By Connor Child
An auditor may choose from several methodologies for determining what to look at when auditing a company. One of the frequently employed techniques is called “haphazard sampling.”
Haphazard sampling is a non-statistical technique used by auditors to simulate a variety of random sampling techniques when testing for potential errors in various accounting populations such as inventory and accounts receivable.
Chris Skousen, an associate professor of accounting at the Jon M. Huntsman School of Business, believes haphazard sampling is misleading for a simple reason — humans are inherently biased and are therefore unable to perform truly haphazard sampling. He recently completed a paper that looked at the potential biases in haphazard sampling. It will be published in an upcoming issue of the academic journal Behavioral Research in Accounting.
For the paper, Dr. Skousen and his co-authors created a study where students and professional auditors had to haphazardly select samples from a given set of accounting data.
“Even with training and an explanation of haphazard sampling, we discovered that they didn’t do it haphazardly,” Dr. Skousen said. “And we discovered their samples were biased toward the easiest items to find. If you had a listing, it was biased to what was listed at the very top or very bottom of a page. They missed what was in the middle.”
“With the risk of financial fraud in today’s world, we need to be fully aware of our financial statements and how they’re being audited,” Dr. Skousen said.
Larry Walther, School of Accountancy department head, has high hopes for the findings in the paper.
“This article has the potential to be impactful to the accounting profession,” Dr. Walther said. “The findings could be highly significant in that the current audit standards do not require use of a valid sampling plan.”
Dr. Skousen said it is not the primary responsibility of auditors to find fraud. Still, he said he wanted accounting professionals to be aware of the dangers of “coming to conclusions based on a faulty methodology.”
Dr. Skousen worked on the paper with Kenneth Hall, Tom Hall, and Bethane Pierce from The University of Texas at Arlington, and Andrew Higson from Loughborough University