed to test for non-randomness in the selected sample observations but there was no evidence of a pattern of non-randomness. The second hypothesis, the amounts of sales selected in the sample were proportional to the amounts of sales in the population on a monthly basis.