Audit sampling techniques may permit errors or dishonesty to go undetected. Audit sampling occurs when a review of less than 100% of a population occurs. Determining how the size of a population is used in an audit sample is a function of “sampling risk”, which is defined as the probability that a sample will not accurately represent an entire population. It means that an auditor might reach conclusions based on a sample size that would be different from conclusions drawn from an entire population.
Audit sampling commonly uses two types of sampling methods: statistical and non-statistical (judgmental). The statistical method is math driven and is based on a random selection of sample items (tables and statistical percentages). The judgmental method is based on professional experience and expertise (professional judgement). What is important to know is that audit sampling, carries advantages and disadvantages.
Among its advantages, statistical sampling saves time and is cost effective during engagements. At times, audit sampling is used by forensic accountants during engagements that entail legal proceedings. Forensic Accountants and the clients must discuss the risks associated with each methodology – i.e., that the statistical sampling is easier to present, but may be more difficult to support. In addition, if statistical data sampling reveals internal irregularities, the scope of an investigation could expand accordingly. The result is that the time and scope of a matter could be more costly. That said, sampling can overlook important data; thereby increasing risks of a material misstatement.
When a CPA is engaged and exclusively utilizes the judgmental method, he or she is removing “probability theory”, and is wholly dependent on “judgment” (the information would still need to be corroborated). A CPA might determine that 100% testing is appropriate if a population consists of a small number of high value components, or if there is perceived risk of a material misstatement. And, what about small value components which when taken together become material?
International Standards on Auditing (ISA) recognizes five principal methods of “audit sampling”. They are:
1. random selection: ensures that all items within a population stand an equal chance of selection by the use of random number generation,
2. systematic selection: divides the number of sampling units within a population by sample size to generate a sampling interval,
3. monetary unit sampling: a value-weighted selection whereby sample size, selection and evaluation will result in a conclusion in monetary amounts,
4. haphazard selection: does not follow a structured technique,
5. block selection: selecting a block (or blocks) of contiguous items from within a population.
Because an objective is to avoid bias and because haphazard and block selection are considered to be non- statistical sampling on their own, they are perceived as not forming a reasonable basis from which to draw conclusions that can be supported, and ensure that all samples are representative of their population.
A CPA engaged for a forensic accounting investigation cannot rely on any single analytical tool. The volume of data in cases in the court system today gives rise to the reality that CPA forensic accountants have to sample the data. This results in judgment calls being made to determine which items to examine. The examination of a partial universe of transactions (less than 100%), leaves room for error. Sampling can often ignore small transactions, thereby missing significant details and patterns which could lead to valid conclusions.
Don’t be a victim of your own making. When hiring a CPA forensic accountant make sure that his or her understanding and experience with audit sampling is equally broad and deep.