Attribute Sampling

Statistical procedure used to study characteristics of a population. An attribute is a qualitative characteristic that a unit of a population either possesses or does not possess. For example, an account receivable is either past due or not; proper authorization for a payment either exists or does not. Thus the population under consideration is composed of two mutually exclusive classes=units possessing the attribute and units not possessing it. See also Variables Sampling.

Statistical procedure used to study the characteristics of a population. Attribute is a qualitative characteristic that a unit of a population either possesses or does not possess. For example, an account receivable is either past due or not; proper authorization for a payment either exists or does not. Thus the population under consideration is composed of two mutually exclusive classes-units possessing the attribute and units not possessing it. The statistical procedure used to estimate the occurrence of a particular attribute in a population is referred to as attribute sampling. This technique can be used by the auditor to substantiate such accounting populations as cash receipts, cash payments, payrolls, sales, and entries posted to the wrong account. In this analysis, the auditor usually determines the expected occurrence rate and the upper precision limit. The occurrence rate equals the percentage of the population having the attribute. Precision is the magnitude of deviation of a sample value from the population parameter being estimated.

Attribute sampling is particularly valuable in estimating the extent of compliance, such as the effectiveness of accounting controls using tests of transactions. Tables are used to determine sample size based on the desired confidence level, upper precision limit, and the expected rate of occurrence. Note that when analyzing a sample, the auditor may test for several different attributes. The exact definitions of attributes and occurrences should be contained in the working papers. See also Variables Sampling.