Sequential sampling is a sampling technique that involves the evaluation of each sample taken from a population to see if it fits a desired conclusion; the auditor stops evaluating samples as soon as there is sufficient support for the conclusion. This approach can result in fewer sampling units being examined, though the sampling will continue if any deviations are found. Consequently, sequential sampling plans work best when few deviations are expected.
A sequential sample is usually comprised of anywhere from two to four groups of sampling units. The auditor uses a computer program to determine the size of each of these groups, based on the tolerable rate of deviation, the risk of overreliance, and the expected rate of population deviation.
The sequential sampling process begins with the auditor examining the first group of sampling units. Based on the results of this examination, the auditor decides whether to:
Accept the assessed level of control risk, without engaging in any additional sampling;
Halt any further sampling, because the planned confidence and tolerable rate of deviation cannot be achieved, due to the presence of too many deviations; or
Engage in the examination of additional sampling units in order to gather more information about whether the planned assessed level of control risk can be supported.
For example, an auditor develops a set of three groups of sampling units, where each successive group contains the same number of units to be sampled. The sampling plan is to continue to the next group of sampling units if the preceding group contains at least one deviation. Several outcomes are:
Scenario 1. An analysis of the first group uncovers no deviations, so the auditor concludes that the sample supports the planned assessed level of control risk. Accordingly, she decides not to examine any additional sampling units.
Scenario 2. An analysis of the first group uncovers two deviations, so the auditor decides to continue with the sampling, using the next sampling group. This second group is found to contain one additional deviation, so the audit continues to the third group of samples in her continuing quest for more information, to see if the increased sample results will eventually support the assessed level of control risk.
Scenario 3. An analysis of the first group uncovers four deviations, which is too many deviations. Engaging in the examination of further groups of sampling units will not improve the situation, so the auditor stops the sampling process.
When it appears necessary to proceed to the next group of sampling units, the auditor should consider the cost-benefit of continuing to engage in testing. It is possible that the auditor will not be willing to proceed through every group of sampling units, and instead will accept the conclusion that the planned confidence and tolerable rate of deviation cannot be achieved.